Back
[00:09:38] <kholz> Hey, I have an LPKF mill/drill here at work and I'm wondering if anyone has gotten anywhere with getting EMC going on one of these.
[00:25:11] <Rugludallur> kholz: not me, someone might though
[00:25:53] <Rugludallur> kholz: might need to wait a bit for an answer, most people only check the screen every now and then
[00:26:13] <robin_z> quite ;)
[00:27:49] <skunkworks> do you know any info on it? what kind of drives or such?
[00:35:52] <kholz> I don't, yet.
[00:36:00] <kholz> Lemme check and see if I can find any.
[00:40:24] <kholz> stepping motors on the XY, and some hydraulic gizmo for the Z drive
[00:40:40] <kholz> Rather, electromagnetic with hydraulic dampening
[00:45:29] <cradek> the Z sounds challenging
[00:46:08] <cradek> is there an existing control?
[00:50:23] <cradek> bbl
[00:51:45] <wb9mjn> Is that one of the PCB mechanical etching machines ?
[00:52:10] <wb9mjn> The RF one has a hockey puck like device to set the depth of cut ?
[00:52:44] <wb9mjn> Very fancy...and accurate so that only the copper gets cut away, leaving an untoched PCB substrate...
[00:53:53] <kholz> Well, one can set depth
[00:53:59] <kholz> there is an adjustment knob
[00:54:26] <kholz> There is software, but it is way weak.
[00:55:32] <wb9mjn> When you said "Mill Drill" I am not sure people knew what you were talking about...
[00:56:02] <wb9mjn> There are machine tools that are half milling machine, half drill press, that are called "mill drill's"...
[00:56:18] <wb9mjn> Is this a Mill Drill, or a PCB fabricating machine ?
[00:57:57] <fenn> does anyone understand how that polygon lathe works? i dont understand how you can make a flat surface with two synchronized cutters unless you move back and forth really really fast
[00:58:12] <wb9mjn> Its common for some Mill Drills to be converted to CNC operation from manual....Rong Fu is a common brand of
[00:58:15] <wb9mjn> Mill Drill...
[00:58:38] <fenn> rong fu is round column which will slip and destroy your workpiece
[00:58:47] <fenn> you want a dovetail column
[00:58:53] <wb9mjn> I have not figured one out Fenn...
[00:59:11] <wb9mjn> Yea...I have a dovetailed square column machine here...small milling machine...
[00:59:40] <fenn> oh i guess rong fu makes dovetail column mills too; i've only ever seen the round ones though
[00:59:45] <wb9mjn> I believe Kholz is talking about a PCB prototyping machine (LPKF is know for those not Mill Drill's) and
[01:00:00] <wb9mjn> just trying to be sure before I get too confused...
[01:00:22] <kholz> wb9mjn: yessir
[01:00:28] <kholz> It is a PCB fab machine
[01:00:33] <wb9mjn> Ok...
[01:00:40] <kholz> Sorry for being in and out so much
[01:00:55] <wb9mjn> That s fine, I am in an out allot too...
[01:01:13] <kholz> I'm at work, and as much as it is work related, I'm trying to get an embedded linux system off the ground too ;[
[01:01:24] <wb9mjn> So, do you want to motorize the Z depth or just set it to the proper oz of copper clading on the PCB ?
[01:01:31] <fenn> hockey puck device?
[01:01:58] <wb9mjn> Yes...the hockey puck glides over the PCB, and the spindle rides on it...
[01:02:06] <kholz> wb9mjn: there is no programmable depth control I don't believe
[01:02:18] <kholz> It has a manual adjustment knob
[01:02:21] <wb9mjn> The cutter is exposed through a hole in the bottom of the hockey puck....
[01:02:28] <kholz> yep
[01:02:40] <wb9mjn> Yep...Kholz, I have seen on-line video of how those work....
[01:03:21] <wb9mjn> Might do it with a flex shaft, and a stepper off to the side of the puck...
[01:03:45] <wb9mjn> Is there a locking nut ?
[01:03:54] <kholz> for what? Bits?
[01:03:57] <wb9mjn> Or is it tight enough to just turn to adjust the cutter depth ?
[01:04:10] <kholz> no lock nut
[01:04:12] <wb9mjn> Locking nut for the adjustment knob...
[01:04:31] <kholz> It clicks up and down between positions.
[01:04:33] <wb9mjn> Does it take allot of force ...would a small NEMA 17 motor do it ?
[01:04:49] <kholz> It takes a decent amount of force.
[01:04:51] <wb9mjn> The Z is on a solenoid that goes up and down....
[01:04:55] <wb9mjn> right ?
[01:04:57] <kholz> yes
[01:05:06] <kholz> One has to manually adjust height per bit anyhow
[01:05:18] <wb9mjn> The depth of cut is adjusted by the knob to expose the cutter out the bottom of the hockey puck...
[01:05:27] <kholz> exactly
[01:05:32] <wb9mjn> Yep....
[01:06:06] <wb9mjn> Do you need the adjustment to be motorized ?
[01:06:10] <kholz> not at all
[01:06:13] <wb9mjn> Ok....
[01:06:31] <wb9mjn> So, you just need support for the solenoid Z down action...
[01:06:34] <kholz> yep
[01:07:06] <wb9mjn> Ok...welp, I am not the guy who can figure that out...but hopefully those that are are reading this and will
[01:07:11] <wb9mjn> chime in when they get time...
[01:07:15] <kholz> :)
[01:07:16] <kholz> k
[01:07:44] <wb9mjn> The machine is pretty simple from a control aspect, compared to what people do on here with EMC...
[01:08:15] <wb9mjn> What makes it expensive is the precision head, and the very expensive BLDC spindle motor and drive...
[01:08:26] <wb9mjn> Which has very good axial TIR (runout)...
[01:09:03] <wb9mjn> And the need for the X / Y to be accurate to under .001, so that .001 combined accuracy can be achieved in practice...
[01:09:19] <wb9mjn> Which is typical PCB specs...
[01:09:22] <fenn> so this hockey puck is made out of delrin or something?
[01:09:34] <wb9mjn> Its aluminum, or steel...shiny metal...
[01:10:03] <fenn> dont you need a motorized Z for drilling?
[01:10:17] <wb9mjn> There are two companies that spcecialise in these machines LPKF and TTech...LPKF has the hockey puck....
[01:10:26] <wb9mjn> They charge like $20K for these machines...
[01:10:41] <wb9mjn> Good point...
[01:10:52] <wb9mjn> I think they have a manual drill press for that....
[01:11:17] <fenn> uhhh
[01:11:17] <wb9mjn> These are for prototypes...and allot of the electronics these days are surface mount...
[01:11:21] <fenn> ummmm
[01:11:23] <wb9mjn> Very little drilling...
[01:11:25] <fenn> that's fucking retarded
[01:11:45] <wb9mjn> The RF patterns can be quite intracate to work....
[01:12:36] <wb9mjn> The dimensional accuracty is very important....otherwise the circuit wont work to properly guide high frequency
[01:12:38] <wb9mjn> signals...
[01:12:58] <wb9mjn> These days, with SATA and USB, some of these RF aspects are needed in digital PCB's as well...
[01:13:34] <wb9mjn> Especially with 3.05 GHz processors...and 1 GB/s front side bus's...
[01:14:33] <wb9mjn> Surface mount electronics usually only need holes in the boards for Via's...
[01:14:46] <fenn> yep yep yep
[01:15:00] <fenn> apparently the small vias are done with a laser
[01:15:08] <wb9mjn> I think the fullup LPKF system that can do multi-layer boards, with Vias is another $40 K beyond the mechanical
[01:15:13] <wb9mjn> etching machine...
[01:15:48] <tomp> hello
[01:15:53] <wb9mjn> Gets crazy exepensive....and it explains why so many people get into amateur CNC to make PCB's...like
[01:15:56] <fenn> howdy tom
[01:15:58] <wb9mjn> Cradek and Jepler...
[01:16:06] <skunkworks> I do pretty well with my gantry.
[01:16:27] <fenn> wb9mjn: you can get custom pcb's for like $100 easily
[01:16:35] <skunkworks> I would love to have something a little smaller to do pcb - but it works
[01:16:41] <fenn> but if you do it a lot or have multiple revisions that gets expensive
[01:17:07] <fenn> heh skunkworks i think your machine takes the prize for "most oversized pcb cutter"
[01:17:31] <tomp> hi fenn
[01:17:34] <wb9mjn> For digital stuff those $100 board proto shops make sense....but
[01:17:48] <wb9mjn> for RF stuff that needs to be tuned in, not really....
[01:17:48] <skunkworks> :) 6'X4'X3' working area - big? ;)
[01:18:03] <fenn> tomp: what do you think of using wire edm end-on to etch copper off of pcb board stock?
[01:18:20] <fenn> raster scan so you always have electrical continuity to the substrate
[01:18:32] <tomp> hmm, hard to guide it if it's so exposed over the roller, but maybe end of tube?
[01:18:43] <wb9mjn> might burn the substrate...its only epoxy fiberglass...and any deposited carbon residue would act like a
[01:19:01] <wb9mjn> stray resistor...not good for 1 G Ohm input Opamp circuts....
[01:19:05] <jtr> Skunkworks: so you can do multilayer boards up to 3ft thick?
[01:19:06] <tomp> oh, then ecm
[01:19:14] <fenn> ecm?
[01:19:21] <skunkworks> jtr - yah - doesn't everyone?
[01:19:28] <tomp> yeh, let me get an url
[01:19:38] <fenn> electrochemical machining
[01:19:58] <wb9mjn> Yep...skunk has like penguin 3D models down inside his PCB's...
[01:20:30] <fenn> sounds even slower than edm
[01:20:45] <fenn> like electroplating in reverse, basically
[01:20:53] <wb9mjn> Not if you have 500 amp supply....
[01:20:55] <tomp> http://www.indoor.flyer.co.uk/ecm.htm yeh can be slow but wont hurt substrate, and very very very lo wear
[01:21:02] <kholz> wb9mjn: we are an RF lab.
[01:21:08] <kholz> I'm making a lot of antennas here
[01:21:13] <tomp> and very stupid simple power supply, no oscillator needed
[01:21:27] <kholz> and building up proto boards
[01:21:40] <skunkworks> http://www.electronicsam.com/images/KandT/servostart/almost.JPG
[01:21:45] <skunkworks> it does ok
[01:21:45] <tomp> it's very much etching, but could be cnc'd
[01:22:29] <fenn> doesnt the copper build up on the electrode?
[01:22:33] <wb9mjn> I have done a few micrstrip patch antennas...but they were high performance with low Er substrates...
[01:22:47] <wb9mjn> Didn't etch the radiators....
[01:23:27] <wb9mjn> Just sliced 2 oz copper...
[01:23:47] <tomp> fenn: no the copper goes into the dielectric stream ( salt water bath flows it away)
[01:23:53] <wb9mjn> I work on high frequency interconnects now....40 Ghz blind mate connectors...
[01:24:06] <jtr> skunkworks: I saw that - nice! might want to use that design someday.
[01:24:53] <wb9mjn> Not sure if you can be accurate enough for Kholz work....
[01:25:05] <wb9mjn> EMC that is...
[01:25:39] <skunkworks> why would emc not be accurate enough?
[01:25:57] <wb9mjn> Fenn...the copper will build up in solution between the electrode and work...so you need to flush fresh ionic liquid through
[01:26:01] <wb9mjn> continously...
[01:26:05] <tomp> uh, this box has ubu 5.1 and i'd like to update to 6.06 (maybe w live cd?) will i loose the apps. files i have already?
[01:26:23] <wb9mjn> +/- .001 " .... typical for 2 to 10 GHz patch antennas....
[01:26:45] <tomp> emc is not what's accurate, your machine tools is whats accurate ( rather the limiting factor )
[01:26:49] <fenn> tomp: preferences and such will be saved.. it will ask you sometimes if you want to delete the preferences and use the new one
[01:27:00] <tomp> thanx
[01:27:08] <wb9mjn> They are somewhat hi Q (as they say), and narrow banded....to get the antenna on the same freq each time takes
[01:27:30] <wb9mjn> accurate cutting...but I am making some assumptions about what Kholz is working on...
[01:27:32] <kholz> wb9mjn: What are you doing?
[01:27:48] <tomp> the ecm process is removing molecules, so it isnt a limit to your precision :)
[01:28:08] <wb9mjn> Oh...thinking electrochemical machining not accurate enough...
[01:28:22] <wb9mjn> EMC the program, is as accurate as you can make the machine...
[01:28:24] <kholz> Nono, chemical would be fine
[01:28:44] <kholz> We're using rodgers though
[01:28:55] <kholz> (in our radio, that is)
[01:31:37] <tomp> using edm would have a tool that is maybe .010 to .0006" larger than the form that you percieve as 'tool' (the tool in edm is a field surrounding the form you use )
[01:31:55] <tomp> so your machine tool better be rigid when using a tool that is that small
[01:32:23] <tomp> if you moved .002 too colose or too far, you loose the process
[01:32:33] <tomp> close
[01:32:43] <wb9mjn> EDM would probably do irepairable harm to the microwave circuit board (rodgers) dielectric material...
[01:33:20] <wb9mjn> Yea...that is why they use the expensive german PCB spindles in those machines....
[01:33:39] <tomp> low energy edm used for the bonding tools in chip building are in the hundreds of joules, not very strong
[01:33:51] <wb9mjn> Kavo or NSK...
[01:34:27] <tomp> but i dont know the material of your di material
[01:34:33] <wb9mjn> You cannot have any sputter in the PCB material...it will make an artificial dielectric and dramatically change
[01:34:50] <wb9mjn> the material dielectric constant....which is what you spent all the money on it, in the first place...
[01:35:27] <tomp> i meant, the stuff like fiberglass, greenglass, etc, the substrate
[01:35:27] <wb9mjn> ECM might work, but would need a special machine just for that....
[01:35:58] <wb9mjn> Not really....there are all sorts of material...dispersed fiberglass in Teflon....
[01:36:12] <wb9mjn> Ceramic dust in polyester....
[01:36:36] <wb9mjn> Rodgers does not say exactly what it is....
[01:36:57] <wb9mjn> Its a big buisness....every cell phone has a little slab of the stuff in it...
[01:38:03] <tomp> no sputter infers no bur, no chips, maybe deplating is a good way ( ecm is like deplating), tho i did wave guides for phillips on edm
[01:38:32] <tomp> wave guides for microwave are tight tolerance, but then have no substrate
[01:38:35] <wb9mjn> when the edm burns away the copper, there will be sputter...at low power levels it will be microscopic dust that is
[01:38:47] <tomp> gotcha
[01:38:53] <wb9mjn> hot when it comes off the copper clading, and embed itself intothe dielectric....
[01:39:20] <wb9mjn> The metal dust in the dielectric is a structure in RF called an "artificial dielectric"....
[01:39:55] <wb9mjn> And dramatically high dielectric constants can be made with the right dust concentration in the dielectirc...
[01:39:59] <tomp> so whats alternative to ecm ( which sems to fill the bill ) grinding, milling, shaping?
[01:40:12] <wb9mjn> So, even a little will change things too much for a RF circuit to work...
[01:40:42] <wb9mjn> Typically you want +/- .05 in the dielectric constant to get your 50 +/- 1 ohm transmission lines...
[01:41:09] <wb9mjn> And antenna are effectively resonators, which are more senatative to dielectric constant variation...
[01:41:38] <wb9mjn> Depending on the material, one sometimes has to design the circuit for a specific PCB grain direction to get the
[01:41:56] <wb9mjn> tuning right ... but that was int he old days....the new PCB's are more anisotropic...
[01:42:34] <wb9mjn> These LPKF machines use special cutters they developed in a Kavo spindle...
[01:43:10] <wb9mjn> The hockey puck mounts the spindle, so that the cutter exposure out the bottom of the hockey puck is very
[01:43:12] <wb9mjn> accurate...
[01:43:54] <wb9mjn> then the hockey puck glides over the pcb...
[01:45:13] <tomp> some way for swarf to be removed from the hockey puck covering the pcb?
[01:45:21] <fenn> wb9mjn: interesting about the metal dust raising the dielectric constant.. never heard of that
[01:45:29] <wb9mjn> With that arrangement even 1/2 oz copper clading can be removed and not damage the substrate...or so they say...
[01:46:01] <tomp> lower? (it got polluted)
[01:46:03] <wb9mjn> They have a vacumm nozle on it Tomp...
[01:46:14] <tomp> ok, and burrs?
[01:46:27] <fenn> seems easier to just expose the board with a laser diode and etch it
[01:46:34] <fenn> but what do i k now
[01:46:40] <wb9mjn> I believe its an upcut spiral milling cutter....pulls the chips up, through the hole in the hockey puck...
[01:46:46] <tomp> must be damn clean for trans lines... i guess negligible
[01:47:35] <wb9mjn> The prototyping thing is just to get something that is close enough to what you are going to get from a etched board...
[01:48:08] <wb9mjn> Even if its laborious to make....you can put the board back on the machine, and sliver off a little here and there...
[01:48:14] <wb9mjn> to get the design right....
[01:48:19] <fenn> oh i see
[01:49:06] <wb9mjn> The vacumm also holds the puck down to the PCB...
[01:49:16] <tomp> what kind of exposure length from tool to hockey puck guide surface? as much as a thousandth?
[01:49:37] <wb9mjn> I think 1 oz clading is around .0007 inch thick...
[01:50:13] <wb9mjn> So, for 1/2 oz, the cutter exposure is probably .00025 to .00030....to cut away the .00035 clading...
[01:50:47] <tomp> so very little, tenths is the domain for 'touch up work', i couldnt put that board back in to do rework.. too ham fisted :)
[01:51:12] <wb9mjn> The screw Kholze was talking about is what sets that cutter exposure...
[01:51:36] <wb9mjn> There are typically registration holes that fit down on pins...
[01:51:51] <wb9mjn> then a vacumm chuck pulls the board down....
[01:52:14] <tomp> and the puck is like a remora (fish)
[01:52:27] <wb9mjn> Maybe not pins....probably edge guides...
[01:52:34] <wb9mjn> kinda...
[01:54:00] <wb9mjn> Check out
http://www.lpkfusa.com/protomat/index.htm
[01:54:36] <wb9mjn> They do have laser versions...very expensive....the high power laser pulse evaporates the copper....
[01:55:50] <wb9mjn> Here is an RF one...shows the hockey puck...
[01:55:54] <wb9mjn> http://www.lpkfusa.com/protomat/m60.htm
[01:56:39] <wb9mjn> Says its accurate to .0003 inch in the X and Y ...
[01:57:33] <tomp> nice spindle, that a kavo? and is that a rack type tool changer on the rear edge of table? ( got older flash plugin, cant watch )
[01:57:37] <wb9mjn> There is another one, the X60, that is accurate to .00004 inch...
[01:57:50] <wb9mjn> Either a Kavo or an NSK....
[01:58:30] <wb9mjn> The one with the tool changer is for general PCB work, not the real accurate RF stuff....
[02:00:05] <wb9mjn> The RF ones have "semi-automatic tool changing"..I think you drop the tool in from behind, with the screw that sets depth
[02:00:11] <wb9mjn> released...and tight it ...
[02:00:20] <wb9mjn> Then the screw offset is reapplied...
[02:03:58] <tomp> thier 'laser-scalpel' sample part looks pretty clean, would a pcb antenna or transmission line need to be cleaner?
http://www.lpkf.com/products/pcb-cutting-technology/laserscalpel.htm
[02:04:33] <wb9mjn> They have a new one that apparently senses and maps the PCB surface variations...the H100....
[02:05:02] <wb9mjn> X/Y accurate to .00001 inch....uses optical fiduciary mark recognition....
[02:12:48] <tomp> i cant read the pdf (text is garbled), and they're from germany, ???
[02:13:33] <tomp> wow, if i copy the lassooed text to clipboard & paste here.. it's fixed ! like this "and prompts for a tool change at appropriate tim"
[02:19:17] <tomp> hmm, it was kpdf, evince is ok
[02:23:40] <wb9mjn> http://www.lpkf.com/ Yep..."headquartered in Germany"....but worldwide...
[02:25:32] <tomp> oh ,here its, ubu 6.06 dapper cd... i'll be gone for a bit, bye bye
[02:49:22] <cradek> back
[02:49:28] <SWPadnos> front
[02:49:47] <cradek> hmm, the guy with the interesting machine left
[02:49:52] <cradek> hope he gets it going
[02:50:15] <cradek> must have some kind of digital out that controls Z - shouldn't be too hard
[02:50:17] <SWPadnos> hmmm - kholz?
[02:50:45] <cradek> yes
[02:51:01] <SWPadnos> oh - that's where the PCB repair link came from
[02:51:06] <SWPadnos> the laser scalpel
[02:55:58] <cradek> I wish I understood all of spamassassin
[02:57:18] <SWPadnos> I guess you could say I wish I knew anything of spamassassin
[02:57:24] <SWPadnos> other than what it is
[02:58:06] <cradek> for 1/1000 messages (really) it gives a result I don't understand
[02:58:10] <cradek> for the rest it works perfectly
[02:58:16] <cradek> you'd think I would be happy
[02:58:30] <cradek> but that's several a week!
[02:58:33] <SWPadnos> that's only one "nine" (I think) :)
[02:58:55] <cradek> I wouldn't be without it, that's for sure
[02:59:04] <cradek> but for this spam it marked it as in my whitelist
[02:59:14] <cradek> but I can't figure out what whitelist...
[02:59:29] <SWPadnos> is there a log of the rules it applied to various mails?
[02:59:36] <cradek> there's auto whitelisting (this sender tends to send me good mail, so let a rare spammy looking message through)
[02:59:59] <cradek> it applies all rules - the matches are listed in a header, the unexpected match this time is USER_IN_WHITELIST
[03:00:30] <SWPadnos> hmmm - you'd think it could tell you what list or what user it thinks it found
[03:00:59] <cradek> hmm let me look at the long report (duh)
[03:01:16] <SWPadnos> heh
[03:01:21] <cradek> oh, there isn't one, because it's "not" spam
[03:01:27] <cradek> duh again
[03:01:44] <cradek> you don't use spamassassin?
[03:02:02] <SWPadnos> I actually don't use any spam filter - only Mozilla's auto-marking
[03:02:14] <cradek> does that work?
[03:02:33] <cradek> I get hundreds (500?) of spams a day - I would be screwed without it
[03:02:35] <SWPadnos> I think my ISP just added a spam filter, which seemed to work for a while (though it had ~10-15 false positives in 500 or so spams)
[03:02:38] <Jymmm> Yep, you teach it.
[03:02:46] <SWPadnos> I only seem to get 20-50 a day
[03:02:55] <cradek> how long have you had your email address?
[03:02:58] <SWPadnos> maybe more - I have 3 accounts that I check continuously
[03:03:05] <SWPadnos> 5 years maybe
[03:03:09] <Jymmm> 8 years
[03:03:14] <cradek> huh
[03:03:26] <Jymmm> I've had my email address 8 years
[03:03:29] <cradek> I refuse to play games, so I probably get more
[03:03:39] <cradek> my email is on usenet, on the web, in mailing list logs, etc etc.
[03:03:49] <Jymmm> SWPadnos 3?! I got 9
[03:04:10] <cradek> in the dns registry (many times)
[03:04:40] <SWPadnos> well, I have two other active ones that I don't check (yahoo and gmail), plus some other stuff like webmaster at linuxcnc.org
[03:04:51] <SWPadnos> actually, I probably get as many spams for that one as any other
[03:05:22] <SWPadnos> I guess I've only had this account for 4 years
[03:05:40] <SWPadnos> and I have 9568 emails in my spam folder
[03:06:08] <cradek> maybe mozilla's filter does work then...
[03:06:13] <SWPadnos> this address has been all over as well - LKML, several Yahoo groups, it's used for many online registrations ...
[03:06:30] <cradek> SA gives a score, so I can devnull the high scores and keep the barely-spams
[03:06:39] <SWPadnos> it seems good. it misses a few, almost never marks legitimate mail
[03:06:59] <cradek> so I can wade through the barelys if I think I'm missing something
[03:07:07] <cradek> it's never there, but it's peace of mind
[03:07:18] <SWPadnos> I guess my ISP is now using "CanIt"
[03:07:39] <SWPadnos> I guess I should check the spam folder on the server, now that you mention it :)
[03:07:55] <SWPadnos> http://www.engadget.com/2006/12/27/hitachi-does-800-x-480-display-for-phones/
[03:09:02] <cradek> wow!
[03:09:09] <SWPadnos> cool, huh?
[03:09:25] <cradek> hitachi introduces a battery belt-strap at the same time?
[03:09:36] <SWPadnos> I didn't notice that part
[03:09:40] <cradek> heh
[03:09:47] <cradek> I want that for a GPS receiver
[03:10:14] <cradek> my $wh = $self->_check_whitelist_rcvd ($self->{conf}->{whitelist_from_rcvd}, $_);
[03:10:24] <cradek> well I think this is the source code...
[03:10:27] <cradek> * cradek grumbles
[03:11:25] <SWPadnos> ah - here it is:
http://www.gizmodo.com/gadgets/gadgets/sanyo-develops-smallest-hd-display-71-inches-208080.php
[03:11:47] <fenn> :9
[03:12:15] <fenn> i want a mems mirror array head-mount-display
[03:12:24] <fenn> such as used in DLP systems
[03:12:28] <SWPadnos> oh, that's a different page
[03:13:25] <cradek> wow
[03:13:59] <SWPadnos> that one could do for a machine display, I think
[03:14:07] <cradek> I always look forward to future display technology
[03:15:43] <fenn> goldmine-elec sells surplus mems arrays for $3 with no specs
[03:15:55] <fenn> i have no idea where to start looking though
[03:16:03] <SWPadnos> it's the color wheel and stuff that's the trouble
[03:16:17] <fenn> i'm going to use an rgb led
[03:16:24] <fenn> blink blink blink
[03:16:53] <fenn> should be colinear enough
[03:17:14] <fenn> have a 3W luxeon rgb led sitting here
[03:17:19] <cradek> hmm I think lerman's changes will break some very strange existing programs: #[SQRT[9]]
[03:17:42] <SWPadnos> oh yeah - there are function names, aren'tthere?
[03:17:51] <cradek> yeah and I think they can go there - let me check
[03:18:19] <fenn> *cough testsuite cough*
[03:18:30] <cradek> yeah #3=1, g0x#[SQRT[9]] goes to x=1
[03:18:50] <cradek> it's definitely crazy, but it is existing behavior
[03:19:12] <cradek> with his changes this will actually be a syntax error I think.
[03:19:24] <cradek> (param names read up to the first ])
[03:19:34] <SWPadnos> well, just the fact that SQRT might be interpreted as a 0 the first time it's encountered is a problem as well ...
[03:20:25] <SWPadnos> "program scope" is also something that needs further thought (we started on that a little bit)
[03:21:46] <cradek> SWPadnos: I think he didn't consider that #[... is valid and means something today
[03:22:36] <cradek> I put my program on the wiki page
[03:22:59] <SWPadnos> ok - cool
[03:23:25] <cradek> oh no, my name was in some cookie so he'll know it was me :-)
[03:23:50] <SWPadnos> well, it's in his IRC log anyway ... :)
[03:26:05] <cradek> I wonder if my class of program will always be #[...[...]...] which you can actually tell apart from #[....] (no nested brackets)
[03:26:38] <SWPadnos> what if I want to use a named variable inside a bracket equation?
[03:26:54] <SWPadnos> hmmm
[03:27:47] <cradek> ATAN[#[asdf]]/[#[jkl]]
[03:28:12] <cradek> there's special code for these functions, I wonder how much he'll have to mess with
[03:29:35] <SWPadnos> out of curiosity, are there any functions that take more than one parameter?
[03:29:39] <cradek> yes ATAN
[03:29:50] <SWPadnos> that was one parameter - A/B
[03:29:55] <cradek> you have to use a /
[03:30:11] <SWPadnos> you can't do ATAN[0.5]?
[03:30:11] <cradek> no, it's really not, it's just special semantics
[03:30:16] <cradek> no I don't think so
[03:30:19] <SWPadnos> odd
[03:30:28] <SWPadnos> that should be atan2 then, shouldn't it?
[03:30:30] <cradek> it's actually atan2, but it's named atan
[03:30:34] <SWPadnos> ok
[03:31:14] <skunkworks> sure will be easier to read with variable names.
[03:31:27] <SWPadnos> depends on the names ;)
[03:31:35] <SWPadnos> #[i]+#[j] ;)
[03:31:45] <skunkworks> dblfeeddrill
[03:32:09] <skunkworks> hey I grew up on for i = 1 to 100
[03:32:16] <SWPadnos> #[unnamed_constant_parameter_for_determining_feed_rate]
[03:32:39] <cradek> you know he can fix all these problems by using a new character
[03:32:45] <SWPadnos> !
[03:32:46] <cradek> there's a lot of ascii leftover that we're not using
[03:32:49] <cradek> #_i_
[03:32:55] <SWPadnos> !I
[03:33:02] <cradek> $I
[03:33:14] <cradek> no it has to demarcate both sides
[03:33:21] <cradek> I think
[03:33:25] <SWPadnos> ?I?
[03:33:29] <cradek> you have to know when to stop reading the variable name
[03:33:30] <fenn> yuck
[03:33:35] <skunkworks> demarcate? jeez
[03:33:44] <jepler> x$ y e s this is nuts
[03:33:52] <jepler> remember, X1 2 3 means the same as X123
[03:33:54] <SWPadnos> well, use brackets for the demarcation, but a different character before the brackets
[03:34:04] <SWPadnos> like ![varname]
[03:34:12] <SWPadnos> or $[varname]
[03:34:24] <SWPadnos> oe, as ken pointed out: $[v a r na me]
[03:34:38] <jepler> is that the same as [varname] or is it different?
[03:34:48] <SWPadnos> same - the interp strips out all spaces
[03:34:59] <SWPadnos> and converts to lowercase
[03:35:12] <SWPadnos> so $[Var Name] = $[varN a m e]
[03:35:22] <fenn> how about comma?
[03:35:36] <skunkworks> ok - I need sleep. still making it up from the holiday. Night
[03:35:40] <jepler> see you skunkworks
[03:35:46] <cradek> g'night sam
[03:36:26] <cradek> crazy cat won't stop headbutting me
[03:36:34] <jmkasunich> heh
[03:36:36] <cradek> get away from me you freak
[03:36:44] <SWPadnos> heh
[03:36:51] <cradek> she's part rhinoceros
[03:37:17] <cradek> #@VARNAME@ to make it look like a failed autconf
[03:38:07] <cradek> I was going to try something but I have no clue now what it was
[03:38:23] <fenn> what's the problem again?
[03:38:33] <cradek> see the wiki page
[03:38:57] <cradek> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?Named_Parameters
[03:40:17] <SWPadnos> I think you're right that a variable name will be exactly enclosed by a set of brackets
[03:40:27] <SWPadnos> (less some spaces, perhaps)
[03:40:40] <cradek> yes I think that might fix it
[03:41:13] <fenn> why do you need brackets?
[03:41:27] <SWPadnos> SQRT wouldn't be interpreted as a name because it's followed by an open bracket, not a close
[03:41:27] <cradek> to know when the var name ends
[03:42:29] <cradek> fenn: remember whitespace is ignored
[03:42:44] <SWPadnos> better yet - whitespace is removed
[03:43:10] <SWPadnos> the parser gets a sequence of characters with no whitespace present, according to the documentation
[03:43:33] <jepler> everyone should just program in gcl.
http://emergent.unpy.net/index.cgi-files/sandbox/example.gcl
[03:43:58] <jmkasunich> I'm afraid to look
[03:44:10] <cradek> or any other language they want, using AXIS filters
[03:44:24] <SWPadnos> hey - that looks a bit like python ;)
[03:44:28] <jepler> no, it's tcl
[03:44:33] <SWPadnos> oh right
[03:44:34] <jepler> interpreter here:
http://emergent.unpy.net/index.cgi-files/sandbox/gcode.tcl
[03:44:36] <SWPadnos> incr and stuff
[03:44:53] <SWPadnos> I have an idea:
[03:44:55] <fenn> jepler: gcl already means gnu common lisp...
[03:44:59] <jepler> fenn: aw poo
[03:45:05] <SWPadnos> let's not, and say we did :)
[03:45:15] <SWPadnos> there you go - use lisp
[03:45:19] <SWPadnos> RMS would be happy
[03:45:28] <fenn> me too
[03:45:41] <fenn> you guys are practically making lisp using square brackets anyway
[03:45:44] <jepler> (progn ((g0 (cons x 0) (cons y 0))))
[03:45:45] <cradek> nothing wrong with lisp or tcl or python or any other language - all of them have named variables
[03:45:57] <SWPadnos> you'd just have to convince a bunch of machinists to learn something that most programming students can't (or don't want to) fathom
[03:45:58] <cradek> haha
[03:46:02] <jepler> or do I have to write (cons 'x 0) ?
[03:46:16] <fenn> yes
[03:46:19] <SWPadnos> m4, baby!
[03:46:26] <cradek> yes if you want to use (assoc) later
[03:46:34] <cradek> we could rewrite the interpreter in lisp too
[03:48:31] <SWPadnos> that certainly killed off the conversation
[03:49:15] <jepler> if my task was to add advanced features to g-code I would do it in a transformative way, which is what m4 or gcode.tcl do
[03:49:24] <jepler> lisp would be a fine implementation language, however
[03:49:32] <jepler> we don't have any lisp in emc2 yet, after all
[03:50:10] <jepler> goodnight guys
[03:50:19] <cradek> bye
[03:50:24] <SWPadnos> see you
[03:50:39] <jmkasunich> goodnight jeff
[03:51:45] <cradek> I added the #_VAR_ suggestion to the wiki
[03:51:56] <cradek> we didn't agree on a character but I think we all sort of agree on the approach
[03:52:04] <fenn> lets see how many more languages we can add to emc
[03:52:04] <SWPadnos> I don't think it's necessary
[03:52:31] <cradek> you might be right
[03:52:34] <SWPadnos> any text that could be interpreted as a variable name should be interpreted "normally" if it isn't followed by a close bracket
[03:52:36] <cradek> you should add that too
[03:52:40] <SWPadnos> ok
[03:52:41] <fenn> already have c, c++, tcl, python, HAL, .INI, rs274ngc
[03:53:48] <fenn> bash, NML (?), ... any more?
[03:54:41] <SWPadnos> no perl yet?
[03:54:48] <cradek> oh god
[03:54:49] <SWPadnos> how about make and configure
[03:54:59] <cradek> if tcl counts, make sure does
[03:55:43] <fenn> whatever happened to the java code generation tools?
[03:55:58] <cradek> they whithered away
[03:56:25] <cradek> withered?
[03:56:30] <cradek> withered.
[03:57:37] <cradek> it was before my time but I think someone decided porting them to emc2 was more trouble than it was worth
[03:57:40] <tomp> help! i tried to update ubuntu5.01 to 6.06 using the ubuntu 6.06 install cd, now have flat brown screen after logging in.... no desktop, no icons
[03:58:26] <cradek> tomp: I haven't ever done that update - you might want to check the ubuntu wiki or forums
[03:58:44] <jmkasunich> the machine is locked up?
[03:58:53] <jmkasunich> or can you ctrl-alt-f1 to a tty?
[03:59:05] <tomp> uh kinda, i can drop to a terminal
[03:59:16] <tomp> then what?
[03:59:21] <jmkasunich> I dunno
[03:59:41] <jmkasunich> I vaguely recall something like that on one of my installs, but I had a hard lockup with a brown screen
[03:59:43] <SWPadnos> no top or bottom panel either?
[04:00:08] <tomp> k, this another box, so i'll ask at ubuntu : SWP yes, a big brown field, thats all
[04:00:23] <SWPadnos> have you tried rebooting?
[04:00:30] <jmkasunich> my panel mounted PC
[04:00:31] <jmkasunich> http://jmkasunich.dyndns.org/cgi-bin/blosxom/index.html
[04:00:40] <jmkasunich> kinda weird looking ;-)
[04:00:42] <SWPadnos> I'm pretty sure I saw that behavior, and it went away at the second boot (and thereafter)
[04:00:46] <cradek> Results 1 - 10 of about 66,800 for ubuntu dapper update blank brown screen
[04:02:13] <SWPadnos> I guess ctrl-alt-backspace worked ;)
[04:02:49] <cradek> jmkasunich: it looks like it will fly away
[04:03:06] <jmkasunich> ;-)
[04:03:35] <jmkasunich> thats what you get when you take a pile of parts, a bunch of scrap aluminum, and no real plan
[04:03:53] <cradek> hope you don't need any full length ISA cards...
[04:04:06] <jmkasunich> same here
[04:04:35] <jmkasunich> * jmkasunich never checked the length of the 5i20 card....
[04:04:51] <cradek> I think it's big, but not that big
[04:05:37] <SWPadnos> one sec
[04:06:01] <jmkasunich> I have 8.5 inches
[04:06:09] <cradek> I'm going to get to bed early tonight. bye all
[04:06:14] <jmkasunich> goodnight
[04:06:16] <SWPadnos> it's just under 6.75
[04:06:20] <SWPadnos> see you cradek
[04:06:21] <jmkasunich> cool
[04:06:26] <jmkasunich> thanks SWPadnos
[04:06:41] <SWPadnos> full height though - sticks ~1/4" above the bracket
[04:06:56] <jmkasunich> thats ok
[04:07:11] <jmkasunich> unexpected, but ok...
[04:07:15] <SWPadnos> heh
[04:07:20] <SWPadnos> isthat an ISA NIC?
[04:07:40] <jmkasunich> no, pci
[04:07:52] <fenn> is the phone really that yellow?
[04:08:23] <SWPadnos> ok - I remember seeing a card in an ISA slot before - I guess you were just making sure the panel was square
[04:08:48] <jmkasunich> fenn: yess, it was exposed to direct sunlight for a long time next to a window
[04:09:05] <jmkasunich> SWPadnos: yeah, I had 2 nics in there at one time
[04:09:25] <SWPadnos> ah
[04:09:56] <SWPadnos> hmmm - I may need more 50-pin cables
[04:10:04] <SWPadnos> might have to place a digikey order now ;)
[04:10:12] <jmkasunich> I have a bunch of old scsi cables from the compile farm
[04:10:17] <SWPadnos> heh
[04:10:20] <SWPadnos> that should work
[04:10:28] <jmkasunich> surely you can dig up some single ended scsi cables from somewhere?
[04:10:42] <SWPadnos> I've got several, but they're mostly long or have 5 or 6 connectors
[04:10:48] <jmkasunich> ah
[04:10:49] <SWPadnos> though I suppose I could cut those in half
[04:10:57] <jmkasunich> yeah
[04:10:59] <SWPadnos> seems a shame to do that though
[04:11:17] <jmkasunich> that reminds me, I wanted to cut the second connector off the IDE cable
[04:11:22] <jmkasunich> so it would lay neater in there
[04:12:07] <fenn> you can re-use idc connectors if you're really that stingy
[04:12:21] <SWPadnos> I'm not - hence the upcoming digikey order ;)
[04:12:28] <SWPadnos> it's the cable you can't reuse though
[04:12:28] <fenn> didnt think so
[04:12:52] <fenn> sure just dump liquid electrical tape all over it :D
[04:13:05] <SWPadnos> heh
[04:13:22] <SWPadnos> I was looking at twisted pair ribbon cable - man, that stuff is expensive at DK
[04:13:34] <jmkasunich> if its cable you need, I have 200 feet of it
[04:13:44] <jmkasunich> the twisted stuff, actually ;-)
[04:13:49] <SWPadnos> I wish the cards used 40-pin cables, I probably have 100 IDE cables sitting arounf
[04:13:51] <SWPadnos> cool
[04:13:53] <SWPadnos> around
[04:14:06] <jmkasunich> I think its like 16" of twist, then 2" of straight
[04:14:10] <SWPadnos> I'll trade you some cable for some connectors
[04:14:12] <jmkasunich> kinda hard to make a 12" cable
[04:14:15] <jmkasunich> deal
[04:14:19] <SWPadnos> 18" is fine
[04:14:26] <jmkasunich> how much you want?
[04:14:42] <jmkasunich> 10 feet, 20 feet?
[04:14:53] <SWPadnos> well, I'm getting several cards - I'd probably want to make 10 cables or so
[04:14:59] <jmkasunich> 40 feet? ;-)
[04:15:05] <SWPadnos> probably 3 foot cables too, so 30 feet maybe?
[04:15:10] <SWPadnos> ok - 40 is good
[04:15:11] <jmkasunich> no problem
[04:15:15] <SWPadnos> how many connectors do you want?
[04:15:38] <jmkasunich> I dunni
[04:15:41] <jmkasunich> I dunno
[04:15:59] <fenn> what's a spool, 100 feet?
[04:16:24] <SWPadnos> you're getting 2 5i20, so that's 6 cables for all the connectors, plus an extra if you want the 7i31 in there as well
[04:17:16] <jmkasunich> I have 8 cables here already (SCSI), about 2 feet each
[04:17:20] <SWPadnos> plus any for the motor driver etc. maybe 15 connectors?
[04:17:22] <SWPadnos> ok
[04:17:33] <jmkasunich> for the shoptask it might be nice to be able to make custom lengths tho
[04:17:49] <jmkasunich> three custom cables, reuse one end, needs three connectors
[04:17:53] <SWPadnos> you can always take up a few inches with a Z fold
[04:18:27] <jmkasunich> for anything less than the existing cable length I can just cut an put one new connector on
[04:18:32] <jmkasunich> (that is untwisted cable)
[04:18:37] <SWPadnos> true
[04:18:55] <jmkasunich> or with the twisted wire, I can make any multiple of 18", but that needs 2 connectors per cable
[04:19:16] <jmkasunich> IIRC, those connectors aren't cheap, maybe $9-$10 each?
[04:19:46] <fenn> wha?
[04:19:53] <jmkasunich> so I don't want to just stock up at your expense - the cable didn't cost me anything
[04:19:59] <SWPadnos> that's what I'm seeing.
[04:20:14] <SWPadnos> I think I found some $2.50 - $3 ones though
[04:20:30] <jmkasunich> oh, thats a lot nicer
[04:20:36] <fenn> are they gold plated or something?
[04:20:40] <SWPadnos> of course
[04:20:46] <fenn> maybe solid gold
[04:20:48] <SWPadnos> the contacts anyway
[04:21:02] <jmkasunich> the really good ones are gold plated all over
[04:21:28] <jmkasunich> they keep getting rejected though, something about short circuits
[04:21:29] <SWPadnos> I don't think these can have a second strain relief though
[04:21:32] <SWPadnos> heh
[04:22:20] <jmkasunich> looks like twisted pair 50conductor is about $1.30 per foot
[04:22:27] <jmkasunich> so 40 feet is about $50
[04:22:43] <SWPadnos> I'm not sure I saw it for that little :)
[04:22:56] <jmkasunich> pick me out $20-25 worth of connectors, that way we split the dumpster bonus pretty evenly
[04:23:09] <SWPadnos> ok - sounds good to me. thanks
[04:23:09] <jmkasunich> $130 for 100 feet at digikey
[04:23:17] <SWPadnos> do you want the strain relief if possible?
[04:23:18] <SWPadnos> ok
[04:23:40] <jmkasunich> get me whatever kind you get
[04:23:50] <SWPadnos> ok
[04:23:59] <jmkasunich> are you in a hurry?
[04:24:02] <SWPadnos> I just need to see what the difference between these two is
[04:24:07] <jmkasunich> we leave tomorrow for 5 days at relatives
[04:24:36] <Jymmm> jmkasunich you poor bastard!
[04:24:44] <SWPadnos> have fun
[04:25:03] <Jymmm> whats the 25p for?
[04:25:30] <Jymmm> whats the 25p cable for?
[04:25:30] <jmkasunich> 5i20 I/O uses 50 pin connectors
[04:25:35] <Jymmm> ah ok
[04:25:52] <Jymmm> and old scsi cables won't cut it?
[04:26:05] <SWPadnos> ok, they're $1.60 in 25 qty, so you should get 15 of them for $25-ish
[04:26:17] <jmkasunich> beauty!
[04:26:42] <SWPadnos> strange - the ones with strain relief are less expensive than the ones without
[04:27:22] <SWPadnos> 609-1745-ND, if you want to take a look
[04:28:56] <SWPadnos> hmmm - these are cool: A3059-ND
[04:29:15] <SWPadnos> good for the other end (the output of the various modules that have 50-pin output connectors)
[04:29:22] <SWPadnos> http://rocky.digikey.com/WebLib/Amp/Web%20Photo/New%20Photos/2-102393-3,%202-102398-3,%202-102448-3.jpg
[04:29:57] <jmkasunich> you get a bonus - I measured out 30 segments of cable, then I measured a segment, and its actually 20"
[04:30:01] <jmkasunich> so thats 50 feet
[04:30:09] <SWPadnos> damn :)
[04:30:28] <jmkasunich> and I still have a box and a half
[04:31:51] <SWPadnos> hmmm - that last one isn't so interesting any more. it looked a bit like screw terminals in the photo
[04:32:32] <SWPadnos> actually, I had a question for you about inrush limiters
[04:32:43] <jmkasunich> "Assmann" cable? ;-)
[04:32:49] <SWPadnos> heh - and connectors ;)
[04:32:57] <SWPadnos> plugs and sockets
[04:33:42] <jmkasunich> feel free to hold my mesa order and toss the connectors in there
[04:33:59] <jmkasunich> I'll ship your cable next week when we get back - its setting on my bench so I don't forget
[04:34:07] <SWPadnos> ok. no problem
[04:34:26] <jmkasunich> now - inrush
[04:34:42] <SWPadnos> I was looking at PTCs with a few ohms "cold" resistance - do you have an opinion on how to choose the appropriate resistance?
[04:35:15] <jmkasunich> assume the capacitors are a short circuit, and choose the resistance so the inrush current doesn't scare you
[04:35:39] <SWPadnos> the transformer is ~2kVA, and runs on 120, so up to ~20A continuous
[04:35:51] <jmkasunich> inrush of 50-100% of the full load running current is reasonable
[04:36:01] <SWPadnos> ok - that low
[04:36:12] <SWPadnos> I was thinking that it could be a bit higher without too much trouble
[04:36:17] <SWPadnos> like 30A or so
[04:36:33] <SWPadnos> even so, two 5 ohm PTCs would give me 12A
[04:36:40] <jmkasunich> probalby can, depends on your switching devices (relay, breaker, whatever you're gonna turn it on with)
[04:36:45] <SWPadnos> sure
[04:36:59] <jmkasunich> remember, I deal with full load currents of 500-1200A, so 50% is still a lot for me
[04:37:03] <SWPadnos> I'm not positive of what I'll use there
[04:37:05] <SWPadnos> heh
[04:37:10] <jmkasunich> gotta recalibrate the brain for lower power
[04:37:36] <SWPadnos> I was also trying to get something with low "hot" resistance as well - they range from ~30 mOhm to ~0.2 ohm
[04:37:38] <SWPadnos> heh
[04:38:00] <jmkasunich> I have no firsthand experience with PTCs
[04:38:13] <jmkasunich> I don't think we use them in any current drives, at least not over a few HP
[04:38:19] <jmkasunich> the low end products might still use them
[04:38:28] <SWPadnos> I suspect a couple of watts is needed to keep it hot anyway, so I'm not too worried about the hot resistance
[04:38:40] <jmkasunich> low end = cheaper, and from 0.5 to 5 HP
[04:38:52] <SWPadnos> interesting. do you use zero-cross switches for the input transformers, or do you just not have input transformers?
[04:39:01] <SWPadnos> (or something else)
[04:39:19] <jmkasunich> no VFDs that I'm aware of use transformers
[04:39:26] <SWPadnos> ah - ok
[04:39:29] <jmkasunich> the inrush is the caps
[04:39:32] <SWPadnos> ok
[04:40:12] <jmkasunich> FWIW, I'm almost certainly gonna use resistor/relay for my project
[04:40:22] <jmkasunich> two relays really
[04:40:43] <jmkasunich> use resistors large enough to limit the charging current to a couple amps
[04:40:54] <jmkasunich> and use a smallish relay (KUP, etc) for the charging phase
[04:41:06] <jmkasunich> then bypass with a contactor rated 15-20A
[04:41:11] <jmkasunich> on estop, open both
[04:41:55] <SWPadnos> that seems a bit complex, but may be less susceptible to reboot problems when hot
[04:42:17] <jmkasunich> its probably overkill, thats what I do it seems
[04:42:27] <SWPadnos> well, I'm in that boat as well, most of the time
[04:42:40] <jmkasunich> but a 0.2 ohm hot resistance at 20A is 80 watts!
[04:43:07] <SWPadnos> I was looking at ones with ~0.08 ohm on resistance
[04:43:09] <jmkasunich> 30mOhm is better, 12 watts
[04:43:36] <SWPadnos> well, I don't think the servos will be using 2 kVA very often :)
[04:43:40] <jmkasunich> true
[04:43:51] <SWPadnos> (though each could use 6KW in short bursts)
[04:43:59] <jmkasunich> and its I^2, so at half load it gets pretty reasonable
[04:44:03] <SWPadnos> yep
[04:44:09] <Jymmm> 1.2 MegaWatts Flux capacitor
[04:44:19] <jmkasunich> I probalby should seriously consider ptcs myself
[04:44:20] <SWPadnos> that's 1.21 gigawatts - get it right! :)
[04:44:31] <jmkasunich> eliminates a relay
[04:44:34] <Jymmm> SWPadnos heh
[04:44:51] <jmkasunich> are the ones you are looking at from DK, or elsewhere?
[04:44:57] <SWPadnos> digikey
[04:45:12] <SWPadnos> I searched for inrush
[04:45:21] <jmkasunich> so did I
[04:45:48] <SWPadnos> ametherm had the relatively reasonably priced ones
[04:46:45] <SWPadnos> 570-1029-ND
[04:48:14] <jmkasunich> 25A?
[04:48:25] <SWPadnos> slight margin
[04:48:50] <SWPadnos> hmmm - my 50A rectifier may not have enough margin
[04:49:06] <jmkasunich> depends - check the datasheet
[04:49:14] <jmkasunich> lots of rectifiers are rated for _average_ current
[04:49:20] <SWPadnos> true
[04:49:24] <jmkasunich> two in a bridge means 12.5A average each
[04:49:37] <jmkasunich> CT bridge that is
[04:50:17] <jmkasunich> "body temperature at maximum current 228C"
[04:50:23] <SWPadnos> heh
[04:50:30] <SWPadnos> I had noticed some of those
[04:50:37] <jmkasunich> don't use eutectic solder
[04:50:50] <jmkasunich> thermakl time constant 232 seconds
[04:50:55] <jmkasunich> don't power cycle too fast
[04:51:13] <SWPadnos> I don't see that on the DK datasheet
[04:51:31] <SWPadnos> though it seemed that the various datasheets had different sets of information
[04:51:35] <jmkasunich> I looked up a 15A 10 ohm one
[04:51:51] <jmkasunich> unfortunately I closed the window, don't know the P/N
[04:52:00] <SWPadnos> ok
[04:52:31] <Jymmm> SWPadnos grab me before you leave.
[04:52:36] <SWPadnos> ewwww
[04:52:50] <Jymmm> you're married, shush
[04:52:55] <SWPadnos> oh, right
[04:53:04] <tomp> hello, i got a brown screen after login, i added a new user, he gets same thing, I got lots of ubuntu cd's. can i repair with a cd?
[04:53:31] <fenn> i doubt you really need an 80 watt rated resistor
[04:53:48] <jmkasunich> fenn: for inrush?
[04:53:50] <fenn> its only for a fraction of a second
[04:54:00] <fenn> yes
[04:54:11] <jmkasunich> the PTCs dissipate non-trivial power at full load, all the time
[04:54:39] <jmkasunich> they disspate a couple kilowatts for a fraction of a second at power up
[04:54:41] <SWPadnos> at 20A continuous, and 0.2ohms when hot, that PTC would dissipate 20 * 20 * 0.2, or 80W
[04:56:49] <SWPadnos> ok, a single half-cycle surge of 400A is allowed for this rectifier
[04:58:26] <jmkasunich> the switch contacts (relay, whatever) are more likely to be the limiting factor
[04:58:33] <jmkasunich> that or flickering the lights in the house
[04:59:02] <SWPadnos> my transformer is a dual-secondary at 44V/22A each, so the total of 44A is close to the rated 50
[04:59:10] <SWPadnos> hence my decision to check it :)
[04:59:22] <jmkasunich> oops, I forgot about the transformer
[04:59:26] <SWPadnos> or maybe it's 46V 22A - don't remember
[04:59:31] <jmkasunich> I was thinking 20A diode current
[04:59:36] <SWPadnos> yeah - this is on the low side :)
[05:00:26] <tomp> is there an apt-get command to reinstall? something like apt-get dist-keepold apt-get upgrade ???
[05:00:40] <SWPadnos> keepold? I don't think so
[05:00:58] <tomp> no, like that idea, i know that literal doesnt exist
[05:01:03] <SWPadnos> more like apt-get dist-upgrade --reinstall (or something similar)
[05:01:19] <SWPadnos> old settings would be kept where possible by default, I think
[05:01:38] <SWPadnos> I'm not sure if reinstall is valid for dist-upgrade though
[05:14:51] <jmkasunich> quitting time here - still gotta copy a rebate form for this monitor (if I don't do it now, I'll forget when we get back) and pack
[05:15:03] <jmkasunich> goodnight all
[05:15:04] <SWPadnos> heh. enjoy your trip
[05:15:19] <jmkasunich> 9 hour drive tomorrow :-(
[05:15:24] <SWPadnos> ewww
[05:15:38] <SWPadnos> that's a time for cellular internet and a wife ;)
[05:15:49] <Jymmm> lol
[05:17:51] <SWPadnos> so Jymmm what's up?
[05:49:40] <tomp-tag> hello, i fixed the login problem (apt-get install ububtu-desktop) now how do i upgrade from ubu 5 to ubu 6.06 ?
[05:49:46] <SWPadnos> would the real tomp please step forward? :)
[05:50:00] <tomp-tag> uh , are there 2 of me again?
[05:50:03] <SWPadnos> is it not 6.06 now?
[05:50:10] <SWPadnos> tomp and tomp-tag
[05:50:30] <tomp-tag> oh yeah I got another system on, i go kill me now
[05:50:36] <SWPadnos> no need
[05:50:46] <SWPadnos> I'm her etwice as well
[05:51:32] <tomp-tag> 'ello, no i have breezy now & want dapper for this box
[05:52:05] <SWPadnos> you should be able to upgrade by running synaptic - it should tell you that there are new versions (though it may want to move you straight to edgy)
[05:53:57] <tomp-tag> is here a way to use the dapper cds that just upgrades the system ( no partitioning, keeping most of old file system) ?
[05:54:36] <SWPadnos> it's possible that you'll be prompted to upgrade when you insert the CD
[05:54:58] <tomp-tag> lets see
[05:55:00] <SWPadnos> breezy may not have done that though
[05:55:57] <SWPadnos> http://ubuntuguide.org/wiki/Ubuntu#How_to_upgrade_from_Breezy_Badger_to_Dapper_Drake_.28experimental.29
[05:56:08] <SWPadnos> that may help, or it may completely trash your system ;)
[05:56:28] <SWPadnos> you may be able to do it with apt-cdrom add, then apt-get upgrade
[05:56:53] <tomp-tag> thanks ( the install vrsn of dapper didnt askif i wanted to u/g) going to the url you sent, just lookin
[05:57:17] <SWPadnos> isn't there an "upgrade" option when you boot from the dapper CD>
[05:57:19] <SWPadnos> ?
[05:57:32] <SWPadnos> (not that I know it works well, mind you)
[05:57:38] <tomp-tag> btw: i mailed usps M.O. today for the group buy, really i did, really :)
[05:57:45] <SWPadnos> ok - thanks :_)
[05:57:47] <SWPadnos> :)
[05:58:28] <SWPadnos> need anything from digikey? :)
[05:59:02] <tomp-tag> oh, the trick with the repositories, i read that in other places, but with a lot more bravado...
[05:59:21] <tomp-tag> digikey?, not today
[05:59:26] <SWPadnos> well, that's pretty easy to do (it's basically a s/breezy/dapper/g on the sources.list
[06:00:23] <tomp-tag> yes the guys said that was the trick... I'm sure i put a lot of additional crap lines in there, trying to get LiVEs etc to work
[06:08:50] <tomp-tag> doin it now..., so closing all apps down, thanks & good nite
[06:20:54] <Jymmm> SWPadnos Hey
[06:22:13] <Jymmm> SWPadnos: Do you know of a way to emulate a kybd from a computer? Meaning sending keystrokes like in a terminal to another computer via PS/2
[06:26:42] <SWPadnos> not offhand, except maybe by bit-twiddling manually (similar to some I2C drivers)
[06:27:16] <Jymmm> Ah, nah I was looking for something fairly simple.
[06:27:22] <SWPadnos> something like connecting parallel port pins to the PS2 connector on the other machine
[06:27:30] <Jymmm> really?
[06:27:39] <SWPadnos> there may be devices that can convert serial to PS/2, for scanners and the like
[06:27:58] <SWPadnos> with a parport, you'd have to write drivers, and I'm not sure how to deal with clocks and stuff
[06:28:47] <Jymmm> Hmmmm, ok. I do have a PS/2 barcode reader, maybe it be easier to just encode it and print it out.
[06:29:01] <SWPadnos> well, there used to be the "keyboard wedge"
[06:29:08] <Jymmm> http://www.edbydesign.com/books/B000A3XRSO.html
[06:29:09] <SWPadnos> used with serial barcode readers
[06:29:18] <Jymmm> Yeah, mine is PS/2
[06:29:28] <SWPadnos> ah
[06:29:39] <Jymmm> That DVd player takes a PS/2 kybd to input the content of each disc.
[06:29:54] <SWPadnos> yep - seen those, didn't buy partly for that reason :)
[06:30:24] <Jymmm> Well 400 DVD jukebox for $290 isn't too shabby and it has a line doubler built in
[06:30:37] <SWPadnos> I wonder if this would work:
http://www.vetra.com/327text.html
[06:30:50] <SWPadnos> they say for mice, but then again, who's to know?
[06:31:53] <Jymmm> $54
[06:32:10] <SWPadnos> that's less than the $149 one that just converts serial -> PS/2
[06:32:11] <Jymmm> Eh, barcode font and paper might be cheaper =)
[06:32:29] <SWPadnos> might not be, depending on the printer ...
[06:32:39] <Jymmm> Work printer of course =)
[06:32:44] <SWPadnos> thief
[06:33:11] <Jymmm> Hey, my job description says the endorse empoyeees working on their own projects.
[06:33:42] <SWPadnos> I wonder if the little adapter from an old serial/PS/2 mouse would help
[06:34:51] <Jymmm> No idea, we're talking serial mouse protocol, not so much terminal which is what I'm looking for
[06:35:56] <SWPadnos> it's got a serial port on one side and ps/2 onthe other - I'm not sure what's in between
[06:37:08] <Jymmm> Yeah, I still have a few of those adpaters around, but I know that PS/2 protocol is a three bytes, ane serial is one byte packet
[06:37:47] <SWPadnos> well, you can send 3 bytes at a time ...
[06:38:16] <SWPadnos> I don't know how much of that converter was software in the mouse and how much is hardware in the dongle though
[06:38:19] <Jymmm> Since serial is almost dead, I was thinking USB to PS/2 sorta thing.
[06:38:36] <SWPadnos> oh sure - there are plenty of those
[06:38:56] <Jymmm> I've come across a few times where it be nice to be able to send out command to another machine via kybd
[06:39:35] <A-L-P-H-A> where's the change log for emc?
[06:39:35] <Jymmm> Like when one doens't have a fdd to boot from. just dump an image file to the PS/2 port
[06:40:24] <SWPadnos> one thing you have to be careful of is grounding. USB and PS/2 connectors have direct connections to ground and 5V
[06:40:36] <SWPadnos> it isn't expected that you'll connect a "foreign" power supply there ...
[06:41:07] <Jymmm> ah, well the usb was just a iface to a built in uart sorta thing.
[06:41:21] <Jymmm> or whatever PS/2 uses that is.
[06:41:22] <SWPadnos> if you have some more modern mice, you could try one of the USB/PS2 adapters that come with them
[06:41:57] <SWPadnos> like this one:
http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=1220387&CatId=469
[06:43:28] <Jymmm> We have a few of those around here, but there is no way to send from a terminalto the usb
[06:44:15] <SWPadnos> http://www.beyondlogic.org/keyboard/keybrd.htm
[06:45:21] <Jymmm> Yeah, but the reverse of that.
[06:45:34] <SWPadnos> sure, but it tells you all about the protocol
[06:46:50] <Jymmm> That's assuming that I wanted to learn that much.... didn't I just say I'm goin for my CCNA ? LOL
[06:47:02] <SWPadnos> wuss ;)
[06:47:07] <Jymmm> Damn Skippy!
[06:47:54] <Jymmm> SWPadnos But.... what I'm thinking is that this jukebox might have IDE interface internally, thaen what I need is a way to change the disc from a computer. Now THAT I'm willing to learn.
[06:48:21] <SWPadnos> I'd be surprised if it were IDE internally
[06:48:47] <Jymmm> What not, Sony makes IDE CDROMS, abd this is a DVD afterall...
[06:48:50] <SWPadnos> it's more likely to be a datastream direct into a decoder
[06:49:04] <SWPadnos> well, here's hoping, but I'd still be surprised
[06:49:41] <Jymmm> If the slight chance is is IDE, it be an AWESOME hack.... 400 disc drive library on a NAS box
[06:49:50] <Jymmm> for $300
[06:49:57] <SWPadnos> that would be very nice
[06:50:55] <Jymmm> I really want to see the inside of this thing... maybe I'll "rent" one for a ocupe of days
[06:51:14] <SWPadnos> hey - now these are cool for emc:
http://www.vetra.com/Encoder2.htm
[06:51:51] <Jymmm> SWPadnos hold on, I got something better....
[06:52:17] <Jymmm> http://www.ultimarc.com/ipac1.html
[06:52:36] <SWPadnos> ah yes - the gaming stuff
[06:53:06] <Jymmm> and still have a passthur as well
[06:53:37] <Jymmm> compeltely programmable and static too
[06:54:41] <SWPadnos> so you could get 3 or 4 of those and a Mesa 5i20, and make a HAL config that takes data in and "presses" the external buttons in the correct order :)
[06:55:18] <Jymmm> I think the ipac has macros
[06:56:09] <SWPadnos> well, this is exactly what you want, but probably 10 times the cost:
http://www.vetra.com/335text.html
[06:56:24] <SWPadnos> $99 for the OEM PC board version
[06:57:37] <Jymmm> barcode wedge it is!
[06:57:47] <Jymmm> brb
[06:57:56] <SWPadnos> I'm headed for bed. see you later
[07:02:35] <Jymmm> hasta
[08:48:36] <aip_tom> cheaper is the keywiz, groovygamegear.com has USB device for $21.
[08:50:08] <aip_tom> I've got one showing up tomorrow for a project
[09:46:06] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[12:35:37] <Guest786> heloo
[12:35:58] <Guest786> please i need patch rtai for kenel 2.4.33
[12:50:10] <Guest786> please i need patch rtai for kenel 2.4.33
[13:25:13] <Guest786> please i need patch rtai for kenel 2.4.33
[13:32:38] <jepler> l~N
[13:32:39] <jepler> oops
[13:32:41] <jepler> good morning :)
[13:34:23] <skunkworks> jepler: good morning ;)
[13:37:29] <skunkworks> So to play with my pluto on my portable - I am thinking of using power from the usb port
[13:40:32] <skunkworks> jepler: did you see your post on fpga4fun..
[13:42:43] <jepler> skunkworks: I saw there were some replies
[13:42:49] <jepler> I should write and say it wasn't ruined after all
[13:43:20] <skunkworks> 1 unit load (100ma) from the usb should be enough to power the pluto and a few led's
[13:43:22] <jepler> skunkworks: as for power -- yeah I've been thinking of the USB port too.
[13:44:32] <jepler> I don't know how much power the acex chip is consuming. obviously when it was getting hot for me, it was consuming a lot..
[13:45:01] <jepler> I was thinking of building one of these between the USB and the Pluto's power input:
http://www.uchobby.com/index.php/2006/11/12/current-limited-pc-power/
[13:46:02] <jepler> to not burn up in current limit mode, T1 needs to be bigger than a TO-92 -- it'll dissipate .9W or so in the short-circuit condition
[13:46:44] <jepler> oh, and R1 is 3.6 ohm, not 36 ohm
[13:47:28] <jepler> I haven't gotten around to doing anything about it, though
[13:48:22] <jepler> * jepler posts a reply to that effect
[13:50:09] <skunkworks> jepler: your going to have a diode drop + what ever voltage across the resister. would that be too low for the pluto?
[13:51:51] <jepler> the dropout voltage of the lm1117 is 800mV, and its output voltage is 3.3V. so anything above 4.1V at the input should be fine
[13:52:06] <jepler> (and that's the dropout at lm1117's top current, 800mA)
[13:52:41] <jepler> er -- oops
[13:52:43] <jepler> the dropout is 1.2V
[13:52:48] <jepler> more like 1.1V at 100mA
[13:52:56] <jepler> so make that 4.4V instead
[13:53:59] <jepler> to sum up: I dunno for sure
[13:54:03] <skunkworks> :)
[13:54:21] <Guest786> please i need patch rtai for kenel 2.4.33
[13:55:09] <Guest786> please
[14:20:13] <skunkworks> Guest786: I don't know -
http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?RtaiSteps
[14:20:22] <skunkworks> ^but that looks promising
[15:37:02] <Guest786> please i need patch rtai for kenel 2.4.33
[15:54:13] <anonimasu> hm..
[15:55:36] <anonimasu> https://www.rtai.org/
[15:57:07] <anonimasu> Guest786: there's a download section
[15:57:12] <jepler> if you're going to promote
https:// as the only way to browse your website, you probably want to go to the trouble of getting a certificate that your users' browsers will accept without complaint
[15:57:57] <anonimasu> hehe
[15:58:13] <anonimasu> jepler: oh, but now that costs money and isnt free as in beer is it?
[15:58:20] <anonimasu> *sarcasm
[15:58:20] <anonimasu> *
[15:58:52] <anonimasu> im pondeing hooking up adaptive feed override to my mill.
[15:59:57] <anonimasu> I wonder how well it'd work..
[16:00:04] <anonimasu> has anyone done it?
[16:00:32] <jepler> mdynac is using it for his edm
[16:00:49] <anonimasu> with a mill?
[16:01:16] <anonimasu> jepler: how well does it work?
[16:01:24] <jepler> I think it works just fine
[16:02:06] <jepler> as long as you're comfortable building emc2 from CVS -- it's not in 2.0.x, you have to get pre-2.1.0.
[16:02:11] <anonimasu> oh
[16:02:14] <anonimasu> I'll do it on the plc..
[16:03:17] <anonimasu> as that's where I get spindle load information from
[16:04:13] <jepler> you would scale the value to the range 0..1 and hook that to the HAL pin motion.adaptive-feed
[16:04:28] <jepler> also you must enable adaptive feed with G50 P1 or something like that
[16:04:41] <anonimasu> how much does adaptive feed tweak my feeds?
[16:04:50] <jepler> 0% to 100%
[16:04:57] <anonimasu> that's scary :)
[16:05:27] <jepler> so if you program F10, leave the "feed override" in the GUI at 100%, and put the value 0.5 on motion.adaptive-feed, the resulting feed rate will be 5 per minute
[16:06:04] <anonimasu> ah ok
[16:06:25] <jepler> if you use both the GUI feed override and motion.adaptive-feed, they're multiplied together. 80% FO and motion.adaptive-feed = 0.5 would give you 4 per minute
[16:07:03] <anonimasu> dunno, I'll probably roll my own stuff.. +/- a % I can set in the plc..
[16:07:24] <anonimasu> I'm running a older version of head right now
[16:08:44] <anonimasu> and well it works great so im reluctant to upgrading/changing my setup for now :)
[16:09:07] <jepler> but at least you're familiar with compiling from CVS
[16:09:10] <anonimasu> yes
[16:09:44] <anonimasu> I havent machined since I got a real keypad, but it seems like what I've been missing for setting up parts and stuff easily
[16:09:51] <SWPadnos> you could always run the input from your spindle load sensor through a limiter block in HAL, so it never brings the feedrate below X percent
[16:10:18] <anonimasu> I were thinking of giving the plc the choice of slowing or feeding +/- a set percentage max..
[16:10:58] <anonimasu> as I can do nml commands from it..
[16:11:10] <SWPadnos> FO is not an NML command
[16:11:12] <jepler> if you want one of two values depending on a bit from the PLC, then you could use a mux2 -- e.g., mux2.0.in0 = .8, mux2.0.in1 = 1.0, mux2.0.sel <= some-plc-pin
[16:11:17] <anonimasu> I an set the feed override..
[16:11:22] <SWPadnos> err - AF is not, FO is :)
[16:11:38] <anonimasu> jepler: it's a real plc..
[16:11:56] <anonimasu> but I guess I could make it connect to hal..
[16:11:58] <anonimasu> :)
[16:12:00] <anonimasu> that'd be cool
[16:13:08] <jepler> if the portion in the PC is in C or Python, you'll find it is pretty easy to make it into a HAL component if you decide to do it
[16:13:14] <anonimasu> it's python
[16:13:24] <anonimasu> :)
[16:13:38] <jepler> cool -- python's my favorite language
[16:13:44] <jepler> what plc is this?
[16:13:48] <anonimasu> www.sigmatek.at
[16:13:52] <anonimasu> stuff we use at work..
[16:14:21] <anonimasu> I've got a spare plc we ran a prototype with 2 years ago
[16:14:46] <anonimasu> jepler: with a real keypad :)
[16:15:10] <jepler> it looks like expensive stuff
[16:15:45] <anonimasu> yeah, though there isnt much options for what we do with them... siemens/other brands dosent do the stuff we need
[16:16:02] <anonimasu> custom hardware would work, but there's a major effort with that also :)
[16:17:00] <anonimasu> free programming software and free support..
[16:18:16] <jepler> so you're using the python "emc" module to send NML messages based on input from the PLC?
[16:18:26] <anonimasu> jepler: yes
[16:18:38] <anonimasu> via the serial port..
[16:18:50] <jepler> if you do a write-up of this, I'd like to link to it from the axis website or linuxcnc wiki
[16:18:56] <anonimasu> sure..
[16:19:32] <anonimasu> I wish python would run on the plc ;)
[16:20:14] <anonimasu> im only doing it one way for now, in extension i'll be doing bidirectional communication off hal or something
[16:20:26] <anonimasu> for i/o
[16:20:37] <anonimasu> but there's so much work with the plc code to do that :)
[16:20:58] <anonimasu> jepler: when my software works as I want it I can release it for other's use..
[16:21:06] <anonimasu> it's hardcoded commands right now..
[16:21:21] <jepler> even if not many people would have this specific plc, it sounds like a good demonstration of extending emc with pypython
[16:21:33] <anonimasu> I were thinking that the same thing can be done with a avr..
[16:21:50] <anonimasu> or anything that can send serial commands
[16:22:08] <anonimasu> yeah :)
[16:22:15] <anonimasu> I'm a python fan nowdays
[16:22:27] <anonimasu> it made a 300 line program into a 40 line one..
[16:22:59] <jepler> I know you don't like the visual ladder logic, but somewhere I ran into a ladder logic compiler for microcontrollers
[16:23:18] <anonimasu> :)
[16:24:03] <anonimasu> I bet there's rs232 panels..
[16:24:16] <jepler> it would be neat to have something like an avr with a PLC-like interpreter .. program it with your specific logic when emc starts, transfer data bidirectionally to HAL
[16:24:30] <anonimasu> yeah
[16:24:31] <anonimasu> really nice
[16:24:44] <anonimasu> there wouldnt be too much work with it..
[16:24:53] <anonimasu> well the plc interpreter..
[16:25:02] <anonimasu> but bacically a p code interpreter would do..
[16:25:10] <anonimasu> though, the hard part is multitasking and such..
[16:25:18] <anonimasu> that's not fun on avr's :)
[16:25:33] <jepler> http://cq.cx/ladder.pl
[16:25:50] <jepler> he has his own bytecode, not sure exactly what it's like
[16:25:55] <anonimasu> cool stuff :)
[16:26:37] <jepler> I usually work with smaller AVRs (2k flash, 128 byte RAM) so stuff like multitasking doesn't come up
[16:27:05] <anonimasu> yeah :)
[16:27:55] <anonimasu> it's still a neat thing to do
[16:28:07] <anonimasu> plc/ladde on a chip
[16:28:16] <anonimasu> for the non technical people out there
[16:28:32] <Lerneaen_Hydra> 'lo
[16:28:34] <anonimasu> hey lh
[16:28:35] <jepler> hi Lerneaen_Hydra
[16:28:45] <anonimasu> I'll be right back.. need to check up on the dog :)
[16:28:55] <anonimasu> parents are babysitting for a while
[16:29:11] <Lerneaen_Hydra> having a nice, uh, what's it called.... hmm, the time between christmas and new years?
[16:30:51] <jepler> something like atmega16 might be a good choice -- use 3/4 of so of the RAM for the logic, 1/4 for the data. the interpreter will fit nicely in the mega8's flash. as long as you can put up with update rates in the kHz range, that is
[16:31:13] <jepler> (emc's classicladder uses a hard-coded 100ms )
[16:35:00] <Lerneaen_Hydra> whoo atmel!
[16:46:21] <jepler> can anyone understand what this person is asking for?
[16:46:21] <jepler> > hello?
[16:46:22] <jepler> > i not know step clk(0~10khz)..
[16:46:22] <jepler> > step input(int0) to pwm convertion?
[16:46:22] <jepler> > thank you.
[16:47:03] <cradek> sorry, no
[16:49:46] <Lerneaen_Hydra> you people always seem to get the most understandable questions. O.o
[16:50:05] <dnisbet2> Hi
[16:50:06] <dnisbet2> Hi
[16:50:12] <dnisbet2> Does this work ???
[16:51:17] <SWPadnos> no
[16:51:32] <jepler> dnisbet2: we can see what you are typing
[16:51:32] <dnisbet2> That's good then
[16:52:34] <dnisbet2> Does anybody know how I change the direction of my Stepper motors using the HAL commands
[16:52:48] <dnisbet2> I have things running, but the wrong ways
[16:53:01] <cradek> a lot of people have answered your question on the mailing list
[16:53:19] <cradek> the answer is put a - sign in your INPUT_SCALE
[16:53:49] <dnisbet2> thanks, I'll go and try it
[16:53:56] <cradek> welcome
[16:54:50] <skunkworks> dnisbet2: did you also post this on the emc users mail list?
[16:55:17] <cradek> yes it's his question we both answered
[16:55:22] <cradek> at the same time
[16:55:34] <skunkworks> :)
[16:55:36] <cradek> and with the same answer! it must be right then
[16:56:31] <skunkworks> I had an example ;)
[16:56:34] <dnisbet2> I'll let you both know in few minutes if your right :-)
[16:56:48] <cradek> skunkworks: yes your answer was better
[16:56:51] <cradek> dnisbet2: we are :-)
[16:57:02] <skunkworks> I was going to say that
[16:57:54] <cradek> dnisbet2: this is a common question. can you tell us where in the docs or on the web you looked for the answer, so we can put it there for the next guy to find?
[17:01:50] <skunkworks> This morning I was actually thinking maybe there should be a comment after it #to reverse axis movement - change the scale to a negative - or something like that.
[17:02:18] <skunkworks> the person would most likely be in the ini to change the scale atlest
[17:02:49] <skunkworks> (at the end of the input_scale line)
[17:09:41] <etla> hi all
[17:09:52] <SWPadnos> hiya Anders
[17:10:19] <etla> anyone remember the formula for calculating the push force with a ballscrew
[17:10:37] <etla> assuming I know the torque on the screw, diameter and pitch and assume some efficiency
[17:11:02] <Lerneaen_Hydra> hi etla
[17:11:11] <SWPadnos> efficiency * torque * 2 * pi / pitch
[17:11:13] <SWPadnos> I think
[17:11:14] <Lerneaen_Hydra> shouldn't be too hard
[17:11:37] <SWPadnos> oops also the diameter should be in the denominator
[17:11:47] <SWPadnos> efficiency * torque * 2 * pi / (pitch * diameter)
[17:12:14] <Lerneaen_Hydra> hmm, why would diameter affect it?
[17:12:35] <SWPadnos> also, pitch may have to move to the numerator depending on how it's specified (ie, mm/rev or revs/inch ...)
[17:12:42] <SWPadnos> durned if I know :)
[17:12:56] <Lerneaen_Hydra> or rather, why should diameter even be there
[17:13:07] <etla> I found a webpage about this, it also talks about a friction coefficient that lowers the force
[17:13:17] <etla> thanks anyway, I'll try to calculate something...
[17:13:18] <SWPadnos> you need to divide torque by some length to get force
[17:13:54] <Lerneaen_Hydra> yeah
[17:14:04] <Lerneaen_Hydra> pitch would be the length, right?
[17:14:34] <SWPadnos> pitch is the distance travelled for one turn of the srew (I think - it could be the number of turns per unitof length in the equation I wrote)
[17:14:55] <Lerneaen_Hydra> yeah, so what does the diameter have anything to do with that?
[17:15:48] <SWPadnos> torque / diameter = force
[17:15:56] <SWPadnos> newton-meter / meter = newton
[17:17:14] <dnisbet2> Thanks to all, I'm now moving in the right direction- I first looked on the EMC home page, under support, then the EMC documents, but that just confused me more, I then tried the Sherline site to see if that may be relevant, also read the WikkI, did a search on Google. My suggestion would be a FAQ page on the EMC web site for these simple things ??
[17:17:23] <SWPadnos> the way I think of the calculation (which I didn't really do beforw writing that equation) is how far and how hard do you have to push to move the table
[17:19:29] <etla> from a bit of googling, it seems that F= 2*pi*eff*Torque/Lead no diameter in there
[17:19:40] <SWPadnos> ok by me :)
[17:19:56] <SWPadnos> hey - that's what I wrote first :)
[17:20:14] <etla> so the continuous torque of my motors corresponds to about 3000 N of force - that should be enough on a minimill ? ;)
[17:20:32] <SWPadnos> cutting forces can be high, but I'd bet you're right :)
[17:24:01] <etla> so anyone have an opinion about the slow-speed question I posted on the list ?
[17:24:17] <etla> am I in trouble with a 4000count/rev system ?
[17:24:28] <SWPadnos> I hope not - I have the same thing :)
[17:24:46] <SWPadnos> you don't need one "tick" per interrupt
[17:25:15] <etla> ok, so how does the PID loop act when there's no movement for one interrupt cycle ?
[17:25:24] <SWPadnos> I think what'll happen is that the PID output will oscillate a bit around the correct value
[17:25:41] <SWPadnos> the I term goes up, the P term stays the same, and D is 0, I think
[17:25:53] <cradek> have you looked at sim/servo? I think you could simulate and play with it
[17:25:56] <SWPadnos> so the longer you go without any apparent motion, the higher the output will go due to I
[17:26:20] <etla> cradek: does sim/servo have some machine dynamics (inertia etc.) built in ?
[17:26:39] <cradek> I'm not sure if it does, but it could
[17:26:56] <cradek> I used a limit block to give inertia to the sim/lathe spindle
[17:27:56] <etla> because if there is no inertia I would assume that an oscillating PID output will give an oscillatory movement
[17:28:19] <etla> but in reality the inertial, or mechanical bandwidth if you want to be fancy, is probably much below the servo thread rate
[17:28:55] <SWPadnos> yep. as long as you get some movement within the mechanical inertia "time constant", you should be OK
[17:29:19] <etla> yes, that sounds reasonable.
[17:29:52] <SWPadnos> consider though that no motion is also a valid setting, and that works
[17:30:14] <SWPadnos> essentially, very slow motion may degenerate into a series of zero motion states at successive encoder count values
[17:31:02] <cradek> what length corresponds to one count?
[17:31:10] <etla> I calculated that about 37mm/min (1.5ipm or so) would correspond to 1kHz count rate
[17:31:31] <SWPadnos> 4000 counts/2.5mm
[17:31:33] <etla> In practice I'd like to use something like 10mm/min but would not need to go much lower
[17:32:03] <SWPadnos> you may get a staircase at 1/1600mm increments instead of a nice smooth motion
[17:32:12] <etla> so the lowest I'd want to use would be about 200Hz count rate, or one encoder count every 5 servo periods
[17:33:45] <SWPadnos> actually, the lowest would be zero counts pe rsecond
[17:33:47] <etla> but as noted earlier, if the PID oscillates at 200Hz, the motor and mechanics probably act as a low pass filter, so the oscillation won't be visible
[17:34:17] <SWPadnos> remember that the position command is also changing. the position will be a smooth ramp which slowly diverges from the feedback position
[17:35:01] <SWPadnos> so at first, the PID will not try to move the motor. once the error exceeds DEADBAND, PID will try to correct. that will probably cause the motor to move one encoder count (maybe more, but then it would be pulled back)
[17:35:22] <SWPadnos> once that happens, the motor is probably past the intended position, so PID will try to move it backwards a little
[17:35:36] <SWPadnos> (unless it's within DEADBAND)
[17:36:11] <SWPadnos> so you may get a little dithering between successive encoder edges, or it may just staircase from edge to edge
[17:36:12] <alex_joni> yeah, but that will happen in the first 2-3 steps
[17:36:31] <alex_joni> after that integral and derivative will takeover a bit
[17:36:55] <SWPadnos> it should happen at every step if the step rate is very low relative to the servo period
[17:37:50] <alex_joni> bbl guys
[17:37:58] <SWPadnos> see you Alex
[17:39:26] <etla> thanks for the explanations - seems I don't have to worry about running the servos slower tha 1 count/servo thread rate
[17:40:04] <SWPadnos> I don't think so. if it were a problem then stopping would be hard :)
[17:43:59] <Lerneaen_Hydra_> haha, what a nice log I've got
[17:44:41] <Lerneaen_Hydra_> etla starts talking about <1 update/cycle, I exit, rejoin, etla talks about how it should all be good
[17:45:10] <SWPadnos> step 2: ... "then a miracle occurs" ...
[17:46:02] <Lerneaen_Hydra_> haha
[17:46:31] <Lerneaen_Hydra_> 1) ask question
[17:46:32] <Lerneaen_Hydra_> 2) ...
[17:46:34] <Lerneaen_Hydra_> 3) get answer
[17:46:35] <Lerneaen_Hydra_> 4) ???
[17:46:37] <Lerneaen_Hydra_> 5) PROFIT!!!
[17:49:59] <SWPadnos> http://www.webamused.com/blogosophy/archives/002064.html
[17:54:36] <etla> SWP: have you worked with VCP ?
[17:56:31] <col> hello
[17:56:40] <SWPadnos> not really
[17:56:43] <SWPadnos> col, hello
[17:57:38] <col> just been to look at the machines im buying today
[17:58:01] <etla> SWP: ok, I'll try to take a look at the code...
[17:58:02] <col> been trying to find info on the servo motors
[17:58:08] <etla> col: what kind of machines ?
[17:58:35] <col> hurco milling machine and churchill lathe (seimens motors and controls)
[17:59:22] <etla> wow, sounds big. any pictures online ?
[17:59:31] <etla> will they be emc-driven ?
[17:59:35] <col> yeah
[18:00:02] <Lerneaen_Hydra_> bbl
[18:00:16] <Lerneaen_Hydra_> going to try to get *nix on my main machine once and for all
[18:00:18] <Lerneaen_Hydra_> whee
[18:00:54] <col> got pics on my fone
[18:00:58] <col> cant get blue tooth working tho
[18:01:09] <col> the lathe has a 12" chuck and hydrostatic bearings
[18:01:17] <col> has 15 degree c axis
[18:01:35] <col> and 10 tool changer and drill
[18:02:29] <etla> is this a for-profit project or just hobby ??
[18:02:59] <col> there machines for me
[18:03:24] <etla> what kind of pc-hardware are you planning to use with emc ?
[18:03:56] <col> the montenc stuff looks good
[18:04:49] <etla> you got all the servo drives with the machines ?
[18:04:53] <col> yup
[18:04:57] <col> there all complete
[18:05:05] <col> the hurco mill has a blown memory board
[18:05:20] <col> and the lathe just has ancient siemens controls
[18:05:36] <col> came out of an aerospace factory for new machines
[18:08:57] <col> lathe looks like this one
[18:09:02] <col> http://www.gtprecision.co.uk/images/machsales/CHURCHILL%20015.jpg
[18:09:21] <SWPadnos> you must have a very large garage :)
[18:09:41] <col> yeah its an old converted barn
[18:10:12] <etla> gotta go, c u later
[18:10:18] <col> laters
[18:10:22] <SWPadnos> see ya
[18:10:38] <col> you reckon the montenc would be a wise choice ?
[18:11:09] <SWPadnos> dunno. I don't have enough real experience with the various hardware EMC2 supports (or the machines) to tell you
[18:12:16] <SWPadnos> I think it works fine, though I believe there may be issues with the way the encoder counters work (praticularly regarding resetting them on an index pulse, which is used for spindle-synchronized motion)
[18:13:14] <col> oh right
[18:13:15] <SWPadnos> (of course, I could be remembering a problem with a completely different piece of hardware :) )
[18:13:48] <col> ill have to see what the general recomendation is
[18:14:02] <SWPadnos> yep. it will also depend on how much of the old system you can keep
[18:14:09] <col> id prefer to have the card pci as i dont think i have any old comuters with isa boards anymore
[18:14:26] <col> i should be able to keep most of the old system tbh
[18:14:27] <SWPadnos> if the motors and drivers are good, then you need to get a controller card that has the correct output for the drivers
[18:14:33] <col> the lathe has all new encoders fitted
[18:14:37] <SWPadnos> PCI is definitely better
[18:14:57] <col> the drivers being the servo amplifiers ?
[18:15:05] <SWPadnos> the Mesa 5i20, Motenc, and I think some servo-to-go are all PCI (also vigialnt, I think)
[18:15:06] <SWPadnos> yes
[18:15:08] <tomp> hello
[18:15:19] <SWPadnos> if they need an analog input, you need something with analog output
[18:15:24] <col> shoudlnt they work on a standard +-10v anyway ?
[18:15:27] <SWPadnos> if they need PWM, you don't need analog ...
[18:15:30] <SWPadnos> could be
[18:15:35] <col> how do you tell
[18:15:56] <SWPadnos> you probably need to look at them and google a bit, or look at the manual if you get one
[18:15:59] <SWPadnos> hi tomp
[18:16:13] <col> iv got the wiring diagrams for the lathe
[18:16:14] <SWPadnos> are you subscribed to the Yahoo group "CAD_CAM_EDM_DRO"?
[18:16:17] <SWPadnos> ok
[18:16:48] <SWPadnos> some folks in the emc crowd and probably more on CCED should know about your hardware
[18:17:15] <col> ok cool
[18:17:37] <col> once i get the machines delivered i can take more detailed pics and remove things for inspection
[18:18:13] <SWPadnos> yep. I'd wait on any hardware purchases until you know more about what you need
[18:19:07] <col> the motors on the mill look like fairly standard dc
[18:19:32] <SWPadnos> depending on the age of the mill, they could be DC brush, stepper, or even brushless DC or AC
[18:19:43] <SWPadnos> the wiring diagrams and amplifiers will tell you
[18:19:45] <col> dc brush i think
[18:19:51] <col> yeah
[18:22:13] <col> http://www.andersonindustrialtech.com/hurco50.jpg
[18:22:17] <col> hurco looks a bit like that one
[18:24:21] <col> only a 10 tool changer
[18:24:30] <col> and slightly smaller but pretty similar
[18:25:30] <tomp> hello, i upgraded ubuntu breezy to dapper... i got probs running emc ( rtapi failed) and the running the cvs built version (cannot find -sec HAL -var HALUI ..) where to begin?
[18:27:09] <SWPadnos> did you change the emc2 repository names to dapper ?
[18:27:57] <tomp> yes, a mere umpteenn million files to download, but it ran ( running i now)
[18:28:08] <tomp> running it now
[18:28:12] <SWPadnos> and also update emc2 / magma kernels?
[18:28:44] <tomp> dunno, a zillion lines scrolled by, how do i check?
[18:28:54] <SWPadnos> you can run the GUI now, right>
[18:28:56] <SWPadnos> ?
[18:29:35] <tomp> yep ( had to force the /dev/input directory into existance, and MAKEDEV would create the nodes, but i forced them )
[18:29:45] <SWPadnos> run synaptic (System -> Administration -> Synaptic Package Manager)
[18:29:47] <tomp> would not create..
[18:29:51] <SWPadnos> search for emc
[18:29:56] <tomp> k
[18:30:06] <SWPadnos> that'll tell you the installed version and any available versions
[18:30:22] <SWPadnos> you may want to reload first, just to be sure your package lists are up to date
[18:31:44] <tomp> emc 1:2.0.5
[18:32:15] <tomp> reloading & rechecking
[18:33:26] <tomp> same result and axis 1:4a0-0.2
[18:34:52] <tomp> mark 'em for re-install?
[18:35:20] <tomp> just for the hellofit?
[18:35:28] <skunkworks> jepler:
http://www.electronicsam.com/images/KandT/servostart/usb5v.jpg
[18:35:39] <skunkworks> wth?
[18:35:43] <skunkworks> jepler:
http://www.electronicsam.com/images/KandT/servostart/usb5v.jpg
[18:35:52] <skunkworks> I must like white space
[18:37:11] <tomp> hmm, it wont give me reinstall option, so ? remove & install ?
[18:38:53] <col> jepler:
http://www.electronicsam.com/images/KandT/servostart/usb5v.jpg
[18:38:57] <tomp> running emc from the menu gets...
[18:38:59] <col> frick
[18:39:00] <col> sorry
[18:39:02] <tomp> Can't write to /dev/rtai_shm - aborting
[18:39:02] <tomp> Realtime system did not load
[18:39:02] <tomp> Shutting down and cleaning up EMC...
[18:41:16] <tomp> so, any ideas on some lower level tests? somethign like bin/loadrt blah to see if any rt stuff is operational?
[18:43:28] <skunkworks> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting does that help you?
[18:44:28] <col> was very impressed when i put that live cd in it all loaded up and worked
[18:44:41] <skunkworks> col: cool huh?
[18:44:41] <tomp> skunkworks: i'll try that thanks: i was looking for the HAL docs, & thought i'd try the halmeter example... cant remember the name of the pdf tho...
[18:45:19] <col> was expecting to have to fiddle around installing nix on a machine then compiling kernels ect
[18:46:44] <col> gonna knock up a machine next week and install the ubuntu and emc straight from the live cd
[18:47:06] <tomp> skunkworks: looks like the right thread, thanks
[18:47:57] <skunkworks> tomp: even a broken clock is right twice a day ;z0
[18:48:00] <skunkworks> :)
[18:52:58] <col> gonna be intersting trying to control the spindle speed on the lathe
[18:53:07] <col> the motor looks like somthing that powers a train
[18:54:24] <tomp> yes, the RTAI_SHM dev wasnt created, but the err is 'an't write to /dev/rtai_shm - aborting' ?? the docs say uppercase, the err sez lower!
[18:54:45] <tomp> trying lower case device creation...
[18:56:29] <skunkworks> col: how many hp? we had some 50hp spindle lathes
[18:57:17] <col> not entirely sure on hp
[18:57:41] <col> says 89 amps 380v on the motor plate
[18:59:33] <Rugludallur> 34kw so 45hp or so
[18:59:45] <jepler> skunkworks: cute -- how's it work?
[19:00:08] <jepler> skunkworks: if I could get over the worry about accidental damage to my USB port I'd do something that simple.
[19:00:10] <skunkworks> jepler: the led lights up... that is as far as I have gotten
[19:00:18] <col> Rugludallur: thanks
[19:00:34] <tomp> Can't write to /dev/rtai_shm - aborting... damn permissions again, rtai_shm rw-r--r-- chgd to rw-rw-rw- to force it to work
[19:02:59] <tomp> sim axis running chips now , but i think this may be munged in ways i wont spot easily, oughta pack up & trash this install
[19:03:31] <jepler> $ ls -al /dev/RTAI_SHM /dev/rtai_shm
[19:03:31] <jepler> ls: /dev/RTAI_SHM: No such file or directory
[19:03:31] <jepler> ls: /dev/rtai_shm: No such file or directory
[19:03:31] <jepler> $ /etc/init.d/realtime start
[19:03:31] <jepler> $ ls -al /dev/RTAI_SHM /dev/rtai_shm
[19:03:34] <jepler> lrwxrwxrwx 1 root root 8 2006-12-28 13:02 /dev/rtai_shm -> RTAI_SHM
[19:03:37] <jepler> crw-rw-rw- 1 root root 10, 254 2006-12-28 13:02 /dev/RTAI_SHM
[19:03:44] <jepler> ^^ here's what should happen when starting realtime.
[19:03:53] <jepler> (breezy system with emc 2.0.5 package installed)
[19:05:39] <jepler> $ cat /etc/udev/rules.d/emc2.rules
[19:05:39] <jepler> KERNEL=="RTAI_SHM" SYMLINK=="rtai_shm" MODE="0666"
[19:06:00] <jepler> ^^^ this file (part of the emc2 2.0.5 deb) makes that happen automatically
[19:08:59] <tomp> thanks, checking emc2.rules and symlink
[19:14:34] <col> gtg cyas later
[19:14:48] <col> thanks for the help
[19:19:48] <tomp> i deleted 2 2 devs i created, then stopped & then started (/etc/init.d/realtime start) and got .
[19:19:48] <tomp> .
[19:19:48] <tomp> .
[19:19:48] <tomp> Can't write to /dev/rtai_shm - aborting
[19:19:53] <tomp> emc2.rules is ok
[19:20:19] <cradek> /etc/init.d/udev restart
[19:20:58] <tomp> long list of permission denied's
[19:21:04] <cradek> sudo
[19:23:24] <tomp> no perm denieds, lotsa insmod: error inserting '/usr/realtime-2.6.12-magma/modules/emc2/hal_lib.ko': -1 File exists then 'cant write to /dev/rtai_shm'
[19:23:53] <tomp> i think the hours last nite & this morning are < reinstall
[19:24:18] <tomp> > (still fuzzy)
[19:24:39] <skunkworks> I thought I was the only one that could really bork up the emc/rt install. ;)
[19:25:46] <tomp> this was upgrade
[19:26:30] <tomp> makin a list & chekin it twice... what should i keep?
[19:26:42] <cradek> I think someone? found it works best to uninstall the rt kernel, rtai, and emc2 packages before doing the major upgrade
[19:27:02] <cradek> if you nuke those packages (use --purge) and then reinstall the dapper versions, it will probably be fine
[19:27:38] <cradek> if you have not yet made a backup (off the hard drive) of your emc2 configuration, you should do that right away
[19:28:23] <tomp> the emc2 & emc2-head dirs are going to a cf chip now thanks
[19:31:27] <tomp> crap the usb not seen, off to unused ide partitions then
[19:51:04] <skunkworks> tomp:
[19:51:06] <skunkworks> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Breezy_Upgrading
[19:51:25] <skunkworks> too late now I suppose ;)
[19:52:01] <tomp> ok, dpkg --purge emc2-axis emc2 rtai-modules-2.6.12-magma rtai done now to install ( skunkworks, thanks )
[19:58:49] <tomp> uh, rtai-modules-* doesnt exit now...i just removed it ok, but now dpkg & synaptic deny it's existance ??
[20:00:15] <SWPadnos> apt-get install emc2 emc2-axis (and optionally emc2-dev)
[20:00:56] <skunkworks> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2 starting at 3?
[20:01:04] <skunkworks> ah what SWPadnos said I think
[20:01:09] <skunkworks> :)
[20:02:24] <tomp> sudo apt-get install emc2 emc2-axis emc2-dev
[20:02:40] <tomp> Package emc2 is not available, but is referred to by another package.
[20:02:50] <CIA-8> 03jepler 07HEAD * 10emc2/debian/extras-Ubuntu-5.10/emc2.files: add udev rules to create /dev/rtf automatically when rtai_fifos is loaded
[20:03:25] <tomp> repository list?
[20:03:38] <CIA-8> 03jepler 07HEAD * 10emc2/debian/extras-Ubuntu-5.10/etc/udev/rules.d/emc2.rules: add udev rules to create /dev/rtf automatically when rtai_fifos is loaded
[20:03:44] <CIA-8> 03jepler 07HEAD * 10emc2/debian/extras-Ubuntu-6.06/emc2.files: add udev rules to create /dev/rtf automatically when rtai_fifos is loaded
[20:03:44] <CIA-8> 03jepler 07HEAD * 10emc2/debian/extras-Ubuntu-6.06/etc/udev/scripts/rtai_fifos.sh: add udev rules to create /dev/rtf automatically when rtai_fifos is loaded
[20:03:44] <CIA-8> 03jepler 07HEAD * 10emc2/debian/extras-Ubuntu-5.10/etc/udev/scripts/rtai_fifos.sh: add udev rules to create /dev/rtf automatically when rtai_fifos is loaded
[20:03:46] <CIA-8> 03jepler 07HEAD * 10emc2/debian/extras-Ubuntu-6.06/etc/udev/rules.d/emc2.rules: add udev rules to create /dev/rtf automatically when rtai_fifos is loaded
[20:04:29] <jepler> tomp: tomp should be
[20:04:29] <jepler> deb
http://www.linuxcnc.org/emc2/ dapper emc2
[20:04:29] <jepler> deb-src
http://www.linuxcnc.org/emc2/ dapper emc2
[20:04:42] <tomp> thanks
[20:04:47] <jepler> also try an "apt-get update" to refresh the package lists
[20:05:37] <tomp> will do b4 i try to get stuff
[20:12:30] <tomp> looks hopeful: the files downloaded & installled... now system restart required... bbl8r
[20:17:20] <tomp> hello, new msgs: cd ~/emc2; scripts/emc ----> find: /usr/realtime-2.6.12-magma: No such file or directory
[20:18:14] <tomp> but i can see it.. perms?
[20:19:01] <lerneaen_hydra> anyone here?
[20:19:02] <tomp> rwxr-x-r-x root:root
[20:19:12] <lerneaen_hydra> nice
[20:19:35] <lerneaen_hydra> my linux install isn't going as smoothly as I had hoped. anyone here know grub well?
[20:20:12] <lerneaen_hydra> or at least better than I, which means basically do you know anything at all?
[20:21:11] <tomp> i use grub on all my systems, what do you se happen
[20:21:31] <lerneaen_hydra> uh, it's complicated, just a sec
[20:21:55] <tomp> grub is complex ( lotsa little pieces)
[20:22:01] <lerneaen_hydra> I have a fresh install of kubuntu 6.10, and am having some bootloader issues. My main rig has 6 harddisks, and I've installed kubuntu on the "last" one. (two are PATA, four SATA, kubuntu installed on sddX). I've set up a small /boot partition (sdd1) and the main root partition (sdd3), and installed the bootloader to sdd (and not hda). When starting grub loads nicely, but I get an error 17, cannot mount partition
[20:22:32] <SWPadnos> good thing it's a really simple setup - NOT! :)
[20:22:37] <lerneaen_hydra> indeed
[20:22:52] <lerneaen_hydra> lots of NTFS bastardisation too (that I hope to get rid of soon too)
[20:23:18] <cradek> tomp: you need to recompile it
[20:23:22] <SWPadnos> did you already have grub installed on HDD1 (or whatever the boot drive is)
[20:23:32] <tomp> out of my experience, i have all ide , up to 3 drives, BUT you can read the grub manuals and, at the command line at boot, seeif you can access that partition at all
[20:23:46] <SWPadnos> tab completion even works in GRUB
[20:23:53] <jepler> tomp: cradek means, you have to recompile emc2 (make clean; ./configure ...; make && sudo make setuid)
[20:24:28] <tomp> ceadek: recompile... what? i was running emc2 from apt-get install.... not from cvs... recompile still apply?
[20:24:29] <jepler> tomp: the compiled ~/emc2 is still trying to use the old realtime kernel
[20:24:43] <tomp> k
[20:24:48] <lerneaen_hydra> SWPadnos, nope, no grub anywhere except for the new install
[20:24:55] <lerneaen_hydra> which was on the last disk
[20:24:56] <jepler> If you're using scripts/emc then you're using the CVS version, not the packaged version
[20:25:31] <cradek> lerneaen_hydra: I'm scared of any machine that has 6 hard disks and is not scsi
[20:26:08] <jepler> cradek: 4 PATA ports is not uncommon
[20:26:18] <lerneaen_hydra> cradek, haha, I routinely back up stuff that I can't easily get a hold of again
[20:26:21] <tomp> really!, then i just invoked it wrong ( using simplest vrsn to test but forgot how to run it as i dont use 'emc2' much, just emc2-head' )
[20:26:36] <jepler> tomp: try /usr/bin/emc to use the installed version
[20:26:41] <tomp> thanks
[20:27:31] <lerneaen_hydra> cradek, the computer I use as my main one has 8 sata ports and 4 pata
[20:28:21] <tomp> cradek: jepler: runs now thanks, i get emc2-head from cvs
[20:31:52] <lerneaen_hydra> so no ideas re the grub issue?
[20:32:16] <SWPadnos> you need to set up your existing bootloader to boot the ubuntu partition
[20:32:18] <lerneaen_hydra> no obvious "check XYZ" thing?
[20:32:28] <SWPadnos> or you could do it as a "chainload" I think
[20:32:28] <lerneaen_hydra> It does that, I've checked
[20:32:39] <jepler> i've never used grub in a system with that many disks
[20:32:59] <tomp> eror 17 = Cannot mount selected partition
[20:32:59] <tomp> This error is returned if the partition requested exists, but the filesystem type cannot be recognized by GRUB
[20:33:23] <lerneaen_hydra> that's strange, because it's EXT3
[20:33:35] <tomp> can you fdisk the drive & see what it reports?
[20:33:48] <lerneaen_hydra> yeah
[20:33:57] <lerneaen_hydra> I'll boot a live-cd and check to see what it is
[20:34:13] <jepler> are you sure grub has the right mapping between (hdX) and the devices?
[20:34:26] <lerneaen_hydra> I would think it does
[20:34:38] <lerneaen_hydra> I'm not sure how to check that though
[20:35:04] <jepler> do you get a grub prompt, or not even that?
[20:35:08] <lerneaen_hydra> yeah
[20:35:11] <tomp> grub hd0,0 = linux hda1
[20:35:15] <lerneaen_hydra> I can choose boot options and stuff
[20:35:27] <lerneaen_hydra> currently it's set to hd5,0
[20:35:32] <cradek> when you boot the livecd, you can chroot into the linux install, then make everything right (grub.conf, device.map, mtab), then install grub
[20:35:39] <jepler> get to the prompt and try: root (hd3,0) or other numbers
[20:35:56] <lerneaen_hydra> which sounds sane, because it's the last harddisk and the first partition
[20:35:57] <lerneaen_hydra> oh
[20:36:01] <lerneaen_hydra> fsck
[20:36:17] <lerneaen_hydra> I want to be loading my / partition, right?
[20:36:18] <jepler> on my one system with sata & pata both, grub knows the sata drive as (hd0)
[20:36:21] <lerneaen_hydra> and not my /boot?
[20:37:00] <cradek> what do you mean "loading"
[20:37:19] <lerneaen_hydra> as in when grub tries to mount a partition to boot the OS
[20:38:25] <lerneaen_hydra> funny, in knoppix the drives are in a different order
[20:38:33] <lerneaen_hydra> sdd -> sdc
[20:40:39] <tomp> ie: root (hd1,0)
[20:40:40] <tomp> kernel/boot/vmlinuz-2.6.12-9-386 root=/dev/hdb6 ro single
[20:40:47] <lerneaen_hydra> looks like I'll just have to test different harddisk numbers
[20:41:14] <lerneaen_hydra> what does the root parameter define?
[20:42:15] <cradek> the hdX is what's defined in device.map
[20:42:26] <cradek> the ,Y is the partition number of the root directory
[20:42:36] <lerneaen_hydra> hmm, I meant more like what is it used for
[20:42:56] <lerneaen_hydra> I can see that the next row defines the drive/partition to boot from
[20:43:08] <lerneaen_hydra> and where the kernel is
[20:43:36] <cradek> I guess I don't know
[20:43:40] <tomp> oh, it's / of the file system you're mounting
[20:43:42] <cradek> (I know it has to be right)
[20:44:02] <cradek> well the kernel mounts the root fs, not grub
[20:44:03] <tomp> (i answered a zillion times but began with / so it didnt appear here )
[20:44:38] <tomp> glossary
http://www.redhat.com/docs/manuals/linux/RHL-9-Manual/ref-guide/s1-grub-terminology.html
[20:44:38] <cradek> lerneaen_hydra: beware the ,Y is zero based, unlike what fdisk shows you
[20:44:44] <lerneaen_hydra> ah
[20:45:11] <lerneaen_hydra> it just seems to me that / is defined twice
[20:45:40] <tomp> grub begins hd0,0 linux begins hda1
[20:46:19] <lerneaen_hydra> oh, so root is the grub root, and the second is kernel root (/)
[20:46:26] <cradek> grub might parse fstab on that root (hdx,y) in order to determine paths like /boot
[20:46:47] <cradek> because I don't see how it would know what /boot means otherwise
[20:47:04] <jepler> on a system with mixed scsi (or sata) and pata, it's not clear to me whether (hd0) is sda or hda. that's what I'm saying.
[20:47:25] <lerneaen_hydra> in the install hdXX came before sdXX
[20:47:25] <cradek> jepler: it just depends what you put in device.map
[20:47:31] <cradek> jepler: (I think)
[20:47:36] <jepler> cradek: if /boot is a separate partition, then the grub configuration would have "root (something that designates /boot); kernel /vmlinuz..."
[20:47:49] <lerneaen_hydra> I have /boot on a seperate partition
[20:47:58] <lerneaen_hydra> thought it would be good
[20:48:00] <cradek> jepler: ohhhh
[20:48:12] <lerneaen_hydra> some old hold habit from RH7.x
[20:48:22] <SWPadnos> I wonder if this question should be asked in #grub?
[20:48:41] <lerneaen_hydra> but...but....
[20:48:44] <lerneaen_hydra> * lerneaen_hydra sniffles
[20:48:46] <SWPadnos> heh :)
[20:48:47] <cradek> be brave!
[20:48:49] <jepler> (at least, that's how it works on my fedora system with separate /boot)
[20:48:49] <lerneaen_hydra> ;)
[20:49:01] <lerneaen_hydra> I guess I can test
[20:49:21] <ATAT> I'm not sure if this is the driver board or the configuration, but my EMC config along with the hobbycnc board is making really strange movements
[20:49:42] <ATAT> the mill moves fast in the negative x direction, but slow in the positive x
[20:49:54] <ATAT> EMC thinks the movement speeds are the same
[20:50:38] <ATAT> any suggestions? I've tried tweaking all the config files I could think of to no improment
[20:51:04] <cradek> ATAT: how fast/slow?
[20:51:29] <tomp> in my example: 'kernel ... hdb3' was the '/' of the system the entry 'root (hd0,1)' is where the '/boot' is located
[20:51:31] <jepler> ATAT: which pinout file are you using? are you sure it matches the pinout of this driver board?
[20:51:34] <ATAT> making the mill move 1 mil in negative x moves it 1 mil, but about 1/4 mil in the positive x
[20:51:55] <ATAT> i checked the pinout of the driver board and made usre it mached the standard_pinout.hal file
[20:52:07] <tomp> and i can boot that filesystem :)
[20:54:58] <cradek> ATAT: does just X do this? what, if any, parts of the setup do you KNOW are good?
[20:55:22] <lerneaen_hydra> tomp, hmm, sounds pretty much like my setup
[20:55:44] <ATAT> all 3 axis do the same thing.. swapping between axis X,Y,Z,A on the driver board makes no difference
[20:56:17] <cradek> ATAT: what does a 1" move do?
[20:56:36] <cradek> ATAT: nothing is immediately coming to mind for any of us, so we need as much information as you have
[20:56:53] <tomp> learnean: when grub appears, have a manual printed out or another computer... you can debug it at that prompt, you can search for a file on a partition with the commands, and that way make sure you're on the right one
[20:56:59] <eholmgren> lerneaen_hydra: is it bad that I still make a 10M boot partition?
[20:57:36] <ATAT> 1" move in negative X is 1", positive X is about 1/4 that value, but not exactly
[20:57:57] <jepler> ATAT: does the speed of the move matter?
[20:58:12] <ATAT> not that I can tell
[20:58:13] <cradek> are you using a laptop?
[20:58:15] <lerneaen_hydra> eholmgren, I don't know, I still make a 100mb one, for what it's worth
[20:58:37] <lerneaen_hydra> tomp, you can search for files on partitions from that prompt?
[20:58:40] <lerneaen_hydra> nice
[20:58:47] <tomp> yes
[20:58:51] <ATAT> cradek: was that question addressed to me?
[20:58:58] <cradek> yes sorry
[20:59:04] <ATAT> cradek: yes
[20:59:09] <jepler> ATAT: turn the Feed override down to 10% and check that you have the same problem
[20:59:10] <ATAT> would that matter?
[20:59:30] <cradek> is positive X high or low on the direction pin?
[20:59:45] <cradek> laptops usually have 3.3v parallel ports which is a little low for something expecting 5v
[21:00:04] <ATAT> ah
[21:00:09] <cradek> but yeah, try what jepler says
[21:00:13] <ATAT> so the solution would be to get a desktop?
[21:00:24] <cradek> no, the solution is to troubleshoot some more first :-)
[21:00:33] <ATAT> jepler: I'll try that
[21:00:42] <ATAT> also whats the point of OUTPUT_SCALE
[21:00:51] <cradek> it's ignored for steppers
[21:00:55] <ATAT> oh
[21:01:03] <ATAT> got it. thanks.
[21:01:09] <cradek> what's your INPUT_SCALE?
[21:01:32] <ATAT> 16000
[21:01:55] <cradek> all the INPUT_SCALES are positive?
[21:02:33] <ATAT> X and Y are positive Z is negative
[21:03:09] <cradek> are you sure Z has the same behavior as X and Y (the positive move is short)?
[21:03:32] <skunkworks> jepler: is there a pinout of the left pluto plug? (actual altera pin numbers?
[21:03:40] <ATAT> I'll go check... ill log out for a second.
[21:04:16] <jepler> cradek may have hit the nail on the head. it looks like the hobbycnc board is based on the allegro sla706x. The datasheet says that the input high voltage minimum is .75*vcc, 3.75V. If it is a 3.3v parallel port, and there are not appropriate pull-ups on the board, the chip might not be seeing "logic high" reliably.
[21:04:25] <jepler> skunkworks: yes, jas
[21:04:41] <cradek> jepler: then I suggested a good test, we'll see
[21:04:45] <SWPadnos> you'd think that would apply to the step pin as well though
[21:04:58] <SWPadnos> so the actual distances would be random, up to the commanded length
[21:05:07] <jepler> SWPadnos: welll yes that is the problem with my theory
[21:05:13] <SWPadnos> heh
[21:05:28] <SWPadnos> you can blame it on cradek - he thought of it :)
[21:05:40] <jepler> skunkworks:
http://emergent.unpy.net/01165081407 but it may be wrong (I established it by trial and error with a continuity-tester, then copied the pin numbers from a dirty napkin to my blog)
[21:05:53] <skunkworks> :) thanks
[21:07:38] <jepler> skunkworks: I am not sure why it's missing from the official documentation -- the more I/Os the merrier
[21:08:04] <skunkworks> jepler: right - I missed that on your site.
[21:08:58] <skunkworks> * skunkworks is getting a bunch of case led's to play with (plugs on the end and everything).. Just need to add a resister in series.
[21:09:14] <skunkworks> * pc case
[21:09:26] <SWPadnos> you could have had so much with just another 7i31 ...
[21:09:38] <SWPadnos> any 7i31, actually
[21:10:14] <cradek> jepler: looks like with a positive scale, dir is low for a positive move
[21:10:23] <skunkworks> I will probably proof things out with the pluto - then it will be downgraded to a lathe ;)
[21:10:58] <cradek> jepler: so I think it's unlikely my guess is right
[21:11:56] <jepler> cradek: darn
[21:15:29] <ATAT> tried with various feed rates, no changes
[21:15:42] <ATAT> the Z axis behavior is reversed though, positive Z is fast, negative z is slow
[21:15:53] <ATAT> probably because of the negative input scale
[21:16:31] <cradek> jepler found that the hobbycnc board's spec says it IS incompatible with a 3.3v parallel port
[21:16:52] <ATAT> alright I'll try using a desktop then. time to assemble one
[21:17:00] <cradek> that combined with the reversed Z behavior makes me think it's the problem
[21:17:58] <cradek> if you have any desktop computer and the live CD, you can test that theory without making a new machine.
[21:18:00] <ATAT> are there any laptops with 5v parallel output?
[21:18:09] <ATAT> you can run the emc off livecd?
[21:18:13] <cradek> it's hard to say. quite likely not.
[21:18:15] <cradek> yes you can
[21:18:18] <SWPadnos> probably only ones that are 5 years old or more
[21:18:31] <SWPadnos> you can also make or buy a breakout board
[21:18:55] <ATAT> hmmm okay
[21:19:05] <ATAT> ill try that too
[21:19:09] <cradek> a desktop machine is going to be the best solution in the long run. laptops often have problems with realtime too.
[21:19:46] <cradek> an old PIII seems currently to be the ideal machine
[21:19:53] <cradek> PIII desktop, I mean
[21:20:12] <ATAT> sighs. the only spare desktop I have is a celeron from a few years back
[21:20:13] <ATAT> ahwell
[21:20:24] <cradek> how much ram?
[21:20:31] <ATAT> about 512
[21:20:32] <SWPadnos> I run emc2 on a celeron 500 with 512M, so they're not all bad
[21:20:36] <cradek> it'll be fine
[21:20:58] <SWPadnos> in fact, it has realtime performance about as good as anything else people have tried (latencies in the 6 uS range)
[21:21:16] <lerneaen_hydra> ok, some progress, my BIOS apparently moves drive positions around when booting, so what I thought was HD5,0 was in fact HD0,0
[21:21:19] <cradek> your scale of 16000 is not very demanding unless you want the machine to move very fast
[21:21:30] <cradek> so yeah, try the celeron
[21:21:47] <lerneaen_hydra> which means my root (hd0,0) now works, but I'm having issues with my other boot command
[21:22:10] <lerneaen_hydra> which defines where the kernel is located
[21:22:40] <tomp> i used cvs & built new emc2-head, failed, emc-debug reports 'could not find -sec HAL -var HALUI...' (and i chose the cfg from emc2-head not emc2)
[21:23:05] <tomp> learnean: kernel is located in the partion you had /boot in
[21:23:23] <lerneaen_hydra> tomp, I have seperate /boot and / partitions
[21:23:37] <cradek> the kernels are still in /boot
[21:23:44] <lerneaen_hydra> cradek, you sure?
[21:23:52] <tomp> me too, my kernel image is in /boot, inside /boot is /grub... same for you?
[21:23:56] <cradek> pretty sure, but look
[21:24:31] <tomp> my /boot is also spearate partition
[21:24:40] <tomp> separate
[21:24:41] <ATAT> cradek: is it possible to remotely run emc?
[21:25:09] <lerman> I'm doing it using vnc. :-)
[21:25:22] <lerneaen_hydra> so that would mean my kernel command should be /.../ /dev/sdd1 ?
[21:25:24] <cradek> ATAT: there are lots of options - the easiest is ssh -X, like any other X application
[21:25:56] <skunkworks> jepler: It is pretty cool you used the indicator led to show the output of up0/down0.. (just going over your documents)
[21:26:03] <ATAT> hmm okay. because I just realized 2 desktops and 1 monitor doesnt work too well =p
[21:26:07] <jepler> skunkworks: it was very nice in testing
[21:26:11] <lerman> :-) because I just got it running. I'm not using X because I'm not setup for X on my windows machine.
[21:26:20] <jepler> skunkworks: I've mostly worked on pluto without real hardware attached
[21:26:31] <lerneaen_hydra> maybe I should just reinstall, have /boot in the same partition in / and pay more attention when installing
[21:26:32] <cradek> ATAT: yay, an excuse to buy an LCD
[21:26:36] <skunkworks> jepler: I figured. very resourceful
[21:29:02] <tomp> lerneaen_hydra; can you see the the images in the /boot partition? like /boot/vmlinuz.... (near 1 meg size)?
[21:29:26] <lerneaen_hydra> I'm not sure, it's hard to read anything at all
[21:29:40] <tomp> clean the monitor screen?
[21:29:42] <tomp> :)
[21:29:44] <lerneaen_hydra> I'm just going to hose it and reinstall, takes under 15 minutes
[21:29:46] <lerneaen_hydra> :p
[21:30:12] <tomp> i'm ready to do that with this emc2-head, but i re-iterate...
[21:30:20] <tomp> i used cvs & built new emc2-head, failed, emc-debug reports 'could not find -sec HAL -var HALUI...' (and i chose the cfg from emc2-head not emc2)
[21:30:46] <cradek> which config?
[21:30:54] <tomp> axis
[21:31:20] <cradek> sim axis?
[21:31:37] <tomp> yes, sim axis, i can get the full path,,,
[21:31:46] <cradek> hmm that works for me...
[21:32:12] <tomp> and tried this too scripts/emc configs/sim/axis.ini with same result
[21:32:26] <tomp> uh, tha wouldnt work...
[21:32:28] <lerneaen_hydra> cradek, is there any easy way to make shortcuts to EMC for a certain config? IIRC you were discussing adding that a long time ago
[21:32:43] <cradek> yes
[21:32:54] <tomp> yes it should ( just doesnt)
[21:33:10] <tomp> sorry
[21:34:04] <jepler> tomp: some of those messages are not fatal. can you put everything from that window in pastebin please?
[21:34:59] <jepler> lerneaen_hydra: in 2.1 there's a checkbutton on the configuration selector that says "Create Desktop Shortcut". If you check it then hit OK, it makes a shortcut file in your Desktop directory, which should be easy for you to access
[21:35:08] <jepler> tomp:
http://pastebin.ca/ if you're not familiar with it
[21:36:05] <lerneaen_hydra> jepler, cool, does it autoload a certian selected config too?
[21:36:35] <jepler> lerneaen_hydra: yes, that "Create Desktop Shortcut" is a shortcut to *that* configuration file
[21:36:44] <lerneaen_hydra> nice
[21:37:10] <lerneaen_hydra> IIRC that was one of the harder things to add when I was über new to emc
[21:37:32] <lerneaen_hydra> except for that mystical thing called HAL of course ;)
[21:37:36] <jepler> tomp: in fact, I get that message (could not find .. HAL ...) in my debug file too, even when emc runs OK
[21:39:02] <jepler> lerneaen_hydra: in 2.1 it will also tell you that you should use a copy of the sample configs, and offer to make the copy for you automatically
[21:39:06] <tomp> reading back,, will try, i was just backing up ready to breakout the weedwhacker :)
[21:39:10] <lerneaen_hydra> hey, nice
[21:39:26] <lerneaen_hydra> lots of people seem to chmod the original files
[21:40:05] <cradek> our packages deal with that as well as possible, but it's still kind of a bad idea
[21:40:52] <ATAT> well
[21:41:15] <ATAT> I'm going to try using a 5+ year old laptop as EMC controller first.. then use a desktop, this way I dont have to lug around a CRT
[21:41:20] <ATAT> hopefully its 5v output on the parallel
[21:41:55] <cradek> ATAT: check it first!
[21:42:24] <ATAT> how? the scope probes I have dont go far enough in
[21:42:45] <ATAT> and I'd need to generate signal right? or is one of the pins always logic high
[21:42:46] <SWPadnos> carefully use small paperclips
[21:42:58] <cradek> some animals use a blade of grass to get ants out of an anthill...
[21:43:04] <SWPadnos> you can run through the HAL tutorial and set up a square wave generator if you like
[21:43:09] <cradek> I saw it on TV once
[21:43:26] <SWPadnos> various apes - small sticks and things ...
[21:43:34] <cradek> yeah that
[21:43:38] <ATAT> mmm that would work
[21:43:39] <tomp> jepler: collected, now off to pastebin... (blade of grass :)
[21:43:58] <ATAT> is there any way to force one of the pins to latch to logic high so I dont even have to power up the scope?
[21:44:15] <cradek> pins 2-9 will latch whatever you printed to them last
[21:44:21] <lerneaen_hydra> ATAT, um, have you done a RT test on the laptop?
[21:44:31] <ATAT> RT test? whats that
[21:44:47] <lerneaen_hydra> laptops usually don't work well with realtime stuff
[21:44:56] <ATAT> ah
[21:45:13] <tomp> http://pastebin.com/846697 that wasnt hard
[21:45:16] <lerneaen_hydra> not well at all. some however work reasonably
[21:45:41] <cradek> tomp: insmod: error inserting '/home/tomp/emc2-head/rtlib/rtapi.ko': -1 Invalid module format
[21:45:50] <ATAT> hmm alright. I'll go assemble the desktop while I wait then
[21:45:59] <tomp> yah, saw that, dunno what to hit tho
[21:46:09] <cradek> did you make clean like jepler said to?
[21:46:18] <jepler> tomp: does "uname -a" say you're using the right kernel?
[21:46:29] <cradek> oh this was a new checkout wasn't it
[21:46:38] <cradek> too many things going on at once today
[21:46:45] <tomp> Linux cncbox.titsnoir.org 2.6.15-magma #1 Fri Jun 9 20:51:19 EEST 2006 i686 GNU/Linux
[21:47:00] <jepler> run "dmesg" and it will have some additional information about the insmod error
[21:48:41] <jepler> the last dozen lines or so are probably all you need to look at
[21:49:00] <anonimasu> hello
[21:49:02] <tomp> uh, lookin, but this .12 vs .15 version , is that a problem?
[21:49:27] <tomp> rtapi: version magic '2.6.12-magma 586TSC gcc-3.4' should be '2.6.15-magma 586TSC gcc-4.0'
[21:49:39] <jepler> 2.6.12 would indicate that you've got something left over from the old "breezy" kernel
[21:49:52] <jepler> 2.6.12 was the kernel version of "breezy", 2.6.15 is "dapper"
[21:51:18] <tomp> ok, it's going in the disposal now, but thanks for the help, "i dont believe in upgrades i dont believe in upgrades i dont belive in upgrades" ( c.lion )
[21:53:28] <cradek> he was almost done...
[21:53:53] <lerneaen_hydra> right, now I can bug you instead
[21:54:36] <lerneaen_hydra> apparently my bios reorders the drives on bootup, becuase what should be my last drive becomes hd0,0
[21:54:54] <lerneaen_hydra> also sda1 seems to contain the kernel image
[21:55:08] <lerneaen_hydra> because the bootloader comes up and looks all nice
[21:55:28] <lerneaen_hydra> however it dies after a while and throws a hissy fit
[21:55:33] <SWPadnos> this may be one of the situations in which disklabels are useful
[21:55:39] <lerneaen_hydra> disklabels?
[21:56:38] <SWPadnos> yep. you can tag disks with a name, not just a number
[21:56:51] <SWPadnos> it stays when the disks are reordered
[21:57:07] <lerneaen_hydra> !!!
[21:57:07] <SWPadnos> other than that, I don't know anything about them (like how to set things up with them)
[21:57:11] <lerneaen_hydra> ME WANTS
[21:57:23] <SWPadnos> have I mentioned #grub recently? ;)
[21:57:34] <lerneaen_hydra> hmm, they didn't mention that
[21:57:53] <SWPadnos> possibly because you need to reformat or something ;)
[21:58:08] <cradek> man tune2fs
[21:58:20] <lerneaen_hydra> oh noes
[21:58:24] <lerneaen_hydra> reformat?
[21:58:29] <cradek> man tune2fs
[21:58:45] <SWPadnos> or you can use tune2fs, I think :)
[21:59:07] <lerneaen_hydra> hey
[21:59:07] <lerneaen_hydra> cool
[21:59:40] <cradek> but that doesn't help grub really - that's for the kernel mounting the root
[21:59:47] <skunkworks> I have a problem with my ntfs drive...
[21:59:51] <skunkworks> ;)
[22:00:00] <SWPadnos> I thought grub could also use disklabels
[22:00:05] <lerneaen_hydra> cradek,, thats what I have having issues with that
[22:00:07] <lerneaen_hydra> err
[22:00:16] <lerneaen_hydra> thats what I have having issues with
[22:00:20] <SWPadnos> skunkworks, try #ntfs :)
[22:00:21] <cradek> skunkworks: try #fu^H^H^H #ntfs
[22:00:56] <lerneaen_hydra> skunkworks, try #bukkak^W #ntfs2ext3
[22:01:55] <SWPadnos> try #OMFG_not_teh_windoze^Wntfs
[22:02:07] <skunkworks> my programmer bought himself a levitron
http://www.levitron.com/ pretty cool. he spent a good 4 hours at home trying to get it working. we spent a good 15 minutes and got it.
[22:02:16] <SWPadnos> they're cool
[22:02:25] <SWPadnos> my wife got me one years ago
[22:02:32] <SWPadnos> or was that my girlfriend way back then>
[22:02:34] <SWPadnos> ?
[22:03:08] <skunkworks> I would not ask
[22:03:21] <skunkworks> ;)
[22:03:22] <lerneaen_hydra> be glad A-L-P-H-A is not here
[22:03:29] <SWPadnos> heh - I was just running that scenario through my mind
[22:04:25] <SWPadnos> interesting - mine is a rectangular block of wood, not that UFO-like plastic thing
[22:05:11] <skunkworks> took us a bit to get the right weight - then leveling the magnetic field.
[22:05:18] <skunkworks> he had it way to heavy
[22:05:24] <SWPadnos> yeah, leveling is pretty important
[22:05:38] <SWPadnos> hmmm - I have a machinists level now - that may be easier :)
[22:05:59] <skunkworks> na- we found the actuall base was pretty crocked.
[22:06:04] <anonimasu> oh..
[22:06:06] <skunkworks> crocked?
[22:06:08] <anonimasu> SWPadnos: good luck :D
[22:06:17] <anonimasu> * anonimasu has bad thoughts about machinist levels
[22:06:21] <SWPadnos> heh
[22:06:32] <lerneaen_hydra> hmm
[22:06:37] <skunkworks> we have one - .0005 per division.. works great
[22:06:47] <skunkworks> per foot
[22:06:52] <SWPadnos> same as mine
[22:06:57] <anonimasu> hehe..
[22:07:03] <anonimasu> I think that's the same as mine
[22:07:07] <lerneaen_hydra> funny, tune2fs complains that it can't find a valid filesystem superblock
[22:07:06] <SWPadnos> cheap, but it works
[22:07:26] <SWPadnos> the high precision Starretts are 0.0001/foot/div, I think
[22:07:39] <skunkworks> I think we ended up having to calibrat ours. it was a bit off
[22:07:44] <skunkworks> callibrate
[22:07:52] <skunkworks> whatever
[22:07:56] <SWPadnos> yep. or you can use the "rotate 180 deg" trick
[22:08:12] <lerneaen_hydra> oh, nm
[22:08:13] <skunkworks> right - we did the rotate 180 trick to calibrate it.
[22:15:22] <lerneaen_hydra> hey!
[22:15:27] <lerneaen_hydra> labels did the trick!
[22:15:44] <SWPadnos> cool!
[22:15:46] <skunkworks> these guys are great :)
[22:15:46] <lerneaen_hydra> boots nicely now!
[22:15:59] <lerneaen_hydra> thanks for all the help :D
[22:29:20] <eholmgren> I had big problems with one of my ntfs drives
[22:29:35] <eholmgren> mainly that it's still running win2k
[22:34:26] <skunkworks> bbl
[22:44:30] <cradek> I identified which part on my car makes the funny noise, but I have no idea what it is
[22:44:38] <cradek> I even have the service manual
[22:44:54] <lerneaen_hydra> a stowaway maybe?
[22:45:16] <cradek> no, it's attached
[22:45:28] <cradek> it even has icons on it, but I don't understand them
[22:45:55] <cradek> I guess I only speak english, not "icon"
[22:45:59] <ATAT> is there any way to upconvert 3.3v parallel to 5v parallel? the breakout boards I see are to terminal blocks
[22:46:57] <cradek> ATAT: any breakout board with optoisolators will probably do that, but it will be expensive
[22:47:21] <ATAT> hmm sighs. im guessing it'd still be cheaper than buying a new hd and cd drive
[22:47:35] <ATAT> checked the old desktop and those two components are dead. sighs.
[22:47:47] <SWPadnos> ATAT: there are several companies that have buffered boards:
[22:48:17] <SWPadnos> http://www.pmdx.com http://www.cnc4pc.com http://www.campbelldesigns.com
[22:48:30] <cradek> wow, there's an access hatch to get to the fuel pump in this car. yay german engineers
[22:49:10] <ATAT> SWPadnos: are there any that have parallel port out and not terminal blocks?
[22:49:32] <SWPadnos> one of the PMDS boards has a centronics connector on one side and a DB-25 on the other
[22:49:40] <SWPadnos> you'd use a standard printer cable with it
[22:49:44] <SWPadnos> PMDX, that is
[22:49:59] <SWPadnos> the PMDX-103, I think
[22:50:10] <ATAT> oh cool
[22:50:24] <ATAT> but expensive.. sighs. I guess I'll go buy a cd drive and a harddrive then
[22:50:27] <SWPadnos> of course, it's about the same priceas the Xylotex board ;)
[22:50:32] <cradek> ATAT: have you confirmed that is the problem?
[22:50:46] <cradek> ATAT: buying things before you're sure is foolish
[22:51:00] <ATAT> cradek: I've nothing here that can output 5v parallel
[22:51:14] <ATAT> and I know for sure that the controller wont work with what I have
[22:51:14] <cradek> no working desktop at all? borrow one?
[22:51:19] <cradek> ok
[22:51:24] <ATAT> well I have a working desktop, just no parallel port =/
[22:51:29] <cradek> just don't want you to waste money based on my guess
[22:51:49] <ATAT> well a desktop is better for this I hear
[22:51:51] <ATAT> so might as well
[22:51:56] <cradek> ah, what a pain
[22:52:21] <cradek> that is definitely true
[22:57:59] <cradek> "Every 2 years: Check roll-over protection with convertible top open."
[22:58:16] <lerneaen_hydra> O.o
[22:58:26] <lerneaen_hydra> sounds rather destructive
[22:58:49] <cradek> yeah, it must mean something I don't know how to do
[23:23:31] <jepler> lerman: while you're fixing the interpreter, you should see if you can make 'G0 X[-#1]' work (that is, like 'G0 X[0-#1]')
[23:25:19] <SWPadnos> even if that doesn't work (which I would consider a bug), the same can be done with G0 X-[#1], can't it?
[23:25:36] <SWPadnos> or even X-#1
[23:25:41] <lerneaen_hydra> 'night
[23:25:55] <SWPadnos> see you LH
[23:28:36] <lerman> jepler: I'll take a look.
[23:31:54] <roltek> what does x-[#1] mean
[23:32:06] <SWPadnos> move to negative #1 in X
[23:32:12] <SWPadnos> it's just a unary minus
[23:32:59] <roltek> why would you do that and not just wright g0 x-1.0
[23:33:16] <SWPadnos> #1 is a variable - it could contain 347.9
[23:33:44] <aip_tom> SWPadnos: here's an even cheaper keyboard switch interface:
http://groovygamegear.com/webstore/index.php?main_page=product_info&cPath=76_80&products_id=200
[23:34:02] <roltek> i have never seen programing like that in shop
[23:34:20] <SWPadnos> oh yeah - I saw that. what jymmm was looking for was a serial to PS/2 adapter though :)
[23:34:20] <aip_tom> $21, USB too.
[23:34:54] <aip_tom> also $21:
http://groovygamegear.com/webstore/index.php?main_page=product_info&products_id=200
[23:35:18] <SWPadnos> roltek, lots of code you wouldn't want to write here: ;)
http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Oword
[23:36:49] <jepler> SWPadnos: X-[#1] seems to be an error too
[23:36:59] <jepler> "Bad number format"
[23:37:15] <SWPadnos> I figured the brackets aren't needed, I gues they aren't allowed ;)
[23:37:25] <SWPadnos> does X-#1 work?
[23:37:46] <jepler> no
[23:38:05] <jepler> there is no "unary minus operator", just binary minus and negative number literals
[23:38:34] <SWPadnos> ok - that last one is a casualty of assuming that anything beginning with a - (or a +, probably) is a number
[23:38:50] <jepler> someone asked me in e-mail a few months ago about these "keywiz" devices. I'm not sure if he got them working or not
[23:38:58] <jepler> (I mean, working with his emc2)
[23:39:02] <SWPadnos> :)
[23:39:26] <jepler> it's not a bug, in the sense that it's not a deviation from the language spec
[23:39:56] <jepler> it's a bug in the sense that almost all languages have unary minus
[23:40:05] <roltek> what kind of machine shops use this kind of programming
[23:40:19] <SWPadnos> in a sense, the variable replacement should be done the same as halcmd - look for any #, paste in the variable value, then pass the result to the interpreter
[23:40:32] <SWPadnos> ones that use subroutines?
[23:53:47] <tomp> hello
