#emc | Logs for 2010-05-01

[00:21:23] <pfred1> * pfred1 relapsed today :(
[00:21:59] <pfred1> I need professional help I'm telling you!
[00:22:28] <MattyMatt> me too. I need an investment banker
[00:22:47] <pfred1> I have a tool fetish or something sometimes i just can't help myself
[00:23:20] <pfred1> today I bought this old beat to death B&D polisher and two crappy old drill indexes at a yard sale
[00:23:43] <pfred1> why? why? WHY!
[00:24:26] <andypugh> A friend of mine had a good catch-all excuse "They are manufactured articles and it is wrong to let them go to waste"
[00:24:47] <pfred1> andypugh yeah but someone else could take care of them why me?
[00:25:23] <MattyMatt> how many $?
[00:25:27] <pfred1> you should see this polisher yo ucould anchor a rowboat with it
[00:25:36] <pfred1> $10 for the lot
[00:25:52] <MattyMatt> old B&D were cast in good potmetal
[00:25:58] <pfred1> but it sort of matches this Dayton angle grinder i have
[00:27:26] <pfred1> I guess i felt sorry for the old guy I saw him on TV once
[00:27:53] <pfred1> on an installment of Travels with Charlie on WBOC
[00:28:22] <pfred1> guy ran a TV repair place up the road from me and he's going out now
[00:29:01] <MattyMatt> shame he can't sell you his experience
[00:29:20] <pfred1> MattyMatt what good would it be who uses tube TVs anymore?
[00:29:47] <MattyMatt> valve amps are coming back
[00:29:57] <pfred1> not for me they aren't
[00:30:36] <MattyMatt> broadcast satellites still use a big tube
[00:31:08] <MattyMatt> and then there's electron microscopes, pretty much a crt
[00:33:14] <mikegg_> http://www.youtube.com/watch?v=bI06lujiD7E
[00:33:17] <mikegg_> cool
[00:37:35] <MattyMatt> when robots did that in comics, I always thought the artist was stupid & lazy
[00:38:54] <pfred1> hmmm they didn't archive the story
[00:45:58] <ewlsey> anyone know about dynapath delta controls?
[00:49:53] <pfred1> ewlsey are you being obtuse?
[00:50:08] <ewlsey> probably
[00:50:41] <ewlsey> I'm looking at a lathe with a delta 40 control, I have no experience with them
[00:50:42] <pfred1> i sos cel es
[01:09:07] <MattyMatt> ewlsey. the solution you'll find here is, rip it out down to the motors and install emc2 :)
[01:09:41] <ewlsey> MattyMatt, I thought I would ask
[01:09:53] <MattyMatt> is it this kind of vintage? http://www.youtube.com/watch?v=2DJWCIZsX6E
[01:09:57] <ewlsey> I don't think it needs to be retrofit
[01:11:20] <ewlsey> that is relic, what I hate about the knee mills is that the controls never seem to control spindle speed
[01:12:44] <MattyMatt> look at 1:55 in that vid :) casette interface
[01:12:45] <ewlsey> here is the machine: http://cgi.ebay.com/SouthBend-Magna-1220-CNC-Lathe-/120554409834?cmd=ViewItem&pt=BI_Lathes&hash=item1c119a4f6a#ht_500wt_1182
[01:13:18] <pfred1> * pfred1 is reminded of a line in the movie Used Cars ....
[01:13:22] <Valen> ewlsey shine
[01:13:28] <Valen> shiny
[01:13:59] <pfred1> that price is too f*cking high we're gonna blow the sh*t out of it!
[01:14:33] <pfred1> I have to see that movie again now
[01:14:38] <ewlsey> MattyMatt, it remember casette interface on the old Emco machines at school
[01:15:33] <andypugh> Casette? There's posh! The wire-eroder at Sheffield Uni was paper tape. (and that was in 2001)
[01:15:53] <pfred1> andypugh that is such a bad idea on so many levels
[01:16:34] <andypugh> Eventually they wired a long serial link to it from a PC.
[01:16:35] <ewlsey> Valen, I like it but it has no tailstock, not sure if that is a deal breaker
[01:16:46] <Valen> pretty expensive
[01:16:46] <MattyMatt> there were times using cassettes I wished they were paper tape.
[01:17:14] <andypugh> Expensive and no tailstock? Sounds largely useless
[01:17:45] <andypugh> Is there scope for a tailstock?
[01:17:55] <ewlsey> it doesn't seem too bad, the lathe is only 11 years old
[01:18:13] <MattyMatt> there's scope for an autochanger, but the ad isn't clear if that's included in the price
[01:18:30] <pfred1> ewlsey yeah whats 11 years of grinding carbide going ot do to a machine?
[01:18:37] <Valen> i just meant its a fair bit of coin, if your using it to make money its probably ok
[01:19:11] <pfred1> Valen you think counterfeiting on it is worthwhile do you?
[01:19:15] <ewlsey> pfred1, it will probably wear it out
[01:19:40] <Valen> if your making krugerands probably ;-P
[01:19:44] <ewlsey> i talked to the seller, said it saw little use
[01:19:58] <pfred1> I remember the carbide scam a while back
[01:20:05] <ewlsey> ?
[01:20:06] <pfred1> where they painted bricks of carbide gold
[01:20:37] <pfred1> or plated them or something
[01:20:51] <Valen> tungsten and gold have the same density
[01:21:01] <ewlsey> maybe I'll save my penies for a used Haas... Decisions
[01:21:04] <andypugh> mattymatt: Are you talking about using casette tapes for Z80 assembler development by any chance? That was hideously slow wasn't it? Paper tape would have been faster I am sure. (And you could spot the mistakes on the medium while you were waiting)
[01:21:15] <pfred1> Valen yeah but just punching it with a punch should be a dead giveaway I'd imagine
[01:21:21] <Valen> a deep gold plate on a tungsten inner and its pretty much indistinguishable from the real thing
[01:21:42] <MattyMatt> or uranium, same density again
[01:21:54] <Valen> actually you would use "heavy metal" 99% tungsten powder with a polymer binder
[01:22:05] <pfred1> well isn't uranium worth more than gold by weight?
[01:22:14] <Valen> nah non enriched is pretty cheap
[01:22:21] <MattyMatt> yeah that strikes me as more of a way to smuggle uranium
[01:22:23] <Valen> they use it for weights in aircraft
[01:22:33] <andypugh> I have speculated what they would do if you dumped a shipping-container sized block of Osmium somewhere. It would be too heavy to lift (20,000 tons) and too hard to cut...
[01:22:44] <pfred1> can't be healthy and why would yo uwant to weight an aircraft?
[01:22:53] <MattyMatt> keel
[01:23:09] <Valen> counterweights for ailerons and such
[01:23:11] <MattyMatt> 747 has one to keep it upright
[01:23:26] <pfred1> andypugh we had a truckload of heavy mat delivered to a job abd i so wanted to pour it with concrete
[01:23:57] <andypugh> heavy mat?
[01:24:07] <pfred1> andypugh concrete reinforcing
[01:24:17] <pfred1> its called mat
[01:24:17] <MattyMatt> that helps explain why the 747s punched clean through the WTC
[01:24:36] <pfred1> MattyMatt not really it was more like cheese hitting a grater
[01:24:48] <Valen> probably more to do with buildings mainly consisting of air and 747's mainly of liquid
[01:25:01] <pfred1> MattyMatt there was 12 inch H beams between every window
[01:25:07] <MattyMatt> but the keel probably acted like a huge penetrator
[01:25:35] <pfred1> my buddy Sammie wrecked that job
[01:25:54] <pfred1> the big orange machines you saw in the video those were his
[01:26:13] <Valen> I don't think a 747 would have a "keel" of DU or some such
[01:26:27] <andypugh> I know somebody at work who is convinced that the WTC was still at 1200 (C) three days later. No amount of telling him that it would have been incandescent at that temperature will convince him otherwise.
[01:26:53] <pfred1> andypugh there were hot spots in the debris that lasted for weeks
[01:26:59] <andypugh> It is part of his belief that it was all a conspiracy and blown up by CIA nukes or something.
[01:27:03] <Valen> andypugh, bits of it probably were
[01:27:10] <pfred1> andypugh there was a substation in one of the towers you know with oil filled transformers etc
[01:27:22] <andypugh> Not 1200C though, nothing stays that hot for that long.
[01:27:23] <Valen> lol you would hit 1200C from a chimny effect burning the paper
[01:27:34] <Valen> thats not paticularly hot really
[01:27:36] <pfred1> andypugh dump that much concrete on it and it does
[01:27:45] <Valen> melt Al, soften steel,
[01:28:14] <pfred1> damned building was filled with paper plastic jet fuel it was a hell of a fire
[01:28:43] <pfred1> I watched it from the top deck of a parking garage i built
[01:29:22] <andypugh> Oh, absolutely. You probably have to hear the whole wierd theory to get the gist of why it doesn't stack up.
[01:29:23] <Valen> yeah, 747s have a "keel beam" but its a "lightweight" al/magnesium jobby
[01:30:28] <ewlsey> off topic, anyone have any ideas on DIY spindle balancing?
[01:30:57] <Valen> acceleromiter, magnet, hall effect sensor and a 2 channel oscilloscope QED
[01:31:19] <pfred1> ewlsey or you could be realistic and just statically balance it between two good bearings
[01:31:35] <ewlsey> I don't want to take it out of the mill
[01:31:47] <ewlsey> i have a scope but it is only one channel
[01:32:13] <Valen> got a seperate trigger?
[01:32:26] <ewlsey> I was going to try adding an eccentric mass (hose clampy maybe) and see if I can experiment a bit
[01:32:57] <ewlsey> I don't have an accelerometer, buy I might be able to borrow one from work...
[01:33:24] <Valen> you can get them pretty cheap from sparkfun or even sample from analog devices
[01:33:25] <ewlsey> I could use EMC as a scope could I not?
[01:33:35] <Valen> probably not fast enough
[01:33:44] <ewlsey> hmm
[01:33:47] <Valen> and you would need an analog input
[01:34:03] <Valen> I havent tried this btw so its at your own risk ;->
[01:34:16] <Valen> you might be able to use a microphone if theres no gears
[01:34:24] <ewlsey> the spindle has a hall effect sensor for speed output
[01:34:36] <ewlsey> it is timing belt drive
[01:34:42] <pfred1> ewlsey if the pulley has an inside depression I htink you'd be best off just throwing a weight in there and moving it about a bit
[01:35:07] <ewlsey> pfred1, i think that will be a good place to start
[01:35:22] <pfred1> ewlsey bit of gum or something to temp hold it
[01:35:23] <ewlsey> It probably won't take much to get it in balance
[01:35:36] <pfred1> ewlsey coins are nice weights too
[01:35:55] <ewlsey> I swapped pulleys, and now it has a nice buzz above 6000RPM
[01:35:55] <pfred1> pennies cut easily enough
[01:36:06] <pfred1> can you rotate the pulley?
[01:36:10] <ewlsey> yes
[01:36:20] <pfred1> thats what we did when tires were really out on rims
[01:36:22] <ewlsey> but i don't know if that will help
[01:36:26] <pfred1> just half twist the tire
[01:36:40] <ewlsey> hmm
[01:36:58] <pfred1> chalk the chaft and pulley of course
[01:37:00] <ewlsey> I can't decide if it is an imbalance, or just the timing belt cogging
[01:37:02] <pfred1> shaft even
[01:37:22] <ewlsey> chalk?
[01:37:30] <pfred1> sure to mark it
[01:37:33] <ewlsey> oh
[01:38:12] <pfred1> have you tried adjusting the belt tension?
[01:38:16] <ewlsey> yes
[01:38:34] <pfred1> timing belts don't have to be so tight you can get a note out of them usually
[01:38:50] <ewlsey> I know timing belts have some noise always, I just don't know how much
[01:39:08] <ewlsey> I used to have a tension guage...
[01:39:17] <andypugh> They don't have to make noise.
[01:39:44] <Valen> http://www.youtube.com/watch?v=bI06lujiD7E
[01:39:48] <ewlsey> andypugh, in my experience they hum as the air in the timing grooves is forced out by the belt
[01:39:54] <pfred1> andypugh of course they have to make some noise they're mechanical parts in motion
[01:40:32] <pfred1> they're noisier than V belts
[01:41:14] <andypugh> Yes, OK. But with the right fit of pulley to belt they can be very quiet. And with the wrong fit they can be "we can't sell this car, cancel production, run about in circles, fly around europe to meeting" DAMHIKIJKOK
[01:41:20] <MattyMatt> http://en.wikipedia.org/wiki/Boeing_747 doesn't mention DU keel, but does describe the 10 fighter plane carrier variant :)
[01:41:20] <ewlsey> valen, i remember doing inverted pendulum experiments in college, the professor would put a glass of water on top of the system and smack it, if you spilled it was a bad deal
[01:41:34] <MattyMatt> ...that never got made
[01:41:51] <pfred1> andypugh well yeah you almost need a PHD to get the right cogs and belt combination it seems
[01:42:06] <Valen> ewlsey, I like the "assist" mode
[01:42:13] <Valen> thats gotta be usefull for something ;->
[01:42:43] <andypugh> We have guys with PhDs for that sort of problem. Lots of them actually.
[01:42:59] <ewlsey> valen, check this out http://www.youtube.com/watch?v=0GBPTgTd8bE
[01:43:21] <pfred1> check this out: http://img443.imageshack.us/img443/754/tow007.jpg
[01:43:21] <ewlsey> I got to see one up close, it is pretty neat
[01:43:26] <Valen> looks like an outsized sex toy lol
[01:43:35] <pfred1> that was about as close as i wanted to be
[01:44:09] <ewlsey> the balance is not great, it never really stops moving around
[01:44:46] <pfred1> ewlsey oh it stopped moving by about picture 009
[01:45:02] <MattyMatt> "massive potential" = runs out of batteries while you're acting smug with a trayful of beer in the pub
[01:45:15] <Valen> if its top heavy it can never really stop
[01:45:54] <pfred1> Valen I'm sure thats what the guys who set the counterweights at the top of the towers said
[01:46:05] <ewlsey> valen, if you think about, you can't really sit perfectly still, so it moves around to try to keep you balance. I like that is moves in to axes
[01:47:51] <ewlsey> wow I barely sound coherent...
[01:48:42] <pcw_home> great, something that wastes power doing nothing
[01:48:44] <pfred1> ewlsey feel sort of like this do you? http://img52.imageshack.us/img52/8474/tow008.jpg
[01:49:05] <MattyMatt> all that energy goes into breaking my chairs
[01:49:22] <MattyMatt> I'm not a static load
[01:50:26] <ewlsey> finished my coolant mister: http://www.cnczone.com/forums/showthread.php?t=102934
[01:51:00] <pfred1> I have a Bijur
[01:52:36] <ewlsey> mine is "fogless", though I never found that too big a big deal on the standard misters
[01:52:55] <MattyMatt> have some more coolant, and don't call me mister
[01:53:01] <pfred1> well it claims if you use their goop its foggless I donno it sits on a shelf here
[01:53:48] <pfred1> I use an oxyclean plastic squirt bottle and squeeze technique
[01:54:30] <ewlsey> good enough for most work pfed1
[01:54:46] <ewlsey> I wanted to pretend I had a realy machine (I don't)
[01:54:50] <pfred1> ewlsey I haven't found anything yet I can't squeeze my way through
[01:55:25] <pfred1> though I am planning on upgrading to an empty dishwashing detergent bottle
[01:55:54] <pfred1> its on my bench now I just haven't filled it with coolant yet
[01:56:10] <ewlsey> lol
[01:56:22] <pfred1> hey I've seen to many rancid sumps in my time
[01:56:37] <pfred1> no thanks for that crap in my garage!
[01:57:45] <pfred1> my Bijur is pretty sweet though solenoid operated and everything
[01:59:18] <ewlsey> i have the solenoid on mine, very handy
[01:59:49] <pfred1> ewlsey thats how my preferred coolant system works by the handy method yes
[02:00:09] <pfred1> if you want more you squeeze your handy a little harder
[02:02:48] <pfred1> this is it: http://www.cnccookbook.com/img/MillStuff/BIJUR1.jpg
[02:04:07] <pfred1> cept i think mine is a sickly greenish color
[02:04:18] <MattyMatt> looks like a fridge pump to me
[02:04:25] <pfred1> can't remember i haven't seen it in years now
[02:04:47] <pfred1> MattyMatt really its like an FRL unit
[02:04:59] <andypugh> Looks very american to me. Being a large lump of cast metal. We are afraid to use metal over this side of the sea.
[02:05:26] <andypugh> It would be plastic and/or pressed tin over here.
[02:06:03] <pfred1> andypugh thing is like a million years old
[02:06:10] <MattyMatt> we need the metal for the war effort. and if there's no war we need one for the metal effort
[02:06:55] <MattyMatt> * MattyMatt commandeers andypugh's railings
[02:07:31] <pfred1> andypugh i saw the stats once Americans use 5 tons of metal a year a piece
[02:08:00] <andypugh> Crikey! How much of that is Coke tins?
[02:08:31] <pfred1> I donno how they spread it out but it seemed awfully crazy to me maybe they took national tonnage and just divided it by population I donno
[02:09:24] <pfred1> everytime i look at my store of metal stock I just can't see it myself
[02:09:33] <pfred1> and i collect the stuff!
[02:10:01] <andypugh> Aye, I reckon most of us here have a heavier than average metal habit.
[02:10:05] <ewlsey> cars, appliances, building
[02:10:18] <MattyMatt> bridge to Key West
[02:10:21] <ewlsey> everything is steel
[02:10:30] <pfred1> MattyMatt ever riden on it?
[02:10:39] <MattyMatt> nope, just seen pics
[02:10:50] <pfred1> I have its cool its like you're super power boating
[02:10:51] <telmnstr> quick question. When I try to configure it, it says checking for RT dir.. configure: error: RT not found
[02:11:03] <telmnstr> What is it looking for? I specified the RT kernel header directory but it didn't seem to like that
[02:11:10] <pfred1> Real Time
[02:11:53] <telmnstr> I upgraded this ubuntu 9.10 to a real time kernel
[02:11:54] <pfred1> configure has options you can feed it the directory on the ./configure line
[02:12:03] <telmnstr> yea, I fed it the real time kernel
[02:12:05] <pfred1> ./configure --help should spit out info
[02:12:11] <telmnstr> but the non-descript error didn't really say what it was
[02:12:25] <MattyMatt> telmnstr: page 1 of the emc2 docs say don't do that
[02:12:53] <telmnstr> Says don't do what?
[02:12:58] <telmnstr> only run it on 8.10 ?
[02:13:44] <ewlsey> yes
[02:13:50] <telmnstr> hmmm page #1 of the EMC docs talk about ubuntu 6.06
[02:13:53] <telmnstr> heh
[02:14:33] <ewlsey> in BIOS can you change the kernel to the RT kernel?
[02:14:38] <telmnstr> the dox talk about running it on 9.10
[02:14:49] <pfred1> telmnstr I want to try the puppy cnc
[02:14:53] <telmnstr> No, you set the grub loader in unbuntu to not be lame and show the list of kernels
[02:15:02] <andypugh> telmnstr: If you have changed the kernel you will need to re-compile EMC2.
[02:15:17] <telmnstr> I haven't compiled EMC2 yet. I'm just trying to compile it the first time
[02:16:15] <pfred1> andypugh I hope EMC doesn't stick with Ubuntu
[02:16:41] <andypugh> telmnstr: http://neo-technical.wikispaces.com/emc2-ubuntu
[02:16:53] <MattyMatt> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl in big letters at the top
[02:17:14] <andypugh> That worked for me as a walkthrough, though not with 9.10
[02:17:50] <MattyMatt> I have installed RTAI in my debian lenny, in anticipation :)
[02:17:59] <telmnstr> Im not upgrading, I installed 9.10
[02:18:13] <telmnstr> Dont have 8.10 :-)
[02:18:23] <MattyMatt> I got the live CD
[02:18:32] <andypugh> telmnstr: Does RTAI work? Can you run the latency test?
[02:18:33] <MattyMatt> installed from that
[02:18:36] <pfred1> MattyMatt what do you feed it?
[02:18:39] <telmnstr> That's what we just discovered
[02:18:49] <telmnstr> busted ass ubuntu, the rt kernel is a different version from the rtai stuff
[02:19:00] <telmnstr> like they totally mismatch
[02:19:43] <andypugh> Yeah, I am not at all sure how you are meant to figure our which RTAI tarball goes with which kernel.
[02:19:45] <telmnstr> moving to rtai kernel
[02:20:02] <telmnstr> oh ubuntu is just debian, apt-cache search rtai
[02:20:51] <MattyMatt> ah so the RTAI package is only half the story?
[02:21:03] <telmnstr> I installed rt kernel but no rtai
[02:21:14] <telmnstr> now I'm moving to a kernel tied to the rtai stuff
[02:21:25] <telmnstr> Anyone compile this on FreeBSD?
[02:21:43] <cradek> did you patch your kernel with the patch that comes in the rtai distribution?
[02:21:56] <telmnstr> I'm using the ubuntu distribution poop
[02:22:22] <cradek> regardless of where you got it, is it an rtai-patched kernel?
[02:22:29] <telmnstr> yes
[02:22:35] <telmnstr> but it overwrote grub with lilo
[02:22:38] <telmnstr> teeehee
[02:22:55] <cradek> that's pretty fascinating
[02:23:02] <telmnstr> yea
[02:23:35] <pfred1> coolcnc looks pretty cool
[02:26:35] <telmnstr> INSTALL IT ALL!
[02:26:47] <telmnstr> Once this machine doesn't boot anymore I'll get off my ass and burn the live CD
[02:27:22] <cradek> seems like you're well on your way
[02:27:51] <telmnstr> brb... maybe
[02:32:18] <andypugh> Night all
[02:36:18] <telmnstr> heh okay I'll go to a livecd
[02:36:36] <telmnstr> rtai drivers aren't in place so the kernel can't see the disk
[02:36:40] <telmnstr> not worth time to keep messing with it
[02:39:21] <MattyMatt> I burned 5 copies of the liveCD
[02:39:37] <MattyMatt> turned out to be bad RAM :)
[02:41:42] <Valen> ouch
[02:42:59] <MattyMatt> no loss in the end. it prompted me to revive a better machine
[02:43:14] <MattyMatt> and spare CDs are handy
[02:44:19] <MattyMatt> I gotta ditch the usb keyb & mouse tho, if what I hear about legacy USB is true doing terrible things to latency
[02:45:17] <MattyMatt> I heard that from a usb 3 hypesperson too. he said usb 2 can use 30% cpu
[02:48:03] <MattyMatt> ah I'll care when it's the thing stopping me milling :) right now I gotta do a bezier integrator
[02:48:17] <MattyMatt> in python
[02:48:50] <MattyMatt> * MattyMatt asks google to code it for him
[02:58:00] <MattyMatt> google obliges, via gamedev
[03:38:39] <cbayly> hi! I'm just starting to look at building/(buying?) a desktop sized CNC for etching PCB's, and I have a few questions about what I can expect for tolerances... In particular what would be the narrowest trace widths and trace to trace gaps I could expect? I realize this has everything to do with the hardware, but I'm not even sure if I'm looking at the 15mil range, or the 5mil range. My end goal is to mill PCB's for surface mount components, SOIC, QFN, & 1206
[03:38:44] <MattyMatt> yehaw. I made some gcode
[03:39:21] <cradek> cbayly: on a good tight machine with antibacklash provision I can easily cut 8mil traces
[03:40:02] <cradek> on a sloppy machine (dremel tool or no anti-backlash) you'd be lucky to do the 15mil.
[03:40:34] <cradek> cbayly: a tiny bit of information on my site: http://timeguy.com/cradek/cnc/pcb
[03:41:13] <cbayly> cradek: Oh cool, that's exactly what I was hoping to hear :) I have a board with a MSP430 and a bit of glue that was originally done at PCBPool @ 6mil traces and a 6mil gap. Awesome, yeah, I'm hoping to build a nice tight machine, at the expense of speed and what not if need be.
[03:41:17] <cbayly> Thanks, I'll check it out
[03:41:43] <MattyMatt> look at the Sable 2015 before building
[03:42:04] <cradek> if you use a V tool you need careful depth control, which means a flat table and a good way to hold the pcb flat
[03:42:21] <cradek> I bet vacuum is best -- but I use double stick carpet tape
[03:42:27] <MattyMatt> if 8" * 6" is big enough, that's a tight machine by all accounts. pretty and good value too
[03:44:23] <MattyMatt> mill a pcb sized pocket in mdf
[03:44:48] <cradek> MattyMatt: is that thing all aluminum?
[03:45:03] <cbayly> Ahh okay, I have a friend who has built a 5' x 6' table who is using a vaccum setup (with a permiable wood between the vaccum and the board, he's been milling 0.1" type boards and larger traces
[03:45:28] <MattyMatt> afaik yeah, sides of the frame are sold Al at least
[03:45:37] <MattyMatt> ask tom3p, he has one
[03:45:56] <cradek> MattyMatt: ah I bet it works pretty well then.
[03:46:29] <cbayly> he's been saying he really likes the vacuum for holding things down, but I'm not sure if I'll be going that way... probably depends on if the machine is inside the house or in the garage, and $$ :)
[03:46:55] <MattyMatt> spraymount
[03:47:24] <cradek> do you have a good air compressor? I'd be tempted to try an air spindle (die grinder). (note, I haven't tried this)
[03:49:40] <cbayly> I don't at the moment, but might be looking at it... The guy with the large CNC also has a 18"Hg vacuum for holding the work... (Some people have ALL the toys :)
[03:50:27] <MattyMatt> what? a piston full of Hg for instant quiet static vacuum?
[03:50:58] <cbayly> Ah, no, that was just the measurement he was using for the suction on it... Hoping there's not Hg in it :)
[03:51:24] <MattyMatt> you've got me thinking tho :) same idea with water
[03:51:44] <MattyMatt> or bellows
[03:52:37] <MattyMatt> water would only work if your workshop is on the 4th floor
[03:52:39] <cbayly> Hehe, an earlier version of his machine is at http://hovercrafters.com/brendin/cnc/
[03:52:54] <cbayly> Heh yeah, or if you want to run a 40' pipe up on your property
[03:53:44] <cbayly> He's not running EMC, but I'm a little more insistant on not running windows if I can at all help it
[03:54:20] <cradek> oh do people still use windows? bizarre.
[03:54:41] <cradek> looks like a nice built-in vacuum. wood is messy otherwise.
[03:55:19] <MattyMatt> 18" Hg pressure is only 0.5 atmospheres. not a very hard vacuum at all. you could beat that with an inverted footpump I'd guess
[03:58:33] <MattyMatt> but yeah if you're running a suction for dust collection anyway, use that
[03:59:39] <MattyMatt> spraymount is pretty tacky tho
[03:59:59] <MattyMatt> 1 can lasts for ages
[04:01:32] <MattyMatt> I compromised with my machine, tried to make it strong enough for metalwork, big enough for woodwork, accurate enough for pcbs, and cheap enough for me :)
[04:02:26] <MattyMatt> I'm not sure how many of those goals I met
[04:03:28] <Dave911> logger_emc:bookmark
[04:03:28] <Dave911> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2010-05-01.txt
[04:06:00] <Srpski> Srpski is now known as Dannyboy
[04:56:13] <cbayly> thanks again for the tips, I'll be back, but I've got about a days worth of reading to catch up on now :)
[07:09:55] <ChanServ> [#emc] "This is the #emc channel - talk related to the Enhanced Machine Controller and general machining. Website: http://www.linuxcnc.org/, wiki at http://wiki.linuxcnc.org/"
[07:10:04] <ChanServ> [#emc] "This is the #emc channel - talk related to the Enhanced Machine Controller and general machining. Website: http://www.linuxcnc.org/, wiki at http://wiki.linuxcnc.org/"
[07:58:22] <sealive> morning
[07:58:33] <sealive> awallin: are you here ?
[08:10:16] <awallin> hello
[08:10:22] <sealive> morning
[08:10:43] <sealive> the visualisation tool ewindow does not close
[08:11:24] <sealive> the cutter diameter is this in inches?
[08:12:18] <sealive> cutter = cam.ConeCutter(0.118, math.pi/6) -> 3mm cone 60deg
[08:12:31] <awallin> either close the window, or hit enter in the python console (or you could omit the iren.Start() at the end if you don't want to rotate and zoom in the view)
[08:12:43] <awallin> there is no notion of units in ocl.
[08:12:58] <awallin> so scale your STL file to whatever units you want to use
[08:14:10] <sealive> what are the cc1 points
[08:15:02] <awallin> cutter contact points, i.e. points where the cutter made contact with the STL surface
[08:15:32] <awallin> it also has a type, VERTEX, EDGE, or FACET which tells you what part of a triangle was contacted
[08:16:24] <sealive> ok
[08:19:20] <sealive> i have to fiond out how i can kill this window
[08:20:54] <sealive> myscreen.Close() does not work
[08:21:29] <awallin> that is defined in lib/camvtk.py
[08:21:52] <awallin> and the VTK stuff is here: http://www.vtk.org/doc/nightly/html/annotated.html
[08:22:03] <awallin> you probably want to close the render window
[08:25:02] <awallin> sealive: try myscreen.iren.TerminateApp()
[08:26:04] <sealive> yes that workt but terminated also python idle
[08:26:42] <MattyMatt> I milled!
[08:26:58] <MattyMatt> with no hands on the jog buttons
[08:26:59] <sealive> :D MattyMatt
[08:27:39] <sealive> i saw yesterday that the repair team of the Golden gate brige also uses EMC2
[08:27:51] <sealive> with ther plasma cutter
[08:30:25] <sealive> awallin: normal cone engraver have a flat on the tip is yours sharp?
[08:32:54] <MattyMatt> cone = countersink = chamfering tool = sharp. bullnose = blunt
[08:37:07] <sealive> awallin: ther shoudt for international or global reasen be a picture of the shape on the google projekt page
[08:37:48] <MattyMatt> http://pastebin.com/ZH8fZVmQ
[08:38:14] <MattyMatt> that's the blender code for exporting beziers to G1s
[08:38:59] <awallin> sealive: a blunt cutter could be done, it would be a cylinder out to some radius, and then a cone after that
[08:39:31] <MattyMatt> truncated cone is easy mathematically
[08:40:07] <sealive> matty nice G-code out of blender
[08:40:36] <awallin> yes, but are there enough optimizations to justify writing yet another cutter-class, or should we make a generic CompoundCutter which can be composed of the basic cutters?
[08:40:43] <sealive> awallin: you shoud also connect to blenerdevelopers for your ocl
[08:41:33] <sealive> for normal users this will do the job
[08:41:47] <awallin> well there is very little functionality in ocl so far... I would like to get z-slice working and maybe the cutting simulation also
[08:42:18] <sealive> did you look at freemill?
[08:42:28] <awallin> maybe if you cut wood/foam with a router these parallel finish passes are enough
[08:42:42] <awallin> I maybe tried freemill a long time ago, don't remember
[08:43:03] <sealive> what os you are using
[08:43:24] <awallin> mostly ubuntu for development, but also win7 on laptop
[08:44:06] <MattyMatt> is ocl yours awallin?
[08:44:18] <sealive> it works perfect on XP with 50 postprocessors but only one tool shape
[08:44:34] <awallin> MattyMatt: in the sense that I have been working on it, yes. it's GPLd (unlike Heeks)
[08:45:29] <MattyMatt> ah. I don't like libs to be GPL
[08:45:52] <awallin> sealive: ocl compiles on Windows (I haven't tried), so someone could write a gui that did similar things like freemill
[08:46:16] <sealive> i try on wx
[08:47:01] <awallin> cool, do you have an OpenGL or other 3D rendering widget for wx?
[08:47:54] <MattyMatt> and I don't like wX because it forces you to use C++ :)
[08:48:19] <MattyMatt> that's the only reason I don't dive into Qt
[08:48:19] <awallin> there are python bindings for wx
[08:48:34] <awallin> and there are python bindings for qt also
[08:48:43] <MattyMatt> any C bindings for either? :)
[08:49:12] <awallin> I would think anytinh beyond "hello world" could/should benefit from oo
[08:49:50] <MattyMatt> you can do oop in C, with a different syntax
[08:50:19] <awallin> and the benefit of avoiding those nasty two plusplus signs is??
[08:50:45] <MattyMatt> avoiding reading Stroustrups book, that's the big one
[08:50:56] <awallin> ah..
[08:51:54] <MattyMatt> and always knowing what + means in the current context :)
[08:52:49] <awallin> some people get kicks out of redefining +
[08:53:15] <MattyMatt> yes, I'm sure they do
[08:53:35] <MattyMatt> I don't need that kind of paranoia while I'm coding :)
[08:54:13] <MattyMatt> now if it let you define new operators from the unicode set, I'd be game
[08:55:57] <sealive> awallin: you got mail
[08:56:21] <sealive> i got to go 1mai dinner
[08:56:53] <MattyMatt> m'aidez
[08:56:59] <sealive> we will talk about this Gui system
[09:02:02] <MattyMatt> I can't decide whether to do sim or interpreter next
[09:04:12] <MattyMatt> if ocl is all in python and you're putting your octree voxels in ocl, then I could probably get that in blender easily
[09:05:41] <MattyMatt> blender has it's own boolean ops, which work quite well apparently, but I'd hate to see a mesh after a tool has been subtracted 10000 times
[09:09:44] <MattyMatt> I'll start on the interpreter
[09:10:07] <MattyMatt> or gcode importer, from the blender perspective
[09:15:49] <awallin> MattyMatt: ocl is in c++ with python bindings
[09:16:18] <awallin> are you importing g-code into blender?
[09:16:19] <MattyMatt> I thought so. I saw it while building Heeks
[09:16:33] <MattyMatt> yeah :)
[09:17:59] <awallin> the octree will condense away unneccessary nodes when they are all cleared
[09:18:14] <awallin> but ofcourse the surface will be complicated once you move your tool 10000 times
[09:19:55] <piasdom> g'mornin all
[09:21:06] <MattyMatt> with voxels down to around the size of the tool, or smaller, it'll work fine :) no node cube will ever usually have more than 3 extra cut polys
[09:22:02] <MattyMatt> so you can code a hard limit
[09:22:27] <MattyMatt> and if you exceed it, clear the node. call it metal fatigue :)
[09:23:03] <MattyMatt> gmorning piasdom
[09:24:23] <MattyMatt> mm openCL
[09:39:49] <piasdom> what a good channel to ask for help with c++ for a beginner ?
[09:40:55] <MattyMatt> do you know C?
[09:41:18] <piasdom> nope
[09:41:27] <piasdom> to me, they're the same
[09:41:52] <MattyMatt> silly question, leading to the "why don't you learn C first?" . I'm a veteran of many flamewars :)
[09:42:20] <toastydeath> i wish my first programming classes were in c, it would have been more useful to me
[09:42:26] <piasdom> because at work they started me with c++ :)
[09:42:50] <piasdom> i'm learning this to help at work
[09:43:21] <MattyMatt> I started with TC++ 2.0 It confused me until I learned what is C and what is that nasty wrapper stuff
[09:44:22] <MattyMatt> anyway ask your C++ (or possibly C) question :)
[09:45:00] <piasdom> how do i get the % of (x*y)-&
[09:45:25] <piasdom> * %
[09:46:14] <MattyMatt> I don't get the question
[09:46:15] <piasdom> i want to subtract a percentage of what i multiplied
[09:46:15] <toastydeath> is that percent or modulo
[09:46:22] <piasdom> percent
[09:46:44] <toastydeath> x*y*p / 100
[09:47:02] <toastydeath> where p is a real number on [0, 100]
[09:47:08] <piasdom> say; i want to do 5 x10 THEN take 30% off that
[09:47:41] <MattyMatt> (x * y) * (100-p)/100
[09:47:43] <toastydeath> then your p value is 100-p
[09:48:02] <toastydeath> (100 - p) * x * y / 100
[09:48:06] <toastydeath> so i agree with MattyMatt
[09:48:23] <piasdom> thanks i'll try that
[09:53:46] <MattyMatt> sweet. my straight fluted carbide endmill just arrived, right after I proved my old one starts fire to the wood at 1mm cut depth
[10:00:20] <piasdom> i tried (10*20)*(100-30)/100. give me the results of 10*20..
[10:02:37] <toastydeath> ?
[10:02:47] <toastydeath> that's the formula, do it on a calculator?
[10:02:52] <toastydeath> if you don't trust the computer?
[10:04:12] <piasdom> k
[10:04:17] <piasdom> thanks
[10:04:25] <toastydeath> i mean talk it out
[10:04:48] <toastydeath> you have 10*20, and then you're multiplying that by 100 percent minus thirty percent
[10:04:55] <toastydeath> so 200 * .7
[10:05:47] <piasdom> don't see where the .7 comes from
[10:06:31] <toastydeath> (100-30)/100
[10:06:38] <toastydeath> 70/100 = 7/10
[10:06:44] <toastydeath> = .7
[10:06:59] <piasdom> cool...thanks again
[10:07:01] <toastydeath> np
[10:09:21] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[10:20:28] <MattyMatt> that real endmill makes all the difference :) clean white wood, lovely finish
[10:22:03] <MattyMatt> I'm now in the wooden articles manufacturing business
[10:39:09] <piasdom> i think i got it now,BUT :) i want the results of (10*20)-30% to be 140(subtracting the 30%) not 60(30% of 200)
[10:41:19] <micges> piasdom: (a*b) - (p/100)*(a*b)
[10:41:29] <micges> p -percents, a,b values
[10:42:18] <piasdom> thanks micges
[10:42:54] <micges> np
[10:43:06] <toastydeath> piasdom, subtracting thirty percent is taking 70 percent of the original value
[10:44:16] <piasdom> toastydeath; i know, but we were using 30 and in my formula i was using 70. thanks
[10:45:57] <toastydeath> micges's formula is algebraically identical to the one you're using now
[10:46:05] <toastydeath> ab - p/100*ab
[10:46:10] <toastydeath> ab-pab/100
[10:46:22] <toastydeath> ab(1-p/100)
[10:46:30] <toastydeath> ab(100/100 - p/100)
[10:46:41] <toastydeath> ab(100-p)/100
[10:47:07] <toastydeath> <toastydeath> (100 - p) * x * y / 100
[11:36:00] <piasdom> ok, i have {cout << "pay: " << (i*base)-(33/100)*(i*base) << endl;} and the result is 915.2. should be 613.18
[12:00:21] <toastydeath> what are the test numbers you're using
[12:00:27] <toastydeath> and if available, the exact wording of the problem
[12:05:40] <toastydeath> piasdom
[12:26:49] <piasdom> the problem is one i made up...(something to do while machine ran) i = 40 base = 22.88
[12:27:39] <piasdom> trying to get the net pay instead of the gross
[12:29:48] <piasdom> i input hours(40) and want the net pay for a week
[12:31:57] <toastydeath> unfortunately I don't know how to debug that code, i don't know why it's giving you i*base by itself with nothing else
[12:32:15] <toastydeath> what happens if you enclose the entire formula with a pair of parens
[12:32:29] <piasdom> toastydeath; thanks anyway
[12:32:41] <piasdom> i tried that...same results
[12:33:16] <piasdom> i tried para in different places with no change
[13:02:50] <awallin> piasdom: depending on how smart your compiler is, it might use integer division
[13:03:02] <awallin> so either explicit casts (double)100
[13:03:11] <awallin> or perhaps 100.0 to give the compiler a hint
[13:03:18] <awallin> do that for all the ints
[13:04:40] <awallin> anyone know "sealive"'s email?
[13:13:16] <awallin> piasdom: any luck with that division?
[13:23:02] <piasdom> i have everything double _____ = ? like double base = 22.88
[13:24:52] <awallin> ok, and you still get strange results?
[13:26:34] <piasdom> the same results (a*b) 915.2
[13:27:53] <piasdom> i actually understand the formula, just don't know WHY it no work
[13:28:17] <ries> what is teh recommended way to create autozero (Z) im EMC2? I found this, but somebody said :unfortunately it relies on using g92 , I am not sure what they means in practise
[13:29:11] <awallin> ries: I guess the clean way is to use tool offset?
[13:29:32] <ries> awallin: but I cannot... I have a normal router with off the shelff bits
[13:31:55] <awallin> I think the comment that g92 is bad means that the clean way of doing part-offset is through coordinate systems (g54 and whatnot) and the clean way of doing tool offsets is through the tool table
[13:32:10] <awallin> I'm not sure if any of those can be done with touch-off
[13:32:30] <awallin> but you can do a g92z0 from MDI and you might be happy with that
[13:34:57] <ries> awallin: ok, I only do it once per g-code file, currently I am not aiming for toolchanges within a single file
[13:35:36] <ries> it was just the somebody sayed 'unfortunately' and I don't understand the implications yet of that.
[13:35:59] <ries> I 'think' I can also make a pyton button for g92z0, right?
[13:37:34] <awallin> what would I know about pyVCP... :)
[13:38:22] <awallin> sorry I haven't followed the latest developments. It could be possible I don't know.
[13:40:36] <ries> ok, np
[13:44:14] <awallin> ries: check this http://www.linuxcnc.org/docs/devel/html/gui_halui.html#r1_2_6
[13:44:51] <awallin> I think you can set up halui to run different MDI commands and trigger them with a halui.mdi-command-XX pin
[13:45:04] <awallin> you can then wire that pin to a pyVCP or a real button
[14:43:50] <piasdom> awallin; i had to use 33/100.0 ... what does the .0 do /
[14:44:27] <awallin> it tells the compiler that 100 is a floating point number
[14:44:38] <awallin> so the result will be in floating point
[14:44:55] <awallin> otherwise it could be as an int, and you would get rounding errors
[14:47:26] <cradek> it's not a rounding error, but 33/100 (in integer division) is 0
[14:47:28] <toastydeath> makes sense, 33/100 ~= 0, 0 * a * b = 0
[14:48:26] <piasdom> cool....Thanks all
[14:49:32] <piasdom> is "double base = 22" and interger ?
[14:49:49] <cradek> no, base is a double
[14:50:08] <piasdom> i have all of thme doubles
[14:50:16] <piasdom> *them
[14:50:27] <piasdom> why would it see it as an int
[14:50:54] <cradek> sorry, I didn't read enough of the conversation to give meaningful answers
[14:50:58] <cradek> I'll let others :-)
[14:51:37] <piasdom> cradek;thanks anyway
[14:51:46] <awallin> ints are smaller and faster than floats or doubles, the compiler likes to see ints (I guess)
[14:53:42] <toastydeath> my guess is that since (33/100) is in parens, it does the integer division before it sees any doubles.
[14:54:53] <piasdom> i'm guessing that's a good guess :)
[14:59:20] <awallin> hah, found my old Psion 3a in a drawer while cleaning (you should never clean...)
[15:11:33] <MattyMatt> I always wanted a 5, but when the opportunity to buy one for 60 quid came up, I couldn't justify it
[15:12:12] <MattyMatt> they really should try them again. that folding keyb was ace
[15:12:44] <awallin> must be from 1995ish... I think it was 300gbp
[15:12:51] <MattyMatt> iirc Sony Ericsson have the licence
[15:13:46] <MattyMatt> they made some blue ones
[15:15:11] <awallin> ah, walkman DD
[15:15:20] <awallin> only cassette player you could run with
[15:15:36] <awallin> must be earlier than 1995?
[15:16:12] <awallin> I should run with it, show it off to all the ipoders
[15:17:23] <MattyMatt> my Tandy one from 82 was good too. iirc DD was about that time
[15:18:32] <MattyMatt> it got stolen before I could compile any long term reliability study :)
[15:18:33] <awallin> http://cache.gawker.com/assets/images/gizmodo/2009/07/wmdd.png
[15:18:42] <awallin> mine is not exactly like that
[15:23:37] <MattyMatt> lots of the old cnc controls had cassette interfaces. I never knew that until yesterday :)
[15:25:01] <MattyMatt> what cpu in psion 3? it's ~40Mhz ARM isn't it?
[15:27:22] <MattyMatt> NDS is similar, but has some FP assist hardware and a touchscreen, so probably wins as a potential emc2 machine
[15:29:56] <sealive> hi
[15:30:37] <MattyMatt> hi sealive
[15:32:16] <sealive> hi matt
[15:37:50] <sealive> awallin: are you here
[15:38:37] <MattyMatt> should I write up my blender gcode exporter on the emc2 wiki?
[15:39:30] <MattyMatt> or is that a bit off topic?
[15:39:43] <sealive> no do so
[15:41:29] <MattyMatt> I should ideally have one page summary on emc2 wiki, one on reprap wiki, both linking to a site that never dies
[15:42:40] <MattyMatt> that ain't gonna happen in a hurry, so I guess I'll just have to duplicate most of it
[15:43:48] <awallin> sealive: hi, I'm going out in a minute. got your email, what were your plans with freemill?
[15:46:19] <sealive> to give you the programm
[15:46:34] <sealive> how can i fill the polys out as you did it in tux
[15:46:50] <sealive> to see the real stl
[15:47:13] <awallin> surface.SetSurface()
[15:47:15] <sealive> stl.SetWireframe()
[15:47:19] <awallin> on the VTK actor
[15:47:40] <awallin> the surface object for VTK is separate from the surface object for ocl. a bit confusing maybe
[15:48:18] <awallin> sealive: if you zip the exe file maybe it goes through gmail?
[15:49:03] <sealive> no no zip with exe accepted
[15:51:14] <MattyMatt> http://filebin.ca/
[15:53:52] <sealive> mat i do not upload a file to a freehoster
[15:55:38] <MattyMatt> this is called freemill? :)
[15:56:48] <awallin> this one? http://www.mecsoft.com/freemill.shtml
[15:58:06] <sealive> yes
[15:58:25] <sealive> ok does not matter
[15:59:04] <sealive> it is so boaring that i cant close the VTK window without loosing the hole python system
[15:59:41] <awallin> the VTK mailing list is pretty good, if you don't find an answer in the VTK docs
[15:59:50] <MattyMatt> open a bigger window in front of it
[15:59:58] <sealive> :D
[16:01:03] <MattyMatt> that freemill is shiny. it's a good job I didn't see that before I got something working
[16:02:10] <awallin> try here also http://www.vtk.org/Wiki/VTK/Examples
[16:02:21] <awallin> I guess the freemill code is not available?
[16:04:48] <sealive> matt as awallin ocl will be to
[16:08:43] <awallin> one question is wether ocl would need its own GUI, or if it's better to work on integration with HeeksCAD
[16:10:14] <sealive> to go both ways woud be better
[16:11:06] <awallin> I like VTK for 3D, and probably QT for the gui part
[16:11:33] <sealive> Xmin,Xmax: (-10.8692, 11.3333)
[16:11:35] <sealive> Ymin,Ymax: (-14.1763, 14.1945)
[16:11:36] <sealive> Zmin,Zmax: (-7.41216, -0.198) there are numbers of may and min in the polydata wy dont you use these ?
[16:12:17] <awallin> this is a high-level question. so far I am interested in the cutter against one triangle.
[16:12:26] <awallin> there are still problems with bullCutter
[16:12:35] <awallin> and there are no compound cutters yet
[16:13:12] <awallin> also think about what to use from VTK and what to include in ocl
[16:13:20] <MattyMatt> I say share GUI with Heeks, but wean Heeks off wX dependency
[16:13:27] <awallin> if there is a desire to not require VTK...
[16:13:59] <MattyMatt> if you don't mind C++ dependency, there's Qt these days
[16:14:13] <sealive> i agree
[16:14:33] <MattyMatt> personally I'd be happy if all GUI was done in gtk
[16:14:47] <awallin> I would hate to "hand write" a gui for certain functions, that then needs to be changed whenever new functions are added
[16:15:04] <awallin> the gui code needs to be smart enough to ask the libraries what services they provide
[16:15:12] <awallin> introspection or whatever
[16:15:27] <MattyMatt> Dan has already had to do that, with that GL object tree
[16:15:29] <awallin> someone whit a solid background in oo/gui programming could help
[16:16:18] <MattyMatt> I have several years of solid bad memories about trying to force stuff to work with a bad GUI :)
[16:16:30] <MattyMatt> so, I know how NOT to do it
[16:17:29] <MattyMatt> oop in C works quite well tho
[16:18:28] <awallin> but these are high level questions. we still need an ocl-type library for the number crunching. and there are IMO a lot of people who talk the talk and few who code the code when it comes to the CAM-algorithms...
[16:22:00] <MattyMatt> I don't think a lib is so vital at first, unless you want to unify the existing ones
[16:22:31] <MattyMatt> e.g. turn ocl into the lib you need
[16:23:01] <awallin> I don't think the open-source machining algorithms exist out there. that is why I started ocl.
[16:24:02] <MattyMatt> there are the ones Dan uses. libactp etc, or are they single function libs written by you?
[16:24:26] <awallin> no I did not write them
[16:24:49] <awallin> actp is probably a remnant from the freesteel guys
[16:25:46] <awallin> and pycam is too slow (that's what I heard anyway)
[16:26:36] <MattyMatt> and a bit primitive by the looks of it.
[16:26:53] <MattyMatt> I don't mind slow too much, if it works in python :)
[16:27:04] <MattyMatt> in blender rather
[16:28:07] <MattyMatt> it's a good way to prototype all this stuff before fast implementation
[16:28:23] <pfred1> MattyMatt it is?
[16:28:47] <MattyMatt> CAM in blender is it
[16:28:57] <awallin> yes, python scripting both for testing and rapid development is good
[16:29:04] <awallin> that should be included in any GUI also
[16:29:08] <awallin> belnder, heeks, etc
[16:29:15] <awallin> blah, typing fail
[16:29:40] <pfred1> * pfred1 is experimenting with a new pointing device so expect screw ups here!
[16:32:45] <MattyMatt> pycam's intermediate stage, where it produces a .py than produces an .ngc . that's inefficient I'm sure
[16:33:16] <pfred1> MattyMatt how can you be so sure?
[16:33:35] <MattyMatt> occams razor :)
[16:33:56] <awallin> someone should come up with an XML standard for these operations. anyone could then write a post-processor into G-code, HPGL or whatever from that
[16:33:57] <awallin> ?
[16:33:59] <pfred1> MattyMatt can you script a parsing program in .ngc?
[16:34:15] <MattyMatt> I see one stage that could be eliminated, or turned into a proper file format
[16:34:57] <awallin> or punch card?
[16:35:33] <pfred1> MattyMatt so what you're saying in effect is you think exchanging one translation for another would be more efficient?
[16:36:40] <sealive> awallin: bounds=polydata.GetBounds()
[16:36:54] <MattyMatt> it would be a choice of either skipping a transltation, or doing that translation in a bidirectional way. like at the moment you can't get back to original from the generated .py
[16:37:26] <awallin> sealive: ok... but how tightly do you want to bind ocl to VTK?
[16:37:44] <pfred1> MattyMatt so what you really want then is another export option?
[16:38:02] <MattyMatt> probably yeah :)
[16:38:34] <MattyMatt> I should mash up pyCAM and blender, and see where the intersect is
[16:39:13] <pfred1> MattyMatt nothing says streamlining to me quite like slapping a few more bells and whistles on here and there!
[16:39:34] <MattyMatt> shiny bells, shiny whistles :)
[16:39:41] <micges> cool: http://smotri.com/video/view/?id=v1370873b25b
[16:40:16] <MattyMatt> blender is a safe environment. the GUI is maintained Somewhere Else
[16:41:19] <MattyMatt> gotta go banking
[16:41:25] <MattyMatt> bbiab
[16:41:39] <sealive> awallin: as tight as it is if you import it so use it
[16:42:09] <sealive> where is the diameter of the cutter ?
[16:42:11] <awallin> sealive: the way Heeks now uses ocl it does not require VTK
[16:42:20] <awallin> cutter.diameter ?
[16:42:22] <awallin> :)
[16:43:13] <pfred1> micges my audio seems out of sync with the video
[16:44:00] <sealive> diameter=cutter.diameter()
[16:44:36] <awallin> in python it should just be a propoerty, not a method call
[16:44:55] <pfred1> russians sure do look and dress funny don't they?
[16:45:25] <sealive> yes you are wright
[16:46:18] <pfred1> sealive I mean what are they all trying to look like pimps out of 70s blacksploitation movies?
[16:47:51] <sealive> bounds=polydata.GetBounds()
[16:47:52] <sealive> minx=bounds[0]-(2*cutter.diameter)
[16:47:54] <sealive> dx=0.2
[16:47:55] <sealive> maxx=bounds[1]+(2*cutter.diameter)
[16:47:57] <sealive> miny=bounds[2]-(2*cutter.diameter)
[16:47:58] <sealive> dy=1.3
[16:48:00] <sealive> maxy=bounds[3]+(2*cutter.diameter)
[16:48:01] <sealive> z=bounds[4]-cutter.diameter
[16:48:03] <sealive>
[16:55:39] <sealive> are you still there
[16:57:49] <awallin> who? :)
[16:59:54] <pfred1> blankman
[17:00:38] <pfred1> these goofs on CNCzone never fail to amuse me i swear!
[17:00:44] <MattyMatt> that 1ms robot is awesome. they should redo the finger stabbing scene in Aliens to match
[17:01:28] <pfred1> one guy in a thread is asking if some boat anchor chinese stepper motors are slow or not
[17:02:30] <pfred1> don't all stepper motors slow down as you scale them up?
[17:02:44] <pfred1> which is why the servo market is so healthy
[17:02:54] <MattyMatt> a bot to answer basic questions would be cool, but they have to be asked occasionally
[17:03:28] <pfred1> this channel at least needs a link translation bot
[17:03:51] <MattyMatt> a bot that finds the closest match to a wikipedia page would be cool
[17:04:14] <MattyMatt> that's google, probably :)
[17:04:20] <pfred1> like subtitle in ##politics
[17:04:59] <pfred1> someone ought to invite subtitle to join this channel
[17:05:31] <sealive> did you all saw the TOYOTA Robot on the EXPO in shanghai playing Beethoven on a violin
[17:05:55] <MattyMatt> everyone should talk to their localbot, which then only lets filtered stuff onto the actual internet
[17:06:09] <pfred1> sealive is it just me or do the Japanese all really make wimpy robots?
[17:07:42] <MattyMatt> IRC would go <bot1> my man says Toyota is worth watching <bot2> Toyota is a bitch. I won't tell my man to watch it
[17:11:07] <MattyMatt> it's part of the plan to infiltrate them into society
[17:11:26] <MattyMatt> they should cower, like good slaves
[17:12:10] <pfred1> MattyMatt thats not how America is doing it though we send all of our robots overseas to destroy mud brick shacks and whatnot
[17:12:30] <MattyMatt> bigdog is a bit wimpy IMO
[17:12:49] <MattyMatt> 2-stroke and not big enough to sit on
[17:12:55] <pfred1> MattyMatt you must not have seen the video where the cruel man kicks it on a sheet of ice
[17:13:29] <MattyMatt> oh it's nimble, but you still get the impression you could disable it with a stick
[17:13:53] <pfred1> MattyMatt so what? we
[17:14:16] <pfred1> MattyMatt so what? we'll probably subcontract the contract to the chinese and they'll make them for like $19.99 each in quantity
[17:14:19] <MattyMatt> I suppose they have to be sure it won't gore the researchers before they give it horns :)
[17:14:58] <MattyMatt> robo ox
[17:15:21] <pfred1> oddly the US Army manual for caring for mules is classified
[17:15:43] <pfred1> one of about 3 that are
[17:16:34] <pfred1> I can just imagine some of their pearls of wisdom for dealing with pack animals they don't want the ASPCA to see
[17:17:07] <pfred1> there's probbaly even recipies for cooking them up if the situation warrents it :)
[17:17:08] <MattyMatt> :)
[17:17:34] <MattyMatt> oh yeah, a whole chapter on buthery and cooking. that's half the point of pack animals
[17:17:58] <pfred1> hey it helped the British hold out for like 6 months in WWI
[17:18:02] <MattyMatt> like the south pole expedition
[17:18:11] <pfred1> the cavalry came to the rescue breakfast lunch and dinner
[17:19:44] <pfred1> this mouse is pretty nice I guess
[17:20:12] <pfred1> I picked up a Wacom pad stylus and mouse combo at a yard sale today
[17:21:06] <pfred1> though I'm not seeing pressure sensitivity stuff going on with it yet like it's supposed to do
[17:24:05] <awallin> Dallur: how's the volcano doing lately?
[17:25:07] <pfred1> awallin one thing i really want to see before i die is an ejecting volcanic eruption
[17:25:26] <sealive> awallin: still here?
[17:25:39] <sealive> can i bring out the lines of ocl
[17:25:53] <pfred1> sealive no they're just really lagged awallin typed that over 20 minutes ago!
[17:26:03] <sealive> or do i have to calculate for my own to get the Gcode
[17:26:36] <awallin> sealive: Dan did something with removing redundant CL-point which all lie on a straight line
[17:26:45] <awallin> otherwise there is no export of lines
[17:27:12] <pfred1> awallin Dan of Heeks?
[17:27:18] <awallin> yah
[17:27:28] <pfred1> he hangs out on this network in another channel
[17:28:28] <Dallur> awallin: was afk :) volcano is just doing what they all do, spewing out lava, not much going on at all
[17:28:54] <pfred1> Dallur oh is it glowing ejecta?
[17:29:24] <Dallur> I think it's ejecting about 10m3 per second of very acidic lava
[17:29:40] <Dallur> that means that the lava is very flaky and "dry"
[17:29:43] <pfred1> but is it incandescant?
[17:29:50] <Dallur> yeah,
[17:29:54] <pfred1> cool!
[17:30:02] <awallin> no, hot.
[17:30:03] <Dallur> hard to get to it for this eruption though
[17:30:08] <Dallur> since it was under a glacier
[17:30:23] <Dallur> you have to scale 300m ice cliffs to get to it
[17:30:31] <pfred1> I'd so it
[17:30:35] <pfred1> do it even
[17:30:42] <Dallur> the smaller eruption we had before was easy to get to
[17:30:52] <pfred1> pfft but it was small
[17:31:14] <Dallur> http://heimska.com/Myndir/2010/04_eldgos_hssk/images/MT8V8766.jpg
[17:31:22] <Dallur> some friends of mine went up to that one on the 1st of April
[17:31:29] <Dallur> http://heimska.com/Myndir/2010/04_eldgos_hssk/images/MT8V8772.jpg
[17:31:38] <Dallur> http://heimska.com/Myndir/2010/04_eldgos_hssk/images/MT8V8934.jpg
[17:31:58] <Dallur> some of the idiots actually went up to it and picked rocks from it, one even peed on it
[17:32:32] <Dallur> this eruption is small-medium, the other one was a micro eruption
[17:32:53] <Dallur> Big eruptions are considerably larger: http://en.wikipedia.org/wiki/Laki
[17:32:54] <pfred1> Dallur oh this second ones so sweet I'm going to set it as my desktop wallpaper thanks!
[17:33:29] <Dallur> http://heimska.com/Myndir/2010/04_eldgos_hssk/
[17:33:34] <Dallur> that's the collection
[17:33:40] <Dallur> got to go, c u later
[17:33:58] <pfred1> Dallur I got a thing for melting stuff I always have
[17:34:27] <Jymmm> Anyone using Ubuntu 8.10 yet?
[17:34:43] <pfred1> Jymmm I use 9.10
[17:35:03] <sealive> no 10.04
[17:35:11] <awallin> huh? you mean 10.04? works fine
[17:35:17] <awallin> not for emc2 though
[17:35:42] <Jymmm> Yeah 10.4, seems bloated
[17:35:52] <sealive> no emc2 at the cnc on xbuntu hardy here on 9.10 without cnc connect
[17:36:28] <sealive> awallin: can i bring the lines out of your ocl or do i have to put thew points to lines together
[17:36:58] <awallin> it's all points so far.
[17:37:04] <sealive> ok
[17:37:05] <awallin> line and arc filters are on the todo list
[17:38:23] <sealive> lines woudt cut it first by 40percent
[17:40:56] <pfred1> now that is some cool desktop wallpaper
[17:46:14] <pfred1> Jymmm Ubuntu in general seems bloated to me
[17:47:44] <pfred1> I hope eventually EMC gets away from Ubuntu
[17:51:01] <MattyMatt> I like the full distro
[17:51:32] <MattyMatt> distro browns seasons and thickens, all in one
[17:51:39] <andypugh> Has anyone used the raw read/write commands in Hostmot2?
[18:38:14] <MattyMatt> what does the fpga give you? better RT performance?
[18:39:02] <MattyMatt> e.g. are PID loops on things like servos closed inside it?
[18:42:28] <MattyMatt> and if stepgen is done in the fpga, how much attention does the cpu still need to give it?
[18:43:06] <MattyMatt> I should ask a mesa salesman these questions :)
[18:43:47] <MattyMatt> over lunch, with lots of beermat drawing
[18:44:09] <pjm> the 7143 upgrade i did to my machine has probably been its best update
[18:44:23] <pjm> took all the control away from parports and give to the fpda
[18:44:27] <pjm> fpga even
[18:44:49] <pjm> the encoder counters are wicked, no problem keeping up with spindle encoder etc, all very nice
[18:44:56] <MattyMatt> is it on a pci card? how does the cpu talk to it?
[18:45:05] <pjm> yeah parport > 7i43
[18:47:48] <awallin> MattyMatt: the pwm-generators and encoder counters are on the fpga. PID-loops are in EMC2/realtime
[18:49:25] <MattyMatt> ah OK
[18:50:40] <MattyMatt> is that by preference? would it be good if PID was done onchip and available for async inspection in registers?
[18:52:15] <awallin> depends on what bandwidth (update rate) the PID needs
[18:52:27] <awallin> standard servo-loop in emc is 1ms or 1 kHz
[18:52:35] <MattyMatt> I noticed the price of fpga library functions :) a multiply cost a few $000
[18:52:49] <awallin> I think we are running 4 kHz with our m5i20. in voltage-mode (no current or speed loop)
[18:53:24] <awallin> but 10kHz or more is not doable with realtime I guess
[18:53:44] <awallin> so if you need loops at 10kHz or faster then putting them in fpga is better
[18:54:05] <awallin> we did something at 200kHz update rate at work with an NI FPGA-card and analog output
[18:55:19] <awallin> http://sine.ni.com/cs/app/doc/p/id/cs-802
[18:55:41] <MattyMatt> so, long-term, it's sth to consider? I preume any cpu/bus improvements will be paced by fpga improvements
[18:55:58] <awallin> this kind of steering in the microscope: http://www.youtube.com/watch?v=9mVh5rMjTgY
[18:56:32] <awallin> in the long-term the mechanical bandwidth of your typical cnc-mill will probably not change much
[18:57:01] <awallin> if you need to move around a 100kg table with conventional servos the 1kHz or 5kHz bandwidth that realtime gives is just fine
[18:57:31] <awallin> if you are steering laser-beams or something where you can actually steer/control very fast then moving things onto the fpga makes sense
[19:00:09] <MattyMatt> good point about the machines not being subject to Moores Law
[19:01:27] <MattyMatt> 18 axes is enough for a humanoid. that's probably the limit
[19:01:53] <MattyMatt> 19 for robodog, so it can wag its tail
[19:01:58] <archivist> I have heard of 256 axes
[19:02:27] <archivist> aircraft machining center multi headed
[19:03:19] <archivist> UK company does the simulation software
[19:03:56] <MattyMatt> ah yes :) not Moore's Law, Parkinson's Law. we'll just plug in more low-bandwidth tools until we've got a problem again :)
[19:04:52] <archivist> simulation is to make sure they dont hit each other etc
[19:05:13] <MattyMatt> yeah that's why it all has to one simulation
[19:05:30] <MattyMatt> the whole factory on one emc machine
[19:06:54] <MattyMatt> maybe sim could be distributed, with each machine calculating a frame of its responsibilities, and passing them to the master
[19:07:20] <MattyMatt> rather like a render farm
[19:07:47] <MattyMatt> although they tend to need the whole scene on each renderer
[19:10:12] <MattyMatt> I'll be able to do complex multiaxis sims in blender real soon now
[19:10:54] <MattyMatt> with an accurate model of my spindle, it should be able to avoid colliding ever
[19:12:09] <MattyMatt> I'm wondering if that kind of foolproofing would be useful at a lower level, or just be plain annoying
[19:13:52] <MattyMatt> I suspect it'll be like any other GUI. great when it works how you want, extra annoying when it doesn't
[19:15:21] <MattyMatt> I was abusing homing until today when I got the hang of touching off :)
[19:18:29] <MattyMatt> this 4 flute endmill has transformed this machine
[19:19:58] <MattyMatt> the one that came with the spindle was like a crown gear with a ring of teeth
[19:21:19] <MattyMatt> now I need a 45 deg countersink for textwriting, and I'm in business
[19:22:26] <MattyMatt> and a real ballmill too, I wont bother with the one I've got
[19:23:28] <MattyMatt> http://cgi.ebay.co.uk/3-mm-LONG-SERIES-MITSUBISHI-Ball-nose-carbide-NEW-/390088501678
[19:24:15] <MattyMatt> expensive, but I think that would last forever doing wood
[19:27:44] <MattyMatt> a 4mm collet would be handy
[19:27:57] <MattyMatt> lots of pro ranges start at 4mm
[19:33:41] <MattyMatt> I'm still gonna grind some broken drill bits, at some opportunity :) it's only wood
[19:47:49] <MattyMatt> 1/8" mills are rare as rocking horses this side of the pond, so I might get my 1/8" collet bored to 4mm
[19:50:08] <MattyMatt> if that costs a tenner I'll recoup it straight away. 4mm long ballnose is £4.99
[20:02:49] <frallzor> yo
[20:03:26] <MattyMatt> yay \o/ I milled my first pocket today
[20:03:59] <MattyMatt> with gcode I made in blender
[20:04:03] <frallzor> congrats
[20:05:26] <MattyMatt> now I'm going to spend all my money on tiny mills
[20:06:02] <MattyMatt> I don't need miracle nano coatings for woodwork, but hey :)
[20:11:01] <MattyMatt> I need to mill the insides of my vice jaws. they're plastic coated
[20:12:04] <MattyMatt> ah, it's good enough for woodwork :)
[20:12:55] <frallzor> I still need to plane my table =P
[20:13:33] <MattyMatt> let the machine do that
[20:13:58] <frallzor> yeah I will, butI need to bolt that bloody thing too =P
[20:14:09] <frallzor> was going to get nuts and bolts today, but of course its May 1
[20:15:13] <MattyMatt> doh. holidays are postponed until Monday in UK, which is sensible but dull
[20:18:28] <frallzor> had a piece of mdf doubletaped to the table so I could plane and mill
[20:18:39] <frallzor> daaamn, to get it off was a bitch
[20:19:33] <MattyMatt> I can imagine. 15lbs per square inch + weight of mdf + adhesive
[20:19:57] <frallzor> well it was a small 1x1' piece =)
[20:20:11] <frallzor> sat like a bitch
[20:21:19] <MattyMatt> I'm doing everything in my drill vice, which I've bolted to the table
[20:21:50] <MattyMatt> just to prove it's a mill not a router :)
[20:23:56] <MattyMatt> that's my next job. layer of mdf and some clamps for large sheets
[20:25:29] <MattyMatt> for mass production
[20:26:55] <MattyMatt> it should be trivial to do identical parts with gcode subroutines, with this blender script
[20:28:19] <frallzor> I still havent decided on clamping =/
[20:28:57] <MattyMatt> for such a large machine, vacuum must be the only option?
[20:29:15] <frallzor> vacuum for this size cost too much
[20:29:16] <MattyMatt> or lots of double-sided :)
[20:29:30] <frallzor> im thinking lots of T-slots
[20:29:42] <frallzor> MDF proved how strong it is today
[20:29:52] <MattyMatt> T slots are versatile
[20:30:03] <frallzor> after this little adventure im pretty sure on that slots in mdf will work =)
[20:30:10] <MattyMatt> you can bolt a vacuum bed to them :)
[20:30:45] <MattyMatt> how about a vacuum bed made of mdf?
[20:31:03] <frallzor> well the bed isnt the problem
[20:31:09] <frallzor> its getting the vacuum =)
[20:31:15] <frallzor> $$$
[20:31:55] <frallzor> when I made a booboo when doing the first cuts, the machine stopped in the table and dragged both pieces along =)
[20:32:28] <frallzor> so I cant really use a dustbuster since it wont do any good and will overheat if doing long jobs =(
[20:33:43] <MattyMatt> industrial vacuum cleaner and a hose
[20:34:34] <frallzor> wont do any good
[20:34:35] <frallzor> http://www.cnc-plus.de/index.php?language=en&gm_boosted_category=Vacuum-pumps&cat=c55&
[20:34:45] <frallzor> lower 2-3 is what I would need
[20:35:08] <MattyMatt> I had a nice all steel wet'n'dry, but it rotted after dealing with a rottweiler's worst
[20:38:04] <MattyMatt> yeah t slots are good. mdf hmm tho
[20:38:27] <MattyMatt> it'll kill your chance of using coolant
[20:38:40] <frallzor> well its not made for using coolant =)
[20:38:48] <frallzor> cant really mill stuff that needs it
[20:39:02] <MattyMatt> there's nothing stopping it yet tho, is there?
[20:39:15] <MattyMatt> what about window glass?
[20:39:32] <MattyMatt> I could do that on mine, if I could handle a bit of water
[20:39:38] <ewlsey> hey guys, how can I copy EMC configuration files (.HAL .INI etc.)?
[20:39:57] <ewlsey> i get an error when I try to copy them to a USB key
[20:40:27] <frallzor> ill stick with wood and plastic and some HSM alu
[20:40:32] <cradek> ewlsey: don't be coy
[20:40:42] <ewlsey> ?
[20:41:10] <cradek> I mean, go ahead and tell us what the error is
[20:42:00] <ewlsey> cradek: says "error making symbolic link, operation not allowed"
[20:42:22] <cradek> is one of the things you're trying to copy a symlink?
[20:42:37] <cradek> dos filesystem (probably what's on your stick) doesn't support symlinks
[20:42:47] <cradek> so, don't copy that, copy the thing it points to instead
[20:42:48] <ewlsey> I am trying to copy a configuration so i can use it on another computer
[20:42:58] <cradek> or, tar it up and copy the tar file
[20:43:19] <cradek> using tar is probably better
[20:43:24] <ewlsey> I tried to open the .HAL file in gedit and save it to the USB an got the same error
[20:43:48] <ewlsey> you might have to dumb it down a little (lot) for me cradek
[20:44:28] <MattyMatt> gedit wouldn't give that error, maybe you didn't close that dialog
[20:45:35] <MattyMatt> "save as" is never a symlink afaik
[20:45:55] <ewlsey> gedit says "unexpected error: I/O error"
[20:46:42] <ewlsey> what is the standard method to move the configuration to a new computer>
[20:46:43] <ewlsey> ?
[20:47:05] <MattyMatt> pretty much what your doing, but doing it right >:)
[20:48:02] <ewlsey> awesome...
[20:48:18] <MattyMatt> I think you rould use the unmount in the usb's File menu. wait a few seconds, remove drive, wait a few, reinsert
[20:48:38] <micges> ewlsey: can yo copy config to another dir on same partition?
[20:49:19] <MattyMatt> I think it's a file locking problem
[20:49:36] <normaldotcom> I have some motor drivers that take a single PWM stream and translates it to forward/off/backwards... anyone know if I can get pwmgen to produce a compatible signal?
[20:49:49] <MattyMatt> which has somehow locked the whole usb stick. I think that happens to me
[20:50:31] <micges> MattyMatt: sudo apt-get install pmount should help
[20:50:36] <ewlsey> MattyMatt, I can open and save other files on the key
[20:50:46] <ewlsey> micges I will try quickly
[20:51:07] <MattyMatt> have you got a broken symlink with the filename you want to use?
[20:51:30] <ewlsey> how would I know that?
[20:51:42] <MattyMatt> 0 length file or sth
[20:51:59] <ewlsey> I can copy the config to other locations on the HD, but not onto the USB drive
[20:53:16] <MattyMatt> if you can't save a text file from gedit, then the problem is between the OS and the USB drive
[20:53:36] <micges> ewlsey: try 'sudo apt-get install pmount' in console and restick you usb drive
[20:54:49] <awallin> yay, got toroidal tools to work... finally ( I think)
[20:56:19] <MattyMatt> Polytopes \o/
[20:56:54] <micges> good night all
[20:58:22] <MattyMatt> normaldotcom didn't wait very long
[20:58:22] <ewlsey> micges i need to reset my sudo password since i forgot it
[20:58:27] <ewlsey> lol
[20:59:38] <MattyMatt> you need your user password to log on to the machine:)
[21:00:12] <ewlsey> no I disabled that a long time ago
[21:00:21] <ewlsey> when i knew the password
[21:00:33] <MattyMatt> ah. that's the only place I'd bother with it
[21:01:12] <ewlsey> hmm
[21:01:32] <frallzor> * frallzor pokes MattyMatt with a stick
[21:01:47] <ewlsey> I'm not going to be able to install anything anyway since the machine is not connected to the internet
[21:01:49] <MattyMatt> unless I had an rfid reader that gets it auto if I'm sat within range
[21:17:34] <ewlsey> had to copy the text from each config file into a new file and then save that to the USB, what a PITA
[21:19:49] <frallzor> hmm large sheets can just be clamped, no shit there, small pieces thats the real issues
[22:10:01] <ewlsey> so are there some PCs that cannot run EMC without realtime errors?
[22:10:23] <awallin> yes
[22:10:32] <ewlsey> this is pretty modern hardware, and I have enabled SMI, and i get a realtime error every time
[22:11:13] <ewlsey> I am 0/2 on selecting the correct PC for EMC
[22:11:21] <awallin> I had a motherboard which produced 155ms delays every 10min
[22:11:27] <awallin> I could never get it to work
[22:11:42] <awallin> do you have a separate graphics card?
[22:11:52] <awallin> what graphics driver are you using?
[22:11:57] <ewlsey> awallin I wish mine would wait 10 min. I get the error immediately
[22:12:03] <awallin> and how bad is the latency?
[22:12:49] <ewlsey> base thread had 14000ns latency, but my base period is set at 40000
[22:13:49] <ewlsey> nvidea geforce fx 5300 graphics card
[22:14:54] <ewlsey> other PC had intel integrated graphics, it would not show the Axis backplot unless you clicked on it
[22:14:56] <tom3p> ewlsey, ive had same for years, just bvefore tossing several boards, i went into bios and turned EVERYTHING off the could be turned off, a
[22:15:01] <tom3p> nd salvaged 8 of 10 mobos. later i enabled a FEW bios options, very caautiously.
[22:15:01] <tom3p> I had NO idea what many options meant and could not find descritptions in internet or manufacturer.
[22:16:09] <ewlsey> tom3p, will I be in trouble if I get into BIOS and delete the wrong thing? I'm not a wiz with computers
[22:16:39] <ewlsey> i really don't want to go to a 3rd PC
[22:16:49] <tom3p> ewlsey,same here, i first set the mobo to 'default' then tried. if it booted, then i could always go back to 'default'
[22:17:23] <tom3p> i could change anything and still go back to 'default'
[22:17:52] <ewlsey> it seems weird that I get the RTAI error as soon as I start EMC,
[22:18:23] <ewlsey> if there was really an issue is seem like it would no cause an error until that system ran again
[22:18:43] <ewlsey> frustrating
[22:19:07] <tom3p> is it 'unexpected realtime delay" ?
[22:19:28] <ewlsey> yes
[22:19:31] <ewlsey> on task 1
[22:19:55] <tom3p> that can happen in just a few milliseconds
[22:20:06] <tom3p> you can see it as 'immediate'
[22:20:12] <tom3p> lights out, gotta go!
[22:22:34] <ewlsey> well can anyone recommend a PC? so far I have struck out with a Dell Optiplex 260 and a Medion Composer 5200
[22:23:02] <awallin> if you don't need the PC for much else then an Atom board is ok
[22:23:45] <ewlsey> I only want to run EMC, I need to do software step generation though
[22:24:00] <awallin> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Latency-Test
[22:25:44] <ewlsey> the optiplex is on the list of recommended hardware...
[22:53:35] <ewlsey> awallin, does the atom board have PCI?
[22:53:50] <awallin> yes, but only one slot
[22:54:20] <ewlsey> ok
[22:55:44] <ewlsey> might be an issue if I want to run a second parallel port
[23:26:47] <andypugh> I am struggling with raw mode in the hostmot2 driver. Should I be able to read back the values I have written to a register?
[23:52:01] <ewlsey> does EMC make any distinction between max. rapid speed, and max. feed rate?
[23:52:17] <ewlsey> I only see max. velocity