#emc | Logs for 2008-05-06

[00:12:35] <JymmmEMC> SWPadnos: What you doin with your artigo?
[00:33:59] <BigJohnT_> shhhhh be quiet
[00:54:44] <Gamma-X> BigJohnT, hey there buiddy lol
[04:30:29] <bigcx2> hey all
[04:30:39] <bigcx2> i was trying to build an rtai live cd
[04:30:47] <bigcx2> based on hardy
[04:31:16] <bigcx2> does anyone know if the rtai hal patches exist for 2.6.24?
[04:31:21] <bigcx2> or where to get them?
[06:02:45] <K`zan> Got it together :). Still needs work, but basically it functions.
[06:03:55] <K`zan> Obvious problems with the EMC config, most likely out of ignorance somewhere. Got lead screw pitch set correctly but it only moves half that distance - may have to do with microstepping 1/2.
[06:04:10] <K`zan> Enough for today, will post pix tomorrow.
[12:11:20] <alex_joni> hi
[12:14:35] <skunkworks_> Hi alex
[12:18:57] <alex_joni> what's up?
[12:21:11] <skunkworks_> reading :)
[13:50:52] <pjm> afternoon, anyone here built their own cnc mill and using emc to run it?
[13:59:26] <archivist> yes
[14:00:57] <pjm> and what do u use for the spindle assembly?
[14:01:09] <pjm> i'm thinking of buying the MT3 one from 'little machine shop'
[14:01:55] <archivist> in my case its a lathe headstock up ended (Hobbymat MD65)
[14:02:36] <cradek> from the point of view of availability of cheap tooling, R8 is probably best and MT3 is second best
[14:03:00] <pjm> yeah i have a load of MT3 stuff for my manual mill already
[14:03:08] <pjm> or i'd have gone the R8 route
[14:03:10] <cradek> I've been hearing good things about Tormach tool holders, which will work with either of those
[14:03:15] <archivist> mine is MT2
[14:04:33] <cradek> you are putting together a small machine I assume? NMTB30/40 is better in every way (except cost) for a larger machine
[14:06:28] <pjm> yeah a small mill
[14:06:57] <pjm> so table travel is about 300X250mm
[14:07:29] <pjm> and Z will do about 750mm but in use of course it'll only need to move 100mm - 200mm or so
[14:08:46] <fenn> why is Z travel so big?
[14:08:56] <cradek> bed mill I assume
[14:09:36] <cradek> it's easy to use up 200mm as tool offsets (a drill chuck and drill bit), 100-200mm of work depth, the rest for the part
[14:11:01] <fenn> one would think there would be R8 drill chucks with the chuck part built into the R8 shank
[14:11:16] <pjm> fenn, Z is big because I had the leadscrew and carrier rails already
[14:11:44] <archivist> re use, I like that
[14:12:08] <fenn> or a bad ebay impulse buy
[14:12:08] <pjm> well so far I have the X axis completly done and working nicely with EMC
[14:12:25] <pjm> it was all the crap reclaimed from a big industrial CNC router
[14:12:31] <pjm> multicam thing
[14:13:44] <pjm> a quick pic of X axis http://pjm.dyndns.org/cnc/xaxistable2.jpg
[14:13:58] <pjm> i'm working on the Y axis in the evenings this week, so it should be ready to roll by the weekend
[14:14:30] <cradek> looks bigger than I expected
[14:14:59] <archivist> sensibly solid, looks nice so far
[14:15:05] <cradek> yes
[14:15:29] <pjm> thanks, yeah it hopefully will be solid
[14:15:36] <cradek> will the spindle run on 120 or 240?
[14:15:41] <pjm> it'll be 240v
[14:15:49] <pjm> but i'll probably pwm it
[14:16:07] <pjm> and possibly run an encoder so I can index it around for tapping etc
[14:16:54] <cradek> emc2 has been able to rigid tap for a year or so - but I think very few people are using it. I hope to one of these days.
[14:17:47] <pjm> yeah it would be a cool facility to build mechanics for
[14:17:59] <fenn> are you going to fill that tube with concrete or sand?
[14:18:20] <pjm> fenn, dunno yet, i will test out the vibration first before deciding
[14:18:36] <pjm> the box section is 100mm square
[14:18:44] <pjm> so pretty big
[14:19:17] <pjm> u can see roughly how it will go at http://pjm.dyndns.org/cnc/trialfit2.jpg
[14:19:26] <fenn> steel has much less damping than cast iron.. it's not as much of an issue for wood as for metal
[14:19:45] <pjm> the idea is to stand the table on 2 further bits of 100mm box which will be at the base 90deg from the vertical
[14:20:08] <cradek> whoa the table is almost square!
[14:20:32] <pjm> hehh almost!!
[14:20:56] <cradek> I like the coolant collectors and drain on the end. mine has that.
[14:21:22] <pjm> yeah i'm not sure what that table was originally from
[14:21:28] <pjm> i got it off the scrap yard
[14:21:32] <cradek> looks like the Y travel is extremely limited
[14:21:46] <pjm> yeah its about 250mm or so
[14:22:08] <pjm> so not particulary large but it'll do to start with
[14:22:59] <pjm> the leadscrews / ball nuts all came from an old XY plotter which i fished out of a skip
[14:23:07] <pjm> and they are not actually in too bad condition
[14:23:14] <pjm> plus they are 5mm travel per rev
[14:23:17] <cradek> must have been quite a plotter
[14:23:52] <fenn> * fenn gives pjm the "dumpster recognition award"
[14:23:58] <pjm> hehh thanks
[14:24:03] <pjm> i'm a hoarder!
[14:24:27] <archivist> we recognise the trait (well I do)
[14:25:29] <pjm> its amazing what people chuck away... some is really worth having
[14:25:33] <archivist> hmm you have a better scrap yard than I do
[14:25:36] <pjm> hehh
[14:25:47] <pjm> well i've been hoarding stuff from there for years
[14:25:58] <pjm> the best bit is that they simply sell it by weight
[14:26:07] <pjm> regardless of what lump of technology u have found
[14:26:25] <archivist> the Aluminium for my mill came that way
[14:27:25] <pjm> its the only way if its to be 'cost effective'
[14:27:42] <pjm> i been playing with limit / homing switches over the weekend
[14:27:51] <pjm> and have got a pair of roller micro switches
[14:28:09] <archivist> my XY came from a local bankrupcy auction at a factory
[14:28:11] <cradek> I miss the days when I could haul all the aluminum I wanted out of the local scrap yard for $1/lb
[14:28:13] <pjm> but cant decide if these are accurate enough or should I use opto's or a proximity switch
[14:28:56] <archivist> some of the microswitches are good
[14:29:09] <archivist> and cheap
[14:29:11] <cradek> they don't even let people in to dig around anymore
[14:29:27] <archivist> they do here
[14:29:33] <skunkworks_> yah- there isn't any cool junk yards around here anymore.
[14:30:04] <pjm> so for the homing i was thinking of putting an encoder on the end of the leadscrew, then 'anding' that wiht the output of the limit switch
[14:30:28] <pjm> so the idea is that u can get precise homing between power cycles etc
[14:31:43] <cradek> that will work, but you will have to home very very slowly
[14:31:44] <skunkworks_> We did that with microswitches - one switch is on the axis - the other is run with a arm off of the motor. those are hooked in series.. (iirc) works great other than every rotation of the leadscrew you get a click :) click click click click.. but it is like homing to an index.
[14:32:45] <pjm> yeah thats a good idea
[14:32:57] <cradek> hmm, also you better make sure you can't get on the wrong side of the switch.
[14:33:01] <pjm> perhaps i should try it all as is before i get too carried away over engineering it!
[14:33:29] <pjm> the switch would get crushed if the table wanted to get to the other side
[14:33:36] <cradek> a home switch should divide the table into only two areas. one area/side has the switch on, one has it off
[14:34:20] <cradek> well imagine you're on the switch but your index signal is not active. if you home then, the axis will move further onto the switch because it doesn't see the home switch signal being active
[14:34:22] <pjm> ahh ok interesting i probably need to read the emc integrators book a bit more
[14:34:30] <pjm> as i'm using the end limits as homing too
[14:35:07] <cradek> with a normal home switch, if you're on it, the axis will pull off the switch when you home, then go back
[14:35:23] <cradek> (just beware)
[14:35:27] <fenn> cradek: do you know what caused the scrapyard to change policy?
[14:35:40] <pjm> probably 'health and safety'
[14:35:48] <pjm> this country has gone bloody mad over it
[14:35:54] <pjm> (England)
[14:35:59] <fenn> that's no excuse, the risk of digging in a scrap yard hasn't changed
[14:36:03] <cradek> fenn: I don't think there was any particular event, except for the slow creeping doubt and insanity that seems to accumulate everywhere
[14:36:27] <archivist> pjm Burton/swadlincote here
[14:36:56] <cradek> fenn: when you ask businesses directly about things like this, the commonest excuse is "our insurance company won't let us"
[14:37:17] <archivist> they dont want to pay for public cover
[14:37:58] <fenn> just give everyone armored bobcats
[14:38:46] <fenn> * fenn begins mumbling about how things are gonna change when he takes over..
[14:38:51] <skunkworks_> Ironman suits.
[14:39:17] <fenn> an armored society is a safe society
[14:39:25] <skunkworks_> fenn: coming to the cnc workshop?
[14:39:32] <skunkworks_> * skunkworks_ thinks he already asked that..
[14:39:36] <fenn> uhh.. maybe
[14:39:49] <archivist> there was a scrap place in Derby where car parts were on shelves ready for you
[14:40:26] <fenn> might make sense if cars were designed for disassembly
[14:40:31] <archivist> * archivist hopes the webcam is better this year
[14:40:49] <skunkworks_> HD cncfest.
[14:41:05] <fenn> webcams are boring, get over it
[14:42:39] <archivist> its the nearest I can get
[14:43:11] <fenn> maybe you can have a tuxcnc-fest :)
[14:43:30] <skunkworks_> INSTIGATOR!
[14:43:39] <fenn> * fenn flees in panic
[14:43:54] <archivist> there is at least 5 that I know of in the UK
[14:45:18] <pjm> btw i see from some of the 'professional' CNC machines that there is facility to swivel the head along the X axis for say drilling holes at an angle, is that worth engineering in during the build?
[14:45:38] <pjm> and i think EMC will support 5 axis
[14:46:10] <cradek> pjm: maybe you would want to combine that with a rotary on the table to get 5 axis.
[14:46:19] <cradek> whether you need 5 axis depends on what you want to do with it!
[14:46:30] <pjm> yeah i see what would probably be better / easier to engineer
[14:46:37] <fenn> pjm: like this? http://jmkasunich.com/vannorman/VN_Home.htm
[14:46:37] <archivist> see credek's vid on youtube
[14:46:38] <cradek> it seems like that makes the machine much (much!) more complex
[14:46:48] <pjm> a 4th axis would be my next build once the XYZ is done
[14:47:05] <cradek> oh did you mean manual swivel or cnc swivel?
[14:47:11] <archivist> my 4th is running before the z
[14:47:28] <pjm> CNC swivel
[14:47:39] <pjm> so Z angle controlled by PC
[14:47:49] <pjm> in the X plane only
[14:48:56] <cradek> that would be a "B" axis
[14:49:27] <cradek> once you have that, adding A or C is simple (a simple rotary mounted to the table either horizontal or vertical)
[14:49:35] <pjm> ah ok i knew there must be a name for it just didnt know what letter
[14:50:07] <cradek> A rotates around X, B rotates around Y, ...
[14:50:13] <pjm> got ya
[14:50:16] <pjm> yeah that is simple
[14:50:33] <pjm> so the other concern i have is detecting a stall of an axis during a cut
[14:50:53] <pjm> so i guess the only option there is to use linear encoders on the table axis
[14:51:01] <cradek> shouldn't happen if you have properly sized steppers...
[14:51:06] <pjm> well yes
[14:51:19] <pjm> the steppers i have were just what was laying around
[14:51:27] <pjm> they are something like 60V 3.2A
[14:51:38] <pjm> and IM465 microstepping drivers
[14:51:41] <BigJohnT> pjm if you stall a stepper you will know it
[14:52:08] <cradek> pjm: a new emc2 retrofit of a 5-axis machine: http://youtube.com/watch?v=mxxdq6y8z8M
[14:52:22] <pjm> well this is what I need to detect in order to press the emergency stop or limit switch etc..
[14:52:53] <fenn> that thing is scary
[14:53:02] <cradek> I suggest getting it to work without first, and see how it goes
[14:53:04] <archivist> BigJohnT, in answer to question the other day 31ft 6" (I hung a tape measure off the bridge) + or - seasonal errors
[14:53:15] <cradek> adding scales or encoders to steppers is not trivial
[14:54:02] <pjm> cradek, yeah that is a bit of a big m/c on that youtube vid!!
[14:54:33] <alex_joni> bbl
[14:54:39] <BigJohnT> archivist: that's a big lift
[14:54:42] <pjm> i have some encoders here from old heidenhain machines
[14:54:54] <pjm> the problem is they are massive numbers of pulses per rev
[14:55:00] <cradek> right, you don't have the hardware to read them
[14:55:01] <archivist> BigJohnT, yup at 4tons per stroke
[14:55:22] <pjm> so i guess i'd have to make some dividers to scale down the quadrature output
[14:55:43] <pjm> hmm probably a pic / other micro would be better than doing it in ttl
[14:56:01] <pjm> but yes i'll get it running first before worrying about this stuff
[14:56:15] <fenn> you'll need a special micro to count above around 100kHz
[14:56:18] <cradek> it's pretty hard to decode quadrature correctly
[14:56:39] <jepler> I'm not sure whether to say it's hard, or just that there are a lot of commonly-made mistakes
[14:57:09] <cradek> is it fair to say it's harder than one thinks at first?
[14:57:32] <pjm> probably the easiest method then would be to find some lower res encoders
[14:57:37] <fenn> pjm: there's a "ready-made" solution for 200 USD, the m5i20
[14:57:40] <jepler> some blog entries of mine on this subject: http://emergent.unpy.net/projects/01149094674 http://emergent.unpy.net/projects/01149271333 http://emergent.unpy.net/projects/01149348342
[14:58:12] <jepler> but yeah, spending money on a mesa card is far from the worst thing you can do if you want reliable, high speed encoder counting with emc2.
[14:58:40] <pjm> ah yes that is a handy page, thanks
[14:59:17] <pjm> typically i have a pile of bluechip DIG64 parallel IO cards but seems they are not supported by much
[14:59:40] <pjm> again rescued from control PC's at the local scrap yard
[15:00:06] <pjm> currently stored in the "could be handy one day" garage
[15:00:23] <fenn> rescuing industrial electronics is a losing game
[15:00:34] <fenn> electrical yes, electronics, no
[15:01:27] <fenn> there's too many ways for manufacturers to lock you in
[15:01:31] <pjm> i should try to write a driver for them so at least I can waggle the pins etc for control of stuff
[15:01:42] <fenn> (or out, in our case)
[15:02:54] <jepler> pjm: if you have programming information it is not hard to write a "hal component" that will present the digital I/Os in a way that you can connect them to stuff in emc, just like the pins on the parallel port.. but I didn't immediately find programming information for the blue chip dig-64..
[15:04:52] <pjm> not much at http://www.bluechiptechnology.co.uk/products.php?part_id=33
[15:05:45] <jepler> ah, found this: http://www.bluechiptechnology.co.uk/products/Data_Acqusition/End_of_Life_Products_-_Non_RoHS/DIG-64/_downloads/04-User_Manuals-DIG_64.pdf
[15:05:55] <pjm> oh ahd a datasheet http://www.bluechiptechnology.co.uk/products/Data_Acqusition/End_of_Life_Products_-_Non_RoHS/DIG-64/_downloads/01-Datasheets-DIG_64.pdf
[15:06:06] <pjm> ah yes they are ISA too
[15:06:13] <skunkworks_> isa
[15:06:16] <jepler> oh yours is a PCI card?
[15:06:16] <skunkworks_> ah
[15:07:01] <pjm> no these I have are ISA too
[15:07:25] <jepler> it looks like that document has enough information to write a HAL component for the card
[15:07:39] <pjm> there is some demo software for win available too
[15:07:58] <cradek> nifty - relays built on to the card
[15:08:31] <pjm> yeah 16 of them
[15:08:32] <jepler> a strange assortment of connectors
[15:08:41] <pjm> anyway i have a pile so if u need one np
[15:08:47] <pjm> they were 'free' anyway
[15:08:53] <jepler> one D-type connector, one 10-position header, 1 20-position header, and a 9-pin mini-DIN
[15:10:25] <pjm> yeah there is 2 backplanes for the card
[15:10:33] <pjm> one has a 37w D type
[15:10:44] <pjm> then the headers on the card connect to 2 other connectors
[15:10:52] <pjm> which are not too hand at the moment
[15:19:03] <fenn> 2.5mA output might not be enough to drive opto-isolators (fyi)
[15:20:50] <pjm> yeah that is a bit feeble
[15:23:10] <xyzee> Has anyone here used gcnccam to generate gcode?.. I can't get it to display a drawing.
[15:24:25] <fenn> xyzee: tomp has, but i couldnt get it to display either
[15:26:43] <xyzee> Are there any other free softwares to make gcode from dxf etc?
[15:27:09] <skunkworks_> ace converter works pretty good - but is is a windows app
[15:27:40] <skunkworks_> http://www.dakeng.com/ace.html
[15:30:08] <fenn> this worked for me: http://www.christian-kohloeffel.homepage.t-online.de/dxf2gocde.html
[15:30:20] <fenn> and there is more at http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam
[16:11:06] <archivist> seems I have LOTS of dc printer servo motors(ex skip but were new stock)
[16:13:37] <pjm> i'm looking for a stepper motor for my Z axis something like 60v 3.5A like my X/Y motors
[16:17:58] <archivist> I have 1 fat stepper here with no writing on it
[16:18:12] <archivist> ex CPC
[16:26:53] <archivist> 73mm dia at the mount body 85m dia 95 long
[16:31:07] <skunkworks_> 85m? yikes. don't accellerate it too fast - it will change the earths rotation
[16:31:44] <dimas_> :)
[16:33:01] <alex_joni> skunkworks_: mm :)
[16:33:14] <skunkworks_> ;)
[16:33:21] <alex_joni> * alex_joni imagines a giant gecko G21000
[16:33:22] <archivist> skunkworks_, I wish it had a label! but its worth guessing a few amp and lots of volts probably
[16:34:39] <skunkworks_> measure the windings
[16:34:51] <SWPadnos> turn it at some speed and measure he voltage you get
[16:34:54] <SWPadnos> s/he/the
[16:34:55] <skunkworks_> well crap - that probably won't work
[16:34:56] <SWPadnos> /
[16:35:36] <SWPadnos> then estimate the speed you'd like to run it at (or the max speed you think it can do), and that's the voltage you want to use
[16:35:37] <jepler> anybody recall the name of that company who has sold stepper motors at past cnc workshops? they had quite a range of sizes, including nema 23 and 34. I didn't see anything as big as 73mm diameter though
[16:35:43] <SWPadnos> keling inc
[16:35:48] <SWPadnos> ?
[16:35:56] <jepler> is this 73mm diameter motor one of those with the huge integrated heatsink?
[16:36:04] <jepler> SWPadnos: sounds right
[16:36:53] <archivist> hmm about .5 ohm coil
[16:37:52] <jepler> er, wait -- maybe I don't know how big 73mm is. damn my american birth.
[16:38:30] <archivist> 2.86 inch thats just the mount circle
[16:39:05] <archivist> body is 3.38"
[16:41:07] <SWPadnos> NEMA 34 then
[16:41:29] <awallin> now I'm running latency-test with everything installed on a PATA drive as opposed to SATA previously. will be interesting to see if the 155ms delay is still there
[16:41:45] <SWPadnos> are you using ext3?
[16:42:00] <awallin> hm, dmesg says: spurious 8259A interrupt: IRQ7
[16:42:21] <awallin> SWPadnos: sure, I think ubuntu defaults to ext3
[16:42:37] <SWPadnos> probably. how big is your SATA drive?
[16:43:05] <awallin> I think it was 160g with about half or more for a winXP partition. The pata drive I have now is 40g
[16:43:59] <SWPadnos> ok. I'm not sure how much it matters, and 155ms is a very long time, but I found that kjournald caused latency spikes every 5 seconds or so
[16:44:08] <SWPadnos> it's probably not your problem, I was just curious
[16:56:16] <awallin> well no luck with the pata drive. same 155ms delay keeps repeating itself every 10min. I'm going to change machines now.
[17:00:19] <pjm> archivist, interestin re your stepper motor... i guess without a part # its quite hard to tell something about it
[17:00:28] <pjm> unless u open it up etc
[17:00:51] <archivist> should be "just big enough" TM
[17:01:13] <pjm> the ones i'm using on my XY are 3.75" dia and 2.75" long
[17:01:32] <archivist> reasonable size also
[17:02:06] <pjm> is there a place online to trade/swap parts of cnc stuff?
[17:02:30] <pjm> for example i have some guide rails etc that i need to trade for something else useful!
[17:04:35] <archivist> hmm guide rails
[17:04:45] <archivist> hmm stepper
[17:05:34] <pjm> exactly!
[17:05:39] <pjm> one sec i will make a foto
[17:05:44] <archivist> I want to increase my axis movement
[17:06:58] <alex_joni> awallin: this is still with the m5i20 plugged in?
[17:09:24] <awallin> alex_joni: yes, the m5i20 was in the machine.
[17:10:41] <pjm> archivist, the guide rail is only 750mm long
[17:10:47] <awallin> I'm now trying on another machine, first without the m5i20...
[17:10:50] <pjm> but the guide bar thing is 115 long
[17:11:41] <archivist> pjm my machine atm only has 90 by 120mm travel, so thats an improvement
[17:12:15] <gene> humm, got firefox prob, needs gtk2 >=2.10, 6.06 seems to be stuck at 2.8. Is fixable?
[17:13:01] <pjm> archivist, http://pjm.dyndns.org/cnc/guide1.jpg
[17:13:10] <pjm> that one is 750mm long
[17:13:54] <archivist> hmm nice
[17:14:00] <pjm> archivist, http://pjm.dyndns.org/cnc/guide2.jpg this is 1150mm long
[17:14:51] <pjm> there is virtually no slop that i could detect in the horizontal rail
[17:16:00] <archivist> google the part number, play is specified for some rails, mine are 10micron play
[17:16:12] <pjm> wow yeah thats not much is it
[17:16:59] <pjm> the block has thk hrw27 nok11 on it
[17:17:04] <archivist> I use 2 for z and "assume" i lose the play in machine build
[17:17:35] <alex_joni> awallin: can you try the same machine without the 5i20?
[17:17:42] <pjm> yeah that rail i took the foto of is the smaller of a pair, and the other is on my Z axis with 2 of the bearing carriers
[17:18:08] <awallin> alex_joni: yes, I probably will later today, now I'm testing another new machine
[17:18:52] <archivist> pjm, you could perhaps make a wider and stiffer col and use both
[17:19:54] <alex_joni> awallin: ok, thought that might go faster
[17:19:56] <alex_joni> bbl..
[17:39:38] <gene> must have gotten lost in the noise, got firefox prob, needs gtk2 >=2.10, 6.06 seems to be stuck at 2.8. Is fixable?
[17:42:05] <rayh> gene, did you get the firefox using apt and the 6.06 repositories or from some other source?
[17:42:33] <archivist> if using the ubuntu repositry they wont give out a firefox without a gtk so if working outside the system you fixup as neede
[17:47:55] <gene> Got it with adept, std repos as far as I know
[17:48:38] <gene> since its dead, I'll remove & reinstall just for effect. brb
[17:49:03] <archivist> hmm thats a windows method
[17:52:02] <gene> but it worked!
[17:57:18] <rayh> There you go.
[17:58:12] <alex_joni> hey ray
[18:30:54] <SWPadnos> JymmmEMC, nothing at the moment (re: the artigo)
[18:37:48] <rayh> Hi alex_joni
[18:38:49] <alex_joni> cool stuff for the workshop it seems :)
[18:38:57] <alex_joni> too bad I'm missing it again
[18:39:26] <rayh> I just spoke with Roland. He plans to get the front page for the web site up in a couple days.
[18:39:45] <rayh> Says he's feeling great but takes so many pills he rattles when he walks.
[18:39:57] <alex_joni> haha, that's both sad and funny
[18:40:17] <rayh> Getting old is that way. But it's better than the other.
[18:41:24] <rayh> How you doing on the big projects?
[18:41:38] <alex_joni> slowly getting up to speed
[18:41:52] <alex_joni> one of them (actually a very small one) is running
[18:42:08] <alex_joni> next week we'll get another one
[18:43:47] <SWPadnos> "I intend to live forever. So far, so good" -- Steven Wright
[18:44:04] <alex_joni> at least it's not Padnos :)
[18:44:17] <SWPadnos> indeed
[18:44:24] <SWPadnos> plus I use "ph" instead of "v"
[18:44:38] <alex_joni> ssdi
[18:44:39] <jepler> you prefer that people call you 'stephe'?
[18:44:48] <SWPadnos> not really
[18:45:34] <SWPadnos> sed s/v/ph/ < "Steven" should yield "Stephen" :)
[18:45:42] <awallin> now my new machine works without real-time delays with the same graphics card and with the m5i20 installed. I'm probably not going to bother troubleshooting the problematic machine anymore - it just has this 155ms mysterious delay every 10 minutes.
[18:45:49] <SWPadnos> (with appropriate here file syuntax of course :) )
[18:46:02] <SWPadnos> -u
[18:46:09] <alex_joni> awallin: ok, fairy nuff
[18:47:06] <alex_joni> awallin: what did you end up installing?
[18:47:11] <alex_joni> 6.06 or 8.04?
[18:47:54] <awallin> alex_joni: 8.04 from CD and then the install-script
[18:47:55] <gefink> awallin: how solved you the problem?
[18:48:11] <alex_joni> gefink: different PC
[18:48:14] <awallin> gefink: got another machine with another motherboard and another cpu
[18:48:16] <gefink> hehe
[18:48:29] <gefink> good solution
[18:48:49] <alex_joni> at least you spend the money on something decent, not on a piece of software :D
[18:49:01] <ds2> it looks like SWPadnos's talk can get UCSD extension credit!
[18:49:10] <SWPadnos> haha - too funny :)
[18:49:31] <awallin> how can I get GRUB to find my windows install and show it as a boot option?
[18:49:34] <SWPadnos> if I give enough talks, I may be able to graduate :)
[18:49:50] <ds2> heh...
[18:50:00] <ds2> * ds2 waits for them to offer speakers honary PhDs ;)
[18:50:21] <alex_joni> awallin: do you know on what partition it is?
[18:50:56] <awallin> alex_joni: hmm, ubuntu just shows me the disk label...
[18:51:10] <gefink> aw: ubuntu found ever all bootable partitions automatic
[18:51:27] <alex_joni> awallin: I have this in /boot/grub/menu.lst
[18:51:30] <alex_joni> http://pastebin.ca/1009604
[18:51:38] <alex_joni> (but it got put there by the installer)
[18:51:55] <alex_joni> awallin: sudo fdisk /dev/hda (then p for print)
[18:52:42] <gefink> sudo fdisk -l
[18:52:50] <awallin> sda is my linux install
[18:52:54] <awallin> so maybe sdb then
[18:53:01] <alex_joni> try sudo fdisk -l
[18:53:04] <gefink> but my hda is alwasy detected as sda from ubuntu
[18:53:05] <alex_joni> as gefink suggested
[18:53:23] <alex_joni> gefink: you can change that if you really want I think
[18:53:35] <ds2> that is because it is going through the idescsi layer
[18:53:37] <alex_joni> boot option all_generic_ide or something like that
[18:54:32] <gefink> alex_joni: i dont want rely, but sometimes hdaparm accepts no sda
[18:55:13] <alex_joni> gefink: sometimes?
[18:55:23] <gefink> check dma ....
[18:55:57] <gefink> some functions are not implemented for scsi
[18:55:59] <alex_joni> HDIO_GET_DMA failed: Inappropriate ioctl for device
[18:56:03] <alex_joni> that one? :P
[18:56:09] <gefink> yes
[18:56:19] <awallin> it's in /dev/sdb1
[18:56:19] <alex_joni> well.. it seems to work ok here..
[18:56:26] <gefink> tare is a patch....
[18:56:28] <alex_joni> (but this is a sata disk anyways)
[18:57:19] <gefink> but i have a patched hdparm who van handle ide-scsi un old suse
[18:57:29] <gefink> van=can
[19:04:31] <alex_joni> bbl
[19:06:23] <gene> I was under the impression that we had a newer hdparm called sdparm?
[19:07:59] <gene> from the above thread, I take it 8.04 is ready to be used now??
[19:08:25] <alex_joni> weeeell.. it should work ;)
[19:08:48] <alex_joni> I heard no really bad experiences involving it..
[19:09:52] <BigJohnT_> alex_joni: I've been running the plasma table on pre 2.3 a lot with no problems
[19:09:59] <BigJohnT_> what's wrong
[19:10:34] <gene> humm, I have the dvd burnt, but its plain ubuntu, and I'm a kde fan, can it be switched easily?
[19:10:54] <alex_joni> gene: I think it's sudo apt-get install kubuntu-desktop
[19:11:02] <alex_joni> (and lots of downloads.. :)
[19:11:15] <gene> ok, how about saving my home tree?
[19:11:28] <alex_joni> gene: any reason for upgrading?
[19:11:34] <jepler> the worst thing to do to a working machine is to change it
[19:11:34] <jeeten> hi guys
[19:11:37] <archivist> synaptic, clicky clicky clicky go...
[19:11:38] <jepler> hi jeeten
[19:11:45] <gene> newer other stuff mostly
[19:11:55] <alex_joni> gene: 6.06 is still more than a year supported
[19:12:04] <jeeten> built my first router in january...expanding to gecko in 1 month...
[19:12:05] <alex_joni> I see no reason to change that with something newer
[19:12:11] <jeeten> cam system homemade
[19:12:12] <alex_joni> jeeten: great
[19:12:25] <jeeten> cam system on matlab.
[19:12:33] <BigJohnT_> BigJohnT_ is now known as BigJohnT
[19:12:43] <jeeten> my quaetion: what would you guys really like in a cam system?
[19:12:49] <awallin> jeeten: what kind of toolpaths can you generate?
[19:12:59] <ds2> cam system in matlab? is that a new ploting functionality?
[19:13:03] <gene> ok, so I leave it be, till at least the stress test says I have some time left, I had a bit of a scare last week, pulse down to 44 & chest pain
[19:13:14] <jeeten> i generate in 2d right now..
[19:13:26] <awallin> jeeten: have you seen my drop-cutter tests?
[19:13:29] <jeeten> i haven't figured out how to solve tool path offset in 3d
[19:13:46] <jeeten> no awallin...i am new in cnc
[19:13:47] <ds2> CNC mill + chunks of machineable wax might make a nice answer to all those fancy graphs the mathematica guys used to show
[19:13:49] <jepler> jeeten: yeah it is not an easy topic .. awallin is the guy in this channel to talk to about it, though
[19:14:06] <gene> Neither the sawbones nor I were real fond of the ekg's we pulled an hour later. But I've piecemeal mowed the front yard & still sucking air, so for the instant, I'm ok.
[19:14:16] <awallin> my idea is to code drop-cutter in c++ and have the c++ library callable from python
[19:14:23] <jeeten> wait..wait...too fast.. i can't keep up with the conversation
[19:14:48] <awallin> jeeten: this is a simple test: http://www.anderswallin.net/wp-content/2008_03dropcut/drop-cutter.png
[19:14:54] <jeeten> awallini...whwre is a website for this ...i don't even know what a drop cutter is
[19:15:02] <alleycat_tn> hey everyone. new guy here. i was just curious about how "active" this irc was. cuz i know i'm gonna need help when i set up my cnc
[19:15:03] <archivist> jeeten, switch client logging on you can read at your leasure
[19:15:42] <archivist> alleycat_tn, depends on time of day but usually someone here
[19:15:53] <awallin> jeeten: here's the cam category of my blog http://www.anderswallin.net/category/cnc/cam/ this is also very good http://www.freesteel.co.uk/wpblog/
[19:16:25] <jeeten> aaaahh...exactly why i dropped out of 3d machining..
[19:16:32] <alleycat_tn> archivist, good to hear and good to see alot of people in here
[19:16:39] <jeeten> i couldn't figure out the algorith to offset the tool in 3d
[19:16:51] <jeeten> :)
[19:17:08] <jeeten> but i don't use it here...we are mostly 2d people for the time being
[19:17:31] <SWPadnos> hi jeeten
[19:17:43] <SWPadnos> you'll be 3D once you start using V-cutters :)
[19:17:45] <jeeten> yo swpadnos my man!!!!
[19:17:55] <jeeten> i am a famous man in my country thanks to you!!!!!
[19:17:58] <SWPadnos> hoe's the weather?
[19:17:59] <SWPadnos> heh
[19:18:01] <SWPadnos> oh no :)
[19:18:08] <SWPadnos> err - how's the weather
[19:18:10] <jeeten> super warm...minister in person came to see me..
[19:18:16] <jeeten> ia m famous or waht!!!!
[19:18:19] <SWPadnos> heh - cool
[19:18:26] <jeeten> thank to you bro..
[19:18:39] <SWPadnos> any time
[19:18:40] <jeeten> i haven't forgotten my promises to you
[19:18:55] <alex_joni> SWPadnos: no that's a nice gratitude :)
[19:19:00] <SWPadnos> what's the weather like (and the crowds) in September / October?
[19:19:05] <jeeten> any time you need vacation...mail me..i'l do the rest
[19:19:30] <jeeten> with global warming..weather gone crazy..
[19:19:33] <SWPadnos> heh
[19:19:36] <jeeten> we had snow
[19:19:40] <SWPadnos> whoa
[19:19:44] <jeeten> tropical country with snow!!!!!!!
[19:19:49] <SWPadnos> what's the highest peak there?
[19:19:51] <jeeten> winter gone..
[19:19:56] <jeeten> summer 365/365
[19:20:28] <jeeten> it did not fall on a peak..in rain catchment aresa...golf ball size things made of ice
[19:20:34] <SWPadnos> oh, hail
[19:20:37] <SWPadnos> even better
[19:20:42] <jeeten> yah...that..
[19:20:57] <awallin> jeeten: do you have a website with your cam code or some screenshots ?
[19:20:59] <jeeten> the snow..like cotton fell on reunion...they were lucky..
[19:21:08] <jeeten> i have never seen snow
[19:21:38] <jeeten> next time i guess
[19:21:38] <SWPadnos> maybe I can bring you some (I'd have to find a well-insulated piece of luggage :) )
[19:21:48] <jeeten> (reunion is an island a bit norther)
[19:21:55] <jeeten> your call..anytime
[19:22:12] <jeeten> well anyway...
[19:22:19] <jeeten> did you see my first router?
[19:22:38] <SWPadnos> there's a possibility that I'll take a trip to Germany then India in september/october, so I might (20% chance maybe) be able to drop in
[19:22:55] <SWPadnos> I think so - you posted some links on the cad-cam list some time ago, right?
[19:23:12] <jeeten> awallini..can i send you the pics?
[19:23:21] <jeeten> its in geckodrive forum
[19:23:35] <SWPadnos> ok, different list - I looked at them ;)
[19:23:39] <alex_joni> jeeten: where are you located?
[19:23:42] <SWPadnos> (couldn't remember which list it was)
[19:23:48] <SWPadnos> http://en.wikipedia.org/wiki/Mauritius
[19:24:04] <jeeten> mauritius..tiny isalnd off coast of madagascar...(the isalnd,not the movie)
[19:24:24] <alex_joni> cool.. got any stamps?
[19:24:40] <jeeten> lots...blue penny...mauritian stamp
[19:24:48] <alex_joni> yeah, that one :P
[19:25:01] <jeeten> seriously...its mauritian..
[19:25:15] <alex_joni> jeeten: add yourself to http://www.frappr.com/emc2
[19:25:26] <alex_joni> (this applies to everyone else too :)
[19:25:27] <jeeten> trust a mauritian to make mistake on 2 letter words...and make millions out of it
[19:25:39] <jeeten> ok alex..
[19:25:57] <alex_joni> jeeten: great to hear about new parts of the world where emc is beeing used
[19:26:23] <jeeten> yup...i was educating them on nuances between realtime and hardrealtime
[19:26:26] <alex_joni> I'm probably closest to you atm .. only 2 timezones away
[19:26:28] <jeeten> :)
[19:26:47] <alex_joni> maybe 3 :)
[19:26:52] <jeeten> actually i didn't even know what i was talking about...but they bought it anyway
[19:27:08] <SWPadnos> heh. from Google maps: "We could not calculate driving directions between mumbai and mauritius."
[19:27:45] <jeeten> you can take the plane from bangalore...its direct flight in 5hr 30 mins
[19:27:50] <jeeten> everyday i think
[19:27:51] <SWPadnos> jeeten, I'll send you an email with a paper I just did on the history of EMC - it has some description of realtime stuff
[19:28:05] <SWPadnos> oh right - I think the conference is in Bangalore anyway, not Mumbai
[19:28:33] <alex_joni> SWPadnos: long swim?
[19:28:37] <SWPadnos> heh
[19:28:54] <jeeten> oh...don't worry..we are sticking to emc...we don't have to worry about these things anymore..
[19:29:04] <jeeten> we are focusing on the novel cam aspect.
[19:29:46] <SWPadnos> as for CAM, it would be nice to be able to load "standard" CAD file formats (STL, DXF, DWG, parasolid, IGES, ACIS ...) and generate toolpaths from them
[19:29:58] <SWPadnos> should be simple :)
[19:30:01] <jeeten> stl can..
[19:30:19] <jeeten> stl is cool..as i am building the FARO arm next
[19:30:22] <gene> stl?
[19:30:26] <alex_joni> SWPadnos: forgot STEP :P
[19:30:27] <jeeten> i will need stl import
[19:30:32] <SWPadnos> stereolithography file
[19:30:47] <alex_joni> bbl..
[19:30:53] <SWPadnos> gotta get back to work, just thought I'd say hi
[19:30:55] <gene> hoo boy, I got lost at turn 1
[19:31:01] <alex_joni> gene: n/m ;)
[19:31:05] <SWPadnos> good luck, I'll email you when I have a chance
[19:31:12] <jeeten> ok then...do remember about your vacation in Mauritius
[19:31:19] <SWPadnos> I'm remembering :)
[19:31:52] <gene> Sounds like an offer you can'r refuse...
[19:31:59] <gene> Sounds like an offer you can't refuse...
[19:32:07] <jeeten> so basically.if you are getting point cloud data from a 3d scanner.. a stl import is super useful
[19:32:31] <jeeten> i couldn't get dxf descriptions enuff to make a dxf importer
[19:32:43] <jeeten> but ascii stl is within my reach
[19:33:13] <jeeten> stl is actually for standard tesselated language
[19:33:26] <jeeten> not stereolithography
[19:33:48] <jymm> SWPadnos: what ya doing with your artigo?
[19:35:02] <SWPadnos> jymm, nothing yet
[19:35:05] <gefink> jeeten: the toolshape-compensation is a problem
[19:35:07] <SWPadnos> haven't even booted it
[19:35:21] <gene> Question re spindle encoder on mini-lathe: How many quadrants would be best since max rpms are about 2500?
[19:35:22] <jeeten> heres is what i do for 3d..rather emulated 3d
[19:35:25] <jymm> SWPadnos: Ah, well it runs knoppix damn fast btw
[19:35:30] <jeeten> consider a v carve bit
[19:35:37] <jymm> SWPadnos: even with only 512mb
[19:36:11] <SWPadnos> one day, I'll attach it to a monitor and start it up
[19:36:15] <jeeten> basically the tip can get in between a valley...that is a v shape path
[19:36:38] <jeeten> but if the angle of the valley is more acute that the v carve..the edges will get eaten away..
[19:36:45] <jeeten> that point we all get that problem
[19:37:19] <jeeten> what i do is i sample a circular reghion around the tooltip coordinate
[19:37:45] <jeeten> and check if any "pixel" falls within the v carve cutting region
[19:38:00] <gefink> gene: lower as 200, but for 200 you need a fast PC
[19:38:41] <jeeten> if any pixel is higher than the v-carve cutting region,well the generated g-code simply is made to give a small raise ...
[19:38:43] <jeeten> thats all
[19:38:44] <gene> gefink: which this isn't, XP-1400 runnng at 1600
[19:39:17] <gefink> gene: i think 100 is realistic. whats your latency
[19:39:22] <jepler> 200 * 4 * 2500 / minute = quadrature edges at 33kHz. It's hard to read that fast in software. The absolute max is 1/(BASE_PERIOD + jitter) but I'd only go half that fast for comfort.
[19:39:29] <gene> 14,000 or so
[19:39:37] <jepler> (and remember that's *edges*, not *cycles*)
[19:39:48] <gene> yup
[19:40:01] <jepler> 200 cycles * 4 edges/cycle * 2500 / minute = 33.3 k edges / second
[19:40:23] <jeeten> do you think my approach is too naive? mathematically speaking?
[19:40:25] <gene> I don't think my base period is that fast
[19:40:55] <gefink> gene: then use a lower resolution
[19:41:28] <jepler> or just promise yourself you'll only thread at lower RPMs
[19:41:29] <gene> I think I'm running at about 50 u-sec atm, but there is room for more, probably down to 25 if I push it.
[19:41:55] <gefink> jeeten: i think its correct, but hard to do
[19:42:05] <gene> 2500 rpm & threading don't mentally fit in the same pix here :)
[19:42:18] <jeeten> actually,its algorithmically simpler than polygon intersection routines.
[19:42:32] <gene> 250 wouod be a lot more like it.
[19:42:37] <jeeten> it wastes a lot of calculation cycles to generate a g-code file
[19:42:42] <gene> 250 woud be a lot more like it.
[19:42:49] <awallin> jeeten: I've been working on a machining algorithm for STL surfaces that uses cylindical, spherical or toroidal tools. Adding V-shaped tools would not be that hard
[19:43:11] <jeeten> are these algos available in emc?
[19:43:14] <awallin> jeeten: if you have some pics or code, you can send it to anders.e.e.wallin at gmail.com
[19:43:16] <jepler> say you are sure the spindle won't ever go above 400RPM during threading -- that changes the calculation to 200 * 4 * 400 / minute = only 5kHz edge rate
[19:43:25] <jeeten> wait...i not that down
[19:43:46] <jeeten> could you send me an email awallin? jeetendra_g10@yahoo?
[19:43:50] <awallin> jeeten: no they are very much experimental right now, but I hope to work on them some more during the next few months
[19:43:54] <awallin> jeeten: sure thing
[19:43:56] <jeeten> your email too long to note down
[19:44:14] <gene> Actually, I had something more along the 36 slots range in mind if emc can interpolate from that?
[19:44:40] <jeeten> mine too...but i have no reference...so i don't know if i am screwing up..
[19:44:46] <jeeten> but its beautiful anyway
[19:44:48] <gene> 36 plus an index of course
[19:44:59] <jepler> gene: no, you want at least several edges per 1ms
[19:45:12] <jeeten> considering i never used a cnc router.. i am kinda proud o my baby..
[19:45:47] <awallin> jeeten: are you familiar with Python or C++, that is the combination I'm going to use next
[19:46:02] <gene> Humm, that sounds like fawncy optical stuff, maybe from a mouse to get that sort of resolution?
[19:46:31] <jeeten> too long ago...but i could brush up if its important..
[19:46:46] <gefink> gene: 100 quadrature pulses would be better.
[19:47:15] <jeeten> normally i code directly in matlab..much much easier than c++
[19:47:15] <gene> Ok, that I might be able to carve with the rotary table
[19:47:42] <awallin> jeeten: ok. I'll post on my blog, here on irc, and maybe on the emc-list when I get something nice done.
[19:47:45] <gene> That would be 50 slots, right?
[19:47:53] <jeeten> when i have validated my toolpath sever ,then i will migrate back to c++
[19:48:15] <gefink> gene: last time the mouses have integrated Pulscoded encoders
[19:48:20] <jeeten> just to note: i will have a toolpath server soon..
[19:48:37] <awallin> jeeten: what do you mean by toolpath server?
[19:48:48] <gene> yeah, and talk to seriel waaay tooooooo damned slow.
[19:48:49] <jeeten> this means you can buy online g-code from original authors
[19:49:22] <jeeten> people who design from my cam software and post their work in catalogue ..then you can buy original art
[19:49:24] <gene> I'd just steal the opt stuff from the mouse
[19:49:41] <gene> s/opt/opto
[19:49:42] <awallin> jeeten: ah, you are going to sell the g-code?
[19:49:51] <jeeten> depends on the authors..
[19:50:04] <jeeten> some will give for free..some will ask for money..
[19:50:05] <gefink> gene: ther is only a integreated 3pin encoder
[19:50:05] <jeeten> depends
[19:50:21] <awallin> I have to run now, catch yall later
[19:50:31] <gefink> gene: poor speed, no exact turn
[19:50:42] <jeeten> also..an ultra highly detailed persian bas relief will not retail for the same price as a small flower
[19:50:42] <gene> I've seen those, and to beat all, they are pnp!
[19:51:09] <gefink> gene: the encoder from printers ar better
[19:51:18] <gene> I used to make amiga mice until they started that BS.
[19:51:38] <gefink> gene: but they sheldon have rotaryencoder
[19:51:41] <Sweeper> awsome! the router bits I got from goldmine electric fit my dremel with no adaptations
[19:51:41] <gene> Humm, suggested brand names?
[19:52:24] <gefink> no, the resolution of actual epson is to high
[19:52:42] <gene> Oh, you mean the carriage position strip, not ultimately usable I'd think. too fine
[19:53:05] <jeeten> (encoder disks are custom made at us digital)
[19:53:29] <gene> For what sort of a sheckle count?
[19:54:08] <gene> swag of course
[19:54:27] <jepler> personally I'd also go for a commercial encoder rather than try to build it myself. http://www.usdigital.com/products/incremental-shaft-encoders.shtml under $100 depending on the model and options you require
[19:54:37] <fenn> gene: could you say that again in english?
[19:55:10] <jeeten> aligning encoders is more complex than building a down counter to reduce counter rate
[19:55:17] <gene> Heck, I can probably make gimp print me one for under $100 :)
[19:55:36] <jeeten> a laser printer can print them
[19:55:51] <jeeten> but you won't get mylar
[19:56:00] <gene> Yes, but I do have a ddual trace 100mhz scope, and in english, that swag is a Scientific Wild Assed Guess. :)
[19:56:09] <jeeten> the plastic they use on slide printers..
[19:57:21] <gene> I had more of a dremel diamond disk in mind, cutting slots in/on a brass disc.
[19:57:38] <gene> I could do that with the mill
[19:57:46] <fenn> now i feel obligated to mention http://fennetic.net/machines/encoders
[19:57:47] <jeeten> why don't you check the laser diy forum?
[19:57:49] <Sweeper> * Sweeper is a good father and cuts security devices off his daughter's clothing with his dremel
[19:57:49] <jeeten> they do that
[19:58:00] <gefink> gene: the deskjet-enoder could have a usable resolution
[19:58:22] <gene> and those fill the landfills...
[19:59:07] <fenn> jeeten: link?
[19:59:12] <gene> I'll have to check and see if any of those POJ's still exist around the tv statrion.
[19:59:34] <Sweeper> gene: yea, the only problem is, they use the actual belt, so you're limited there
[20:01:02] <gene> Oh, the belt is the encoder, good grief...
[20:01:04] <jeeten> gimme a minute..
[20:01:23] <jeeten> diy_laser_cutting on yahoo groups
[20:01:31] <alex_joni> good night all
[20:01:42] <gefink> alex_joni: good night
[20:01:53] <jeeten> gnite alex_joni
[20:01:58] <gene> I'll grab the stuff from that link above and see if my printer can do them, thanks.
[20:02:11] <gene> gnight Alex
[20:02:33] <jeeten> gene..are you trying to cut an encoder disk ?did i understand correctly?
[20:02:51] <gene> Thats the general idea now that I have a working A axis
[20:03:28] <gefink> gene: many optodecoder use a gate so that the resolution must match exactly
[20:03:29] <jeeten> right..you mean an indexer?
[20:04:09] <jeeten> also ..an encoder has 2 photodiodes that are spaced precisely to match the encoder disk/strip
[20:04:17] <fenn> not cut it, print laser toner on clear plastic film
[20:04:23] <Sweeper> gene: no, the encoder is on the print head
[20:04:32] <jeeten> otherwise the phase channels a and B don't give a valid phase count
[20:05:08] <fenn> i think you can move it in/out a bit to align the phases
[20:05:14] <gene> At the resolutions of a mouse opto unit, that should be doable with one of the little dremel diamond disks. If they are 3 pin, I can figure that out.
[20:05:19] <jeeten> gene..what problem are you having with commercial encoders?
[20:05:25] <gene> cost
[20:05:50] <jeeten> they are 60 $..
[20:06:00] <gene> plus I like to be able to say "I made it myself'
[20:06:00] <gefink> gen: forget the 3-pin optos. they are bad
[20:06:16] <fenn> i for one don't have $360 to spend on encoders..
[20:06:25] <gene> are they more than just a common supply pin?
[20:06:55] <jeeten> well..are you making that for servos?
[20:07:00] <jeeten> gecko g320?
[20:07:03] <gefink> gene: vcc,gnd,and bidirectional data
[20:07:22] <gene> bi-directional data?
[20:07:31] <fenn> gene: i thought you had stepper motors already?
[20:07:38] <gene> yup
[20:08:11] <jeeten> so you are trying to track the axis with a homemeadfe encoder ?
[20:08:33] <archivist> heh
[20:08:39] <fenn> maybe linear scales would make more sense then
[20:08:51] <gene> Or, now that steppers have been mentioned, how about using one of those for the quadrature generator?.
[20:09:17] <gene> Seems I didn't stete this, its for the mini-lathes spindle encoder
[20:09:18] <gefink> gene: wait 1 ms send pulse wait for leftpulse, wait 200us, send pulse wait for rightpulse .....
[20:09:38] <K`zan_emc> New EMC :-)!
[20:09:48] <gene> so emc can track the rotation of the lathes spindle
[20:10:05] <jeeten> oic..
[20:10:30] <jeeten> to pulse modulate the RPM...not to keep angular position?
[20:10:39] <jeeten> is that it?
[20:10:52] <fenn> to keep track of position, for threading
[20:10:55] <gefink> gene: a stepper produce an autputvoltage while moving
[20:11:16] <jeeten> impossible.(just my practical opinion)
[20:11:48] <gene> and have an idea, within a few degreees, where its at., no, juxt to track it, although I may put in a relay so I can run the spindle with my PMDX-106
[20:12:28] <fenn> well, as long as the spindle is moving above some threshold velocity (voltage is ~ proportional to velocity)
[20:12:43] <gene> As for steppers as generators, if the motor is high voltage, it will be very very hard to move it slow enough to loose a count
[20:12:48] <jeeten> either way..the sensor,the disk must match physically
[20:12:56] <gene> s/loose/lose
[20:13:44] <gefink> gene: no. you loose it for sure if get to slow
[20:13:51] <gene> And I don't belieeve this controller can move it that slow, even its fit and noise moves are 30 degrees at a whack
[20:14:45] <gene> I have one motor that when turned by hand, generates about 20 volts at creep along speedss.
[20:15:21] <jeeten> use the spindle as a router...
[20:15:37] <jeeten> and use two motors to rotate the object and move laterally
[20:15:57] <gefink> gene: it can work. but what want you to do?
[20:16:01] <gene> In any event, that would give me 200 quads per rev, and the PMDX-106 can move it a whole lot more precisely than the hand knob can
[20:16:03] <jeeten> that way...you are guaranteed not to depend on the spindle angular control to thread correctly
[20:16:28] <gene> I expect to have an index pulse too, to reset thinhgs
[20:16:49] <jeeten> gnite guys c ya..
[20:17:50] <gene> Humm, didn't mean to run jeeten off
[20:19:00] <gefink> gene: thread can be possible with this
[20:22:57] <gene> I saw that on a turbocnc thread, but I'd think that would have to imply a much stiffer spindle speed in order to interpolate properly?
[20:28:40] <anonimasu> stiffer?
[20:29:01] <gene> speed control
[20:29:39] <gefink> gene: need you only speed-control or also position-control
[20:30:22] <gefink> for the first use an simple optical interuptor and emc-pid
[20:30:34] <gene> The single flag derived index pulse doesn't seem so usable if the bit engagement slows the spindle, so it seems to me we need to know where the spindle is in the rotation
[20:31:18] <anonimasu> it'd work with a single pulse off the spindle then calculate the speed needed for the thread
[20:31:35] <gefink> gen: i mean phototransostor and 50 to 100 slots
[20:31:36] <gene> The other thought is to make the PMDX-106 contol that, but that seems much more complex in terms of the control bandwidth
[20:32:47] <gene> 72 would make it an even per rev divisor.
[20:33:03] <gefink> yes,
[20:33:05] <gene> Lemme go look up the motor I have in my hand, brb
[20:34:21] <gene> and google pleads no hits, damn
[20:35:31] <gene> This is a 24 volot, 80 ohm coil motor, but I believe it to be a 7.5 degree/step motor.
[20:36:19] <gene> Turning it by hand, it can generate enough to mildly shock :)
[20:37:51] <gene> Run its output through an LM339 and that will give 48 per turn, 96 per half step etc.
[20:38:45] <gefink> gene: use som hysteresis to avoid chaos while motor stops
[20:39:01] <gefink> and us protective diodes
[20:39:10] <cradek> gene: I'd like to see your schematic when you get it working - I think I'd like to try it too
[20:39:19] <gene> I'l drag the scope out, chuck it up and see how slow I can still get a good signal from it. bbl later if I don't get rained out, yes to mall that6 of course. later
[20:39:51] <gefink> gene: lm393 is also good
[20:43:43] <K`zan_emc> Hummm, in the new EMC, x axis won't set to zero with the home axis button :-/.
[20:44:56] <cradek> K`zan_emc: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?CoordinateSystems
[20:45:04] <K`zan_emc> Need to make up a dongle of some sort to be able to jog outside of EMC to get thing alligned.
[20:46:05] <K`zan_emc> Humm, at this point that is pure Greek to me...
[20:46:23] <K`zan_emc> easiest to just turn off the drive and crank the screws by hand.
[20:47:55] <anonimasu> hm
[20:48:18] <K`zan_emc> interesting, now if I quit and come back up X is stuck at one - prev version wherever it was, it came back up at zero.
[20:48:39] <K`zan_emc> When will I ever learn not to upgrade unless there is a really compelling reason, sigh.
[20:48:41] <anonimasu> K`zan_emc: no home switches?
[20:48:49] <K`zan_emc> Not yet.
[20:48:54] <anonimasu> fit them as soon as you can ^^
[20:48:59] <cradek> K`zan_emc: you have an offset. it is not because of your upgrade.
[20:49:03] <K`zan_emc> Getting the thing alligned at this point.
[20:49:05] <cradek> that's why I pasted that link
[20:49:28] <K`zan_emc> Wonder how I managed to set and offset?!?
[20:49:30] <cradek> see number 3 for help clearing the various offsets
[20:49:40] <cradek> probably with the touch-off button
[20:50:00] <K`zan_emc> just been jogging the axis to adjust the table, but the button is close I might have.
[20:50:35] <K`zan_emc> Dunno, but I took the 0.0 touch off and it reset it nicely :)!
[20:50:40] <K`zan_emc> THanks folks!
[20:51:12] <K`zan_emc> Fair warning :), now that I have the hardware basically together I'll try not to ask too many dumb questions....
[20:51:41] <gefink> good night
[20:51:51] <K`zan_emc> Heh, of course I need a 2" piece for the head extension, guess which one between 1 and 10 I don't have LOL!
[20:53:22] <gene> LM339 can do 2 axises, how many channels does a 393 have?
[21:09:19] <gene> looked it up, half as wide as the 339. And I have a very good supply of lm339's :)
[21:09:43] <gene> Otherwise similar, max offset 2.5 mv.
[21:10:42] <gene> And I'm getting rained out, so I guess I better go close up the other buildings
[22:16:33] <tomp2> http://imagebin.ca/view/2EKEb16.html loadrt pci_8255 io="111100001010" dir="0xB001"
[22:25:38] <tomp2> swap those 2 params http://imagebin.ca/view/2EKEb16.html loadrt pci_8255 dir="111100001010" io="0xB001"
[23:30:49] <lewin2> lewin2 is now known as lewing__