Back
[01:27:20] <Loetmichel> mornin'
[01:29:53] <nmz787> what is the max speed in khz that linuxcnc can send out steps?
[02:22:26] <Valen> nmz787: linuxcnc doesn't have a limit
[02:22:33] <Valen> its limited by your hardware
[02:22:45] <Valen> and add a mesa card and even that pretty much goes away
[02:41:12] <nmz787> ahh i thought linuxcnc primarily used the parallel port
[02:43:04] <Jymmm> Many use mesa hardware to get past paraport limitations
[02:56:55] <Valen> I've never used a pport linuxcnc machine lol
[02:57:03] <Valen> linuxcnc is too long, i miss emc
[02:58:11] <Jymmm> heh
[02:59:24] <psha[work]> Valen: emc lawyers are at your door
[02:59:38] <Valen> i know i know
[02:59:40] <Valen> wankers
[02:59:45] <psha[work]> clean up irc history!
[03:13:46] <nmz787> how does mesa solve step speed issues?
[03:14:02] <nmz787> does it just upload a precomputed path to the micro/fpga?
[03:14:17] <nmz787> or does it stream it there in chunks?
[03:14:29] <Valen> in between those
[03:14:34] <Jymmm> Mesa *IS* a fpga
[03:14:50] <Valen> the way emc works is it has a servo thread and a motion thread
[03:15:17] <Valen> the motion thread is responsible for putting out steps at a rate told to it by the servo thread
[03:15:36] <Valen> mesa basically puts the motion thread into the FPGA
[03:15:45] <Valen> its pretty minimal what its used for really
[03:15:49] <Valen> "step at this rate"
[03:16:04] <Valen> then that rate is updated by the servo thread (default 1khz i think)
[03:16:21] <nmz787> i dont see anything on mesanet.com about steppers
[03:16:49] <Valen> they have stepper driver boards, you link them to an anything io board
[03:16:58] <Valen> or use geckos or whatever
[03:17:16] <nmz787> emc?
[03:17:25] <Valen> the hostmot2 firmware is what gets loaded by EMC into the anythingio board
[03:17:34] <Valen> emc = old name for linuxcnc
[03:17:37] <nmz787> o
[03:17:55] <nmz787> so do you use emc instead of linuxcnc?
[03:18:03] <Valen> damn lawyers contributing to the eventual demise of society
[03:18:14] <Valen> they are the exact same thing
[03:18:34] <Valen> somebody did a find/replace for EMC/LinuxCNC in the source code
[03:18:37] * Valen ducks ;->
[03:19:21] <Jymmm> s/Valen//
[03:19:46] <Jymmm> s/Valen/\0/
[03:20:04] <Jymmm> Valen: You've been NULLed =)
[03:20:30] <Jymmm> Valen: You are not even nothing, you are less than nothing =)
[03:21:04] <Valen> sif
[03:21:14] <Valen> nulls are a pain in the ass and *far* too much work
[03:21:28] <Jymmm> Valen: Turn in your LinuxCNC ISO and don't left the swarf hit you on the way out!
[03:21:37] <Jymmm> let
[03:23:11] * Valen loves it when a null unexpectedly creeps into code
[03:23:53] <Jymmm> as the hidden char?
[03:23:57] <nmz787> does it matter which anything io board you use?
[03:24:29] <nmz787> does linuxcnc allow you to specify accelleration profiles?
[03:24:53] <nmz787> and change step frequency on the fly (changing from microstepping to full stepping as you speed up)
[03:25:08] <nmz787> i would think a chirp function would be the best accell profil
[03:25:40] <Valen> changing step size is generally a bad idea
[03:25:48] <Valen> your more likley to hit a resonance
[03:26:13] <Valen> is there any paticular reason for your fascination with step rate?
[03:26:26] <Valen> i'm not a stepper person, but its generally not a limit is it?
[03:27:11] * Jymmm returns to practice exam
[03:27:48] <nmz787> why would you hit a resonance if timed correctly?
[03:28:19] <nmz787> if I don't change step rates, at max speed that i want, it would be 342khz
[03:29:04] <nmz787> and I've read that full-stepping is overall more reproducible than microstepping, due to the magnetic field not following a sine wave perfectly
[03:30:10] <Jymmm> nmz787: Reading ALL of this may be helpful to you (you can download the PDF)
http://www.geckodrive.com/support/step-motor-basics.html
[03:30:21] <nmz787> btw
http://diyhpl.us/laser_etcher/laser_etcher/#index7h1
[03:30:34] <Valen> uhhh "it depends"
[03:31:15] <Valen> 40 tpi is pretty fine
[03:32:18] <Valen> i believe generally the rule of thumb is your "resolution" is going to be about 1/2 a step, the microstepping is actually generally used to keep things smoother
[03:32:46] <Valen> if you want it to be *fast* then don't use steppers
[03:33:07] <Valen> also dont use acme
[03:33:19] <nmz787> why not acme?
[03:33:32] <Valen> lots of friction
[03:33:50] <Jymmm> nmz787: is that your page?
[03:33:53] <Loetmichel> Valen: that depends on the nut
[03:33:53] <Valen> how many RPM are you planning on running that leadscrew at over what distance and what diamiter is it?
[03:34:23] <Valen> Loetmichel: some, but if you anti-backlash at all its going to be lots more than a ballscrew, and these guys seem interested in speed
[03:34:46] <Jymmm> nmz787: If you are building a laser. why even use leadscrews at all? Use belts
[03:34:49] <nmz787> Jymmm: yes
[03:34:52] <Loetmichel> if you use a delron/pom nut or better a special plastic with talc in it you have not nuch more friction than a ball bearing thread
[03:35:12] <Loetmichel> delrin
[03:35:14] <nmz787> Jymmm: belts seem like they'd be even less predictable
[03:35:25] <Valen> except he is going to be doing eleventy million RPM with a 40 tpi acme
[03:35:38] <Valen> something like a 16 tpi ballscrew if you want speed
[03:35:47] <Jymmm> nmz787: Most 250W or less laser use belts.
[03:35:49] <Loetmichel> i would opt to a acme or ballscrew wit a very low tpi
[03:36:02] <Valen> are you rastering or vector cutting?
[03:36:10] <Loetmichel> for a lasercutter
[03:36:12] <Jymmm> Valen: both
[03:36:23] <Loetmichel> its more predictable/rrigid than a belt
[03:36:32] <Valen> has anybody solved the EMC + raster issue?
[03:36:36] <Jymmm> Loetmichel: My laser uses belts
[03:36:39] <Loetmichel> if you have high accelereations and g-forces
[03:36:59] <Jymmm> Loetmichel: It's moving a mirroro with zero force involved.
[03:36:59] <Valen> Loetmichel: *much* higher inertia though
[03:37:19] <Jymmm> Loetmichel: there are no cutting forces involved.
[03:37:19] <nmz787> 3280 RPM
[03:37:20] <Loetmichel> i have seen pick&place bots using ballscrews wiht 1tpi
[03:37:22] <Valen> if you want fast raster type operation inertia would be your limiting factor
[03:37:52] <Loetmichel> but i recon: the murror/lens head of a laser is MUCH lighter than a pichup hhead
[03:38:05] <Valen> btw nmz787 be wary of VXB for bearings
[03:38:12] <nmz787> I am using CW laser with vector type stuff
[03:38:19] <Valen> so no raster?
[03:38:24] <nmz787> I am using a blu-ray writer diode
[03:38:27] <nmz787> no raster
[03:38:35] <nmz787> i do not want the laser to turn off
[03:38:37] <Valen> heh, your not going to be going that fast anyway ;-P
[03:38:46] <nmz787> ?
[03:38:57] <Valen> if you want constant laser power you need *lots* of acceleration
[03:39:05] <nmz787> what do you mean?
[03:39:13] <Valen> people with 20W lasers don't move that fast ;->
[03:39:13] <nmz787> i should have more than enough power
[03:39:33] <nmz787> I'm focusing to a <=5 micron spot size
[03:39:50] <nmz787> and cutting about 50-300 microns thick of material
[03:40:17] <Valen> this a uni project or something?
[03:40:34] <Jymmm> Valen: This was rastered
http://i54.tinypic.com/k2mo7n.jpg
[03:40:42] <Valen> Jymmm: nice
[03:40:59] <Valen> heard you can use spray on moly to etch stainless etc with bog standard lasers?
[03:41:15] <Valen> its whats in that $100 a can "ceramic" stuff
[03:41:20] <Jymmm> Valen: thanks
[03:41:36] <Valen> how long did that take?
[03:42:09] <Jymmm> Valen: CO2 Laser under 250W can not mark metal, but it can "cure" a spray on permanent coating for metal.
[03:42:18] <Jymmm> Valen: to engrave?
[03:42:49] <Valen> http://www.louispickering.com/index.php/tutorials/10-laser-engrave-metal-with-a-low-power-laser
[03:42:55] <Jymmm> Valen: 20minutes run time, 12"x12"
[03:43:34] <Jymmm> Valen: That's the "permanent coating" I was speaking of
[03:43:53] <Valen> still rather cheaper than the $100 a can stuff
[03:43:54] <nmz787> Valen: not a uni project, but it was born out of my uni studies
[03:44:23] <Valen> If you want a high speed machine I would build it thusly
[03:44:34] <Valen> heavy (physically heavy) base
[03:44:46] <Valen> lowish TPI ballscrews
[03:44:59] <Valen> with light antibacklash loading in them
[03:45:08] <Valen> servo driven
[03:45:22] <Jymmm> Valen: I'll try the moly, I know of "other" things that will mark, but more of a grey/whitish color
[03:45:43] <Valen> I think they do the same thing with my stainless steel pans
[03:45:47] <Valen> its pretty permanent there
[03:46:28] <nmz787> the ragged walls like in this paper are what i want to avoid
http://diyhpl.us/~nmz787/jmm9_3_037002.pdf
[03:47:08] <nmz787> Jymmm: I think there is cheaper moly spray for bike maintenance
[03:47:37] <Jymmm> Thermark is HEAVILY patented, to even say XYZ will mark metal (commercially) will get you a letter from a lawyer
[03:47:58] <Valen> because they know their only real ingredient is moly? lol
[03:48:05] <nmz787> well i was planning on just moving the material around, and it will probably only be about 20 or 30 grams
[03:48:38] <Jymmm> Valen: It's not moly per-se but a mineral based
[03:49:15] <nmz787> I know cdroms have pretty low TPI, but they also have error correction built in on the data side
[03:49:41] <nmz787> so they can get away with overshooting then tracking back with a fine adjustment
[03:52:14] <Valen> in metric units, what resolution and repeatability do you need?
[03:52:25] <Valen> also have you confirmed your feed rates?
[03:53:22] <nmz787> I desire <=5 microns resolution, preferably <=1micron
[03:53:44] <nmz787> does repeatability mean from full job to full job
[03:53:58] <nmz787> or repeatability from parallel line starting point to the next?
[03:54:40] <nmz787> repeatability should be <=1 micron
[03:54:56] <nmz787> feed rates meaning the max speed I want and cutting ability?
[03:55:18] <nmz787> I haven't confirmed this, but it should be fine from the calculations I've done
[03:56:00] <nmz787> I could slow up the feed rate, or possibly increase the laser power as well, if tests show that is needed
[03:56:47] <Jymmm> Valen: Thanks, I'll check that out
[03:57:31] <nmz787> I am aiming for rapid iteration through designs, so ideally I'd wait ~15 minutes per job, but I think 1 to 1.5 hours is the max before other methods start to be competitive
[03:57:37] <nmz787> (non CNC methods)
[03:58:20] <Valen> right your accuracy requirements are *way* higher than most CnC equipment
[03:58:37] <Valen> by a decimal place
[03:58:51] <Valen> your in big boy territory
[04:00:04] <Valen> I'd suggest doing something similar to what we have done, we have .001mm glass scales on our mill, use that for positioning, and have something with ~10x that for PID loop control
[04:00:05] <nmz787> so is 3k rpm too high for most motors? stepper or not?
[04:00:21] <Jymmm> Valen: Heh, OEM750's have 50K microstepping =)
[04:00:24] <Valen> I don't know steppers that well but i think so
[04:00:40] <Valen> Jymmm: yeah and what accuracy to they claim? ;-P
[04:00:53] <Jymmm> Valen: that's the driver
[04:01:17] <Jymmm> Valen: used in wafer fab
[04:01:25] <nmz787> your microstepping is pretty much limited by your DAC technique
[04:01:25] <archivist_> nmz787, for speed use servos
[04:02:14] <nmz787> archivist_: i'm not sure if 2 meters per minute is fast or not
[04:02:29] <nmz787> and precision trumps speed
[04:02:36] <nmz787> well, precision and accuracy
[04:02:51] <archivist_> you also need closed loop control which is far better with serve
[04:02:56] <archivist_> servos
[04:03:34] <archivist_> microstepping does not mean improved accuracy also
[04:04:32] <archivist_> http://www.micromo.com/microstepping-myths-and-realities.aspx
[04:04:33] <nmz787> Valen: i know of hemocytometers for small grids
[04:08:38] <nmz787> Valen: can you get what you're talking about in a 6 inch (10 cm) square?
[04:11:30] <nmz787> valen
http://www.thorlabs.com/thorProduct.cfm?partNumber=R1L3S3P
[04:14:28] <nmz787> has anyone here used interferometry as feedback?
[04:14:53] <nmz787> that would really be the way to go, except that i'd need my lab to be in some granite cave
[04:33:40] <archivist_> nmz787, voice coil acuator and lvdt feedback maybe
[06:38:22] <gmagno> (THIS TOTALLY OFFTOPIC) hey, sorry to ask this here, but I know there are german ppl here that may help me. THere's a chance of me going to work to southern germany. I'm not sure about the salary being good or not. For a 3 years experienced electronics enginneer, would a 42000euro/year be good? How much money would it means, after taxes?
[06:45:19] <theos> gmagno join ##deutsch
[06:45:59] <gmagno> theos: ah, great! Thanks
[06:46:06] <theos> :D
[07:49:59] <V0idExp> Hi all
[07:51:57] <V0idExp> i have a problem with an axis
[07:52:26] <V0idExp> it is a stepper driven axis
[07:52:48] <V0idExp> i have a backlash error to be compensated
[07:54:29] <V0idExp> the problem is that the error depends on relative distances and not on absolute ones
[07:55:15] <jthornton> see COMP_FILE
http://linuxcnc.org/docs/html/config/ini_config.html#_axis_lt_num_gt_section_a_id_sub_axis_section_a
[07:55:24] <V0idExp> in other words if i go to 100 starting from 50 is different from going to 100 starting from 99
[07:56:49] <V0idExp> the comp file is for relative distances (i.e. deltas) or for absolute positions along axis?
[07:57:12] <V0idExp> or it is possible to choose the type..
[07:57:41] <jthornton> absolute positions iirc
[07:58:50] <V0idExp> so it compensates errors the same way if i go from 5 mm far or 50 mm far
[08:00:18] <V0idExp> i need something that take into account the relative distance (little error for little distance, greater (asinthotic error) error for greater distances)
[08:16:06] <JT-Shop> you might have to write a comp to do that
[08:18:14] <voidexp> Hi all
[08:18:32] <voidexp> my connection drop down
[08:19:06] <voidexp> have you wrote anything about my problem?
[08:19:16] <voidexp> could you help me?
[08:27:55] <archivist_> voidexp, fix the scale if it is a scale error
[08:28:19] <voidexp> it is not an error scale
[08:28:32] <voidexp> scale error..
[08:29:23] <voidexp> it depend from the delta (distance from the last point and the target destination of the axis)
[08:29:38] <voidexp> little delta little error, greater delta greater error...
[08:30:09] <voidexp> but the error never rise over 0.3 mm
[08:43:32] <archivist_> sounds like creep
[08:44:18] <voidexp> what creep means?
[08:46:50] <archivist_> slip between the drive and driven member
[08:47:26] <archivist_> creep is more definable than slip though
[08:48:26] <archivist_> eg a rubber roller miving something will creep because the drive pressure changes the effective roller diameter
[08:48:30] <voidexp> is there any software solution in order to compensate this kind of error?
[08:48:34] <archivist_> driving
[08:49:28] <archivist_> feedback roller to measure the real movement
[08:50:46] <voidexp> i have see the problem by measuring some axis position with a 3d scanner system (romer absolute arm)
[08:51:38] <voidexp> so i'm sure about the error and i can measure it very well but i've not any way to mechanically correct it
[08:52:04] <voidexp> so the only way it would be a software solution
[08:52:41] <archivist_> a fiddle to the scale will be a partial solution
[08:53:00] <archivist_> feedback is better though
[08:54:01] <tjb1_> r00t4rd3d:
[09:25:09] <treats> What does EMC stand for in this context?
[09:26:11] <L84Supper> treats,
http://en.wikipedia.org/wiki/LinuxCNC
[09:32:24] <skunkworks> that isn't quite correct... The inital system wasn't linux but NT based
[09:34:02] <Valen> voidexp: can you locate the source of the error?
[09:34:35] <Valen> that is perhaps a better use of your time than fixing it in software
[09:36:30] <treats> L84Supper: thanks
[10:37:54] <archivist_> rob_h, any idea where round steel suitable for old automatics can be found (was ground iirc) seems missing from macready's (acenta)
[10:43:25] <archivist_> found some but not to the collet size I have
http://www.steelexpress.co.uk/engineeringsteel/EN1A.html
[11:19:43] <DJ9DJ> tag
[11:25:04] <tjb1> your it
[11:28:28] <tjb1> I will win the speaking to supervisor challenge
[11:28:33] <tjb1> They cant keep me on hold forever
[11:29:20] <Jymmm> "Thank you for holding, your estimated wait time is 14 days"
[11:30:52] <L84Supper> "you are now 35,678.945th in line for the next manager to assist you"
[11:32:15] <Jymmm> "If you would like to continue to hold, please press 4759875135598732179*64448843222254#1668877B465767C119973AAA within 15 seconds."
[11:33:21] <Jymmm> (and yes, A B C and D are valid DTMF tones, just not on phones =)
[11:33:31] <tjb1> bitch hung up on me
[11:33:48] <Jymmm> Did you call her a bitch?
[11:33:55] <tjb1> I told her im not calling an international number for save $1 and she hung up
[11:34:24] <Jymmm> and who/what are you calling in the first place?
[11:34:32] <tjb1> This crap card company
[11:34:37] <Jymmm> card?
[11:34:42] <tjb1> $3 fee for paypal use
[11:34:59] <Jymmm> prepaid card?
[11:35:07] <tjb1> No the card our school forces us to use for refunds
[11:35:35] <Jymmm> ah
[11:35:37] <L84Supper> yeah, they make it difficult to research the fees when you open a new account at several banks
[11:36:09] <L84Supper> nothing in writing, and you get different answers based on which branch or manager you ask
[11:36:20] <tjb1> I cant speak to someone without listening to 5 minutes of recorded
[11:36:23] <tjb1> voice...
[11:36:45] <jdh> I have not.
[11:37:19] <pcw_home> Thats the first level filter...
[11:37:39] <tjb1> Ha, I said "Can I speak to someone that knows english…" hungup...
[11:39:44] <L84Supper> heh
[11:45:42] <skunkworks> why would you do this?
[11:45:44] <skunkworks> setp toggle2nist tmax 3607
[11:49:30] <JT-Shop> I have no idea why you would do anything but set it to 0
[11:52:29] <skunkworks> Isn't tmax a parameter that shows you the longest the comp took to run?
[11:52:38] <JT-Shop> http://linuxcnc.org/docs/html/hal/basic_hal.html#_hal_components
[11:52:40] <JT-Shop> yep
[11:53:08] <skunkworks> I think it should be asked what that line supposidly fixed?
[11:53:41] <skunkworks> look - it is in the manual.
[11:53:44] <skunkworks> heh
[11:54:06] <JT-Shop> I asked that queston some time ago and put the answer in the manual
[11:54:13] <skunkworks> It is rw but again- I don't know why you would set it other than to reset it.
[11:54:33] <tjb1> Dirty pricks, if any of your kids go to college and they try to refund to a "HigherOne" account say no.
[11:55:23] <tjb1> If you leave any amount of money in the account for months unattended they will start a $10 a month fee
[11:56:04] <skunkworks> JT-Shop, I started a reply... DId you?
[11:56:19] <JT-Shop> yes
[11:57:12] <skunkworks> Ok. - go for it.
[11:57:44] <JT-Shop> all ready did a while ago
[11:58:45] <skunkworks> oh
[12:00:02] <skunkworks> Got it
[12:04:58] <IchGuckLive> Hi all
[12:22:55] <IchGuckLive> Folks can you all pleas have alock at the New JHTML page i made if the Carousell turns with the pictures of the XmasCircus ->
http://www.landauer-weihnachtscircus.de/lwch1/ht_foto_1213.html
[12:23:27] <IchGuckLive> Thanks
[12:24:09] <IchGuckLive> if you move the mouse over the pic it shoudt stop moving
[12:24:11] <cncbasher> yes it works here
[12:24:19] <IchGuckLive> B) :D
[12:25:38] <andypugh> Woks in Safari/Mac too.
[12:26:02] <Jymmm> andypugh: You are using Safari?
[12:26:24] <andypugh> Yes.
[12:26:49] <Jymmm> andypugh: LOTS of security holes, fyi.
[12:27:07] <JT-Shop> works for me but makes me dizzy
[12:31:10] <IchGuckLive> JT-Shop: Artist love dizzy
[12:33:29] <JT-Shop> lol
[12:40:53] <IchGuckLive> What is the English word for the Squarehole inside a Drill to get the Force thrue the connection
[12:41:14] <IchGuckLive> the tool that makes the squarhole
[12:41:40] <IchGuckLive> my translater strike on this
[12:42:21] <IchGuckLive> Loetmichel: ? Räumnadel eng.
[12:43:48] <IchGuckLive> BroachPushChipBreakers is this wright
[12:47:51] <andypugh> Morticing chisel?
[12:48:17] <jthornton> broach?
[13:01:16] <IchGuckLive> broch seams to be wright
[13:07:57] <nmz787> IchGuckLive: chuck?
[13:08:46] <pcw_home> bobuck
[13:10:18] <IchGuckLive> nothing on ebay to found on this
[13:10:39] <nmz787> IchGuckLive:
http://www.newmantools.com/jacobs/Image159.gif
[13:11:00] <IchGuckLive> http://es.wikipedia.org/wiki/Archivo:BroachPushChipBreakers.jpg
[13:11:06] <nmz787> IchGuckLive:
http://www.newmantools.com/jacobs/info.htm
[13:11:34] <IchGuckLive> nmz787: no
[13:11:51] <IchGuckLive> get a lok at the wikipedia image
[13:11:55] <nmz787> guess I'm not clear on what force you're talking about
[13:12:03] <nmz787> a drill applies rotational force
[13:12:08] <nmz787> to the drill bit
[13:12:19] <IchGuckLive> the connection between motor and pulley
[13:12:31] <IchGuckLive> with the square metal pice
[13:13:05] <IchGuckLive> http://es.wikipedia.org/wiki/Archivo:BroachPushChipBreakers.jpg
[13:13:37] <IchGuckLive> isent there a chinese how sels this for some USD
[13:13:47] <IchGuckLive> i need 3,4,5mm
[13:13:59] <nmz787> IchGuckLive: see the diagrams of drill presses towards the bottom
[13:13:59] <nmz787> http://bbs.fobshanghai.com/thread-3914206-6-1.html
[13:14:34] <rob_h> archivist_, im not sure i dont know many with swiss or other sliding heads to ask , are you looking for a grade of metal?
[13:15:28] <nmz787> IchGuckLive:
http://www.garagejournal.com/forum/showthread.php?t=51384
[13:15:49] <nmz787> "Drive out the rear quill lock sleeve (10) from the front then drive out the front sleeve from the rear................."
[13:15:57] <nmz787> maybe that?
[13:16:15] <nmz787> Remove the feed stop collar (37)..................
[13:16:20] <IchGuckLive> nmz787: it6 has nothing to do with a drillpress
[13:16:26] <nmz787> oh, i meant "Slide the quill/spindle assembly out...................
[13:16:26] <nmz787> "
[13:16:44] <IchGuckLive> NP O.O
[13:16:57] <IchGuckLive> i give up fortoday
[13:17:05] <IchGuckLive> BY
[13:17:35] <Jymmm> Talk Like a Pirate Day!!!
[13:35:57] <skunkworks> arrrrr
[13:50:44] <archivist_> rob_h, the earlier swiss lathes needed precision ground bar, later have a newer type of support collet that can take some variation in bar size
[13:51:22] <rob_h> yea i saw that when i was looking for machine info other day
[13:51:36] <rob_h> but surly it must still have to be fairly close for the speeds they run at
[13:52:47] <archivist_> some have a rotating collet, mine is stationary and the bar is spinning in it
[13:52:54] <rob_h> what is silversteel like? it that close enuth for you on toll?
[13:53:06] <archivist_> probably yes
[13:54:10] <rob_h> like one job im doing now, using en16t, if i could get that in a good finish id be over the moon lol
[13:54:30] <rob_h> instead i have to centerless grind it... which has to be sent away
[13:54:32] <archivist_> just been freeing off the drilling attachment was siezed, may be usable
[13:54:51] <rob_h> grate
[13:55:06] <rob_h> i saw ur post on mail list about it
[13:55:38] <archivist_> the concurrent uncoordinated moves are the fun part
[13:56:18] <archivist_> lots to learn, I need some work for it
[13:56:48] <rob_h> yea i can see that being the problem on it
[13:57:24] <rob_h> same problem someone would have with machine with sub spindle and two or more turrets
[13:57:35] <archivist_> I happen to have 5mm push and support, and suppliers seem to think 4 and 6 are the right stuff!
[13:59:24] <rob_h> ask these people
[13:59:24] <rob_h> http://toolsteels.co.uk/index.php
[13:59:33] <rob_h> i broght silversteel from them alot quite good prices too
[13:59:44] <rob_h> might be able to help you.. worth a try
[14:00:44] <archivist_> yup thanks
[14:01:54] <archivist_> machine has a 4.5 metre bar feed, may shorten a bit as it is not straight
[14:02:24] <rob_h> most steel in 3m anyway
[14:03:15] <rob_h> i guess u might have to get soeone to centerless grind u stock?
[14:03:20] <archivist_> I have a local supplier too, will drop in to them tomorrow
[14:04:08] <rob_h> have to make a nice short magazine barfeed for it :)
[14:05:19] <archivist_> they did have a magazine feed for the later version, would probably suit this one too
[14:07:57] <rob_h> they cost so much tho even 2nd hand, we want one for the chnc
[14:09:34] <rob_h> gota pop out for abit now speak later
[14:11:40] <archivist_> tea time for me as well seeya
[14:14:30] <Jymmm> Speaking of Tea... anonymous binary file sharing
https://FileTea.me/
[14:43:03] <tjb1> hey Jymmm -y
[14:43:40] <Jymmm> Arrrrrrrrr
[14:43:50] <tjb1> Tonight I am gonna try to put my limit switches on
[14:44:04] <tjb1> What do I need to do in linuxcnc to make it read one home switch and home both x and a motors
[14:47:10] <JT-Shop> home is shared iirc
[14:48:01] <jdh> x and a are the same axis?
[14:49:26] <tjb1> yes
[14:55:12] <tjb1> I envy people that can build a large 4x8 type machine out of nothing but aluminum extrusions
[14:55:47] <jdh> I envy people with enough room for a 4x8 machine.
[14:56:33] <tjb1> lol
[14:56:39] <tjb1> Have you seen mine?
[14:56:44] <jdh> well, unless they lived more than 30 miles from the ocean.
[14:56:55] <jdh> don't think so.
[14:57:17] <voiedexp> hi all
[14:57:34] <voiedexp> i am not able to solve my cnc problems
[14:57:38] <tjb1> http://www.youtube.com/watch?v=pD3zaRatH5M
[14:57:41] <tjb1> thats it right now jdh
[14:58:14] <jdh> nifty, did you make the frame?
[14:58:33] <tjb1> yes
[15:00:29] <voiedexp> the problem now is the following: if i do a closed loop more than one time I see that the loop move to increasing X and the moving step is near to 1mm
[15:01:40] <voiedexp> could anyone help me in finding a solution..
[15:02:38] <skunkworks> voiedexp, more information...
[15:02:44] <voiedexp> the problem is that one axis (Y) work very well while the other (Y) gives me that problem
[15:02:58] <voiedexp> Y is well X is bad
[15:03:04] <voiedexp> sorry
[15:05:00] <voiedexp> if I make a rectangular loop with the rectangle parallel to the axes, i can see that the second time i perform the loop, one couple of side are still in the first position while the other couple of side have moved toward +Y of about 1mm
[15:06:24] <archivist_> stepper? are you accelerating too fast and losing steps
[15:06:28] <voiedexp> i.e. one axis repeat its behaviour loop by loop while the other seems to translate the coordinates toward Y+
[15:06:55] <archivist_> slow your acceleration and top speed see if it gets better
[15:07:12] <voiedexp> i've think about it by the behaviour is incredibly regular... 1mm each loop
[15:08:13] <voiedexp> i think it is diffucult to loose exactly the same amount of step each loop...
[15:08:54] <archivist_> dont assume test, measure
[15:09:19] <archivist_> slipping couplings too
[15:09:34] <voiedexp> and however my speed/acceleration is very slow and very similar to the other axis (same motor) that performs very well
[15:10:07] <archivist_> mass being moved is not the same though, another bad assumption
[15:10:40] <voiedexp> but the bad axis s the one with the lower mass to be moved
[15:10:46] <archivist_> specially if the gearing is different too
[15:11:19] <voiedexp> i'm trying to move with velolicity and acceleration very low
[15:11:33] <archivist_> directional friction is interesting too
[15:11:55] <voiedexp> i've not gears, stepper motors are directly connected with their guides
[15:14:07] <archivist_> show us pictures Im used to steppers driving the leadscrew or belt, guide?
[15:14:40] <DJ9DJ> gn8
[15:16:02] <voiedexp> the steppers guide two belt guides
[15:16:44] <archivist_> guide is the wrong word
[15:17:32] <voiedexp> sorry, my english is not good
[15:18:44] <archivist_> is the belt too loose and it jumps one tooth
[15:19:48] <voiedexp> reducing speed and acceleration do not provide any result
[15:19:51] <voiedexp> :(
[15:21:03] <archivist_> "no result" ?
[15:22:54] <voiedexp> no "good" results...
[15:23:02] <voiedexp> poor results...
[15:29:47] <mrsun> yey my new vfd control card works perfectly =)
[15:30:11] <mrsun> it was even good enough for me to switch gnd and output and it didnt fry :P
[15:30:26] <Tom_itx> which card?
[15:31:05] <mrsun> homemade
[15:31:29] <mrsun> opto isolated card to control 3 digital pins on the vfd and pwm to analog =)
[15:32:57] <jdh> what did you use for the pwm filter?
[15:33:24] <mrsun> two low pass RC filters
[15:33:31] <mrsun> into a rail to rail input output opamp
[15:33:34] <jdh> how do you pick values?
[15:33:42] <mrsun> haha :P
[15:33:45] <mrsun> dont ask ..
[15:33:46] <mrsun> =)
[15:34:29] <jdh> guess & try again?
[15:34:44] <mrsun> guess and it worked good :P
[15:35:17] <mrsun> and some help from pcw_home i think it was =)
[15:35:27] <mrsun> or alot of help =)
[15:36:32] <mrsun> also added a LM317 to get 10.37V without drawing current from the vfd and got some AC opto couplers and added option for how a bob works
[15:37:02] <mrsun> if it draws stuff low or if it wants to push stuff high (+5V common or GND common)
[15:37:09] <mrsun> kinda happy with the design =)
[15:48:22] <voiedexp> there is something that make me sure that the mechanics work well.. changing the loop i have note that the "slip" is rise up to about 20 mm...
[15:49:56] <voiedexp> do you know any parameter that can provide this kind of translation between each loop? Note that on the AXIS gui i see only one loop and not many translated loops..
[15:51:17] <JT-Shop> no parameter that I know of
[15:51:28] <JT-Shop> are you missing steps?
[15:52:59] <voiedexp> that are not steps in my opinion
[15:54:08] <voiedexp> depending on the loops, with the same load on axes (very low) paths are very similar but translated
[15:54:34] <JT-Shop> can you try and explain what you mean my translated?
[15:54:52] <andypugh> voiedexp: I have seen some people have this problem when a pulley is slipping on a shaft, it's worth marking them with a pen. However it would not normally be so regular.
[15:55:45] <andypugh> I wonder if is is possible that electrical problems are sending the pulses for one axis to both?
[15:57:30] <andypugh> JT-Shop: Talking Swiss, the chap making the scale Ferrari engine has a swiss-style steady casting for his Hardinge. It bolts to the saddle, and takes the carbide-faced collets. It sits very close to the tool tip.
[16:00:17] <andypugh> I wonder if Ichgucklive was looking for "Key"?
[16:02:46] <JT-Shop> andypugh: I'd love to see a photo of that
[16:04:10] <andypugh> He's not got it mounted at the moment (it came with the lathe) but if you imagine a conventional travelling steady which is an adjustable collet, then that is how it works.
[16:05:55] <JT-Shop> does it mount to the back of the carriage by the turret encoder? I'm not seeing in my mind how that would work...
[16:09:40] <andypugh> He has a different (4 way) toolpost. I am not sure if you can still do toolchanges with it fitted. My impresson is that the face of the collet is meant to be right up against the tool tip.
[16:10:33] <JT-Shop> I've seen a 4 position turret and yea I can imagine it if you don't spin the turret while it is on
[16:11:38] <andypugh> It's HXL
[16:11:40] <andypugh> http://www.f1-2000.co.uk/index.php?f=hardinge_hxl2
[16:17:34] <JT-Shop> yes, quite different from mine and in some ways the same
[16:19:50] <tjb1> there a tutorial somewhere on how to setup limit switch to control 2 steppers or wiring all limit/home switches to one input and making it home one at a time?
[16:20:26] <JT-Shop> just the manual
[16:20:36] <tjb1> user manual or do I need to use HAL?
[16:20:50] <JT-Shop> integrator manual
[16:20:54] <JT-Shop> ini section
[16:21:49] <andypugh> tjb1: It's basically a case of giving them all a different homing sequence, setting HOME_IS_SHARED and also HOME_IGNORE_LIMITS
http://www.linuxcnc.org/docs/html/config/ini_homing.html#cha:homing-configuration
[16:22:39] <tjb1> So it is possible to wire them all into the same input
[16:22:52] <tjb1> And use them as homing and limit also
[16:23:02] <andypugh> Yes.
[16:23:32] <andypugh> Can be very handy on a parport machine as you can get by with only using one pin for all limits and homes.
[16:23:43] <tjb1> G540 only gives me 4 inputs
[16:23:45] <tjb1> :)
[16:27:43] <tjb1> andypugh: NPN is sinking correct
[16:28:00] <tjb1> And when sensor is activated its basically connecting the ground
[16:28:11] <andypugh> Don't ask me, I barely understand transistors :-)
[16:29:15] <tjb1> Heh
[16:31:56] <skunkworks> seems though you need some hal trickery for shared home and limit switches (all shared) because if you home one axis - the other axis sees the limit trip...
[16:32:21] <skunkworks> iirc
[16:34:09] <JT-Shop> yea you need some tecno-fu
[16:34:27] <tjb1> I can separate them but then I will have no inputs left
[16:34:33] <voidexp> sorry i have lost my connection
[16:34:45] <voidexp> anyone could help me?
[16:34:54] <voidexp> have you see the question?
[16:35:29] <voidexp> i have identified my problem but i don't know how i coyuld solve it
[16:36:24] <voidexp> only under certain conditions one of the motors does not start (no movement no sound..)
[16:36:34] <andypugh> The only trickery for ignoring limits when homing is the HOME_IGNORE_LIMITS flag
[16:37:18] <andypugh> voidexp: Is it that your drive is shutting down when stationary, and not waking up in time?
[16:38:05] <andypugh> Or possibly the amp-enable and direction are switched? (I think that would be a less subtle problem)
[16:38:37] <andypugh> As a basic starting point for trouble-shooting, assume that LinuxCNC works :-)
[16:38:48] <voidexp> the motor works in each direction, only in some cases it drop off
[16:39:01] <JT-Shop> seems the G540 has more problems than it is worth
[16:39:14] <JT-Shop> tjb1: got room for another parallel port card?
[16:39:34] <tjb1> Yes but it shouldnt be needed, 3 home/limits and 1 input for the arc ok
[16:40:40] <tjb1> I actually have one but I got it off ebay and I dont think it is going to work
[16:40:45] <tjb1> Plus I think I lost the driver cd
[16:46:03] <tjb1> How repeatable should a cheap proximity be?
[16:46:45] <tjb1> Im thinking I need to put a toggled limit on the z for the floating torch
[17:01:21] <r00t4rd3d> limits are for girls
[17:01:43] <tjb1> My machine is too big to work by hand ;)
[17:02:24] <r00t4rd3d> are you compensating for something?
[17:02:55] <tjb1> Yes, your lack of...
[17:02:56] <r00t4rd3d> do you also own a redneck sized pickup?
[17:03:01] <tjb1> No I drive a car
[17:03:06] <r00t4rd3d> aww
[17:03:17] <tjb1> 2008 mazda 3 i
[17:05:12] <tjb1> r00t4rd3d: I am gonna try to get limit switches on tonight
[17:05:22] <r00t4rd3d> u got that machine working yet?
[17:05:29] <tjb1> Do I look like godzilla
[17:05:55] <r00t4rd3d> more like the swamp monster
[17:06:45] <tjb1> thanks :)
[17:06:57] <tjb1> Route those pieces for yours yet lazy?
[17:07:50] <r00t4rd3d> nah im doing other stuff
[17:08:02] <tjb1> Lazy!
[17:08:25] <r00t4rd3d> anything but
[17:08:54] <Jymmm> http://i54.tinypic.com/2zgyq05.gif
[17:13:00] <tjb1> get to work r00t
[17:15:03] <r00t4rd3d> http://i.imgur.com/1tPNj.jpg
[17:16:08] <r00t4rd3d> i just got done working
[17:17:05] <r00t4rd3d> 7-5
[17:18:09] <tjb1> well now you better work from 5 - 12
[17:18:27] <r00t4rd3d> more like 8-1
[17:19:00] <tjb1> gonna make one of them at least?
[17:19:00] <r00t4rd3d> i did clean my machine when i got home
[17:19:20] <r00t4rd3d> i dont have the aluminum on hand
[17:19:31] <r00t4rd3d> i got more wood stuff to cut though
[17:20:03] <tjb1> Make aluminum
[17:21:19] <r00t4rd3d> for the cost of it, easier to buy
[17:22:58] <tjb1> I want to see aluminum pieces by friday :p
[17:23:15] <r00t4rd3d> well you better get on it then
[17:24:01] <r00t4rd3d> id probably cut it with a jig saw before my router
[17:24:20] <tjb1> Ah so lazy
[17:24:28] <tjb1> haha
[17:25:20] <r00t4rd3d> i dont want to break shit trying to cut metal
[17:27:23] <tjb1> never know until you try
[17:29:09] <r00t4rd3d> i did a small pocket in some aluminum
[17:29:19] <r00t4rd3d> slowly
[17:32:54] <andypugh> If it breaks, it needs an upgrade.
[17:34:00] <r00t4rd3d> Iam going to McDonalds and getting a McFat with extra Fat.
[17:35:41] <andypugh> r00t4rd3d: Don't forget the Lard Sauce
[17:40:26] <FinboySlick> I don't know if it's the different rules here, but McDonalds is a lot leaner in Quebec than in the US... At least the places I tried in the US.
[17:43:00] <andypugh> I think McDs is rather over-criticised.
[17:43:27] <FinboySlick> They indulged our Quebec culture too and gave us McPoutine :)
[17:43:45] <FinboySlick> Do they have McHaggis in Scotland?
[17:45:36] <andypugh> No, but I think they should.
[17:46:10] <FinboySlick> In the east here, they also have McLobster.
[17:47:19] <jdh> McDonalds cheeseburger & shake is good at 2am while drunk. That's about the only time though.
[17:48:06] <FinboySlick> Every few months, I get a craving for a quarterpounder, eat it, then feel glad it won't be another few months before I try that again.
[18:43:06] <Tom_itx> http://www.youtube.com/watch?v=FuwFlW3n_3E&feature=related
[18:49:01] <andypugh> As a counterpint to the tiny micromoeter:
http://www.imperial.ac.uk/centenary/memories/cornforth.shtml
[18:49:17] <andypugh> (you might want to scroll to the last photo)
[18:56:46] <Tom_itx> :)
[19:07:04] <andypugh> They still have it i assume. We certainly kept it safe in my tenure.
[19:07:42] <Tom_itx> you missed a good bike video yesterday
[19:08:51] <Tom_itx> http://www.youtube.com/watch?feature=player_embedded&v=_vo4VAxtZSs#!
[19:10:05] <andypugh> You tosser
[19:10:27] <andypugh> I do _not_ want to see that sort of stuff unannounced
[19:11:18] <andypugh> What part of that was "good"?
[19:11:26] <Tom_itx> he lived
[19:11:30] <Tom_itx> read the posts
[19:11:32] <andypugh> <quite angry>
[19:12:06] <Tom_itx> i doubt most bikers are that stupid
[19:12:10] <andypugh> "lived" != happy
[19:12:55] <Tom_itx> broken arm and something else.. i forget
[19:13:16] <andypugh> And a huge bill for a new bike
[19:13:44] <Tom_itx> it was a 3rd rider that went between them
[19:13:54] <andypugh> I repeat, what about that was "good" for anyone involved?
[19:14:33] <Tom_itx> He survived the crash with only a broken arm and broken leg
[19:14:39] <Tom_itx> and i bet he thinks twice
[19:16:22] <andypugh> Camera bike left him nowhere to go, as far as I can see
[19:16:46] <Tom_itx> he could have slowd down like the one on the right had he been alert
[19:16:59] <andypugh> Though I have no idea why the plan A of "just stopping" wasn't tried.
[19:17:06] <Tom_itx> me either
[19:23:07] <andypugh> I crashed my bike 6 weeks ago on a race track, gently. I broke no bones, but I am still not right. That guy with the broken arm may never get back to where he was.
[20:36:47] <PCW> hmm www.linuxcnc.org is down
[20:38:27] <Tom_itx> so it seems
[21:00:09] <L84Supper> it's up again!
[21:12:44] <r00t4rd3d> http://i.imgur.com/fxMG5.gif
[21:29:26] <jdh> setp hm2_5i25.0.stepgen.03.steplen 1
[21:29:27] <skorket> has anyone here done pcb routing?
[21:29:34] <jdh> does that 1 mean something special?
[21:30:10] <jdh> (other than 1 nS)
[21:51:18] <skunkworks__> it means exactly 1ns.. what do you mean?
[21:52:22] <skunkworks__> it means the step signal will be on for 1ns.
[21:53:08] <jdh> PCW answered it. I was thinking older versions used # of base periods instead of nS, but that might have been something else.
[22:10:51] <pcw_home> I think the way the hardware and driver work 1 ns will end up one ClockLow so 30 nS on a 5I25 (0 is not allowed)
[22:16:41] <pcw_home> The current software stepgen times are also scaled in ns, but if you set steplen to 1 ns you will get the minimum
[22:16:42] <pcw_home> time which (if reset is not enabled) is one base thread time, so 1 ns steplen will work for the software stepgen
[22:16:44] <pcw_home> but will very likely not work for the Mesa hardware stepgen
[22:18:26] <jdh> he just took his pport config and changed the name.
[22:25:33] <pcw_home> The hardware stepgen gets a little closer to what you specify...
