#linuxcnc | Logs for 2013-09-03

Back
[00:02:47] <nspiel> ok now i switched it to 30 and it move continuously and does not oscillate
[00:03:04] <nspiel> now to learn how to tune p and i
[00:03:05] <pcw_home> For a voltage mode drive (a HBridge) I would set D first to the
[00:03:05] <pcw_home> largest stable amount and then back off about 30% then do the same with P
[00:03:15] <pcw_home> (tune D with a bit of P say 10)
[00:04:04] <nspiel> what do u mean witha bit of p
[00:04:16] <pcw_home> You will be able to use more D if you increase the servo thread rate
[00:04:25] <pcw_home> say 10 for P
[00:04:37] <nspiel> oh ok
[00:05:16] <nspiel> what about i
[00:07:12] <pcw_home> After this tune FF1 and then FF2
[00:07:13] <pcw_home> FF1 is tuned for minimum following error during a high speed cruise
[00:07:15] <pcw_home> FF2 is tune for minimum following error during accel and deccel
[00:07:16] <pcw_home> Note that tuning will change if you chang the motor power voltage
[00:08:05] <pcw_home> so theres not to much point (except learning) unless you have the final power supply
[00:09:09] <pcw_home> I would leave I at 0 until every thing else is as close as it can get
[00:09:11] <pcw_home> or it will just make things worse
[00:09:31] <nspiel> but dont i want 0 steady state error?
[00:10:15] <pcw_home> Any I term at this point will make the rest of the tuning impossible
[00:10:36] <pcw_home> (or at least a lot more difficult)
[00:12:07] <nspiel> so how do i tune the final i value once everything else is set
[00:14:14] <pcw_home> Once everything else it tuned will you can add I but note that it will typically
[00:14:15] <pcw_home> improve static and cruse performance at the expense of dynamic performance
[00:14:54] <nspiel> yeah but isnt the point of i to add zero steady state error? how do i tune for the right value
[00:18:54] <pcw_home> Right for what? any amount of I will result in 0 steady state error (after a time),
[00:18:55] <pcw_home> more force the error to 0 faster and the response to disturb (changing load)
[00:18:57] <pcw_home> faster at the expense of stability, its a tradeoff
[00:19:10] <pcw_home> forces
[00:20:57] <nspiel> right for my system
[00:21:38] <nspiel> do you increase i until unstable then do 30% as well?
[00:25:55] <pcw_home> I would not in general without a lot of testing. I would add just enough to zero static error in a reasonable time
[00:25:56] <pcw_home>
[00:37:28] <nspiel> how do u measure zero static error
[00:49:23] <pcw_home> axis.N.f-error
[00:50:05] <pcw_home> You will need to plot this (with halscope) for tuning
[00:55:53] <nspiel> ok cool thanks
[01:04:16] <nspiel> what will happen if i try to run the linuxcnc gcode script now
[01:53:56] <CaptHindsight> https://www.birchwoodtechnologies.com/store/trutemp.html True black magnetite finish at 200°F
[01:54:06] <CaptHindsight> ^^ anyone try these?
[02:06:11] <Loetmichel> mornin'
[02:19:54] <DJ9DJ> moin
[07:46:52] <false_> hey everyone
[07:57:40] <false_> So, after unclogging the entire auto lube system (thanks for the tips!), we're back to tuning now. Anyone care to comment on this: http://tinypic.com/r/rvkf43/5 . I have removed the servo from the mill to make sure it's not a mechnical problem with the mill itself. I cannot get rid of the spikes at the beginning, adding more P helps, but I have to add D to keep it from oscillating. Any amount of FF2 makes the spikes worse.
[07:59:08] <skunkworks> Tuning without the servo in the machine is usually fruitless...
[08:01:12] <false_> Yes, so I read, but the spikes are the same when the motor is attached, which makes me wonder if there is some kind of filtering on the drive input? (A06B-6035, fanuc velocity drives)
[08:06:22] <false_> It's not my intention to tune the servo and than put it back on the machine, but I wanted to be sure it wasn't mechanical. Now I can see the problem is exactly the same without load attached, which makes me suspect the drives?
[08:10:36] <Valen> backlash somewhere?
[08:11:32] <Valen> you could also try -ve values for ff2 perhaps?
[08:11:39] <false_> That's why I removed the servo from the machine, the encoder is mounted on the shaft, there is no backlash one the connection between them
[08:12:07] <false_> I thought I should not use negative numbers as PID parameters
[08:13:23] <Valen> you can on the FF stuff, you shouldn't need to but it might be worth a try
[08:13:31] <Valen> I found the FF's to be *very* sensitive
[08:13:39] <Valen> I was using .000something
[08:13:56] <false_> Yes adding as much as 0.0005 FF2 makes the spikes go bonkers
[08:14:10] <false_> I'll try the negative ff2
[08:43:41] <dosas> I'm trying to add a touch off plate, and I'm using this famous example here http://forscience.nl/?p=197. I can startup without error, the probe-in pin works fine. but when i press the touch off button nothing happens
[08:44:38] <jthornton> never heard of it so not so famous I assume
[08:44:53] <archivist> I never heard of it either
[08:44:59] <dosas> okay not famous
[08:45:12] <dosas> but the best thing i found out there
[08:45:24] <archivist> is it some random internet source where you find errors ?
[08:45:55] <archivist> did you miss a step
[08:46:18] <jthornton> or two
[08:46:29] <archivist> diagnosis of those two may find the error (or three)
[08:47:41] <jthornton> I wonder why he used classicladder and made it so complicated
[08:50:55] <jthornton> I would just use a G38 then a G10 in a subroutine file called with the MDI command from the push button
[08:51:50] <jthornton> or G92 if your not using the tool table like I do on my plasma table
[08:53:14] <jthornton> well he left
[09:36:55] <jdh> someone gave me a 0.375" ballscrew & nut. I wonder if I can use it on my 9x20 for the X in the same place as the original leadscrew.
[09:37:33] <archivist> yes if it will fit
[09:37:50] <archivist> probably...only you can see
[09:38:59] <CaptHindsight> jdh: whats the pitch? Will that work for you?
[09:41:11] <jdh> looks like 8TPI (with a tape measure)
[09:42:27] <jdh> 0.3 thou/step with half stepping and 1:1
[09:42:40] <archivist> you can scale it in software so it pitch does not matter unless you want a dial on the shaft
[09:44:17] <jdh> I'd ditch the dial/handle. It just might be cleaner to put it in the same spot as the old one. I bought a 12mm screw/nut that I was going to mount on the side.
[09:47:54] <false_> well this is the best I can do: http://tinypic.com/r/2n287a/5 . How does that look, I don't really know what I should be expecting in terms of recovery times. This is accelerating to 1000mm/m @ 830mm/sec/sec, with no load attached
[09:49:36] <false_> deceleration looks like this: http://tinypic.com/r/3004hw6/5
[09:52:36] <cradek> are those inches or mm?
[09:52:52] <false_> sorry forgot to mention, mm
[09:53:40] <cradek> it still looks like something's totally wrong
[09:53:54] <false_> I was afraid you would say that :P
[09:55:05] <false_> I was thinking there might be some kind of intelligence on the drive? Some kind of input filter, or maybe accel ramps?
[09:55:21] <false_> Could that explain this behavior?
[09:55:59] <cradek> sure, a velocity loop that responds poorly (no matter what the reason) could explain this
[09:56:12] <pcw_home> Sure doesn't look like its in velocity mode
[09:56:39] <false_> Yeah you said that last week, you thought torque mode right?
[09:57:56] <false_> But the drives require a tacho feedback, that's not logical or common is it?
[09:58:22] <cradek> remind me - do you have real tachometers?
[09:59:05] <pcw_home> Also the plots are a little confusing sine the velocity feedback is not shown (it must be oscillating)
[09:59:11] <pcw_home> since
[09:59:22] <false_> I will plot another one
[09:59:49] <false_> You want the hm2.encoder.velocity?
[10:04:29] <pcw_home> Do the oscillations (on the same motion) disappear if you decrease P?
[10:05:16] <false_> http://tinypic.com/r/2s9yas3/5
[10:06:30] <cradek> what is y-vel-fb?
[10:07:33] <cradek> (is it tinypic that converts these to jpg? that's obnoxious)
[10:07:50] <false_> Im uploading png so I would say yes
[10:08:04] <cradek> they're insane then
[10:08:06] <cradek> :-(
[10:08:11] <cradek> what is y-vel-fb?
[10:08:13] <false_> y-vel-fb is coupled to encoder.velocity
[10:09:00] <false_> Would a plot of the pwmgen that generates the tacho be helpfull?
[10:10:03] <false_> pcw_home: reducing p to 60 does not make a difference, not a huge one at least
[10:10:05] <cradek> it doesn't start moving when you command it
[10:10:15] <false_> the drive having a deadband?
[10:12:02] <cradek> it would be interesting if you'd give your amp a short 1v square pulse and plot the resulting velocity
[10:12:18] <cradek> no pid tuning or anything else affecting it
[10:12:47] <cradek> your amps just aren't working right and working on tuning this pid is barking up the wrong tree
[10:13:22] <false_> Ok, hang on, I'll do that
[10:16:49] <skunkworks> this isn't the fanuc control that needed a 'tach' signal from the control card? (the control took the encoder and calculated the velocity and sent that back to the amp)
[10:17:25] <false_> Yes correct, a F/V converter on the mainboard converted encoder pulses to voltage for the drives
[10:17:39] <skunkworks> So - how are you doing it?
[10:18:25] <false_> Reading encoder counts via 7i33 and scaling and then output via another 7i33
[10:18:38] <false_> servo thread is at 5khz
[10:18:39] <skunkworks> ah - ok
[10:18:56] <false_> hang on need to do some rewiring
[10:18:58] <skunkworks> (I have done something similar.. works good if setup )
[10:20:22] * JT-Shop puts the last of Woody's things away in storage
[10:20:39] <skunkworks> cat?
[10:21:54] <JT-Shop> yea
[10:22:28] <JT-Shop> didn't greet me last Wednesday morning like usual
[10:22:58] <skunkworks> aww - sorry to hear that.
[10:23:30] <JT-Shop> yea she was a lap cat too and would sit in my lap while I worked on the computer
[10:24:51] <false_> that sucks
[10:25:11] <false_> cradek: what would you like in you plot sir?
[10:26:39] <cradek> how about the output command pulse (make it long enough for it to settle), encoder position and velocity
[10:29:01] <false_> you mean you want me to emit a 1V pulse from within linuxcnc?
[10:29:30] <cradek> I guess that's what I was picturing - were you doing to do it some other way?
[10:29:57] <false_> I was going to hookup a lab psu and emit the pulse from there :P, but your way sounds easier
[10:30:20] <cradek> both ways are valid tests but they test different things
[10:32:29] <false_> I get it, so as not to waste your time: I should disable the PID loop by setting everything to zero, then from halshow set a pwmgen.value that will generate a 1V pulse, then after 2 seconds or so disable the pwmgen?
[10:33:03] <cradek> sure that sounds good
[10:33:18] <cradek> will your limit switches stop it ok if it goes all wrong?
[10:33:18] <false_> *working
[10:34:48] <false_> the motor is disconnected from the mill right now, I wrote that earlier, but you came to the rescue and a forgot to tell you. I was still having the spikes I had last week and wanted to make sure it wasn't something mechanical. Have I wasted even more of your time?
[10:35:18] <cradek> ouch, a bare motor can't be tuned
[10:35:40] <cradek> well it can, but it can't then be hooked to the machine...
[10:36:15] <false_> You'll have to help me on that last part?
[10:36:45] <cradek> if you change what the motor is connected to (its inertia?), the tuning changes
[10:37:35] <false_> Yes, I get that, funny thing is that the output shown on hal scope shows exactly the same anomalies when it's connected to the mill
[10:38:31] <cradek> hmm
[10:38:44] <false_> I have the Y hooked up on my workbench and the X is still connected, same problem there, just with different tuning parameters of course
[10:39:01] <false_> So I'm suspecting the drives like you
[10:39:03] <cradek> I think you need to put it on the machine, make sure the limit switches work for safety, then test that your velocity loop is working right
[10:39:25] <cradek> you can do that by briefly commanding a steady velocity, and record what it does
[10:39:54] <cradek> ideally you'd command it using the usual dac output, so if you have wiring/ground problems or something, you'll see the real results
[10:40:12] <false_> You mean just jogging for a second or 5?
[10:40:21] <cradek> you will know VERY fast whether you have torque or velocity mode response
[10:40:41] <cradek> not jogging, a steady (like 1v) dac output
[10:40:52] <false_> ok the halshow route you mean
[10:41:05] <cradek> yeah, or use oneshot or siggen or whatever
[10:41:36] <false_> ok, I'll do this on the X, which is still connected
[10:46:34] <dosas> hello, i'm trying to get a touch off plate to work. i'm using this setup http://wiki.linuxcnc.org/cgi-bin/wiki.pl?ClassicLadderExamples#Single_button_probe_touchoff i can start linuxcnc fine and the probe-in pin works(tested in hal meter) but when i press the custom touch off button nothing happens
[10:47:21] <dosas> does anyone know how to debug classic ladder. When i start axis and click the clssicladder symbols tag it says no signals are connected
[10:51:20] <dosas> i have a pin called halui.mode.is-mdi
[10:51:36] <dosas> i suppose it should be true for the classic ladder to work how can i achieve this
[10:51:55] <dosas> ah
[10:52:11] <cradek> I think doing that with halui and classicladder is pretty crazy
[10:52:44] <cradek> you can do it in gcode easily enough
[10:53:04] <cradek> in touchy you could use a macro, or you could run master and remap a gcode to do it
[10:56:29] <false_> cradek: 0.5V also ok, the limitswitches are not functioning at the moment because there are some repairs going on in that department, I do have a portable estop so I will be ok :$
[11:07:12] <JT-Shop> dosas: I wonder why he used classicladder and made it so complicated
[11:07:12] <JT-Shop> <jthornton> I would just use a G38 then a G10 in a subroutine file called with the MDI command from the push button
[11:07:12] <JT-Shop> <jthornton> or G92 if your not using the tool table like I do on my plasma table
[11:07:12] <JT-Shop> <jthornton> well he left
[11:08:10] <cradek> yes O-call a sub is all the touchy macro does. I guess you could also do that with a button to halui
[11:08:31] <cradek> you could set g92 or g5x or the tool length
[11:10:15] <JT-Shop> from my plasma touchoff.ngc subroutine
[11:10:16] <JT-Shop> G38.2 Z-1.75 (Probe to find the surface)
[11:10:16] <JT-Shop> G92 Z-0.070 (Set Z Probe Switch Offset)
[11:10:38] <JT-Shop> the rest is plasma specific
[11:11:24] <cradek> you don't use the result of the probe
[11:11:49] <JT-Shop> no, I just use to to stop the Z
[11:11:54] <JT-Shop> it to
[11:11:56] <cradek> where it stops after the decel is done isn't particularly reliable, because it depends how fast you're going
[11:12:10] <cradek> so you have different results depending on how far above you started
[11:12:15] * cradek shrugs
[11:12:25] <JT-Shop> I probe slow, but your correct I should do it the proper way
[11:12:37] <JT-Shop> just never thought about it till now
[11:12:46] <dosas> who left
[11:12:54] <archivist> you
[11:12:56] <JT-Shop> you
[11:13:00] <dosas> aha
[11:14:11] * JT-Shop ponders how to use the result of the probe even though for a plasma it does not need to be very accurate
[11:15:41] <JT-Shop> is #5063 the absolute position for Z when the probe trips? The manual does not say
[11:15:42] <cradek> you could just add G0 Z[#5063 + .1] and then adjust your G92
[11:16:04] <cradek> successful probing, parameters 5061 to 5069 will be set to the coordinates of X, Y, Z, A, B, C, U, V, W of the location of the controlled point at the time the probe changed state.
[11:16:10] <cradek> After...
[11:17:44] <JT-Shop> I see a link is needed in the Numbered Parameters section 5061-5070 - Result of G38.2 Probe (X Y Z A B C U V W)
[11:19:26] <JT-Shop> that makes more sense now, thanks cradek
[11:21:25] <dosas> what is the #5063
[11:21:58] <CaptHindsight> https://www.birchwoodtechnologies.com/store/trutemp.html True black magnetite finish at 200°F. Anyone ever try this type?
[11:24:06] <pcw_home> Their pictures look nice...
[11:25:57] <JT-Shop> dosas: http://www.linuxcnc.org/docs/html/gcode/overview.html#_numbered_parameters_a_id_sub_numbered_parameters_a
[11:26:26] <JT-Shop> and http://www.linuxcnc.org/docs/html/gcode/gcode.html#sec:G38-probe
[11:27:16] <dosas> ah okay
[11:27:22] <dosas> so s.th. like a return value
[11:29:07] <JT-Shop> CaptHindsight: it took 10lbs to neutralize the acid tank, guess there was a bit more volumne than my WAG
[11:32:49] <dosas> can i acces these values somewhere in the gui
[11:35:53] <JT-Shop> explain what you want to do with the value
[11:40:59] <dosas> have a look at it
[11:42:22] <JT-Shop> use a debug message
[11:42:41] <JT-Shop> http://www.linuxcnc.org/docs/html/gcode/overview.html#sec:debug-messages
[11:43:38] <JT-Shop> in the MDI box say (DEBUG, Z touch off = #5063)
[11:47:52] <false_> cradek: Using: setp hm2.pwmgen.value (0.5V, nice and slow because the lack of limits at the moment): http://tinypic.com/r/qpfjbd/5
[11:48:23] <false_> look familiar :(
[11:54:53] <cradek> what is pwmgen.05?
[11:55:12] <cradek> oh it's your generated velocity feedback
[11:55:14] <false_> thats the tacho output
[11:55:40] <false_> Just for fun I'm gonna try the same experiment on the benchtop servo
[11:56:00] <cradek> ok so this is going 18mm/sec. what was your voltage output for this velocity?
[11:56:00] <false_> Or won't that give us any more info?
[11:56:20] <false_> 0.5038V
[11:56:38] <cradek> this took 0.2 seconds to settle
[11:56:59] <cradek> what rapid speed are you going for? 360mm/sec seems pretty fast
[11:58:08] <false_> The machine is capable of 166.67mm/sec, the dac output is normalized
[11:58:21] <false_> so max output is set to 166.67
[11:58:26] <cradek> but you're off by a factor of 2ish?
[11:58:52] <cradek> 0.5v should have given you something like 166/20 = 8mm/sec but you got 18
[11:59:26] <false_> hmmm
[11:59:30] <false_> you are very correct
[11:59:58] <cradek> check scaling, and then you've got to figure out why this velocity response is so bad
[12:00:14] <cradek> you've got to fix the crazy overshoot and long settle time
[12:00:32] <false_> Gonna do that, problem is that there is no documentation for this old stuff, I was going by info I got from a Fanuc guy.
[12:01:33] <cradek> velocity loop gain is sometimes a knob, but sometimes jumpers or even R/C you have to solder in, but there's an adjustment somewhere
[12:01:47] <cradek> I bet you just have to turn it down and everything will get better
[12:02:21] <false_> Yes, there is a gain pot, but I was told, or I understood I shouldnt turn it until I was very close, which clearly I'm not
[12:02:23] <cradek> although it DOES look like you're in velocity mode, so that's good
[12:02:36] <false_> Well at least we got that settled
[12:02:50] <cradek> yep
[12:03:03] <cradek> now continue going one step at a time and you'll get there :-)
[12:03:31] <false_> Probably want to move to fast, thanks again for your help!
[12:03:39] <cradek> welcome
[12:37:00] <IchGuckLive> hi all B)
[12:56:35] <archivist> grmbl...losing steps on x axis, rebuild it, add balance weight to make no twisting moment with the way A and B are at the moment, another run...just the same
[12:57:27] <archivist> X slide is just not big enough for the weight
[13:02:18] <IchGuckLive> archivist what speed are you going at what weightr
[13:02:45] <IchGuckLive> action fails on strait or on curving
[13:02:50] <archivist> its very slow cutting steel
[13:03:13] <IchGuckLive> microstepping
[13:03:19] <archivist> hell no
[13:03:40] <IchGuckLive> is it a servo
[13:03:54] <archivist> stepper no it is seizing
[13:04:15] <IchGuckLive> ah ok version 2.5.2
[13:04:52] <archivist> when I rebuilt it last week it was about 40 pounds to shift from stationary, down to about 5 pounds
[13:05:15] <archivist> I know it is a mechanical problem
[13:05:20] <IchGuckLive> ok
[13:05:45] <archivist> just frustrated
[13:05:56] <IchGuckLive> timingbelts are mutch better as a direct drive
[13:06:48] <archivist> ffs it is not a drive slipping problem
[13:07:17] <IchGuckLive> ok so you know what to do and are only frustrated ;-(
[13:08:04] <archivist> I know another rebuild adding ball slides is the answer
[13:08:23] <IchGuckLive> are you in the USA
[13:09:12] <archivist> you dont actually read much in here do you
[13:09:35] <archivist> UK
[13:09:38] <IchGuckLive> i shoudt read the logs bevor poting
[13:09:48] <IchGuckLive> im in germany
[13:09:54] <archivist> I know
[13:09:59] <FinboySlick> Easy boys, you already had your war...
[13:10:10] <FinboySlick> ;)
[13:10:11] * archivist giggles
[13:10:19] <IchGuckLive> http://stores.ebay.de/cnc-discount
[13:10:40] <IchGuckLive> this is my supplyer
[13:10:55] <IchGuckLive> its more near to the south of ingland then to me
[13:11:01] <IchGuckLive> England
[13:11:07] <archivist> but 0 income from cnc work in the last 6 months, means no new parts
[13:11:34] <IchGuckLive> ah ok the plasmas here are running near 24/7
[13:11:44] <IchGuckLive> its the hell of payoff
[13:12:09] <IchGuckLive> all people want firertowers to there garden or steel logos
[13:12:44] <IchGuckLive> 1K Eur to build 5K Euro first month
[13:12:51] <IchGuckLive> income
[13:13:11] <IchGuckLive> i now got 3 in the shop
[13:13:20] <IchGuckLive> ofcause its not my sghop
[13:13:30] <IchGuckLive> but i got them working
[13:13:39] <IchGuckLive> and percentige of earning
[13:13:46] <IchGuckLive> thats the best
[13:13:54] <IchGuckLive> no work only income
[13:14:12] <IchGuckLive> ok some little g-coding
[13:14:28] <IchGuckLive> pic or text to g-code
[13:15:15] <IchGuckLive> FinboySlick: its not over till its over
[13:16:12] <IchGuckLive> ok im off by till tomorrow
[13:23:26] <CaptHindsight> anyone have a good source for crossed roller bearings? ~50mm inner/shaft dia?
[13:24:11] <archivist> crossed roller ?
[13:24:37] <CaptHindsight> http://www.ikont.co.jp/eg/product/needle/ndl0801.html like these
[13:25:04] <FinboySlick> CaptHindsight: I thought you meant the linear kind.
[13:28:03] <CaptHindsight> low speed, clean environment, high rigidity, low radial and axial run out
[13:28:39] <archivist> the diagram looks like ball but the title says needle....
[13:34:44] <archivist> better drawing http://www.ina.de/content.ina.de/en/products_services/rotativ_products/crossed_roller_bearings/crossed_roller_bearings.jsp
[13:35:30] <CaptHindsight> http://www.isa.org/FileStore/Intech/WhitePaper/CrossedRollerPaper.pdf
[13:36:42] <archivist> looks like most of the usual bearing companies have them
[13:36:54] <FinboySlick> I wonder if they make recirculating versions of the linear ones.
[13:37:09] <FinboySlick> Seems to me like it would be a lot more rigid than the ball types.
[13:37:44] <archivist> and could withstand higher loading
[13:37:47] <CaptHindsight> yeah, but sometimes people here have found good sources
[13:38:20] <WalterN> that linear bearing...
[13:38:41] <FinboySlick> I guess if you can sacrifice the space/travel ratio, they'd still be viable, especially on larger machines which are typically a fair bit larger than their envelope.
[13:39:17] <FinboySlick> (than their work envelope)
[13:39:27] <CaptHindsight> this is to support a rotary axis with mainly radial loads and very little axial forces
[13:40:18] <CaptHindsight> for example, one side of a trunnion
[13:41:29] <CaptHindsight> and you don't really require an air bearing
[14:31:44] <CaptHindsight> amamzon of all places http://tinyurl.com/kk6ept6 THK Cross Roller Bearing
[14:32:52] <archivist> I might prefer to haggle for the same item direct from a THK distributor
[14:33:18] <CaptHindsight> the main warehouse is 40 minutes from me :)
[14:33:59] <archivist> I need a coupler if I use the ball screw I have, seems china , but that would be a couple of weeks
[14:36:15] <syyl-> amazon has crossroller bearings
[14:36:19] <syyl-> sick world :D
[14:40:00] <Loetmichel> archivist: hopefully not these "couplers" -> http://www.cyrom.org/palbum/main.php?g2_itemId=14196
[14:40:16] <Loetmichel> ... the screws are there because of this: http://www.cyrom.org/palbum/main.php?g2_itemId=14193
[14:40:41] <Loetmichel> ... and that was after not even a month of work on the chinese 6040 CNC router ;-)
[14:40:48] <archivist> fleabay 400394187335
[14:41:22] <Loetmichel> ok, thats the better varian with a spiral
[14:41:31] <Loetmichel> not slottet from both sides ;-)
[14:42:30] <archivist> I have one here that is 1/4" to 1/4" that I could drill/machine out
[14:42:35] <CaptHindsight> I think I pay ~$1 ea for those in China
[14:43:15] <archivist> I assume most cost is postage anyway
[14:43:45] <CaptHindsight> yes, with "free" international shipping
[14:44:30] <jdh> shipping is subsidized by .cn gov?
[14:44:49] <CaptHindsight> http://www.automationtechnologiesinc.com/products-page/cnc-parts/coupling-14-and-14 $12 + shipping
[14:44:50] <archivist> fill a container
[14:45:13] <CaptHindsight> jdh: EMS out of China is pretty cheap
[14:45:54] <CaptHindsight> archivist: do you get hit with a high VAT with parts from China?
[14:46:03] <jdh> is anything cheap enough to cost $1 shipped?
[14:46:04] <archivist> there seems little sense in the shipping prices in various countries
[14:46:39] <archivist> CaptHindsight, on that they would probably be free vat and customs
[14:47:30] <CaptHindsight> yet when I ship to China it's 10% of the value or if it's marked samples, then 10% of the shipping cost
[14:47:31] <archivist> there was no charge for some brushes I got from Hong Kong
[14:48:17] <CaptHindsight> whichever is higher
[14:48:28] <archivist> costs more to collect than its worth for goods below X
[14:49:36] <archivist> got away with it on a second hand ball screw assembly as well
[14:53:29] <archivist> my nearest place to collect is http://www.arceurotrade.co.uk/Catalogue/Stepper-Motors/Oldham-Couplings but a little pricey
[14:58:44] <Loetmichel> CaptHindsight: you are lucky
[15:00:18] <Loetmichel> in germany it is 19% VAT on value AND shipping and then some 0 to 35% customs duty depending on the item
[15:01:27] <jdh> angela needs the cash
[15:01:29] <Loetmichel> ... but only on the value of the part IIRC
[15:01:42] <Loetmichel> but the VAT is due on value AND shipping
[15:02:02] <archivist> and duty if they get a chance ?
[15:03:48] <joebog> morning all
[15:04:26] <joebog> capthindsight, you busy ?
[15:05:38] <joebog> morning andy
[15:05:44] <andypugh> Evening
[15:05:49] <joebog> ya have 10 minutes ?
[15:06:20] <joebog> more beginners questions that I hope are not too silly
[15:06:29] * archivist trains joebog in IRC, just ask the real question :)
[15:07:48] <joebog> so the minimum I need is a power supply for motors a separate supply for opto couplers some motor current driver boards a breakout box/connection and a parr cable
[15:07:54] <joebog> with the software of course
[15:08:21] <joebog> ?
[15:08:23] <andypugh> You may not need the breakout.
[15:08:39] <joebog> I have one, and it makes it neat and easy
[15:08:53] <andypugh> And the inputs to many stepper drives are already optocouplers
[15:09:19] <joebog> they are, but the boards I have from the cipher unit have a separate supply
[15:09:22] <andypugh> So the _minimum_ you need is a parallel cable and the stepper drives + PSU
[15:09:32] <andypugh> Anything else is jam.
[15:09:33] <archivist> just need to buffer the parallel port sometimes
[15:10:03] <joebog> the stepper supply is 45 volts the opto supply is 4.5 volts ( I suspect thats just a sick 5 volt supply)
[15:10:37] <joebog> and to buffer the parr port ? archivist ?
[15:10:43] <andypugh> Sounds likely. You could consider stealing 5V from a USB connector
[15:10:45] <joebog> what required ?
[15:11:17] <andypugh> ULN2007 is beefy.
[15:11:26] <joebog> and I have em
[15:11:31] <joebog> just need to find em
[15:11:38] <archivist> I use either a 74xx open collector buffer or a ULN200x driver chip
[15:12:16] <andypugh> On the ininternet you will find circuits showing how to drive tiny steppers with just a parport and a ULN2007
[15:12:36] <joebog> OK then, Ill start with the power supply and a lump of aluminium as a base board
[15:13:04] <joebog> mount up my breakout board and the drivers
[15:13:35] <joebog> now, will I damage the // port if it doesnt have sufficient current drive ?
[15:15:18] <andypugh> Wire it as current-sinking.
[15:15:34] <joebog> OK
[15:15:41] <andypugh> They can typically sink 15mA, but some can only source 3mA
[15:15:49] <joebog> anything else I need to pay special attention to ?
[15:16:48] <joebog> the motor supply doesnt need regulation, does it ?
[15:16:50] <andypugh> Maybe. It's been so long since I was a clueless noob I can't recall the pitfalls :-)
[15:16:53] <joebog> just a decent cap?
[15:17:06] <andypugh> Yeah, you want a linear passive PSU.
[15:17:13] <archivist> yes
[15:17:56] <joebog> I have a 1000 watt inverter transformer thats 30 volts AC @ about 50 amps plus a 12 volt 2 amp ac supply
[15:18:03] <archivist> you can use a switching psu if you have one lying around
[15:18:15] <joebog> so .1F on the 45 volts and a 7805 on the opto supply will do ?
[15:18:15] <CaptHindsight> syyl-: even more strange, they carry that bearing at Sears http://www.sears.com/thk-cross-roller-bearing-re4510-outer-rotation-45mm/p-SPM7810914811?prdNo=2
[15:18:34] <CaptHindsight> joebog: I have a few minutes
[15:18:50] <joebog> thanks Capt :0
[15:19:03] <joebog> most questions have been answered already :)
[15:19:08] <joebog> read above to catch up
[15:19:39] <joebog> add anything the others missed :)
[15:22:27] <andypugh> Do you think they have the proce right on this motor? http://www.zappautomation.co.uk/en/integrated-dc-brushless-motors/706-bl3573-24ie-p2-12800.html
[15:24:17] <archivist> hmm any feedback from the motor
[15:24:39] <joebog> its only a small motor
[15:24:45] <joebog> 1 1/2 inch !!
[15:24:59] <joebog> your price equats to about $60 aus
[15:25:03] <joebog> thats about right
[15:25:36] <archivist> bah useless rar format docs
[15:26:07] <jdh> warezed d0cs?
[15:26:10] <andypugh> It is cheaper than a similar 35mm DC brushless motor that _doesn't_ have integrated drive electroinics.
[15:26:31] <andypugh> It's the integrated driver bit that has me puzzled.
[15:26:41] <joebog> "little" robotic motors like that have become very cheap
[15:26:48] <joebog> and theres hundreds available
[15:27:28] <joebog> try a search in china, ya should turn up heaps and heaps of dream ideas
[15:27:50] <andypugh> I actually already have more motors than I know what to do with :-)
[15:28:22] <joebog> a chip with integrated current limited MOSFET drives are common in robotics
[15:28:37] <archivist> do they have a typing problem no hole sizes http://www.zappautomation.co.uk/en/flexible-couplings/282-bf-d25l34-flexible-coupling.html
[15:28:44] <joebog> and "usually" fit into a TO220 case
[15:29:39] <archivist> ah had to find the customise...no 7mm
[15:30:13] <joebog> 1 pound per mm !!
[15:30:18] <joebog> is that cheap ? :D
[15:30:32] <joebog> sorry
[15:31:20] <andypugh> It's fairly cheap for 12" round gold bar. What were you looking at?
[15:31:53] <joebog> http://www.zappautomation.co.uk/en/flexible-couplings/282-bf-d25l34-flexible-coupling.html
[15:31:55] <joebog> that
[15:32:18] <kwallace> This BBB is just a little larger than an Arduino UNO.
[15:37:12] <jdh> I booted mine once.
[15:37:18] <jdh> I have no HDMI monitors though
[15:37:59] <joebog> if ya have a fairly modern TV it should be HDMI
[15:38:08] <andypugh> I plugged mine into the TV
[15:38:14] <joebog> mine has about 4 inputs
[15:38:14] <jdh> I do, but it is downstairs and other people watch it.
[15:38:25] <joebog> but I dont have anything to plug into it
[15:38:30] <kwallace> I just plugged the USB cable into my PC and it's blinking.
[15:38:52] <andypugh> They can stiill watch it with a BBB plugged in to it. They just don't get to change channels.
[15:39:03] <joebog> heh heh heh
[15:39:24] <jdh> the Pi on a 60" teevee was amusing for a few minutes.
[15:39:58] <joebog> I occasionally read a bit about Pi
[15:40:11] <andypugh> kwallace: You can probably VNC into it. Check your router for the IP it has been given, and have a go.
[15:40:12] <joebog> its certainly changed since my 2650 days :)
[15:40:36] <jdh> I had to nmap mine
[15:41:48] <kwallace> It looks like it's trying to setup a net conn through the USB.
[15:51:13] <kwallace> Cool, I p[lugged an Ethernet cable in, found the address and browsed it. It came back with a web page saying "Your board is connected".
[15:51:46] <andypugh> It's like magic isn't it?
[15:52:07] <kwallace> I'll try ssh.
[15:54:03] <kwallace> It's just like a Linux PC.
[15:59:48] <kwallace> andypugh: What is the easiest way to VNC into the BBB? Use Applications - Internet - Remote Desktop Viewer ?
[16:07:18] <andypugh> I think mine might. If I end up makeing one.
[16:08:39] <andypugh> I am currently designing a CNC attachment for the Rivett :-) The lathe is famous for the breadth of attachments, a CNC power-feed seems perfectly in keeping.
[16:09:19] <joebog> once I get the basic setup going is this http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Universal_Stepper_Controller the easiest way to go ?
[16:10:07] <andypugh> Either that or the Mesa 5i25 / 7i76 combination.
[16:10:37] <joebog> OK whats the consensus with the guys here watching this thread ?
[16:10:57] <andypugh> I have no experience with the Pico card.
[16:11:22] <andypugh> I wrote the drivers for the Mesa cards. I am not in a position to give unbiased advice.
[16:11:34] <joebog> :D
[16:11:49] <joebog> are the prices equivelent ?
[16:11:57] <joebog> or similar at least ?
[16:13:07] <andypugh> Mesa plug-and-go combo is $200. Pico is $250: http://pico-systems.com/osc2.5/catalog/product_info.php?products_id=30
[16:13:22] <kengu> what is in mesa plug-and-go?
[16:14:12] <andypugh> kengu: There are 4 of them, combinations of 5i25 or 6i25 and 7i76 or 7i77, both with a known-good connection lead
[16:14:53] <andypugh> http://www.mesanet.com/fpgacardinfo.html
[16:15:20] <kwallace> They are similar in that they both use an FPGA to provide fast signals. The packaging is different. The Pico card has some common connections on the FPGA board, Mesa has extension boards of different configurations.
[16:15:33] <andypugh> The choice probably comes down to whether you want to connect via parallel port, PCI or PCIE
[16:16:16] <andypugh> Despite appearances the 5i25 is _not_ a parallel port card. Though it can emulate one with the right software loaded.
[16:17:26] <joebog> where am I gonna find a motherboard ???
[16:18:47] <joebog> I gather the 5i25 plugs into the motherboard ??
[16:19:02] <JT-Shop> you can get advice on Pico cards on the forum and mailing list
[16:19:03] <joebog> of the PC
[16:19:46] <kengu> is there some convention on the mesa naming of products as there is some logic but I cannot see it
[16:23:40] <kengu> but yeah. I am still in the process of converting this plasma cutter to linux cnc and thus replacing the current io-card that is some .. umm.. windows only afaik.
[16:23:59] <kwallace> The 5i25 is a PCI card that plugs into your LinuxCNC PC motherboard.
[16:24:38] <JT-Shop> kengu: steppers?
[16:24:43] <kengu> steppers
[16:24:51] <kwallace> It has two connectors that are kind of like the parallel or printer port, only different.
[16:24:56] <JT-Shop> 5i25 7i76 combo and THCAD
[16:25:04] <joebog> so the two locater slots cut into the board are just to confuse things huh ?
[16:25:14] <JT-Shop> is what I use on my plasma cutter
[16:25:24] <kengu> JT-Shop: something like that could be my choice also
[16:26:00] <kwallace> The slots are part of the PCI slot specification.
[16:26:01] <andypugh> Aren't slots normal on PCI cards?
[16:26:19] <joebog> ONE is
[16:27:04] <joebog> I dont think I have seen two B4
[16:27:14] <joebog> it threw me thats all
[16:27:24] <andypugh> It's a 32bit 5V / 3.3V card
[16:27:34] <kwallace> I seem to recall the slot position depends on the card voltage.
[16:27:45] <andypugh> Yes, it's a universal card
[16:28:02] <joebog> to be honest Ive never paid much attention to whats inside the box
[16:28:15] <joebog> since I chucked windows out nothing has died
[16:28:56] <joebog> last PC I built would have been 5 or 6 years ago
[16:29:02] <andypugh> Plenty of chance for that to change now you are hooking up hundreds of volts to the Parport
[16:29:16] <joebog> I just revived a couple that I have laying about without paying any attention at all
[16:29:54] <joebog> hundreds of volts ???
[16:30:03] <joebog> I know about 45 volts :)
[16:30:43] <andypugh> Ah, but it is a slippery slope to industrial AC servos. Mine are running at 300V :-)
[16:31:06] <joebog> I have one machine to build so far :)
[16:31:15] <joebog> I will use existing steppers :)
[16:31:37] <joebog> the learning curve is BIG already without going to servos
[16:33:01] <kwallace> A port scan on the BBB gives ssh, www, htps and unknown at 3000, so no VNC out of the box.
[16:33:40] <kengu> JT-Shop: what is it with the two versions of thcad, 0-10V and 0-300V.. and ..umm.. i have no experience of plasma cutters.
[16:34:41] <andypugh> Some plasmas have a built-in interface that gives the arc voltage as a 0-10V signal. And some don't. So on the ones that don't you need the high-voltage THCAD
[16:37:55] <kwallace> andypugh: BTW, you're an adventurer, I found that out my neighbor is an adventurer too. http://en.wikipedia.org/wiki/Rex_Pemberton
[16:39:04] <andypugh> I don't think I am really in that category :-)
[16:41:44] <andypugh> Anyway, goodnight folks
[16:54:48] <spack> hey
[16:54:55] <spack> my y axis is going the wrong way
[16:55:31] <spack> i thought i found the line to invert the axis but i guess that's not it
[16:55:36] <spack> because it's having no effect
[16:56:04] <spack> i'm changing setp hm2_5i25.0.gpio.010.invertoutput true
[17:02:51] <PCW> which configuration?
[17:03:57] <spack> the one for the probotix
[17:04:00] <spack> forget what it's called
[17:04:27] <spack> probotix-rf?
[17:04:43] <spack> prob_rfx2?
[17:04:49] <PCW> Y dir is on GPIO8
[17:05:02] <spack> hmm
[17:05:09] <PCW> Z is GPIO 10
[17:05:36] <PCW> x,y,z,a = stepgens 0,1,2,3,4
[17:05:41] <spack> i can jog the y axis back and forth
[17:05:46] <spack> it just goes the opposite way i want
[17:05:51] <spack> i can also jog z back and forth
[17:05:58] <spack> so whatever i have in the config is right
[17:06:04] <spack> other than the dir
[17:06:07] <spack> of y only
[17:06:35] <spack> so i'm confused :S
[17:06:39] <PCW> so is Y connected to stepgen 1?
[17:07:17] <spack> yeah
[17:07:32] <DJ9DJ> gn8
[17:07:34] <PCW> so thats GPIO 8
[17:08:09] <PCW> (on the PROB_RF config)
[17:08:29] <PCW> PROB_RFx2
[17:08:36] <spack> does that show up in the .hal file or anything?
[17:08:57] <PCW> dmesg should show the pinout
[17:09:52] <spack> IO Pin 018 (P2-14): PWMGen #1, pin Out0 (PWM or Up)
[17:10:15] <spack> IO Pin 010 (P3-07): StepGen #2, pin Direction (Output)
[17:10:35] <PCW> #2 is Z
[17:11:12] <spack> oops, i pasted 18 instead of 8
[17:11:38] <spack> 8 is stepgen #1
[17:11:45] <spack> pin direction
[17:11:48] <spack> so far so good
[17:11:56] <PCW> which is Y
[17:12:11] <spack> so if i change the .010 invert to .080 invert, it should invert y?
[17:12:17] <spack> .008
[17:12:18] <spack> ?
[17:12:44] <spack> too much caffeine over here
[17:13:09] <PCW> Should if stepgen 1 is Y and stepgens 1's dir pin is on GPIO8
[17:13:29] <spack> you're right
[17:13:30] <spack> that worked
[17:13:55] <spack> ^5
[17:14:15] <spack> it's been a long time since i set this thing up and i haven't used it since then either
[17:14:21] <spack> thanks
[18:00:41] <Jymmm> Does anyone have 36' Bolt cutters?
[18:01:05] <Jymmm> inch*
[18:56:22] <spack> Jymmm: i do
[18:56:44] <jdh> they have nothing in common
[20:00:48] <Jymmm> spack: what the largest youve cut with them?
[20:01:17] <spack> i'm not sure
[20:01:33] <Jymmm> spack: I *JUST* bought a pair for less then $20 out the door w/ lifetime warranty from HF
[20:01:35] <spack> crappy 1/4" chain links?
[20:02:10] <Jymmm> spack: I think almost 5/8" hasp
[20:02:29] <spack> you cut that?
[20:02:52] <Jymmm> spack: Not yet, I LITERALLY just got them 20 minutes ago.
[20:03:24] <Jymmm> spack: I'mnot syaing that I wont need a 6ft cheater pipe =)
[20:03:32] <CaptHindsight> http://imagebin.org/269800
[20:03:55] <Jymmm> lol cute
[20:05:38] <Jymmm> spack: I also had a 20% off coupon http://www.harborfreight.com/36-inch-bolt-cutters-41150.html
[20:06:23] <Jymmm> JT-Shop: Actualy something from HF half way worth a damn AND a lifetime warranty =)
[20:06:51] <spack> yeah
[20:07:01] <spack> i tried to get a spot welder there a few weeks ago with the 20% off coupon
[20:07:12] <spack> apparently it doesn't cover welders anymore
[20:07:21] <Jymmm> spack: ah, bummer
[20:07:36] <Jymmm> HF is hit and miss on the products. Mostly miss
[20:08:05] <Jymmm> the bolt cutters actually have 5 star reviews, wow
[20:08:11] <CaptHindsight> they just had their Giant Tool Sale and I didn't even see one giant tool
[20:08:42] <Jymmm> lol
[20:09:25] <Jymmm> Damn... http://www.homedepot.com/p/Klein-Tools-36-in-Bolt-Cutters-63336/100647781
[20:09:25] <spack> haha
[20:10:13] <Jymmm> I dont need them THAT bad
[20:10:31] <Jymmm> http://www.youtube.com/watch?v=1ruMqX_cwHg
[20:10:32] <Tecan> (1ruMqX_cwHg) "Combination padlock vs. 36 inch bolt cutter" by "sezardon" is "Howto" - Length: 0:01:36
[20:11:14] <Jymmm> oh shit... HAHAHAH
[20:11:27] <Jymmm> ***SNAP***
[20:14:24] <Jymmm> 42" but still... http://www.youtube.com/watch?v=TXQOO3RMXNM
[20:14:25] <Tecan> (TXQOO3RMXNM) "Bolt cutters vs Master Lock Shrouded Boron Alloy" by "Brute4rce" is "Education" - Length: 0:04:31
[21:30:41] <CaptHindsight> http://www.youtube.com/watch?feature=player_embedded&v=xZOiNdkJ8SU#t=44 Trunnion Table positioning accuracy of up to +/-0.005mm +/-0.0002"
[21:30:42] <Tecan> (xZOiNdkJ8SU) "5-Axis Machining, 5-Axis Mill" by "MDAPrecision" is "Tech" - Length: 0:06:43
[21:33:03] <CaptHindsight> any idea what control that is @5:00 ?
[22:23:20] <joebog> quick question guys, the stepper morors that I have at present hace a 45 volt supply to the motor board without load, how critical is this ?
[22:23:25] <joebog> can I reduce the voltage ?
[22:24:35] <joebog> wow my spelling is attrocious !!
[22:24:46] <joebog> actually its my typing :)
[22:34:16] <CaptHindsight> http://www.selene.co/Blog/2011/07/stepper-motor-voltages-explained/
[22:35:06] <joebog> thanks Capt
[22:35:35] <CaptHindsight> whats the maximum voltage input of your stepper driver ?
[22:36:01] <joebog> 48V
[22:36:18] <CaptHindsight> 45V is higher than any of the Toshiba stepper IC's I've noticed
[22:36:37] <CaptHindsight> ah then never mind
[22:36:48] <joebog> I thought that was flying pretty close ??
[22:37:11] <joebog> its not my "normal" method, but thats what I found in the Cipher box
[22:37:42] <joebog> ans my motors DONT have makers or ratings plates !!!
[22:37:58] <joebog> removed by my greedy supplier !!! at Cipher
[22:40:19] <CaptHindsight> http://www.geckodrive.com/gecko/images/cms_files/Step%20Motor%20Basics%20Guide.pdf
[22:40:34] <CaptHindsight> ^^ covers it pretty well
[22:41:15] <joebog> still reading first link :) but Im logging them all so I can go back
[22:42:08] <joebog> so for my test bed setup it would be OK to run say 35 volts ?
[22:42:22] <CaptHindsight> yes
[22:42:35] <joebog> I know the speed will suffer, but for testing it should be OK or am I going to have problems ?
[22:42:54] <CaptHindsight> should be fine
[22:43:22] <joebog> thks
[22:43:59] <CaptHindsight> just to be safe in case they are some goofy design or spec, check the temp after a few seconds and then again after 30 sec and a few minutes
[22:44:18] <joebog> I am contemplating a MESA setup for these tests
[22:44:58] <joebog> the 5i25 and daughter card is that acceptable ?
[22:45:22] <joebog> its a kit thats available ( as you probably know)
[22:45:33] <CaptHindsight> for step generation but the driver will actually control the voltage and current to the windings
[22:45:49] <joebog> yep I understand
[22:46:15] <joebog> the Mesa kit is my idea of buffering
[22:46:43] <joebog> I have a limited budget :)
[22:47:05] <joebog> so I am trying to get the best bang for the buck so to speak
[22:47:31] <joebog> and without winding a specific unit for the tests I thought to use something I have off the shelf
[22:48:42] <joebog> I am trying not to strip the Cipher unit, The customer wants to try sell it on
[22:50:16] <CaptHindsight> snoozing time
[22:50:24] <joebog> k thanks again
[23:09:51] <uw> anybody work with titanium?
[23:09:58] <uw> or at least know a good supplier?
[23:11:33] <Valen> titanium joe perhaps?