#linuxcnc | Logs for 2012-11-14

Back
[00:12:45] <mazafaka> https://www.youtube.com/watch?v=__2YND93ofE Massachusetts Institute of Technology has its channel on Youtube, free lections available
[00:34:04] <mazafaka> https://www.youtube.com/watch?feature=player_embedded&v=8ox1Cb9uoy8 just watch, it's about RC steam trike
[01:53:13] <DJ9DJ> moin
[02:05:10] <Loetmichel> mornin'
[02:10:51] <Loetmichel> *grrr*... either i should get a diet or buy a new desk chair... thats annoying... *stand up* *pull lever* *let chair get up again* :-(
[02:48:46] <mazafaka> Loetmichel: https://www.youtube.com/watch?feature=player_embedded&v=8ox1Cb9uoy8
[02:48:52] <mazafaka> i need to reboot
[04:46:40] <alex_joni> cool http://www.3ders.org/articles/20121107-voxeljet-builds-aston-martin-models-for-james-bond-film-skyfall.html
[05:58:01] <Gabriel__> Regarding servo systems, I can't realize the advantage of velocity or torque control over the step/dir position control (in other words, close the position loop inside the controller rather than inside the servo drive). Can someone explain this to me? Thanks.
[06:04:40] <asdfasd> I think that the key is acuraccy and quality of control. Im using step/dir servo, in order to control, the pid loop need error, with perfect pid settings you have +- 5 to 10 counts error, this is normal
[06:05:58] <asdfasd> that mean constant move on one axis and start of acceleration on another axis will couse bad sinchronisation between the axes
[06:07:20] <asdfasd> that can be corrected if they are controlled together considering the required trajectory
[06:07:32] <jthornton> with step and direction LinuxCNC ASSUMES everything is ok with closed loop LinuxCNC KNOWS if there is a following error
[06:12:24] <asdfasd> with external pid controller that error remain invisible for linuxcnc and cannot be corrected, but for me even with that error the accuracy is far enough
[06:13:37] <Gabriel__> But the LinuxCNC PID routines also are fed with position commands and acts in order to minimize the errors (by nature of PID). They also seems to be implemented independently for each axis (at least on HAL level, not sure about the upper levels). Is there any perception by LinuxCNC if the error of an axis grow much more than for other axis, and is any countermeasure taken (I mean in overall
[06:13:38] <Gabriel__> axis control, not for each axis individually)?
[06:19:27] <Gabriel__> I mean something like forcing the more advanced axis to wait for the axis that lagged behind for some reason (due to having experinced an unusual cutting resistance in that direction, for example).
[06:21:26] <asdfasd> I guess it is too late to correct during move, the better way is to make the corrections while linuxcnc is planning the move
[06:24:16] <Gabriel__> asdfasd: I think I understood. So the position feedback is also taken into account in layers above the HAL PID, such as in the trajetory planner? Is this correct?
[06:24:53] <asdfasd> Im not sure but I think that is the correct way, but it is possible to be something much better
[06:25:55] <asdfasd> when you drive with step/dir the trajectory planner is using only ramps which you define for each axis
[06:26:41] <asdfasd> if the pid is internal then you have much more data available, you can measure responce time, error, and many more data
[06:29:59] <Gabriel__> Do doubt about the much more data available, I just was not sure which actions the controller could take with advantage over the drive. Maybe not too much on the HAL PID level, but now that you hinted me about it being taken into account in the trajectory planner, it's now clear
[06:30:56] <Gabriel__> I will research the trajectory planner code to understand how it's taken into account. Thanks for helping.
[07:08:44] <Gabriel__> asdfasd: I still need to study the code much more, but considering the way the value motor_pos_fb is used inside control.c, maybe it's possible to work accurately in position mode once I feed the axis.n.motor-pos-fb pin with encoder data. This way the trajectory planner will take into account the encoder data, whilst having no PID inside the HAL layer (avoiding expensive and sensitive analog
[07:08:44] <Gabriel__> electronics and taking advantage from the drive high speed PID controller)
[07:09:19] <Gabriel__> This will require some experiments, and I will report the results later.
[07:13:45] <asdfasd> http://www.linuxcnc.org/index.php/component/content/article/2-technical-articles/42-emc-internals#figure2
[07:13:56] <asdfasd> Motion Controller EMCMOT section
[08:06:24] <archivist> Gabriel__, there is a wiki page or two on the trajectory planner
[08:07:58] <Gabriel__> archivist: thanks, I will check them
[08:15:05] <SWPadnos> Gabriel__, the HAL PID has both error terms and feedforward terms (FF0-FF2). The feedforward terms can be used to give an analog torque drive a little extra push when there's a programmed velocity change
[08:16:12] <SWPadnos> step/dir drives can only react to new position updates (steps), whereas PIDFF can use (some) knowledge of the immediate future
[08:18:25] <SWPadnos> additionally, step/dir drives generally need to have pretty high gain (either I or P, I don't recall), so that they will actually react to a step input quickly. that may not be what you want when you're doing e.g. a long sloping move on two axes
[08:20:10] <SWPadnos> an analog servo/drive can be run at a very slow speed, so you'll get a straight line, but a step/dir drive will move in discrete steps - no motion for a while, then a step, then no motion, then a step etc. This could result in a visible difference in the surface finish of the part, though it wouldn't be very pronounced
[08:23:59] <Loetmichel> SWPadnos: depends on the step width
[08:24:27] <Loetmichel> if step widht < wavelength of light: you cant see anything ;-)
[08:24:47] <SWPadnos> let me know when you buy a machine with sub-angstrom resolution ;)
[08:25:15] <Loetmichel> hrhr
[08:25:33] <archivist> it is a discrete distance, and all the microstepping stuff is not accurate either
[08:25:57] <Loetmichel> light wavelength is in the 500s of nm
[08:26:01] <Loetmichel> not angstrom
[08:26:08] <Loetmichel> thats perfectly doable ;-)
[08:26:28] <Loetmichel> and if everything fails: srive the steppers with sinii
[08:26:31] <SWPadnos> yeah, I was exaggerating your exaggeration
[08:27:10] <archivist> microstepping is close to a sine wave.....
[08:28:00] <archivist> http://www.micromo.com/microstepping-myths-and-realities.aspx
[08:29:45] <SWPadnos> gotta run, see you
[08:36:21] <Gabriel__> SWPadnos: Thanks. I really didn't pay attention on the behavior at low speeds; it could be a potential problem. I understand that the drive response on position mode will be proportional to the error value (in terms of accumulated "not-fulfilled" steps), so slow stepping shouldn't cause exagerated acceleration between the target points. But I agree totally that the response to an instantaneus
[08:36:22] <Gabriel__> analog input will be much faster than to accumulated steps, and will result in inproved accuracy. I was also not aware about the HAL PID's feedforward terms. I will read some about it now to realize how it works.
[08:37:55] <jdh> when linuxcnc notices a servoed axis falling behind and compensates, is it always coordinated with the other axes? Does it slow down other axes to match?
[08:38:01] <jdh> or does that even make sense?
[08:43:38] <archivist> would not make much sense if the other axes carried on at full speed if one slowed due to load
[09:09:54] <mazafaka> archivist: Can brushless DC motor giving out its highest torque return errors on tiny revolutions - the ones like microstepping?
[09:11:43] <mazafaka> I think there was once a set up when machine couldn't follow the trajectory under load. I used in fact experimental equipment and felt myself upset when errors occurred.
[09:12:32] <archivist> if slow then you have a similar discrete error from the encoder, else check your settings etc
[09:14:39] <pcw_home> the limit on torque applied vs error measured ( gain) in a given system is stability
[09:14:52] <mazafaka> We have a motor from the axis, which revolves the spindle (up to 450 RPMs) using some ratio, gears or belt.
[09:21:17] <mazafaka> people wanted form me to use 40-50 mm in diameter 4-flutes end mill bits, and 60 RPM on machine where maximal RPMs on spindle are 450, and I finally got that we can work ***starting from 90 RPM*** with end mills up to 30 mm in diameter, with 4 or better 6 flutes. Finally we used 25mm and 20mm HSS end mill bits, sharpening its tip on regular grinder for cutters. They paid me much, really much for the machine which could work 24/7 instead of several
[10:42:20] <L84Supper> http://www.surpluscenter.com/item.asp?item=1-201-8-4-C&catname=powerTrans great price but a little to large a footprint
[10:44:32] <archivist> those are for rough work not posh machine tool work
[10:45:03] <L84Supper> has anyone seen the prices on bearing support for ball and lead screws?
[10:45:16] <L84Supper> yes, a bit of axial slop
[10:46:47] <L84Supper> http://www.automationtechnologiesinc.com/products-page/ballscrew-supports 3x price difference for free vs fixed ends
[10:46:48] <archivist> they are self aligning bearings little control of where the inner is with reference to the mounts
[10:48:12] <archivist> ball screw bearings are in a different class with preloaded bearings
[10:48:45] <L84Supper> sure, I'm just looking at the actual blocks here
[10:49:41] <L84Supper> I can't seem to find any decent 1/2" blocks in stock
[10:50:04] <L84Supper> i might try the AT ^^ blocks since they are local
[10:51:17] <archivist> but some cheating with a couple of cheaper bearings and a shim/adjustment you could make your own
[10:51:35] <L84Supper> debating as usual
[10:51:39] <L84Supper> time vs buy
[10:52:33] <L84Supper> low speed <1rpm, 2 N load
[10:53:10] <L84Supper> could use a bronze bushing and not wear it out for years
[10:54:30] <L84Supper> http://www.ebay.com/itm/Precision-Ball-Screw-Support-Bearing-25TAC62-/170847902554?pt=LH_DefaultDomain_0&hash=item27c7541b5a bingo and there a bakery there i like
[10:56:59] <archivist> a lead screw I have here has two http://www.ebay.com/itm/Brand-new-vctrc-F3-8M-Thrust-Bearing-For-T-REX-450-SE-V2-SPORT-PRO-/121019839616?pt=LH_DefaultDomain_0&hash=item1c2d583480
[10:57:10] <archivist> inner and outer
[10:57:43] <archivist> not that size but you get the idea
[10:57:44] <L84Supper> heh, Hong Kong, too bad I'm in the midwest for couple of weeks
[10:58:18] <L84Supper> it's getting harder to find parts in the US without 1w+ lead times
[11:00:49] <archivist> that size hobby king in the states :)
[11:01:30] <L84Supper> grainger, mcmaster, reid and few others, but the prices are always really high
[11:02:00] <archivist> $1.85 for two little ones at hobby king
[11:02:19] <archivist> 1/4 in inner /me just found
[11:04:31] <L84Supper> just figured out the magic on ebay, found several in the area for <$5
[11:05:14] <L84Supper> sometimes it odd how parts get listed
[11:06:13] <archivist> depends on the writer of the search engine too
[11:40:12] <Loetmichel> re @ home
[11:44:36] <IchGuckLive> Hi all
[11:44:43] <tjb1> ello.
[11:44:48] <IchGuckLive> B)
[11:45:02] <tjb1> Are you in Germany?
[11:45:07] <IchGuckLive> yes
[11:45:13] <IchGuckLive> dark outside
[11:45:15] <tjb1> What time is it right now
[11:45:21] <IchGuckLive> 18:33
[11:45:52] <tjb1> Ah
[11:45:57] <tjb1> Hows the weather
[11:46:08] <IchGuckLive> Cold near Snow
[11:46:22] <IchGuckLive> today i called the USA
[11:46:37] <IchGuckLive> San antonio texas
[11:46:49] <IchGuckLive> you all have been wright by warning me
[11:47:51] <IchGuckLive> as the university here has a partnership with them (I dident know that ) they invide from time to time stuff from here to teatch there
[11:48:39] <IchGuckLive> they got real MASTER and Bachelor there also not only Religios stuff
[11:49:54] <IchGuckLive> tjb1: how is it going
[11:50:10] <tjb1> it is going
[11:50:40] <Loetmichel> HRHRHR
[11:50:41] <jdh> I wasn't warning you about the school, I was talking about religion in general.
[11:50:45] <Loetmichel> germans ;-)
[11:51:18] <IchGuckLive> Loetmichel: hope we are not the onlyones O.O
[11:52:14] * Loetmichel too
[11:52:26] <Loetmichel> but hopefully with better grammar ;-)
[11:53:28] <IchGuckLive> LOL
[11:55:16] <IchGuckLive> Loetmichel: did you see the sun today at your place
[11:55:46] <IchGuckLive> here it wars so foggy that even the planes coudent see the follow me in front of them
[11:56:59] <Loetmichel> no. foggy all day
[11:57:13] <IchGuckLive> ;-)
[11:57:53] <DJ9DJ> blue sky with lots of sunshine here!
[11:57:56] <DJ9DJ> the whole day long
[11:58:22] <IchGuckLive> locator
[11:58:43] <DJ9DJ> near Münster, NRW :)
[11:58:51] <IchGuckLive> Nice
[12:26:40] <adb> same as DJ9DJ , near Lausanne, suisse
[12:26:58] <DJ9DJ> :)
[12:27:44] <IchGuckLive> adb did you got alot of rain as Italy got
[12:28:30] <adb> yes, few days ago
[12:31:00] <adb> ~ 60 l/m2 in 30 hours
[12:32:14] <IchGuckLive> they got 200 in 24hr
[12:32:36] <adb> i see ..
[12:42:38] <rizo> is there a simple pygtk code for single hal pin to blinking led display available. I found one example halgttk py example at wiki, but it seems to complicated to start with.
[12:52:05] <andypugh> Blinking is the tricky part.
[12:52:35] <rizo> ok, just pin treo ->led on, pin false -> led off
[12:52:43] <andypugh> One solution is to set up a Slow 50% PWMgen in HAL and link that to anything that you want to blink.
[12:52:47] <rizo> ok, just pin true ->led on, pin false -> led off
[12:53:27] <andypugh> Physical LED or Virtual?
[12:53:49] <rizo> virtual, on python interface
[12:54:11] <IchGuckLive> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?GladeVCP_Custom_Widgets
[12:54:20] <IchGuckLive> did you load the HAl_LED
[12:54:33] <IchGuckLive> to the glade wiget groups
[12:58:13] <rizo> i tried this example: http://www.linuxcnc.org/docs/2.4/html/hal_halmodule.html, but i get an error in "while 1": expected an identifier block
[12:59:36] <IchGuckLive> why arent you using pyvcp or glade instet
[13:00:34] <IchGuckLive> what is the main goal to be reatched MFG
[13:00:48] <rizo> i will try...i just wanted to start with something simple - short code :)
[13:01:04] <IchGuckLive> pyvcp
[13:02:59] <IchGuckLive> rizo: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?PyVCP
[13:03:15] <mevon> hi andypugh, just wanted to let you know that I made some progress on EMC2Arduino, and seems it was my limit switches that prevent my mill to move, I was assuming default states in emc were properly set
[13:04:37] <mevon> also, I started off from the simulation 3 axis .ini and added the custom.hal
[13:04:48] <IchGuckLive> rizo: look at the end bottom first to start
[13:05:27] <andypugh> It is the choice of the machine builder which way the limit switches operate. Connecting the pin to 0V is actually rather common, and arguably superior.
[13:05:53] <andypugh> I think all mine are 5V for open and 0v for closed.
[13:06:18] <IchGuckLive> 5V is better for error
[13:06:35] <IchGuckLive> no power error B)
[13:07:45] <andypugh> Short to ground or 0V setting a limit trip sounds safer to me.
[13:08:33] <IchGuckLive> GND can also be a fail switch so 5V is better
[13:08:43] <mazafaka> when machine reaches the limit switch, it shorts the power to axis to the ground&
[13:08:53] <mevon> indeed mine are too but seems limit switches work with toggling effect
[13:09:02] <pcw_home> Right so NC switch to power is arguably safer (open or short to ground are both faults)
[13:09:41] <IchGuckLive> but they are more expensive
[13:10:11] <rizo> IchGuckLive, thank you, will do that
[13:10:11] <Aero-Tec> found and fix my lathe spindle speed problem
[13:10:15] <rizo> bye
[13:10:41] <IchGuckLive> Aero-Tec: the Hyanyang on 7i77
[13:10:49] <mevon> mazafaka, did u need a capacitor for deboucing?
[13:11:08] <IchGuckLive> mevon: best to do it into hal
[13:11:08] <Aero-Tec> is it just me or did anyone else find 2.5 more sensitive to grounding and noise?
[13:11:26] <mevon> IchGuckLive, I see
[13:11:44] <IchGuckLive> mevon: it is easy and got to work
[13:11:49] <Aero-Tec> 2.4 worked fine, but 2.5 went nuts
[13:11:59] <Aero-Tec> yes to cap
[13:12:10] <Aero-Tec> and more grounding
[13:14:10] <mevon> ok I dont know much about milling bots but I build a reprap...
[13:14:16] <Aero-Tec> the last one to track down was grounding the shielding of the Pport cable, the bob I was using did not tie the Pport cable shielding to the grounding board mounting pads
[13:14:29] <mevon> I find cnc and mills are way more complex
[13:15:18] <mevon> but Im pretty sure it can work of the usb
[13:15:41] <IchGuckLive> mevon: did you see the querry ?
[13:16:17] <Aero-Tec> is the 2.5 being more sensitive to noise only something I have found or has this also happened to others?
[13:16:46] <mevon> I did now thks for telling me :S
[13:17:17] <mevon> Im over usb so debouncing is onboard
[13:17:22] <Aero-Tec> one thing that did contribute to the problem is the new drive run at a higher switching speed
[13:17:26] <IchGuckLive> i changed to master 2.6-pre as 2.5 gave me some hits and did move on its own without getting the PC and mashine to a hold had to Push the master power off
[13:18:07] <IchGuckLive> Aero-Tec: so increase all numbers
[13:18:22] <andypugh> Aero-Tec: It is almost impossible that 2.5 could be more sensitive to noise than 2.4
[13:18:27] <IchGuckLive> did you check the max steps per sec to get the savest rate to travel
[13:19:18] <Aero-Tec> it was not a motor problem
[13:19:28] <Aero-Tec> it was a input problem
[13:19:42] <Aero-Tec> index is the only input I used
[13:19:56] <mevon> received noise from nearby signal?
[13:20:16] <Aero-Tec> and it went nuts with out the grounding the filtering
[13:20:37] <Aero-Tec> yes, the higher speed drives made more noise
[13:20:55] <mevon> shorter signal bursts emit a lot of parasitic frequencies
[13:20:58] <Aero-Tec> thing is 2.4 did not react to the noise like 2.5 did
[13:21:20] <mevon> specialy if its sharp
[13:21:22] <Aero-Tec> just found it funny
[13:21:45] <Aero-Tec> when the drive were not active every thing was fine
[13:22:24] <Aero-Tec> but as soon as the drives were activated the spindle speed went nuts
[13:22:43] <mevon> an infinitly short signal burst has an infinite number of spectral harmonics
[13:22:54] <pcw_home> If you don't have debouncing and you have not changed your base thread it seems really
[13:22:56] <pcw_home> unlikely that the LinuxCNC version could have any possible effect on sampled input data
[13:23:39] <IchGuckLive> im off By Gn8
[13:23:44] <Aero-Tec> one would think that
[13:23:47] <pcw_home> though something as simple as re-arranging power cables could
[13:24:10] <pcw_home> Ill bet something else changed
[13:24:50] <Aero-Tec> that is possible as I did move them around about the time I updated
[13:25:14] <Aero-Tec> so it is possible
[13:25:59] <Aero-Tec> it is fixed now, just wanted to ask if anyone else had the same thing happen
[13:26:06] <mevon> gn IchGuckLive
[13:26:28] <mevon> Aero-Tec, Im newby noob
[13:26:47] <mevon> but I believe pcw_home is right
[13:27:11] <Aero-Tec> I am also a nood to EMC
[13:27:16] <mevon> it is crazy to think software wont behave the same twice
[13:27:31] <mevon> unless configs have been changed or so
[13:28:18] <Aero-Tec> it is 2 versions, and a large upgrade so some thing could have changed in the input section of the code
[13:28:45] <mevon> could of but seem software quality is incredible
[13:28:52] <Aero-Tec> just saying, 2.4 worked fine, 2.5 did not
[13:29:04] <mevon> have you switched back to 2.4^
[13:29:06] <mevon> >
[13:29:07] <mevon> ?
[13:29:12] <mevon> sry flood
[13:29:56] <Aero-Tec> it is a very good software package, no complaints on my end
[13:30:12] <Aero-Tec> I would not know how to switch back
[13:30:36] <mevon> :S ok me neither
[13:31:00] <mevon> but a backup could done the job maybe
[13:31:24] <Aero-Tec> if I had a backup to fall back to
[13:32:04] <mevon> Im wondering if I really want to update my ubuntu
[13:32:12] <Aero-Tec> it is working now, was up late last night tracking down the problem
[13:32:29] <mevon> I bet your bot is even better now
[13:32:35] <Aero-Tec> I just updated EMC, not linux
[13:33:09] <mevon> ok so youre still on 10.04? updating the os is recommanded?
[13:33:16] <Aero-Tec> running a P4 CPU it may not handle the linux update
[13:33:34] <Aero-Tec> I am 8.04
[13:33:38] <mevon> indeed
[13:33:57] <mevon> wow ok bow to you^^
[13:34:19] <mevon> :P
[13:34:36] <Aero-Tec> I was going by what I was told
[13:35:00] <mevon> ubuntu is a bigger os
[13:35:02] <Aero-Tec> and the recommendations given to me
[13:35:20] <Aero-Tec> so far I like it
[13:35:53] <mevon> i am very familiar with debian and the linuxcnc workspace is just like it
[13:35:58] <mevon> its perfect
[13:36:07] <andypugh> mevon: If you are running 10.04 and LinuxCNC upgrading the OS is _not_ recommended as we don't have a working version of LinuxCNC for the later kernels yet (but it is very close)
[13:36:08] <Aero-Tec> now to finish the tool table for this next project and making chips
[13:36:44] <mevon> andypugh, ok thanks Im always tempted with sys updates to upgrade... will wait on that thks
[13:40:34] <yoyoek1> hi
[13:40:52] <mevon> hello
[13:42:18] <yoyoek1> where can i find gladevcp ?
[13:47:55] <andypugh> http://www.linuxcnc.org/docs/html/gui/gladevcp.html
[13:48:27] <andypugh> It should just be there, and there ought to be demo configs in the configuration picker.
[13:49:13] <yoyoek1> I don't have gladevpc :( why ?
[13:49:23] <yoyoek1> gladevcp I mean
[13:49:24] <andypugh> What LinuxCNC version?
[13:49:55] <yoyoek1> 2.40~pre
[13:54:37] <kwallace> I believe Glade VCP needs Ubuntu 10.04 and LinuxCNC 2.5. See -> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?GladeVcp
[13:56:01] <yoyoek1> ok ;( I'm scared that the update will kill my computer
[13:56:05] <yoyoek1> it is so old
[14:06:19] <mevon> anyone has a good link or tutorial on how to setup limit switches properly?
[14:06:46] <tjb1> wiring or in linuxcnc?
[14:06:55] <mevon> linuxcnc
[14:06:57] <mevon> pls
[14:07:13] <mevon> but reading on the wiring would be helpful too
[14:07:21] <tjb1> home and limit or just limit?
[14:07:40] <mevon> I dont want to work with home switches :S is that bad?
[14:08:11] <tjb1> Cant you use Stepconf to define limits as inputs?
[14:08:46] <mevon> I did but the logic I saw from that is that the switches are toggled?
[14:09:24] <mevon> like somekind of litteral lever that gets tripped when passed
[14:09:59] <tjb1> So what are your limits?
[14:10:11] <tjb1> Proximity sensors?
[14:10:13] <mevon> plain switches
[14:10:39] <tjb1> http://www.superdroidrobots.com/images/TE-044-000.jpg
[14:10:40] <tjb1> ?
[14:11:06] <mevon> tact switch, inverted, 0V is closed
[14:11:18] <mevon> exactly
[14:11:39] <tjb1> You have each switch wired to its own input or in series?
[14:11:40] <mevon> I have a week pull-up res
[14:11:47] <mevon> eache saperated
[14:11:59] <mevon> sry bad spelling
[14:12:26] <tjb1> 1 switch for both limits?
[14:12:44] <mevon> 6 sw for 3 axis, min max
[14:13:07] <mevon> I have a mantis basicly
[14:13:24] <tjb1> net pos-limit-x <= parport.0.pin-10-in-not
[14:13:35] <tjb1> net neg-limit-x <= parport.0.pin-11-in-not
[14:13:45] <mevon> so -in-not
[14:13:47] <tjb1> Change 10 and 11 to your input numbers
[14:13:50] <tjb1> The not inverts it
[14:14:05] <tjb1> I need the not on mine but you may not
[14:14:20] <tjb1> brb
[14:14:26] <mevon> 0V is the pressed switch state
[14:14:39] <mevon> thks
[14:24:58] <tjb1> You also need this
[14:25:21] <tjb1> net neg-limit-x => axis.0.neg-lim-sw-in
[14:25:22] <tjb1> net pos-limit-x => axis.0.pos-lim-sw-in
[14:26:39] <tjb1> axis.0 should be your x
[14:29:43] <mevon> ok got it nice thanks I just want to make sure I don't rip everything apart
[14:29:54] <mevon> on first times its good to be scared
[14:32:30] <mevon> so 20 ms is a good debouncing delay?
[14:39:52] <Loetmichel> mevon: 20 ms is a good rule of thumb for classic key switches
[14:40:10] <Loetmichel> not for big plunger switches ;-)
[14:40:22] <Loetmichel> or contactors
[14:40:32] <Loetmichel> they can ring MUCH longer
[14:45:18] <mevon> Loetmichel, nice thanks
[14:52:46] <mevon> Im in an infinite loop of overthinking here, about switches and software limits :S
[14:54:00] <mevon> I always come down to this conclusion that only one min/home switch would be needed per axis
[14:59:49] <andypugh> mevon: Yes, one min/max/home switch per axis works perfectly well.
[15:00:11] <andypugh> That is what my machine has, even though it has lots of available IO pins.
[15:00:21] <pfred1> no switch works pretty good too
[15:00:30] <pfred1> :)
[15:03:13] <mevon> think ill go with sw limits and wire the switch output to the axis enable pin directly
[15:04:33] <andypugh> Home switches are worthwhile.
[15:05:08] <mrsun> hmm how far between the bearings for the gantry using V bearings for good rigidity ...
[15:05:35] <andypugh> What is the idea of connecting the limits to enable? That sounds rather eccentric
[15:05:40] <pfred1> how come V groove bearings cost so much?
[15:05:48] <mrsun> pfred1, angular contact
[15:05:55] <mrsun> i guess
[15:05:59] <frallzor> ahoyhoy
[15:06:04] <mrsun> omg no :/
[15:06:04] <mevon> andypugh, yeah somekind of direct hardware safety I guess
[15:06:15] <pfred1> mrsun I'm using plain old skate bearings
[15:06:20] * frallzor cuddles with mrsun
[15:06:31] <mrsun> pfred1, well they arent angular contact and can have play :P
[15:06:36] <frallzor> Whats the discussion about today?
[15:06:47] <pfred1> mrsun I haven't noticed any
[15:06:56] <mrsun> "hmm how far between the bearings for the gantry using V bearings for good rigidity ..." for me .. for others i dont know :P
[15:08:23] <frallzor> My question of the day; a material like "tooling foam" but cheaper
[15:08:28] <frallzor> any suggestions
[15:08:50] <mrsun> tooling foam ?
[15:09:08] <jdh> tooling air
[15:09:24] <mevon> ive seen ppl making molds out of insulation foam (home depot) pasted togheter on multiple layers
[15:09:25] <frallzor> foamed polyurethane
[15:09:33] <andypugh> frallzor: MDF glued together in layers. Not as good as model board though.
[15:09:50] <mrsun> frallzor, for milling out of or what ?
[15:09:55] <frallzor> yes senor
[15:10:03] <frallzor> I do like MDF, but not in this case
[15:10:12] <mrsun> what is it for ? test cuts or ? :)
[15:10:18] <andypugh> You do get obvious layers, and it is fuzzy.
[15:10:19] <frallzor> want the rigidity of MDF/tooling board but lighter and cheaper
[15:10:23] <DaViruz> i like trains.
[15:10:38] <mrsun> i like pancakes
[15:10:47] <mevon> i like french toats
[15:10:49] <frallzor> mrsun "test cuts" for a wall-decoration system
[15:10:51] <frallzor> sort of =)
[15:11:02] <mrsun> machinable wax? :)
[15:11:05] <mrsun> homemade that is :P
[15:11:23] <mevon> clay?
[15:11:23] <mrsun> no wear on tools and you get perfect molding material at the same time
[15:11:26] <frallzor> drillable and possible to screw things into it
[15:11:27] <mrsun> as its self releasing
[15:11:33] <frallzor> ?
[15:11:36] <mevon> dried clay?
[15:11:55] <frallzor> want it as light as possible for tests
[15:12:05] <frallzor> and still strong
[15:12:20] <mrsun> aerogel ? :P
[15:12:44] <pfred1> frallzor sounds to me you want a miracle
[15:12:45] <frallzor> got a few tooling boards but thats not cheap to usew =/
[15:13:02] <mevon> a dried clay block is kinda light and strong
[15:13:18] <andypugh> I made a foundry pattern out of a lot of layers of plywood glued together
[15:13:20] <mevon> nothing linke platstic
[15:13:30] <mevon> melted milk jars
[15:13:48] <pfred1> andypugh I made mine out of pine I still haven't gotten around to casting it yet though
[15:13:53] <frallzor> I guess I have to use my tooling boards for the test =)
[15:14:13] <pfred1> I finished mine with varnish it is very smooth should release well
[15:14:32] <andypugh> Insulation foam will work to check the shape. But the finish will be bad
[15:14:45] <andypugh> pfred1: You have to paint them red, that's the rules!
[15:14:48] <mrsun> frallzor, sånt där som man stoppar blommor i då ? gröna porösa helvetet som bara går sönder ;P
[15:15:00] <mevon> id go with cardboard and white glue
[15:15:08] <pfred1> andypugh mine is natural
[15:15:12] <archivist> I did not paint my pattern!!!
[15:15:13] <frallzor> mrsun too fragile =P
[15:15:16] <mrsun> meh
[15:15:33] <pfred1> I wonder if I have a picture of it someplace it is a step pulley
[15:15:34] <frallzor> but I only need 4 blocks machined so tooling board isnt that bad in the end
[15:15:36] <andypugh> Pattterns just don't work if they aren't red!
[15:15:38] <mrsun> http://www.freemansupply.com/productimages/MachinableWax3.jpg
[15:15:44] <mrsun> rox and not that fragile =)
[15:15:53] <mrsun> but still fragile ... :P
[15:16:03] <andypugh> Reusable too
[15:16:06] <mrsun> yep
[15:16:08] <mrsun> and homemadeable
[15:16:21] <frallzor> how to make? link me =)
[15:16:26] <mrsun> just parafin and LDPE
[15:16:28] <archivist> mine did work! http://www.collection.archivist.info/showresult.php?prog=1&srcprog=searchv13.php&srcdata=title&Type=PD&Accn_no=5333&subject=14612
[15:16:51] <mrsun> frallzor, melt parafinwax (ikea candles) and add plastic shopping bags until no more will disolve :P
[15:17:00] <mrsun> simplest recepie :P
[15:17:10] <pfred1> anyone want to see a picture of my Y axis almost done? well half done?
[15:17:11] <mevon> mrsun, youre a genius
[15:17:15] <frallzor> heh =P
[15:17:16] <frallzor> cool
[15:17:17] <archivist> was the cheapest pallet wood
[15:17:30] <pfred1> archivist pallet wood FTW!
[15:17:31] <mrsun> http://repraplogphase.blogspot.se/2010/06/nice-recipe-for-machineable-wax.html
[15:17:47] <pfred1> hit the skids
[15:17:48] <mevon> ouuuuuuuuuuuuuuuuuh
[15:17:57] <frallzor> will try that even if I dont use it for this
[15:17:58] <mevon> nice likes
[15:18:00] <mrsun> so got use for those peasky shopping bags
[15:18:07] <mrsun> frallzor, its real nice to mill =)
[15:18:15] <mrsun> can cut it in bandsaws etc to size
[15:18:17] <frallzor> possible to color in with food coloring or similar?
[15:18:19] <andypugh> https://picasaweb.google.com/lh/photo/iQeBNoJzloy1Gy-ua9oghNMTjNZETYmyPJy0liipFm0?feat=directlink
[15:18:27] <mrsun> remember to cool real slowly when poured into a block tho
[15:18:31] <mrsun> or it will sink like crazy
[15:18:45] <mrsun> frallzor, just use for example the red parts of shopping bags ;P
[15:18:52] <andypugh> https://picasaweb.google.com/lh/photo/abzo-Io3-DMkbvox4DLo19MTjNZETYmyPJy0liipFm0?feat=directlink
[15:18:54] <pfred1> mrsun melting polyethyene?
[15:19:03] <Aero-Tec> the spindle speed problem is still there
[15:19:19] <Aero-Tec> last night it looked like I had it fixed
[15:19:21] <archivist> andypugh, supposed to be black and the red bits being the cores
[15:19:25] <mevon> mrsun, is remeltable? as in it can be cooked in an oven and washed away?
[15:19:39] <mrsun> washed away ?
[15:19:40] <andypugh> archivist: It's the other way roud isn't it?
[15:19:43] <mrsun> remeltable yes
[15:19:43] <mevon> like wax
[15:19:46] <mrsun> just use the scraps
[15:19:51] <mrsun> if they are not contaminated
[15:20:14] <Aero-Tec> what are you melting, aluminum?
[15:20:24] <mevon> to make casts out of plaster?
[15:20:39] <mrsun> huh ?
[15:20:58] <mrsun> machine it, put some latex on it in layers, rip of the mold and fill that with plaster? :P
[15:21:00] <mevon> like for casting metal in molds
[15:21:05] <archivist> andypugh, were black at the museum, the workshop used to make its own spares at the pumping station, were 50 ish patterns left
[15:21:10] <mrsun> its self releasing, you can cast anything in it and it wont stick :P
[15:21:16] <Aero-Tec> I always thought it would be great to reuse the chips and cast them into something useable
[15:21:53] <mevon> would make great suport material for 3d printing
[15:22:06] <pfred1> melting thin aluminum can be less than rewarding
[15:22:27] <Aero-Tec> one just has to do it right
[15:22:28] <archivist> andypugh, but Kew thought the other way I see http://www.collection.archivist.info/showresult.php?prog=1&srcprog=searchv13.php&srcdata=title&Type=PD&Accn_no=335&subject=9623
[15:22:35] <pfred1> although I've made beercanium
[15:22:43] <andypugh> archivist: COmplex rules: http://books.google.co.uk/books?id=NOotk64Grx0C&pg=PA39&lpg=PA39&dq=foundry+pattern+colours&source=bl&ots=Oc7_QJRTNa&sig=hAWxV23415tI17b0OlGl5VwkwzM&hl=en&sa=X&ei=QQikULSRLaGn0QWbioCIDg&ved=0CDAQ6AEwAg#v=onepage&q=foundry%20pattern%20colours&f=false
[15:23:02] <mrsun> for melting alu chips i would recomend adding them to a alu can or something and wack it "closed" and plunge it to the bottom of an already half filled crucible
[15:23:08] <mrsun> or you will get a buttload of oxides
[15:23:21] <mrsun> you will anyways but you should get alotless
[15:23:29] <Aero-Tec> if you can shield the AL from Ox then it should work out well
[15:23:38] <pfred1> mrsun yeah I like to use the submerge method myself
[15:24:03] <mrsun> but cast alu isnt hard to get hands on so i never save the chips
[15:24:04] <pfred1> Aero-Tec all aluminum exposed to air has an oxide layer
[15:24:14] <mrsun> i just go out and take the sledge hammer and wack myself a new piece of soemthing :P
[15:24:30] <andypugh> NIST / US rules are different again: http://gsi.nist.gov/global/docs/vps/csfiles/cs_19-32.pdf
[15:25:03] <Aero-Tec> true, but when heated to almost melting it really Ox
[15:25:17] <Aero-Tec> the trick is to keep it to a min
[15:25:48] <archivist> andypugh, book I have scanned is not fussy http://www.collection.archivist.info/showresult.php?searchstr=BK170&srcdata=subj&Type=BK&Accn_no=170&srcprog=searchv13.php&dir=&file=bk170p015.jpg&subject=2969
[15:26:28] <mrsun> use a reducing flame
[15:26:54] <pfred1> nicest home brew casts I've ever seen have been done with plaster and burned out foam patterns
[15:27:09] <andypugh> pcw_home: What are the plastic stars in with the 7i49 for?
[15:27:15] <pfred1> as long as you don't mind the foam texture in your casts
[15:29:16] <Aero-Tec> anyone got some ideas on why my spindle speed bar is going whacked?
[15:29:31] <Aero-Tec> put in power filter cap
[15:29:38] <Jymmm> HAHAHAHA you know how they call guys "Tools"? Read the category... http://sfbay.craigslist.org/sby/tls/3409907403.html
[15:30:03] <pfred1> Jymmm being a tool is gender neutral
[15:30:16] <Aero-Tec> debounce cap
[15:30:38] <Aero-Tec> extra gounding
[15:30:52] <Aero-Tec> thought it was fixed and now it is back
[15:31:23] <pfred1> Aero-Tec I've used a cap to induce hysteresis but never really to debounce a circuit
[15:32:17] <andypugh> Aero-Tec: Put an alignment mark on the spindle and machine. Zero the encoder with the lines aligned. Run for a while, see if you see an integer number of turns when the lines are aligned again.
[15:32:19] <Aero-Tec> it is more of a RF filter then debounce
[15:32:28] <andypugh> If you do, then you are not losing counts.
[15:33:02] <andypugh> I have found that the encoder velocity output isn't very clean. You might just have to put a low-pass in HAL on the value.
[15:33:13] <Aero-Tec> not loosing counts at all, spiking like mad
[15:34:55] <Aero-Tec> then sig is high things are fine, but during the low when light is passing through encoder the speed maxes out
[15:35:38] <pfred1> so you have an optical sensor?
[15:36:07] <Aero-Tec> even with the spindle not moving but the encoder wheel stopped at the index hole of the encoder it is still maxed out
[15:36:12] <Aero-Tec> yes
[15:36:39] <Aero-Tec> home brew opto
[15:36:48] <pfred1> not a stray light issue?
[15:36:49] <Aero-Tec> was working good
[15:36:58] <Aero-Tec> turned lights off
[15:37:02] <pfred1> OK
[15:37:05] <pfred1> rules that out
[15:37:09] <Aero-Tec> still doing the same thing
[15:37:14] <PCW> andypugh stars are keys for terminal block plugs, they slide in the receptacles You nip the little green tabs on the plugs to match
[15:37:18] <pfred1> homebrew so you know what the exact device is?
[15:37:25] <pfred1> like is it a phototransistor?
[15:38:07] <Aero-Tec> it is a part from a old printer
[15:38:12] <pfred1> ah
[15:38:32] <pfred1> one of those U shaped things?
[15:38:45] <Aero-Tec> one part that has both LED and opto
[15:38:50] <Aero-Tec> yes
[15:39:03] <pfred1> emitter and detector pair
[15:39:08] <Aero-Tec> yes
[15:39:37] <pfred1> well I've fooled around with optocouplers a little and the devices aren't very clean
[15:39:48] <Aero-Tec> the voltage drops to 0.88 at low
[15:39:50] <pfred1> as in they don't really put out godo digital signals
[15:40:21] <pfred1> if you run yours through this device called a schmitt trigger it really helps
[15:40:43] <pfred1> google 74LS14
[15:40:43] <Aero-Tec> I put a scope on it and it looks clean
[15:40:49] <pfred1> hmmm
[15:41:01] <Aero-Tec> has some RF ring but I killed that
[15:41:11] <pfred1> that is strange because phototransistors really can't change all that well
[15:41:35] <Aero-Tec> things were working well for years
[15:41:55] <pfred1> do you think it might have gotten contaminated?
[15:41:55] <Aero-Tec> mach and then 2.4
[15:42:02] <pfred1> like you have a bit of dust on it now?
[15:42:12] <Aero-Tec> possible
[15:42:27] <Aero-Tec> can blow it out and see if that helps
[15:42:28] <pfred1> that does mess printers up paper jizz
[15:43:09] <pfred1> yeah if it did work and now it doesn't I'd assume some kind of failure
[15:43:27] <Aero-Tec> will try a cleaning and see if that fixes it
[15:43:33] <pfred1> could be an environmental change
[15:43:34] <andypugh> PCW: Ah! Nice. And Possibly rather handy. Are they a Phoenix part? It wouldn't hurt to have the same on the 7i64
[15:44:10] <pfred1> I wouldn't go re-engineering it if it used to work
[15:44:23] <PCW> I forget where we got them (had to buy 10K or so of them)
[15:44:24] <Aero-Tec> with the scope on it and the lathe turned on the sig looks good on the scope
[15:44:52] <PCW> what about at the Parallel port pin?
[15:45:17] <PCW> (that's where it matters)
[15:45:29] <andypugh> Aero-Tec: Is the spindle speed actually _used_ for anything?
[15:45:29] <pfred1> I had horrible experience using opto for parallel port output
[15:45:51] <andypugh> If it is just an indicator, thenlowpass filter it in HAL and move on.
[15:48:57] <Aero-Tec> I use it for controlled feed
[15:49:18] <Aero-Tec> x per rev
[15:49:38] <Aero-Tec> some of the G code uses it
[15:51:36] <pfred1> Aero-Tec well I'm still a big believer in filtering optical parts through schmitt triggers
[15:53:20] <Aero-Tec> will try that
[15:53:44] <Aero-Tec> I have a good collection of ICs here
[15:54:07] <Aero-Tec> guess one could use a 555 for that as well
[15:54:13] <pfred1> well 7414 is common
[15:56:26] <pfred1> Aero-Tec a schematic snippet http://i.imgur.com/Fs19Z.png
[15:58:00] <andypugh> Aero-Tec: It will still work OK for feed-per-rev if filtered, and threading uses position (which you say is correct?)
[15:58:44] <pfred1> andypugh there is nothing wrong with working with a clean signal
[15:59:40] <Aero-Tec> I do use it for threading and for x per rev feed control
[15:59:47] <andypugh> pfred1: True. I have a vague feeling that there is something wrong with the encoder velocity signal. I think it comes down to the module believing the period it is given, when it isn't actually accurate.
[16:00:19] <pfred1> yes I can see that
[16:00:37] <Aero-Tec> what voltage is needed for a low?
[16:00:47] <Aero-Tec> it is running .88
[16:00:51] <pfred1> definition is .8 but it varies
[16:00:56] <Aero-Tec> tad high for my liking
[16:01:07] <pfred1> yeah evne .8 is considered marginal
[16:01:38] <pfred1> what schnitt triggers do it they lock swings down
[16:01:44] <pfred1> that is their reason for being
[16:01:49] <pfred1> garbage in clean out
[16:01:57] <Aero-Tec> would the 7514 see .88 as high or low?
[16:02:22] <pfred1> well you can adjust it with the pull up resistor
[16:02:28] <DJ9DJ> gn8
[16:02:44] <Aero-Tec> tried that
[16:02:53] <Aero-Tec> same thing
[16:02:55] <pfred1> but on the output side it'll be straight up and down high and low square waveform
[16:03:20] <pfred1> there will be no inbetween
[16:03:56] <Aero-Tec> the sig looks very clean on the scope
[16:04:09] <Aero-Tec> very nice and clean
[16:04:19] <Aero-Tec> just low is .88
[16:04:27] <pfred1> yeah that isn't good
[16:04:58] <Aero-Tec> even with a high val resister as a pullup it still is .88
[16:05:39] <PCW> What does it drive? thats marginal for TTL thresholds but probably OK for 5V CMOS
[16:05:48] <Aero-Tec> almost looks like junction voltage of photo trans
[16:06:06] <PCW> must be photodarlington
[16:06:59] <Aero-Tec> so why has it worked for years and now is a royal pain?
[16:07:28] <PCW> a high value pullup make the signal susceptible to spike noise (best thing is to buffer at the OPTO)
[16:09:02] <Aero-Tec> true, and I am using a already high value of 4.7K
[16:09:55] <Aero-Tec> went to 10K and still the same thing
[16:10:00] <PCW> is this just index? if its a wide enough slot you could try debouncing the signal
[16:10:19] <Aero-Tec> just index
[16:10:21] <PCW> Yeah optoDarlington
[16:11:09] <PCW> high gain but slow and high saturation voltage
[16:11:28] <Aero-Tec> wonder if I did a debounce in 2.4 to get it working?
[16:11:52] <Aero-Tec> I know I did in mach
[16:12:42] <Aero-Tec> played with debounce to get it working
[16:13:27] <Aero-Tec> where would debounce be and would a new install not use the same setting?
[16:13:43] <Aero-Tec> could be why 2.4 worked and 2.5 did not
[16:15:02] <andypugh> If it was in your HAL then it should still be there.
[16:15:18] <Aero-Tec> I had messed with 2.4 on and off in the lathe for awhile so do not remember every thing I have done with it
[16:15:45] <Aero-Tec> would that be in the config dir for the machine?
[16:15:57] <Aero-Tec> if so it should be using the same hal
[16:16:38] <Aero-Tec> unless I need to change a location setting
[16:16:56] <Aero-Tec> like linuxcnc instead of EMC
[16:17:11] <andypugh> No, it should still be in the config. Have a look?
[16:22:27] <alex4nder> yoh
[16:31:30] <andypugh> PCW: Did you see the 5i25 / 7i39 query on the mailing list?
[16:37:30] <Aero-Tec> http://pastebin.com/QghaPXrp
[16:37:41] <Aero-Tec> hal seems to have debounce
[16:37:56] <Aero-Tec> not sure why the same debounce did not work
[16:38:14] <Aero-Tec> every thing looks good
[16:38:37] <Aero-Tec> and it must be finding the hal as it moves the axis
[16:39:02] <andypugh> That's a lowpass, not a debounce, but it might be interesting to see what decreasing the gain does.
[16:39:45] <Aero-Tec> any suggestions?
[16:40:05] <Aero-Tec> would adding a high pass help?
[16:40:25] <andypugh> Make the lowpass gain 0.001 and see what happens.
[16:42:09] <Aero-Tec> will do
[16:49:42] <Aero-Tec> tried 1 0.1 0.01 0.001 0.0001 all the same except 0.0001 took a long time to increase
[16:50:43] <Aero-Tec> rpms with the spindle off and the encoder sig low is bouncing between 7K and 10K
[16:51:19] <Aero-Tec> the fastest standing still spindle in the world.....LOL
[16:53:09] <Aero-Tec> could just add a sig trans to drop the voltage of low down
[16:56:35] <andypugh> If the speed wobbles with the spindle stopped, then I don't think it can be software. Does Halscope see pulses coming in?
[16:57:06] <Aero-Tec> looks like it does
[16:57:08] <pfred1> Aero-Tec what do you use as a power supply?
[16:57:30] <Aero-Tec> but hal scope is all over the map
[16:57:56] <Aero-Tec> it will float between
[16:58:19] <Aero-Tec> 5V supply from main power supply
[16:58:34] <pfred1> so out of a PC PSU?
[16:59:27] <pfred1> is the PSU old now?
[16:59:46] <Aero-Tec> I have a 95 v out, 220 in 3000 watt main and it has a 5V regulated output as well
[17:00:28] <pfred1> I mean you say the thing has worked for years
[17:00:36] <Aero-Tec> it is the power supply for the motor drives
[17:00:41] <pfred1> PSUs are known to go bad over time
[17:00:43] <Aero-Tec> yes
[17:01:19] <Aero-Tec> thyroid wound transformer
[17:01:19] <pfred1> your filter caps in the PSU could have developed high ESR by now
[17:01:45] <pfred1> a toroid?
[17:01:53] <Aero-Tec> oops
[17:01:54] <Aero-Tec> yes
[17:01:56] <Aero-Tec> lol
[17:01:56] <pfred1> :)
[17:02:43] <pfred1> any electrolytic over 5 years old is suspect
[17:03:11] <pfred1> high end equipment regular maintainence is to simply replace at that interval
[17:03:24] <Aero-Tec> they are not that old 3 to 4 years
[17:03:28] <pfred1> OK
[17:03:38] <pfred1> depends on the brand some fail in that time
[17:03:50] <pfred1> they aren't TEAPO are they?
[17:03:55] <Aero-Tec> there are 4 of them
[17:04:06] <Aero-Tec> not sure
[17:04:12] <pfred1> I'm not sure but I think TEAPO means cheap garbage in Chinese
[17:04:22] <Aero-Tec> what is teapo?
[17:04:29] <Aero-Tec> lol
[17:04:31] <pfred1> it is a very lousy brand of capacitors
[17:04:41] <Aero-Tec> yes, cheapo china
[17:04:44] <pfred1> whenever yo usee one just replace it
[17:05:00] <pfred1> they're total crap
[17:05:35] <Aero-Tec> I have scoped the index sig and it looks solid
[17:05:52] <Aero-Tec> so I do not think it is the main filter caps
[17:05:58] <andypugh> So, it looks good on a normal scope, and bad on Halscope?
[17:06:20] <andypugh> That probably tells somebody clever exactly what is wrong. I am not that person.
[17:06:53] <Aero-Tec> and I added a cap to the index supply right at the pull up resistor
[17:07:47] <Aero-Tec> yes, scope is good hal scope bad and drifting sig
[17:07:58] <pfred1> parallel port input?
[17:08:18] <Aero-Tec> will see if I can get the .88v low sig down lower
[17:08:27] <Aero-Tec> yes pport
[17:08:30] <pfred1> yeah that sounds not so good to me
[17:08:39] <Aero-Tec> low end simple bob
[17:08:51] <pfred1> but buffered right?
[17:09:05] <Aero-Tec> just pport con and wire terms
[17:09:10] <Aero-Tec> nothing
[17:09:11] <pfred1> oh
[17:09:38] <Aero-Tec> cheapo cheapo
[17:10:00] <pfred1> I made my own BOB http://www.instructables.com/id/Parallel-Port-Break-Out-Board-BOB/
[17:10:04] <Aero-Tec> got a real nice bob for the mill,
[17:10:11] <PCW> .01 uF cap from parallel port pin to gnd?
[17:10:17] <pfred1> I like to think mine's real nice :)
[17:10:50] <pfred1> the real trick is they're AHCT ICs
[17:11:02] <pfred1> that makes a difference
[17:11:39] <Aero-Tec> thats a new one
[17:12:01] <Aero-Tec> been awhile since I designed and electronics
[17:12:22] <pfred1> me too I've been working on building this CNC machine for too long now
[17:12:59] <pfred1> but I want to see if it'll work
[17:13:21] <pfred1> I almost have the Y axis finished
[17:14:43] <Aero-Tec> so what is a AHCT IC?
[17:14:58] <pfred1> CMOS
[17:15:00] <Aero-Tec> what it's claim to fame?
[17:15:15] <pfred1> super low input current
[17:15:42] <pfred1> not bad output either
[17:15:50] <Aero-Tec> I used CMOS 75XX but they did not have that code in them
[17:16:10] <Aero-Tec> also 40XX
[17:16:22] <Aero-Tec> they were CMOS as well
[17:16:31] <Aero-Tec> but low out put power
[17:16:38] <pfred1> I forget all the good bits about AHCT now but when I looked at them they are nicer than your average IC
[17:16:55] <pfred1> AHCT will output 25ma
[17:17:29] <pfred1> a regular parallel port good for 2-12 or so ma
[17:17:35] <pfred1> it varies
[17:17:48] <Aero-Tec> been about 20 years since I did any designing with ICs
[17:17:49] <pfred1> which is why running right off a port sucks
[17:18:16] <pfred1> the original IBM PC had a spec but after that things got a little fuzzy
[17:18:28] <Aero-Tec> I am having problems with a input not a output
[17:18:45] <pfred1> still nice to buffer signals
[17:19:06] <pfred1> but not essential
[17:19:59] <pfred1> would you consider regulating the supply to your sensor?
[17:20:09] <Aero-Tec> will see about adding a switching sig trans like a 2n2222 to see if that helps
[17:20:16] <pfred1> I mean beyond what it is currently being regulated
[17:20:22] <Aero-Tec> or track down the 7514
[17:20:50] <pfred1> 7414
[17:20:56] <Aero-Tec> explain
[17:21:02] <pfred1> it is a hex schmitt inverter
[17:21:19] <pfred1> so if the inversion really bothers you you can invert it twice
[17:21:33] <pfred1> but if you use one don't foret to tie off the unused inputs
[17:21:42] <pfred1> reduces package noise
[17:22:37] <Aero-Tec> why would one add a reg to a sully that had one already?
[17:22:55] <Aero-Tec> oops
[17:22:57] <Aero-Tec> supply
[17:22:59] <pfred1> well I just make little regulators for circuits anymore that way if I need to tweak the voltage I do it at the source and i don't have to worry about transmitting regulated voltages over distances
[17:23:11] <pfred1> well all regulators work by using feedback
[17:23:27] <pfred1> so if your regulated supply is a distance from where you are using the current
[17:23:36] <pfred1> well you should be able to see the troubles there
[17:24:02] <pfred1> pretty much all regulation is bust after about a foot of line or so
[17:24:12] <Aero-Tec> I am sure the problem is the .88v low
[17:24:28] <pfred1> yes and adjusting your puput voltage should fix that
[17:24:32] <pfred1> input even
[17:24:47] <Aero-Tec> the supply at the encoder is bang on 5V
[17:24:53] <pfred1> hmmm
[17:24:58] <pfred1> well that is good
[17:25:18] <pfred1> your low is still too high though
[17:25:35] <Aero-Tec> the opto must be a darlington
[17:25:40] <Aero-Tec> yes
[17:25:43] <pfred1> it could be
[17:26:01] <pfred1> that might be why you're seeing such nice wave out of it
[17:26:13] <pfred1> because a plain phototransistor has rounded shoulders
[17:26:17] <Aero-Tec> junction voltage as the voltage did not change much from 4.7K pull up to 10K
[17:26:40] <pfred1> I mena it is ugly
[17:27:11] <Aero-Tec> the sig look great
[17:27:15] <Aero-Tec> very good
[17:27:29] <pfred1> yeah and like I said a plain phototransistor really can't do that
[17:28:08] <Aero-Tec> and getting 7+K rpm sitting still with the opto at low means that is where the problem is
[17:28:45] <Aero-Tec> when sig is high, rpm is 0
[17:29:05] <pfred1> your low is being interpreted as a high and a low
[17:29:20] <Aero-Tec> that would be my guess
[17:29:29] <pfred1> yeah you've built an oscillator
[17:29:38] <Aero-Tec> lol
[17:29:46] <pfred1> well you have
[17:29:56] <PCW> you could add a diode in series with the output (stiff pullup to opto --> diode --> weak pulldown)
[17:29:58] <Aero-Tec> looks like
[17:30:11] <pfred1> yeah a diode will drop .7 v
[17:30:22] <pfred1> well most of them do
[17:30:46] <Aero-Tec> where would I put it?
[17:30:50] <pfred1> a glass 4148
[17:31:00] <pfred1> just put it in series
[17:31:07] <pfred1> like it is a resistor
[17:31:11] <Aero-Tec> can not see how adding another .7 drop would help
[17:31:23] <pfred1> but with the polarity correct
[17:31:32] <pfred1> no right on your putput
[17:31:35] <pfred1> output
[17:31:50] <pfred1> that should knock your .88 down to .18
[17:31:57] <pfred1> because of the junction drop
[17:31:58] <Aero-Tec> lol
[17:32:07] <Aero-Tec> ok now I follow
[17:32:09] <Loetmichel> pfred1: more or less. SI-diodes drop around 0,7V. until loaded up to the max
[17:32:12] <pfred1> it will affect your high too though
[17:32:26] <Loetmichel> a 1n4007 will drop around 1,1V at 1A
[17:32:38] <pfred1> some diodes only drop .3 too
[17:32:40] <Aero-Tec> past max and you let out the magic smoke
[17:32:50] <Loetmichel> pfred1: schottky
[17:33:16] <pfred1> yeah they're not your every day things
[17:33:23] <PCW> How about this:
[17:33:24] <PCW> put a stiff pulldown on the parallel port input and the connect your opto to 5V --> input with pulldown
[17:33:50] <PCW> then connect
[17:34:14] <pfred1> Aero-Tec what is your high signal voltage?
[17:34:33] <pfred1> I mean if you have the .7 to lose i say go for it
[17:34:42] <Aero-Tec> 4.6 I think
[17:34:46] <pfred1> you got it
[17:34:48] <Aero-Tec> will have to check
[17:34:51] <pfred1> you only need 3.5
[17:35:29] <PCW> 2.4V is legal TTL high
[17:35:30] <Aero-Tec> sound good, move pull up to pull down and rewire the opto
[17:35:41] <pfred1> yeah go with the diode to drop
[17:35:58] <Aero-Tec> I should not need it with that
[17:36:08] <Aero-Tec> the low should be 0
[17:36:13] <pfred1> when I was making PCBs we had a whole case of 1N4148s
[17:36:19] <Aero-Tec> and the high around 4
[17:36:22] <pfred1> that was literally a million of them
[17:36:31] <PCW> no diode nedded if you reverse wire it
[17:36:55] <Aero-Tec> I have a whack of sig diodes as well
[17:37:02] <pfred1> so the boxx gave me a box of 1,000
[17:37:11] <pfred1> boss even
[17:37:19] <PCW> but you may need a stiff pulldown since the parallel port may have a pullup
[17:37:44] <pfred1> that is what I hate about parallel ports you never know what is in them
[17:38:26] <Tom_itx> not quite parallel?
[17:38:35] <pfred1> oh it is all parallel
[17:38:45] <pfred1> but parallel what
[17:39:15] <Aero-Tec> why would they put in a pull up?
[17:39:21] <pfred1> I read everything I could find on the net about them before I built my BOB and by the time I was done I knew less than when I thought I did when I started
[17:39:36] <Aero-Tec> lol
[17:39:46] <PCW> I think the original IBM port had pullups
[17:39:51] <Aero-Tec> not every thing on the net is correct
[17:40:13] <pfred1> PCW that is the only port you can actually find reliable information about
[17:40:21] <pfred1> after that all bets are off
[17:41:32] <pfred1> Aero-Tec as far as technical electronics information goes the Internet is pretty good
[17:41:57] <pfred1> I wish I had this stuff when I was a kid that's for sure
[17:42:09] <Aero-Tec> lol
[17:42:11] <Aero-Tec> me to
[17:42:25] <Aero-Tec> I have walls covered in spec books
[17:42:26] <pfred1> I spent years buying esoteric books about electronics
[17:42:32] <pfred1> yeah me too
[17:43:21] <pfred1> today I can just put a part number into google and bang! I get the data sheet
[17:44:48] <archivist> sometimes the books are still needed :)
[17:45:09] <pfred1> oh yeah I have some arcane books that the data simply never made it on the net
[17:45:38] <pfred1> stuff is so obsolete no one cares enough today
[17:45:53] <archivist> as I catalogue mine (net visible) I sometimes get requests
[17:46:04] <pfred1> often the Internet treats the time before the Internet almost as if it didn't happen
[17:46:41] <archivist> I have just been scanning the Wireless World computer from 1967
[17:46:50] <pfred1> heh
[17:47:30] <pfred1> does the magazine Popular Electronics even still exist?
[17:48:03] <PCW> There are some odd things about parallel port control outputs as well
[17:48:04] <PCW> These are usually open collector with a 4.7K pullups in standard and PS2 mode
[17:48:06] <PCW> but are push-pull in EPP/ECP mode
[17:48:24] <pfred1> yeah port modes are a gotcha
[17:48:50] <pfred1> what mode does LinuxCNC expect?
[17:49:07] <archivist> I had my cnc working changed pc a week ago and had to add one pullup to get it working
[17:49:07] <pfred1> I'm prettyt sure it expects one I just forget what it is
[17:49:34] <pfred1> archivist what PC did you get?
[17:49:39] <JT-Shop> hard to type with cold fingers...
[17:49:39] <pfred1> I want to get an AMD
[17:49:54] <pfred1> ma nthose things got low latency
[17:50:03] <archivist> just another off the junk pile, and old Compaq
[17:50:14] <pfred1> archivist ah my favorite kind
[17:50:23] <PCW> all PP modes should be OK but if you need output drive from the control signals you need EPP/ECP mode
[17:50:41] <PCW> (high drive)
[17:50:45] <pfred1> archivist my LinuxCNC box I got at a garage sale there was this pile in the garage and I asked the lady what is up with that stuff? she said, oh that's all garbage
[17:51:15] <pfred1> oh really?
[17:51:19] <Aero-Tec> that pullup maybe a problem
[17:51:26] <Aero-Tec> just read up on it
[17:52:22] <pfred1> archivist I have a thing for scrap Gateways
[17:52:33] <pfred1> Gateway used to make an OK PC
[17:52:48] <archivist> not had one of those yet
[17:52:58] <pfred1> well I have 2 they're not bad
[17:53:05] <Aero-Tec> I have a good size stack of old CPU boxes as well
[17:53:31] <pfred1> I wouldn't have paid you for one but as far as prebuilts go they're very clone like
[17:53:36] <Aero-Tec> some are still working
[17:53:39] <archivist> I had a vegetable pc arrive today (celeron)
[17:53:43] <Aero-Tec> 386 and 486
[17:53:44] <Aero-Tec> lol
[17:53:56] <Aero-Tec> I think I have a old 286 as well
[17:54:11] <Loetmichel> Aero-Tec: sounds familliar
[17:54:14] <pfred1> I have a Z-80 called a Northstar I think
[17:54:31] <pfred1> dual 5.25 floppy drives!
[17:54:37] <Aero-Tec> I have the Z80 ICs
[17:54:39] <Loetmichel> have my linuxcncPC in an old 286 pizzabox case
[17:54:50] <pfred1> I have those too but this is a whole machine
[17:54:56] <Aero-Tec> 8088 and a few other oldies
[17:55:09] <pfred1> I think i have a 4004 kicking around someplace
[17:55:20] <Loetmichel> my oldest is a 80286 board
[17:55:21] <pfred1> they're worth big bucks
[17:55:40] <Aero-Tec> what about the 6502?
[17:55:47] <Loetmichel> and my oldest running is a 486-33 notebook ;-)
[17:55:49] <Aero-Tec> I have a few of them as well
[17:55:50] <pfred1> ah not so much
[17:56:05] <pfred1> NASA might be interested
[17:56:16] <pfred1> they buy obsolete parts off ebay
[17:56:21] <Loetmichel> beside the varoius C64/Sinclair spectrim/etc in the cellar
[17:56:34] <pfred1> I still have my SInclair
[17:56:38] <Aero-Tec> have a old Sinclair single board computer with hex keypad
[17:56:49] <andypugh> We had a guy at work keeping out PDP-11s running with eBay parts until the year before last.
[17:57:04] <pfred1> I modded mine to take a keyboard
[17:57:26] <andypugh> Sinclair with a Hex keypad? What was that then?
[17:57:28] <pfred1> andypugh I scrapped a PDP 11/34 for parts i couldn't keep it around
[17:57:28] <Loetmichel> andypugh: pdp11 need a "running keeper"? since when?
[17:57:39] <archivist> andypugh, MK14
[17:57:40] <Aero-Tec> not sure if I still got it, been awhile since I last saw it
[17:57:42] <pfred1> I mean it needed a whole room dedicated to it
[17:58:19] <pfred1> I've scrapped 2 minis for parts the PDP and some NEC monster that was built into a desk
[17:58:31] <pfred1> the desk was the PC
[17:58:57] <Loetmichel> until a few years ago my sinclair zx81 was still running with a new rom as the cpu for a oil central heating of a friend
[17:59:11] <andypugh> It seems odd that I had never heard of the Mk14. I am aware of the calculators, watches and tiny radios
[17:59:33] <pfred1> andypugh you like old calculators?
[17:59:43] <Loetmichel> then the furnace water pocke had rusted through
[17:59:53] <Loetmichel> pocket
[17:59:56] <andypugh> Not really. I use one, though, as I bought it as a schoolboy and it still works.
[18:00:27] <pfred1> I picked up a TI-30 at a garage sale last summer
[18:00:46] <pfred1> it is from 1974
[18:01:35] <Loetmichel> pfred1: i have a Ti-30 in my toolcase. its the old one i used in school ;-) (back in the days) ;-)
[18:01:37] <archivist> typing moonlander into the MK14 was a pain....
[18:02:32] <Loetmichel> the one with the red led display
[18:02:42] <andypugh> He didn't learn from the C5 did he? http://www.sinclairzx.com/spec-x-1.html
[18:03:14] <Loetmichel> andypugh: c5 was to low, to dangerous
[18:03:15] <archivist> he has never learnt from his failures
[18:03:31] <archivist> and he has had a few
[18:03:35] <Loetmichel> this one looks better. and is not to early for its time ;-)
[18:04:01] <pfred1> I put up an article about my Timex hack http://www.instructables.com/id/Sinclair-Surplus-Keyboard-Graft/
[18:04:11] <andypugh> One thing I didn't realise was his is the Seadoo diver tow things
[18:04:14] <pfred1> man I was thrilled when it worked
[18:05:41] <Loetmichel> "Reserve for £100
[18:05:42] <Loetmichel> (delivery estimated July 2011)"
[18:05:51] <andypugh> I spent ages grafting a full-size keyboard to my Spectrum after the ribbon-cables died. Then I found out it was a capacitative keyboard, and what I thought were contacts were insulated aluminium discs.
[18:05:56] <Loetmichel> looks like he is gone off the market already
[18:06:20] <pfred1> andypugh mine worked but it really played havoc with TV reception
[18:06:33] <Loetmichel> andypugh: hmm?
[18:07:00] <Loetmichel> the spectrum 48k had a normal foil keyboard under the rubber IIRC
[18:07:15] <Loetmichel> like the "modern" pc keyboards
[18:07:21] <Loetmichel> +cheap
[18:07:36] <mevon> :( what is "cannot unhome while moving, joint 0" error?
[18:07:51] <pfred1> the genius of SInclair is he made the whole system in a PROM
[18:07:58] <Loetmichel> mevon: axis x is still runnning!
[18:08:22] <mevon> :( no unfortunatly it is not moving :(
[18:08:36] <Loetmichel> the system thinks it is.
[18:08:45] <mevon> hummm
[18:08:47] <andypugh> On the bottom of each of the keys of my donor keyboard was a foam "spring" and on the bottom of that was a disc of aluminium. There was a matrix of tracks on the PCB. I soldered wires to that PCB, and then to the Sinclair motherboard. Then found that the "contacts" were some sort of capacitor arrangement.
[18:08:59] <andypugh> mevon: Is the DRO changing?
[18:09:08] <Loetmichel> andypugh: ah, the DONOR keyboard was capacitive. i see
[18:09:25] <Loetmichel> that was a missunderstanding
[18:09:28] <Loetmichel> -s
[18:09:31] <pfred1> I got my keyboard at radio shack
[18:09:37] <mevon> what is the DRO? sry ...
[18:09:47] <pfred1> it was bare guts I think it cost $3.95
[18:09:50] <JT-Shop> Digital Read Out
[18:10:06] <JT-Shop> mevon: what are you doing when you get this error?
[18:10:38] <mevon> after loading the machine, on my first jog
[18:11:16] <mevon> kind of stop right after beginning the move then it trips the e-stop
[18:11:41] <mevon> when i try to f1+f2 to power it up again, it give that answer
[18:12:01] <JT-Shop> http://www.youtube.com/watch?v=ZOJlAcPbmCM
[18:12:32] <mevon> I get a "joint 0 following error" just before that
[18:12:41] <andypugh> mevon: Is the Arduino still sending pulses?
[18:12:44] <JT-Shop> stepper?
[18:12:44] <mevon> I read that theres deboucing that could help
[18:13:05] <Loetmichel> mevon: did the machnie EVER run with this setup?
[18:13:08] <andypugh> This would be so much easier if you had chosen known-good hardware
[18:13:08] <abetusk> I believe I am having problems with my X accuracy. You can see there is variable width on the vertical lines (but the horizontal lines look fine) here: http://imgur.com/LBEHT
[18:13:14] <abetusk> any suggestions on how to debug this?
[18:13:37] <Loetmichel> sounds to me like a endswicht/limit switch set to wrong polarity or open
[18:13:55] <mevon> Loetmichel, it is running if I send the command on serial link
[18:14:27] <JT-Shop> abetusk: that looks like a mechanical issue like back lash
[18:14:55] <andypugh> Or possibly flex.
[18:15:05] <abetusk> JT-Shop, It's a lead screw system on two linear rails with an anti backlash nut installed
[18:15:16] <JT-Shop> did you test it?
[18:15:27] <mevon> andypugh, Im gonna try a longer distance to move see if it stops at the same moment but I'm guessing the panic/estop state is cutting out the power
[18:15:43] <abetusk> JT-Shop, test it how?
[18:16:08] <andypugh> mevon: f-error is because the arduino is not reporting back position.
[18:16:36] <mevon> oh
[18:16:45] <JT-Shop> put an indicator on the axis and move in one direction till the indicator moves then make small moves in the other direction till the indicator starts to move
[18:16:53] <andypugh> Do you know for a fact that anyone has used this system? Everything I have ever believed about LinuxCNC indicates that a buffered USB device won't work.
[18:17:09] <abetusk> JT-Shop, ok, let me try
[18:17:28] <pfred1> I always thought LinuxCNC was anti-USB
[18:17:33] <mevon> andypugh, I know
[18:17:41] <mevon> pfred1, youre right
[18:17:43] <JT-Shop> I use USB with LinuxCNC
[18:17:58] <pfred1> JT-Shop for your keyboard and mouse?
[18:18:07] <andypugh> So do I, but not to control the steppers
[18:18:08] <JT-Shop> joypad
[18:18:14] <pfred1> OK that too
[18:18:49] <pfred1> PS/2 forever!
[18:18:58] <mevon> but parallel ports don not report their position on a step/dir
[18:19:20] <pfred1> that is my big quest anymore PS/2 optical wheel mice are rarer than hens teeth anymore
[18:19:44] <mevon> im hijacking this "net xpos-cmd axis.0.motor-pos-cmd => arduino.axis0-cmd"
[18:20:27] <andypugh> mevon: No, but the stepgen keeps track of the pulses it has made, and reports that back as feedback. You don't know that the motor has moved, but you know the correct number of pulses have gone.
[18:20:46] <andypugh> What do you have connected to axis.0.motor-pos-fb ?
[18:21:06] <mevon> its a python script
[18:21:12] <mevon> fb^
[18:21:15] <mevon> ?
[18:21:17] <mevon> humm
[18:21:21] <mevon> ill check
[18:22:10] <mevon> i thought since im running off the sim .hal it would be wired by default
[18:23:36] <mevon> I have Xpos
[18:24:16] <mevon> afk brb maybe we can discuss this another time thks anyways
[18:24:40] <mevon> when wife makes dinner i better eat...
[18:26:15] <andypugh> I wonder how it is possible to get an f-error with xpos wired directly back to the fb?
[18:27:36] <pfred1> forget USB for stepper control
[18:30:12] <JT-Shop> http://www.youtube.com/watch?v=soCh9U2TU48&feature=related
[18:30:40] <JT-Shop> seems the tool is always below center
[18:39:01] <mevon> pfred1, what you just invested in old crippled PC that can barely handle ubuntu 8.04 ?
[18:39:24] <pfred1> mevon I don't run Ubuntu
[18:43:14] <mevon> in this forum post http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-questions/21127-getting-cannot-unhome-while-moving-errors-withou?start=12 "setp debounce.0.delay 25" is debouncing what actually?
[18:44:38] <mevon> I can put code in the arduino to send feedback, it knows where it is at all times, but what needs debouncing in the precedent example?
[18:44:44] <pfred1> all switch contacts bounce
[18:45:17] <pfred1> digital logic circuits see it as thousands of contact openings and closings
[18:45:22] <andypugh> JT-Shop: Did you see this tool? http://www.youtube.com/watch?v=IkFdJwW_0GI&feature=share&list=UUV_oqll5MjoOlQb-FJc8W5g
[18:45:53] <mevon> pfred1, on what signals?
[18:46:19] <pfred1> mevon any switch on any signal
[18:46:31] <pfred1> all mechanical switches have bounce
[18:46:35] <andypugh> mevon: I see no need for any debounce with your system
[18:46:39] <mevon> all switches are arduino strings sent over usb
[18:46:53] <pfred1> well those aren't mechanical switches
[18:47:45] <pfred1> that doesn't mean an arduino cannot transmit signal bounce though
[18:47:47] <mevon> I believe JRhode in the link had no mechanical switches neither
[18:48:03] <pfred1> because it can
[18:48:27] <mevon> pfred1, i wrote the code i know it doesnt send rogue random string on the usb serial com
[18:48:54] <pfred1> mevon so you've debounced your inputs with software?
[18:49:27] <mevon> i wonder if the simulation without my custom.hal reacts the same way
[18:50:12] <mevon> pfred1, since its usb com, I only send new state of switches when it changes
[18:50:38] <pfred1> mevon switch bounce is by nature a state change
[18:50:42] <andypugh> If you can write Arduino code then you ought to be able to write a 1-step HAL to do what you want.
[18:51:00] <pfred1> off and on high and low
[18:51:03] <mevon> I know arduino, idk linuxcnc hal
[18:51:18] <andypugh> There are only 3 commands, how hard can it be?
[18:51:35] <pfred1> andypugh binary is only 2 states and that gets complicated
[18:52:26] <mevon> sry to in a situation where everyone dropped the ball
[18:52:50] <mevon> didnt want to make you irritated
[18:53:46] <andypugh> I am not irritated tonight. I might have been last night :-)
[18:53:54] <mevon> hal is not complicated, its all the sub module and inner workings that I dont understand
[18:54:04] <mevon> or the week before :P
[18:54:26] <andypugh> You don't need them. All you need is all the stuff that is in your two HAL files put into one rational HAL file.
[18:55:29] <pfred1> all you need is HAL
[18:55:46] <pfred1> mevon andypugh is the 5th Beatle
[18:56:15] <mevon> I bet your a big star too :P
[18:56:27] <pfred1> nah i wanted to be but I have a tin ear
[18:59:54] <mevon> andypugh, you think the links made and remade with the simulation machine is what somehow creates a need for feedback and since i do not update or know how to update Xpos, it doesnt work
[19:00:42] <andypugh> Xpos is a signal. I assume that is connected to axis.0.motor-pos-cmd?
[19:01:07] <andypugh> But if it is also connected to motor-pos-fb an f-error should be impossible,
[19:01:19] <andypugh> do you still have stepgens being loaded?
[19:03:46] <mevon> andypugh, its not connected: axis.0.motor-pos-cmd ==> xpos-cmd
[19:03:56] <andypugh> Ah
[19:04:14] <andypugh> That would do it
[19:04:55] <mevon> then I have arduino.axis0-cmd <== xpos-cmd
[19:05:01] <andypugh> Every net command has a signal immediately after "net" That can be called anything you like.
[19:05:50] <andypugh> You can then have zero or more "pins" but only one pin connected to any signal can be a "writer" or output pin.
[19:05:57] <mevon> so the feedback is not wired to the right signal
[19:06:04] <andypugh> The signal can appear in as many net commands as you want
[19:06:46] <andypugh> arduino.axis0-cmd <== xpos-cmd isn't right. It needs to have the signal first.
[19:07:19] <ynos_> Hello all
[19:07:25] <mevon> somewhere i should have axis.0.motor-pos-cmd-fb <== xpos-cmd
[19:07:28] <mevon> ?
[19:07:50] <mevon> where is Xpos from?
[19:07:52] <andypugh> Yes, but no. The signal name is the forst thing after the net command.
[19:08:01] <mevon> ok
[19:08:18] <andypugh> xpos is just a signal. Possibly an unused one
[19:08:39] <andypugh> As I keep saying, you need to rationalise and sort out your HAL file.
[19:09:03] <pfred1> KISS
[19:09:21] <pfred1> Keep It Simple Silly
[19:09:26] <mevon> :P
[19:11:15] <ynos_> Hey, i just came into a haas TM1 and was wondering if LinuxCNC can copy gcode over the serial port via xmodem? i am looking for something to use as a toolpath designer and someone said i should give this a look.
[19:12:16] <mevon> linuxcnc interprets gcode
[19:12:23] <ynos_> it is building right now, but thought i'd come pester y'all for a few and re-familiarize myself with the irc undebelly
[19:12:34] <ynos_> ahh, so it doesnt generate gcode?
[19:12:39] <pfred1> ynos_ there are a number of terminal applications in Linux I'm not sure if xmodem is the best
[19:12:51] <mevon> it can i believe from DXF
[19:15:31] <ynos_> hmm.
[19:15:47] <ynos_> wow.. how many GUIs are there..
[19:16:01] <andypugh> It's possible that parts of LinuxCNC could send G-code over serial, but there are probably far simpler ways.
[19:16:29] <andypugh> LinuxCNC normally reads G-code and controls the machine.
[19:16:33] <pfred1> I thought he just wanted to process then send data over the serial line
[19:16:56] <ynos_> my options right now are manual entry(ugh), floppy drive(would have to install one on a machine somewhere) and serial.
[19:17:07] <pfred1> serial is your best bet
[19:17:11] <ynos_> yup
[19:17:27] <andypugh> pfred1: Yes, quite. You could use Axis for the preview, then get rid of just about everything downstream.
[19:17:38] <pfred1> yeah LInux can control serial lines like nobody's business
[19:17:50] <pfred1> best in class
[19:18:15] <ynos_> i copied a little prog down using hyperterm today.
[19:18:24] <ynos_> was stoopid easy
[19:18:37] <pfred1> was a long time ago I messed around with any of that though back in the dial up days
[19:18:40] <ynos_> i am just a bit intimidated by creating toolpaths
[19:20:07] <andypugh> There are some toolpath generation packages (it's generally called CAM )
[19:20:08] <pfred1> I kind of recall a program called minicom though
[19:20:25] <pfred1> it might do what you want
[19:20:55] <ynos_> yup, the only reason i used hyperterm is the only laptop with serial i had with me at the time was running xp
[19:20:58] <andypugh> I would think you could just copy the file to the LPT device?
[19:21:25] <mevon> on linux you can use screen command
[19:21:34] <mevon> in terminal
[19:21:40] <ynos_> cli isnt an issue, would just be badass if it could be easy enough my grandpa(old machine shop guy) could come over and not have to use cli.
[19:22:11] <pfred1> minicom is pretty clunky
[19:23:11] <ynos_> looks like a simple script tied in with 'sx' would work.
[19:24:11] <ynos_> sx and cu
[19:24:48] <kwallace> Does tool orientation, front and back angle in the lathe tool table do anything, or is it just to improve the tool graphic in AXIS?
[19:25:27] <andypugh> Just preview AFAIK. No gouging detection or anything.
[19:29:08] <kwallace> andypugh: That's what I'm thinking also. I'm wondering if orientation could be handy to detect whether to auto-create g-code on the near or far side of the workpiece.
[19:29:24] <ynos_> hmm.. am i missing something here? i run linuxcnc and it comes up with a configurator.. then nothing, should it be starting something after that?
[19:29:30] <andypugh> Possibly, yes.
[19:30:26] <andypugh> ynos_: Ignore the configurator initially, start LinuxCNC and then choose one of the "sim" demo configs
[19:30:33] <ynos_> ahh, was the configurator setting something other than coms?
[19:30:33] <mevon> andypugh, how can I load AXIS in the first place, seems if I take my files from stepconf wiz, everything from axis is missing
[19:30:48] <ynos_> i was using the sim/axis demo config
[19:31:11] <ynos_> there we go
[19:31:12] <andypugh> ynos_: Then you should get a GUI and stuff.
[19:31:30] <ynos_> i had to run it 3 times.. then it worked.
[19:31:33] <andypugh> mevon: I don't even understand the question
[19:31:40] <ynos_> are there any services it relies on?
[19:32:03] <andypugh> ynos_: Well, it needs the right kernel. How did you install?
[19:32:52] <mevon> somehow HAL is not loading any AXIS components
[19:32:52] <ynos_> used this - http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Build_A_Simulator_Manually
[19:33:23] <andypugh> Ah, OK, that ought to work (at least a bit)
[19:33:45] <andypugh> mevon: What do you mean by AXIS components?
[19:34:08] <mevon> i get this error at startup "custom.hal:85: Pin 'axis.0.neg-lim-sw-in' does not exist"
[19:34:13] <andypugh> Do you mean AXIS (the GUI) or "axis" the motion pins?
[19:34:28] <pfred1> that is confusing
[19:34:56] <andypugh> Why do you persist on messing about with a custom.hal?
[19:35:13] <mevon> why not
[19:35:19] <pfred1> some moths are just drawn to the flames
[19:35:27] <mevon> if i want to learn i need something to learn on
[19:35:42] <mevon> how have you leanr HAL?
[19:35:51] <andypugh> I am going to say this one more time and go to bed. Go through the HAL by hand making one, consistent, functional HAL file
[19:35:55] <pfred1> I ask andy :)
[19:36:06] <mevon> andypugh, where can I learn to do that
[19:36:35] <mevon> cmon stop being negative
[19:37:12] <andypugh> The HAL manual would be a reasonable start. http://www.linuxcnc.org/docs/html/hal/basic_hal.html and http://www.linuxcnc.org/docs/html/hal/tutorial.html
[19:37:53] <mevon> ok so where are listed all the components and links I need to make a machine?
[19:38:29] <mevon> I dont think anyone is born knowing these things
[19:38:34] <kwallace> At one time there was a HAL tutorial that started at the command line and had one load a thread, load a component, attach function to the thread and so on. It may still be around. ... looking.
[19:39:24] <andypugh> All the components and links you need should be in the two HAL files. I am just saying that you need to combine them and delete the parallel port cruft.
[19:39:47] <mevon> Im trying to do just that now
[19:40:23] <mevon> thing is i get error about things i dont even know of
[19:40:38] <mevon> or even knew they existed
[19:40:54] <pfred1> google is your friend
[19:41:16] <mevon> and sry but I come here when my friend fails me
[19:43:51] <kwallace> http://linuxcnc.org/docs/LinuxCNC_Integrator_Manual.pdf and http://linuxcnc.org/docs/HAL_User_Manual.pdf
[19:44:13] <mevon> kwallace, thks really appreciated
[19:46:04] <pfred1> andy is still the god of HAL
[19:46:14] <kwallace> the "axis.0 " message says it can't find an axis.0. It may for non obvious reasons because the axis.x stuff is created, guessing off hand, by "loadrt motion"?
[19:47:54] <kwallace> I haven't had to mess with this for a while. The [AXIS_x] sections in the .ini also help create the axis.x's.
[19:50:14] <mevon> ok
[19:50:48] <pfred1> grep -n axis.x *.ini
[19:51:00] <mevon> so its from a motion component?
[19:51:30] <kwallace> Oops I think motion and axis functions are loaded by "loadrt [EMCMOT]EMCMOT base_period_ns..." then the .ini file is checked for [AXIS_x] entries.
[19:52:01] <pfred1> then grep -in
[19:53:21] <mevon> ok kwallace i see that in core_sim.hal too
[19:53:39] <mevon> got some ddt hypot comp and or2
[19:53:50] <mevon> bet those are for calculating
[19:54:52] <mevon> ok so I was using servos
[19:54:56] <mevon> this is bad
[19:55:01] <mevon> i have steppers
[19:55:13] <mevon> they dont give feedback
[19:55:51] <pfred1> there is that
[19:55:56] <kwallace> I'm not sure it's bad. The motion setup is the same for servos and steppers.
[19:56:34] <kwallace> In that with steppers the feedback is linked back the the feed back input to motion.
[19:57:38] <kwallace> With servos the feedback loop goes through encoders/encoder component first.
[20:00:10] <mevon> ok so if I have no good feedback signal created, its normal that I get errors
[20:00:38] <mevon> loadrt stepgen step_type=0,0,0
[20:00:51] <mevon> creates 3 stepmotors?
[20:01:14] <kwallace> In the core stepper.hal :
[20:01:17] <kwallace> # connect position feedback from step generators
[20:01:17] <kwallace> 26 # to motion module
[20:01:17] <kwallace> 27 net Xpos-fb stepgen.0.position-fb => axis.0.motor-pos-fb
[20:03:39] <mevon> hummm
[20:03:40] <kwallace> loadrt stepgen creates three step generators. These need to get input from motion and output goes to the motors and also to position feedback.
[20:03:54] <mevon> ok
[20:04:32] <kwallace> The more I recall the trickier it gets, it may take a while to get a hang of it.
[20:09:55] <kwallace> I'm looking at this:
[20:09:58] <kwallace> http://git.linuxcnc.org/gitweb?p=linuxcnc.git;a=blob;f=configs/common/core_stepper.hal
[20:13:26] <kwallace> Two loadrt's are used to load motion and the step generators. Then the functions in the components are connected to threads using addf's.
[20:13:47] <kwallace> See at the bottom what the functions are:
[20:13:53] <kwallace> http://www.linuxcnc.org/docs/2.5/html/man/man9/motion.9.html
[20:16:02] <kwallace> Also here:
[20:16:05] <kwallace> http://www.linuxcnc.org/docs/2.5/html/man/man9/axis.9.html
[20:18:38] <mevon> my trouble is, Im doing this
[20:18:57] <mevon> net Xpos axis.0.motor-pos-cmd => axis.0.motor-pos-fb ddt.0.in
[20:19:01] <mevon> then i do that
[20:19:17] <mevon> net xpos-cmd axis.0.motor-pos-cmd => arduino.axis0-cmd
[20:19:31] <mevon> theres an unlink between
[20:19:48] <mevon> #unlinkp axis.0.motor-pos-cmd
[20:19:53] <mevon> but i commented it out
[20:20:23] <mevon> I should end up with
[20:20:38] <mevon> net Xpos axis.0.motor-pos-cmd => axis.0.motor-pos-fb ddt.0.in arduino.axis0-cmd
[20:20:39] <mevon> ?
[20:25:01] <kwallace> I haven't been following what you are doing with the Arduino.
[20:28:56] <mevon> i will not tell you sry because if i do you will stop helping me :P
[20:29:21] <mevon> specially if I say Im making an usb cnc controller
[20:29:30] <mevon> with an arduino
[20:30:45] <kwallace> I looked at the core-sim.hal,
[20:30:50] <kwallace> 28 # create HAL signals for position commands from motion module
[20:30:50] <kwallace> 29 # loop position commands back to motion module feedback
[20:30:51] <kwallace> 30 net Xpos axis.0.motor-pos-cmd => axis.0.motor-pos-fb ddt.0.in
[20:30:51] <kwallace> 31 net Ypos axis.1.motor-pos-cmd => axis.1.motor-pos-fb ddt.2.in
[20:31:37] <mevon> so if I have a custom.hal in my .ini and I have this line
[20:31:41] <kwallace> So it looks like you are tapping the Arduino to an existing signal.
[20:31:51] <mevon> net xpos-cmd axis.0.motor-pos-cmd => arduino.axis0-cmd
[20:32:22] <mevon> think it messes up my axis.0.motor-pos-cmd
[20:32:37] <mevon> and feedback is lost
[20:33:14] <kwallace> Do you have "xpos" and "xpos-cmd" ?
[20:33:54] <mevon> yes i just did blindly what was told to do
[20:35:42] <mevon> and also because the example is with stepgen not servo-thread
[20:36:19] <mevon> but Im guessing itd make more sens with servos
[20:36:50] <mevon> since Im not really stepping but sending a position command
[20:37:20] <mevon> im not with step/dir drivers
[20:37:38] <mevon> I have l298 hbridge
[20:37:50] <mevon> directly on the arduino
[20:40:25] <mevon> guess id have to send feedback when pos is reached or something like that
[20:41:26] <kwallace> The command doesn't care if it is commanding a stepper or servo at this point.
[20:43:56] <kwallace> I would tend to just try to send the command to the Arduino and either also connect the feedback to the axis feedback or have the Arduino send feedback. But I supposed that is what you have tried to do.
[20:49:13] <kwallace> Can you get a stock sim or stepper configuration running? Then just add the Arduino command input pin to the end of the existing xpos signal line?
[20:52:42] <mevon> kwallace, exactly what Im working on
[20:53:10] <mevon> I started off the EMC2Arduino project
[20:53:26] <mevon> mod it to add my own stepper drivers
[20:53:38] <mevon> now working off the switches
[20:54:01] <ynos_> thanks for your help earlier, i am off to bed! i will likely be in here again some other time.
[20:54:22] <mevon> in the project, its said to stepconf wiz an ordinary parallel port config with your settings
[20:54:30] <mevon> then add the custom.hal et voila
[20:54:33] <mevon> but not really
[20:54:49] <mevon> ynos_ np gn
[20:55:06] <mevon> too late lols
[20:56:05] <mevon> but then the custom.hal doesnt work with the configs generated by stepconf
[20:56:28] <mevon> seems no motion module is created by default
[20:56:52] <mevon> but then again maybe Im wrong about that
[20:57:24] <mevon> maybe because I didnt unlink before
[20:57:35] <mevon> as I commented it out
[20:58:27] <mevon> but to me seems more logic getting the simulation
[20:58:45] <mevon> then modify to work with arduino hijacks
[20:58:52] <mevon> I mean HAL
[20:59:51] <mevon> Im still wondering what are things like servo-thread and CYCLE_TIME
[21:00:45] <kwallace> I don't know anything about the EMC2Arduino project, but adding a custom.hal doesn't seem to be a problem if you know how it fits in.
[21:03:10] <mevon> need to read more on the net command
[21:03:39] <kwallace> Everything starts with the .ini file, it has a section to call .hal files. If you want to add custom.hal just add the call out to the .ini.
[21:05:30] <kwallace> As far as threads go, thy are the clock ticks that cycles the functions. The functions may be loaded, but fi thy are not addf'd to a servo or base thread they won't get run.
[21:07:00] <kwallace> net ties a single signal name to one output pin and any input pins.
[21:07:06] <kwallace> See, simple.
[21:07:53] <mevon> lol really thks
[21:08:00] <mevon> need info on unlinkp
[21:08:21] <mevon> its unlinks every net containing the signal?
[21:09:03] <kwallace> I think it only unlinks a single pin.
[21:10:05] <mevon> so every other pins formerly connected to it remains the same?
[21:13:12] <mevon> ill figure that out
[21:14:00] <kwallace> I don't know for sure. I see something about "Unlinkp <pin name 1> <pin name 2> " by searching the wiki.
[21:14:19] <kwallace> I haven't used it.
[21:19:25] <kwallace> I'd pastebin your .ini, .hal and any error messages and link it to a message on the e-mail list. This way we can see the context the problem lives in.
[21:19:36] <kwallace> Gotta go eat.
[21:19:41] <mevon> ok thks man
[21:19:46] <mevon> np take it easy
[21:20:05] <kwallace> Good Luck
[21:20:07] <mevon> ill temper the isht out of it
[21:20:21] <mevon> :P
[22:00:49] <r00t4rd3d> http://i.imgur.com/6ZEbv.jpg
[22:03:02] <mevon> gn everyone thanksfor the help
[22:26:29] <abetusk> JT-Shop, still around?
[22:31:09] <r00t4rd3d> i highly doubt it
[22:32:28] <r00t4rd3d> he is normally afk by 8/9
[22:32:42] <abetusk> so I've put a dial indicator on the spindle bit and have set it to go back and forth between x0 and x-0.254 and I'm noticing a "drift"
[22:33:52] <abetusk> it goes from 0 to 10-11mils (which is already bad if it goes beyond 10) but then it starts drifting from the 0 mark. I've been running it for a minute and it's already 1.5 mils further away from 0 then when it started
[22:33:58] <jdh> https://picasaweb.google.com/m/viewer?dc=gorganic&source=mobileproducts#photo/112430417093824344570/5754485947799001697/5810889733017829058
[22:34:02] <jdh> craigslist $50
[22:34:17] <r00t4rd3d> nothing pictured
[22:34:30] <r00t4rd3d> Scrapbook Photos (0)
[22:34:43] <abetusk> what's up with that jdh?
[22:35:38] <jdh> https://picasaweb.google.com/m/zoom?dc=gorganic&source=mobileproducts&uname=112430417093824344570&aid=5754485947799001697&id=5810889733017829058&viewportWidth=320&viewportHeight=416
[22:35:47] <Tom_itx> abetusk, are you loosing steps?
[22:35:50] <jdh> how about that?
[22:36:15] <abetusk> jdh, did you do some exif snooping?
[22:36:20] <jdh> weird tablet cut paste
[22:36:32] <abetusk> Tom_itx, I'm going slow, but otherwise I'm just not sure
[22:36:42] <jdh> no. I bought them.
[22:36:44] <Tom_itx> scale set right?
[22:38:18] <Tom_itx> it should at least return to the start point even if it doesn't make it to the destination. the difference in the destination value is your backlash compensation value
[22:38:38] <abetusk> Tom_itx, so you suspect I'm loosing steps?
[22:38:52] <Tom_itx> no idea
[22:39:00] <Tom_itx> maybe something is loose
[22:39:22] <Tom_itx> make sure the axis is tight when you set the zero point
[22:39:51] <Tom_itx> ie, move in a positive direction then set it then move away from it in a negative direction and take the measurement
[22:40:55] <Tom_itx> if it still drifts, you are loosing steps or something is amiss in the axis setup
[22:41:12] <Tom_itx> pulley driven, it could be a loose pulley
[22:49:15] <abetusk> it's lead screw
[22:49:57] <Tom_itx> couplers good and tight?
[22:51:37] <abetusk> I just retightened them...
[22:52:12] <abetusk> I'm also getting drift in the y direction. It's not as bad, only .5 mil and it seems to be holding steady, but still...
[22:52:14] <Tom_itx> dial indicator secured tight?
[22:52:47] <abetusk> I have two metal vises on them...
[22:52:56] <abetusk> tried to make it tight as I could
[22:53:08] <Tom_itx> tied to the spindle?
[22:53:23] <Tom_itx> is the spindle turning slightly when you measure it?
[22:54:02] <abetusk> the dial indicator is tied to the table
[22:54:13] <abetusk> The spindle does not appear to be moving
[22:54:45] <abetusk> arg...maybe it is the dial indicator not secured tightly enough. I just pushed it a little and it threw it out of whack
[22:54:55] <Tom_itx> :)
[22:55:07] <Tom_itx> doesn't take much
[22:55:23] <abetusk> how "should" one secure the dial indicator to test?
[22:55:57] <Tom_itx> i put my last word in the spindle when i did mine
[22:56:10] <Tom_itx> used a block secured in a vise to butt up against it
[22:56:23] <Tom_itx> backed off from it and told it to return then measured the difference