Back
[00:47:32] <RyanS> archivist were you able to see that second link I sent you for the image of the motor?
[00:48:31] <RyanS> http://s210.photobucket.com/user/Mingzhus/media/drillpressjunctionbox_zps5b116975.jpg.html
[01:25:28] <archivist> photobucket is terminally broken
[01:25:54] <RyanS> bugger
[01:26:17] <archivist> use imagebin or something
[01:27:08] <RyanS> ok
[01:28:13] <RyanS> http://imagebin.org/280077
[01:28:26] <RyanS> Much easier
[01:29:20] <RyanS> I'm assuming the left column goes to the motor
[01:30:10] <archivist> yes, only 3 connections, so no rewiring to 220 without surgery
[01:30:28] <RyanS> fuck
[01:30:39] <RyanS> How many terminals would be easy?
[01:31:01] <archivist> can be just as easy to get a second hand motor
[01:31:28] <archivist> there should be 6 wires going to the motor coils for easy
[01:32:40] <archivist> I am used to taking the motor apart and finding the star point, cutting and rewiring
[01:32:48] <RyanS> right, I was quoted $500 .... If it's the easy way. How much work is involved the hard way?
[01:33:42] <RyanS> $1000 hard or $600 hard if you get my drift
[01:34:13] <archivist> to me $500 is far too expensive
[01:34:54] <archivist> go to a local motor rewind shop and ask
[01:35:11] <RyanS> a VFD, having to test and research it myself would cost 300...
[01:36:28] <archivist> finding another motor can cost less
[01:36:44] <RyanS> Yeh, this is a motor joint "specialising in three-phase conversion"
[01:37:36] <RyanS> Do you have a link on 220 surgery, or what exactly should I google?
[01:37:56] <RyanS> I could get my electrical engineering friend to do it
[01:39:21] <RyanS> http://www.ebay.com.au/itm/3-Phase-Motor-/111158942497
[01:40:32] <archivist> you need to see the delta sign and 220 on the rating plate as well
[01:41:30] <uw> you should probably just plug whatever wire into wherever
[01:41:45] <archivist> and note the mounting yours is likely foot mount not end as that ebay one is
[01:42:46] <RyanS> Actually isn't changing the motor a pain involving machining the pulley
[01:43:23] <archivist> not is diameters are the same
[01:43:29] <archivist> is/if
[01:49:34] <archivist> this is the surgery
http://madmodder.net/index.php?topic=7773.0
[01:50:12] <RyanS> hmmm if I did it myself, with a new motor.. Providing one follows the VFD manual is easy to mess up or do they generally default to safe settings ?
[01:52:28] <archivist> the right motor is trivial, the wrong one is like that link
[01:55:35] <RyanS> Brand-new isn't expensive
http://www.ebay.com.au/itm/Electrical-motor-Three-phase-415v-1-5kw-2-0HP-2800rpm-shaft-24mm-/171065144243?pt=AU_B_I_Electrical_Test_Equipment&hash=item27d446f3b3 will have a look at the manufacturer to see the badge
[01:55:42] <archivist> a six terminal box like
http://www.diynot.com/forums/electrics/understanding-motor-name-plates.317415/ has straps to be moved
[01:57:15] <RyanS> i see, on the second image you can rearrange the straps
[01:57:54] <RyanS> yeah rewiring the motor inside is out of my league
[01:58:24] <archivist> the label on the motor on that last ebay shows both voltages so will have straps, the text does not say also its 2800 not 1400 rpm
[01:59:54] <RyanS> No idea what my current motors RPM is, doesn't have a bloody badge
[02:01:00] <archivist> 1400/1450 is more common
[02:01:57] <RyanS> It . Shouldn't be that important? Depending on how you set the belt drive
[02:04:12] <archivist> it would mean you setting a slower speed and also often using the vfd on slower settings
[02:04:42] <RyanS> ahh yeah , easier to overheat
[02:04:53] <archivist> some motors can over heat if used for too long when run well under rated speed
[02:05:39] <archivist> that seller has 100% just ask them for exactly what you need
[02:06:24] <RyanS> but modern motors generally take account of the fact that people run them on VFD?
[02:07:34] <archivist> those that are dual marked yes
[02:07:51] <archivist> but for drilling you never use them that hard
[02:08:05] <RyanS> Have you had any experience with Chinese motors versus big name European,US, Japanese motors? Three-phase motors aren't complicated devices
[02:09:51] <archivist> I have a chinese motor on my weedy single phase drill, the rest are old british etc
[02:11:02] <archivist> and disappearing to a job now
[02:11:25] <RyanS> And I guess the current motor I have is Chinese anywhow, sounded super quiet in the shop
[02:30:02] <_DJ_> moin
[03:35:55] <mrsun> any good "tutorials" on how to setup and stuff for limit/home switches? :)
[04:26:10] <kengu> mrsun: like the documentation,
http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Homing_And_Limit_Switch ?
[04:30:25] <RyanS> The length of a motor excluding the shaft is irrelevant to IEC spec right?
[07:53:58] <tvc> Hello, can anyone point me in a direction to see an example where a hal float pin is set or changed through python? is this even possible?
[08:08:32] <mrsun> yeah! my limit switches and touch of mono 3.5mm connector has come! yeah! =)
[08:25:32] <archivist_herron> tvc, one puts the net connection in postgui somewhere
[08:43:29] <jdh> 20% off everything and free shipping at Enco
[09:05:51] <tjtr33> tvc,
http://linuxcnc.org/docs/html/hal/halmodule.html may be of use
[09:39:28] <tvc> THANKS!!
[09:50:33] <skunkworks> logger[psha],
[09:50:33] <logger[psha]> skunkworks: Log stored at
http://psha.org.ru/irc/%23linuxcnc/2013-12-02.html
[09:50:52] <mrsun> http://www.ebay.com/itm/FOUR-BEARING-2-2KW-WATER-COOLED-SPINDLE-MOTOR-ER20-INVERTER-VFD-/261136014224?pt=LH_DefaultDomain_0&hash=item3ccceb5b90 anyone has any info about the quality etc of these cheap china spindles? =)
[09:53:32] <skunkworks> do a search on cnczone
[09:59:39] <JT-Shop> skunkworks: do you have a EMCO PC TURN 50
[10:03:48] <JT-Shop> http://www.linuxcnc.org/index.php/english/forum/26-turning/27240-emco-pc-turn-50-help
[10:08:13] <skunkworks> No - it is a compact 5 - and I don't have a tool changer
[10:08:22] <skunkworks> (I was reading that...)
[10:08:41] <skunkworks> he does need an index - which was probably there hopefully still
[10:10:27] <tjtr33> tvc merry christmas!
[10:12:26] <JesusAlos_> hi
[10:12:44] <JeAl> hi
[10:25:33] <JeAl> http://imagebin.org/280199
[10:25:43] <JeAl> I think this pc is good for linuxcnc
[10:26:15] <JeAl> but appear "Real time delay error" each sesion
[10:40:53] <pcw_home> you may have to lower your base thread rate to avoid the real time error messages
[10:53:19] <Jymmm> Why you shouldn't digitally sign your name...
http://www.harborfreight.com/protectingcustomers
[10:53:32] <Jymmm> ink rules!!! =)
[10:58:23] <JeAl> I put the same base treath value that figure in latency test
[11:08:10] <pcw_home> which value?
[11:54:50] <IchGuckLive> hi all B)
[12:07:56] <IchGuckLive> all quiet today here after the shopping realy this wekend in the USA
[12:20:49] <kengu> hum
[12:26:21] <_DJ_> renamd
[13:10:50] <Tom_itx> pcw_home you around?
[13:11:08] <Tom_itx> hm2/hm2_7i43.0: board has FPGA '3s200tq144' but the firmware in hm2/7i43/SVST2_4_7I47SSS.BIT is for FPGA '3s400tq144'
[13:11:23] <Tom_itx> but i've got the 400k gate chip, XC3S400
[13:11:29] <Tom_itx> any ideas?
[13:12:12] <Tom_itx> i had an issue with something and reinstalled the firmware zip and after that i got the error
[13:12:28] <Tom_itx> when running it in linuxcnc
[13:12:51] <Jymmm> Tom_itx: did you try it?
[13:13:16] <Tom_itx> Jymmm, notice the last statement
[13:14:11] <Tom_itx> it compiled fine in xilinx but i got the error when running lcnc
[13:14:36] <Tom_itx> i did recompile it using the 200k gate option but haven't had a chance to try it
[13:14:47] <Jymmm> So, it's not actually firmware, but a driver?
[13:14:55] <Tom_itx> it's a bit file
[13:15:23] <Jymmm> That doesn't tel me anything. is it flashed directly to the mesa card?
[13:15:31] <Tom_itx> no
[13:15:37] <Tom_itx> it's loaded from the config line
[13:15:43] <Tom_itx> upon startup
[13:16:13] <Jymmm> Ok, and you haven't actually tried the 200K yet?
[13:16:31] <Tom_itx> not yet, it was getting late
[13:16:36] <Jymmm> which is what I asked you the first time =)
[13:16:38] <Tom_itx> but i don't have that chip
[13:17:14] <Jymmm> Mae a backup, try it, if it doesn't work, restore to where you were previously
[13:17:17] <Jymmm> make*
[13:17:36] <Jymmm> no arm, no fowl =)
[13:17:50] <Jymmm> no hotwings =)
[13:20:09] <Jymmm> Tom_itx: or call and ask =)
[13:21:08] <Jymmm> Does anyone know if a window regulator has stops/adjustments for left/right movement of the glass?
[13:21:54] <Tom_itx> the arm is likely broke
[13:21:56] <Tom_itx> or bent
[13:22:31] <Jymmm> I pulled the old one and didn't see anything. It looks exactly ike the replacement I jsut received.
[13:22:57] <Tom_itx> honda?
[13:23:01] <Jymmm> yeah
[13:23:43] <Jymmm> It could be the door itself, but I just don't know.
[13:24:10] <Jymmm> brb, let me compare them side-by-side...
[13:39:05] <ler_hydra> cradek, hey, do you still have your lathe with homebrew servo power amplifiers?
[13:39:26] <cradek> it's not in use, but yeah, I still have it
[13:41:33] <ler_hydra> cradek, ok, uh, want to be a guinea pig? ;) I've developed what I think is at least a somewhat decent servo controller, and I could use some testing and some exposure (
http://www.rabidmantis.se/flexible-servo-controller/)
[13:42:31] <Jymmm> Tom_itx: Ok, the ONLY thing I can see is the where the scissor part of the arms pivot there is maybe an 1/8" total play, where in the new one there is none
[13:42:41] <ler_hydra> IIRC your homebrew drive(s) were pretty much just power amplifiers, with the position loop closed inside of linuxcnc, so they're not really equivalent to what I linked to, which is a full single-channel PID controller.
[13:43:03] <Jymmm> Tom_itx: All the other arms are the same.
[13:43:56] <ler_hydra> anyway, if you're interested I'll assemble two additional ones when I make the next batch (after new years or so) and send them your way. If you think there's someone else that would be more interested/have a larger use for them that would be fine as well.
[13:44:34] <ler_hydra> (and just to clarify, by 'to give' I mean free)
[13:46:02] <ler_hydra> in the end I think it would be useful to make a derivative version without any internal PID controller, essentially letting the device respond to queries as to the current encoder position (allowing for high encoder speeds) and setting a current setpoint in the motor.
[13:47:04] <cradek> I was wondering why you made them so smart if they're for use with linuxcnc
[13:47:41] <ler_hydra> at the moment it's more of a general-purpose controller to be used with "dumb" hosts
[13:47:41] <cradek> I guess it saves you buying fpga hardware to count the encoders? but you'd still want fpga hardware to generate steps anyway...
[13:48:16] <cradek> aha
[13:48:29] <cpresser> does anybody know a source for (non-metric) ball-bearings in europe (or better: germany)?
[13:49:04] <ler_hydra> where by dumb I mean embedded, so the host would be able to at least generate steps at relatively high rates
[13:49:12] <cradek> hmm, wonder about using them for my (languishing) delta project:
http://timeguy.com/cradek-files/emc/rdelta-workvolume.png
[13:49:34] <cradek> I'd have to check what those servos are - I think 24v nominal
[13:49:57] <ler_hydra> that looks like a fun setup =)
[13:49:59] <cradek> I have a nice 4x V supply that currently runs steppers on my mini mill...
[13:50:20] <cradek> actually that's not what I'm planning - but I can't find a pic of the uspide-down one
[13:50:30] <ler_hydra> the components on the drive I linked to are rated at abs. max. 50V, so that should be fine
[13:50:52] <ler_hydra> I'm guessing the grey boxes are servos with a relatively hefty gearbox?
[13:51:27] <cradek> yes they're like rotary tables without the table, I got a few surplus
[13:51:53] <cradek> they don't seem to have any measurable backlash - I don't know what's inside them
[13:51:54] <ler_hydra> are they relatively OK wrt. backlash?
[13:52:03] <ler_hydra> nice!
[13:52:22] <cradek> yes I really should take one apart...
[13:53:37] <cradek> I have an up-to-date model plan somewhere ... building ja4 to look for it :-)
[13:54:06] <ler_hydra> ja4?
[13:54:18] <cradek> the joints_axes4 branch
[13:54:32] <ler_hydra> ah, ok
[13:59:01] <Jymmm> cradek: why the double arms?
[13:59:19] <Jymmm> cradek: reduced height?
[13:59:43] <cradek> Jymmm: have a look at delta robots working on youtube and you'll get a feel for what the double strut is for (it maintains a parallelogram shape)
[14:00:15] <Jymmm> cradek: I understand that, but I mean your choice for that design compared to others
[14:00:49] <Jymmm> cradek: That "elbow" seems like a weak point
[14:01:16] <Jymmm> compared to the linear actuator ones I've seen
[14:01:25] <cradek> ler_hydra:
http://timeguy.com/cradek-files/emc/upside-down.png is the setup I'm really hoping to build (kinematics are already done and working)
[14:01:58] <ler_hydra> cradek, nice setup! What are you aiming to do with the platform?
[14:02:00] <Jymmm> cradek: are those motors, or gear boxes of sorts?
[14:02:08] <cradek> the pink square representing working volume is 8" on a side, giving some idea of the servo size (the top cylinders)
[14:02:19] <cradek> the working volume is an 8" cube
[14:02:39] <ler_hydra> those servos are pretty big...
[14:02:40] <cradek> ler_hydra: high speed spindle for engraving - probably mostly PCBs
[14:03:11] <Jymmm> cradek: are those motors, or gear boxes of sorts?
[14:03:14] <cradek> yeah and they're geared 3:1ish before the worm, so lots of power
[14:03:26] <Jymmm> hmmm
[14:03:39] <cradek> Jymmm: the motor is the top cylinder, then there's a pulley set on the outside, then a worm gear inside the lower box
[14:03:48] <Jymmm> cradek: k
[14:03:55] <cradek> I guess the motor pulley is not drawn, the other one is
[14:04:05] <Jymmm> cradek: hell of a lot of torque on those
[14:04:13] <ler_hydra> interesting. I haven't looked much into the delta-structure, does it give you some significant benefit over a traditional "trivkins" (mill-esque) setup? (moving mass seems to look like one benefit)
[14:04:27] <cradek> no sliding bits needed
[14:04:52] <Jymmm> cradek: how are you planning on attaching at the elbows?
[14:05:02] <ler_hydra> ah, yeah, true, only rotational joints. very enticing that...
[14:05:40] <cradek> Jymmm: a joint made out of chromed balls and bronze seats - haven't tried it yet
[14:06:06] <cradek> ler_hydra: also you don't need a table 3x as long as the working length
[14:06:25] <cradek> 8x8x8 on a portable desktop machine is a fairly impressive volume
[14:07:02] <ler_hydra> ah, yes. clearly several benefits to that structure
[14:07:47] <cradek> yeah that's why I want to experiment with it. the kinematics both ways aren't too bad either.
[14:08:35] <skunkworks> cradek, I can't wait to see it running :)
[14:08:55] <ler_hydra> do you know roughly how much current you'd estimate you'd like to be able to supply?
[14:09:01] <cradek> well I have all the parts - so it's virtually done, right?
[14:09:22] <Jymmm> cradek: When did you move to Kansas dorthy?
[14:09:38] <cradek> ler_hydra: I measured stall current but forget what it was :-/
[14:09:56] <ler_hydra> cradek, hehe, hate it when that happens ;)
[14:10:23] <cradek> guessing ~ 3A
[14:10:40] <Jymmm> cradek: Do you have a way to measure the force placed on the elbow?
[14:11:08] <ler_hydra> oh, that's not a whole lot, I guess there's no need for lots of torque outout?
[14:11:37] <skunkworks> you proabaly want 3 times that peak...
[14:12:01] <ler_hydra> skunkworks, wasn't that stall current though?
[14:12:09] <Jymmm> I know nuttin, but the elbow just doesn't seem "right" to me for some reason.
[14:12:24] <Jymmm> I can't put my finger on it.
[14:12:54] <Jymmm> ok, off to install the new regulator and expect it NOT to fix the problem. *sigh*
[14:12:55] <cradek> Jymmm: for engraving there's little or no cutting force. I sure agree that the design of the joints will be the thing that constrains cutting force
[14:13:48] <cradek> ler_hydra: but 10A 42V sounds like plenty and then some, right?
[14:13:52] <Jymmm> cradek: It's not just cutting forces that hits me wrong. something else, but I can't put a finger on it. Think like over-extending your own elbow.... something along those lines.
[14:14:18] <Jymmm> ok, back in a bit
[14:14:30] <cradek> Jymmm: did you watch some videos? that really does help the intuitive grasp of the kinematics
[14:15:00] <cradek> you have to pick the strut lengths to keep you away from degenerate situations
[14:15:00] <ler_hydra> cradek, well, the drive I've linked to can supply upwards of 25A, so it's pretty overkill when driving at most 3A, but yeah, there's plenty of electrical power available
[14:15:48] <cradek> wonder if my 5i25 can generate quadrature steps (surely so)
[14:16:23] <ler_hydra> if so it would be really easy to test the drives
[14:16:41] <cradek> I had not worked out what to use for servo amps for this yet - a bit of a stumbling block
[14:17:02] <cradek> I think the encoders are just single-ended - is that ok?
[14:17:28] <ler_hydra> if you know of some communications protocol that could be implemented between linuxcnc and the drive, essentially just making it a dumb power amplifier and servo counter, it would (probably) be pretty easy to implement
[14:18:09] <ler_hydra> it only accepts single-ended encoder inputs so that's fine
[14:18:11] <cradek> heh can you use an existing driver like ppmc?
[14:18:24] <ler_hydra> I'm assuming they're 5/3.3V compatible?
[14:18:38] <ler_hydra> hmm, do you have to link to the protocol spec.?
[14:19:24] <cradek> er maybe I should've said 7i43, the register map is published
[14:20:11] <ler_hydra> cradek, hmm, google's not giving me any reasonable results for "7i43 "register map""
[14:20:19] <ler_hydra> got any pointers?
[14:20:40] <cradek> hmmmm
[14:21:25] <JeAl> pcw_home: 8062
[14:21:45] <JeAl> http://imagebin.org/280199
[14:22:59] <cradek> ler_hydra: I don't know what I'm talking about
[14:24:07] <ler_hydra> cradek, hahaha, no worries. If you happen to figure out some hypothetical comms protocol that would work just let me know and I'll code something together
[14:24:08] <cradek> ler_hydra: ask in -devel and maybe someone smart will have advice
[14:25:15] <ler_hydra> cradek, ideally this would be something that could run relatively slowly, IE could be bit-banged at the base thread rate and give enough data to the servo thread rate for semi-decent systems
[14:25:56] <cradek> sounds like you're going to end up doing EPP, for better or for worse
[14:26:31] <cradek> not sure what other hardware interface you could find that's cheap-or-free
[14:27:08] <ler_hydra> yep, probably not much that can be done to get around that
[14:28:45] <ler_hydra> I've been toying with the idea of some UART/RS232-esque serial interface with binary data; the host shifts in say 8 bits of data and receives 8 bits; where the sent 8 bits generate a setpoint motor current and the received 8 bits are the encoder position modulo 2^8
[14:29:44] <cradek> I don't think rs232 is fast enough is it?
[14:30:28] <cradek> (but fwiw I think you can manage to have 9 data bits on 16x50 uarts...?)
[14:30:48] <Tom_itx> _any_ backlash in a delta is gonna be amplified
[14:30:57] <JeAl> gn
[14:31:14] <cradek> http://pico-systems.com/motion.html says the register specification is "on this web site" but I can't find it
[14:31:54] <ler_hydra> well, not necessarily rs232, in fact something like SPI (with TX, RX, and CLOCK) pins might be better, as long as n*base_thread < servo_thread then n bits of data can be transmitted
[14:31:56] <andypugh> ler_hydra: That sounds a lot like the Mesa "Smart Serial" Interface, but that runs at 2.5Bb
[14:32:09] <andypugh> 2.5Mb, that is.
[14:32:32] <cradek> it'd be lovely if you could use an already-invented wheel that's at least mostly round
[14:32:35] <ler_hydra> andypugh, hmm, that would be hard to bit-bang on a RT-slow machine
[14:32:59] <andypugh> At least one company is making their own hardware to the Mesa specs.
[14:33:01] <ler_hydra> but the protocol itself might well be able to be mangled
[14:33:18] <ler_hydra> hmm, interesting. Got any references as to the smar serial interface?
[14:34:19] <andypugh> It's described in the manuals of the devices:
http://www.mesanet.com/pdf/motion/8i20man.pdf for example
[14:34:52] <andypugh> The basic driver code would work at any speed.
[14:35:03] <andypugh> (ie, the HAL end)
[14:35:08] <cradek> ler_hydra:
http://pico-systems.com/univstep_regs.html
[14:35:27] <PCW> so 8.062 usec is your jitter but what is the base thread set for (I don't know enough about stepconf to know if it sets this automagically)
[14:36:22] <cradek> ler_hydra: I'd check with him to make sure this is up-to-date; I know index handling has changed over the years, and so has velocity estimation (but I'm not sure whether univstep does either of these)
[14:37:02] <PCW> something SPI like might be OK
[14:37:52] <ler_hydra> cradek, hmm, though there would be significant overhead in using this protocol (for one, my drive can only handle one encoder, connect to one motor, and has no real GPIO), but something akin to that could probably be done
[14:38:12] <PCW> SPI out a parallel port at the base thread
[14:38:14] <PCW> (with framing done by a timeout to save pins)
[14:38:15] <Jymmm> PCW: Tom_itx is hunting you down; needs info on 200K vs 400K gate driver
[14:38:35] <PCW> common clock for n channels
[14:39:03] <ler_hydra> at the moment the PID loop consumes about 50% of the execution time, the processor is clocked at 32MHz and has an internal data bus of 8 bits, so there's (enough) processing power for basic stuff
[14:39:38] <Jymmm> Tom_itx: NFC, installed new regulator and seems to work fine. But so did the original when I R&R it. Still need to reinstall the door panel
[14:39:38] <ler_hydra> PCW, yeah, essentially MISO/MOSI/SCLK and independent !CS pins for each device means 3+n pins are needed
[14:40:10] <ler_hydra> though a common clock and independant MISO/MOSI pins would allow for simultaneous communications
[14:40:53] <Jymmm> 2013-12-02.10:54:55 Tom_itx: pcw_home you around?
[14:40:53] <Jymmm> 2013-12-02.10:55:13 Tom_itx: hm2/hm2_7i43.0: board has FPGA '3s200tq144' but the firmware in hm2/7i43/SVST2_4_7I47SSS.BIT is for FPGA '3s400tq144'
[14:40:53] <Jymmm> 2013-12-02.10:55:28 Tom_itx: but i've got the 400k gate chip, XC3S400
[14:40:54] <Jymmm> 2013-12-02.10:55:34 Tom_itx: any ideas?
[14:41:03] <PCW> or run from a RPI or BBB that has SPI
[14:41:55] <ler_hydra> cradek, anyway, it seems I'm rambling (again), what I had intended to ask was; (1) do you want two (or maybe three) of these drives? I'll give them away, mostly I'd just like for them to become seen and used, and (2) if not, do you think some other linuxCNC dev could use a pair of them?
[14:43:12] <Jymmm> ler_hydra: how are they tuned?
[14:43:19] <cradek> aside from feeling bad about talking you up to 3, I would use them for this project, and it might be a kick in the pants to actually get it done. if I fail to use them in a reasonable time I'd be happy to pass them on to someone else of your choice.
[14:43:39] <cradek> Jymmm: serial console
[14:43:50] <Jymmm> cradek: ah, cool.
[14:44:04] <ler_hydra> Jymmm, with the current software PID tuning is done over a serial console, position/velocity/acceleration steps are plotted in ascii-art
[14:44:44] <ler_hydra> Jymmm, bottom right of cover page;
http://www.rabidmantis.se/flexible-servo-controller/
[14:45:51] <Jymmm> now to write a tuning sketch =)
[14:46:11] <PCW_> most of the sserial remote manuals have the protocol in the back
[14:46:26] <ler_hydra> cradek, hehe, that's quite alright. I'll probably end up skimping on the power terminal block (crimp ring connectors can be used instead for incoming power and motor connections), but otherwise they're relatively inexpensive, and I'll be assembling several myself. It's past time for me to contribute something to linuxCNC anyway =)
[14:46:49] <ler_hydra> PCW_, I'll have to keep that in mind!
[14:47:11] <Jymmm> ler_hydra: ASCII ART FTW!!! lol
[14:47:16] <andypugh> Tom_itx: That FPGA size error message can be misleading, you also get that if there is no response at all from the card.
[14:47:39] <ler_hydra> Jymmm, hehe, yep, part of the reason for making it in the first place ;) (and it works pretty well too)
[14:48:08] <Jymmm> ler_hydra: I have "The Draw" if you want to reverse that plot into tuning parameters =)
[14:49:04] <ler_hydra> Jymmm, hmm, what's that? Something along the lines of input a step response to get approximate PID gains for some configured closed loop response?
[14:49:28] <Jymmm> ler_hydra: You KNOW you want to turn an animated ascii art rocket ship into plot params...
http://en.wikipedia.org/wiki/TheDraw
[14:50:02] <Jymmm> ler_hydra: (Yes, there is/was actual animated ascii art =)
[14:50:09] <ler_hydra> Jymmm, HAHA, it would be GREAT to have a servo amp that has a rocket!
[14:50:27] <ler_hydra> at the moment I'm stuck with just a lame ASCII-art logo on boot
[14:50:44] <Jymmm> ler_hydra: There ya go... id DOES output binary, pascal, and a few other formats
[14:50:49] <Jymmm> it*
[14:51:17] <Jymmm> ler_hydra: I think there is a deb version in the archives too
[14:52:11] <Jymmm> Alright, now to reinstall the door skin and see if it still works....
[14:52:35] <ler_hydra> Jymmm, as an aside, TheDraw is a really strange name
[14:56:52] <ler_hydra> cradek, anyway, I'll let you know when I've assembled the boards and need an address to send them to. If you think of some improvement to the drives I
[14:57:42] <ler_hydra> *I'd appreciate it =)
[14:57:56] <cradek> awesome, thanks
[15:32:18] <Jymmm> ler_hydra: Well, realize this was monochrome/bbs/cga days.... there was no VGA or Photoshop at the time =)
[15:32:50] <ler_hydra> Jymmm, hehe, still, it's more than just 7-bit ascii ;)
[15:33:02] <Jymmm> well, ascii, ansi, etc
[15:33:16] <Jymmm> ZMODEM
[15:33:53] <Jymmm> Heh, I actually used xmodem about a year ago, works pretty good =)
[15:34:26] <ler_hydra> what for?
[15:34:50] <Jymmm> had to download a file from embedded debian systems
[15:41:21] <NickParker> PCW: Could you explain the ModuleID array in my PINS.vhd file to me? Tom_itx was talking about instances and some other stuff in it last night, and I didn't really understand.
[15:42:34] <Jymmm> ler_hydra: WITH wifi...
http://www.logicpd.com/products/system-on-modules/dm3730-torpedo-wireless-som/
[15:43:33] <ler_hydra> Jymmm, not bad!
[15:44:08] <Jymmm> ler_hydra: Yeah, just took a standard debian ISO, nothing fancy
[15:46:30] <Jymmm> ler_hydra:
http://electronicdesign.com/embedded/how-boot-linux-under-one-second
[15:47:23] <PCW> NickParker:
[15:47:25] <PCW> type ModuleRecord is -- probably need an alternate way for smart modules
[15:47:26] <PCW> record
[15:47:28] <PCW> GTag : std_logic_vector(7 downto 0);
[15:47:29] <PCW> Version : std_logic_vector(7 downto 0);
[15:47:31] <PCW> Clock : std_logic_vector(7 downto 0);
[15:47:33] <PCW> NumInstances : std_logic_vector(7 downto 0);
[15:47:34] <PCW> BaseAddr : std_logic_vector(15 downto 0);
[15:47:36] <PCW> NumRegisters : std_logic_vector(7 downto 0);
[15:47:37] <PCW> Strides : std_logic_vector(7 downto 0);
[15:47:39] <PCW> MultRegs : std_logic_vector(31 downto 0);
[15:47:40] <PCW> end record;
[15:47:48] <ler_hydra> Jymmm, is this something you're developing as well?
[15:50:26] <andypugh> I want a socket set screw (grub screw) with no threads, so I can put my own peculiar threads on it....
[15:50:34] <Jymmm> ler_hydra: No, it was just a sub-assembly of a product I was working on.
[15:50:43] <Jymmm> working with*
[15:51:02] <Jymmm> andypugh: JB Weld?
[15:51:05] <ler_hydra> Jymmm, ok, still, seems like a nice module to know about
[15:51:17] <ler_hydra> do you have an idea as to price?
[15:51:37] <Jymmm> ler_hydra: iirc $400'ish USD
[15:52:20] <Jymmm> $304 USD
[15:52:36] <Jymmm> https://avnetexpress.avnet.com/store/em/EMController?term=SOMDM3730-30&x=-1271&y=-54&action=products&langId=-1&storeId=500201&catalogId=500201&hbxSType=&N=0&filterButton=true
[15:52:39] <ler_hydra> =O... that's... a lot, for what (to me) seems to be essentially three IC's
[15:52:55] <ler_hydra> plus some passives and other stuff
[15:52:59] <Jymmm> http://www.digikey.com/product-search/en?vendor=0&keywords=SOMDM3730-30
[15:53:12] <Jymmm> ler_hydra: When size/weight matters
[15:53:59] <Jymmm> http://www.digikey.com/product-detail/en/SOMDM3730-30-1780AKIR/460-3517-ND/3114404
[15:54:13] <ler_hydra> Jymmm, i'm trying really hard not to make a lewd reference to that...
[15:54:45] <Jymmm> I'm trying to figure out where "weight" plays in your lewd comment
[15:55:18] <ler_hydra> different people have different preferences? err, yeah, anyway...
[15:55:26] <Jymmm> rotf
[15:55:43] <ler_hydra> still, a nice module. Is it made as a turnkey widget?
[15:56:28] <Jymmm> What you see is what you get. You have to supply/design the board it plugs into
[15:56:42] <Jymmm> for the I/O that is
[15:57:23] <ler_hydra> well, yes, absolutely, and for some power supply as well I'm assuming (though it looked like there was some step-down converter onboard?)
[16:00:58] <Jymmm> ler_hydra:
http://player.vimeo.com/video/38106663
[16:03:09] <andypugh> Jymmm: I am hanging my CNC axis off of it, I think JB weld might be sub-optimal :-)
[16:03:56] <Jymmm> andypugh: Then I guess you dont want to hear my suggestion of hot melt glue then =)
[16:06:01] <uw> whats so great about this logic pd
[16:06:19] <Jymmm> A ways back someone told me to use charcoal lighter fluid to remove lithium grease, it works great btw. But, I found that you can use waterless hand cleaner to remove grease, oil, etc without having to use a lot of solvent, it's gelled so it stay sin place, and you barely use any at all . Great for cleaning up tools too.
[16:06:29] <Jymmm> uw: it's the size of a US Dime.
[16:08:18] <uw> "meh"
[16:09:02] <NickParker> hey PCW or andypugh how does this pin file look for trying to make a config with 3 (4 output table mode) stepgens and 1 pwmgen?
[16:09:04] <NickParker> http://pastebin.com/FV7XHQZJ
[16:09:59] <NickParker> I put the proper instances in, and replaced lines in the moduleID array for sserial and qcounter that I'm not using with null lines
[16:17:18] <PCW> you need lines like this for unused pins
[16:17:20] <PCW> IOPortTag & x"00" & NullTag & x"00", -- I/O 16
[16:17:21] <PCW>
[16:17:55] <PCW> because even unused pins are available as GPIO
[16:19:10] <Jymmm> ler_hydra: does your thermal controller accept iambic paddles =)
[16:19:59] <ler_hydra> Jymmm, haha, that would be a nice feature ;)
[16:20:01] <Jymmm> ler_hydra: I'll hookup my bencher paddle to it...
http://upload.wikimedia.org/wikipedia/commons/2/2d/Bencher_paddle.jpg
[16:20:38] <andypugh> Woohoo! Rog finally has his Hardinge HXL dialled in to his (rather exacting) standards:
https://plus.google.com/photos/108164504656404380542/albums/5951771387407954465?authkey=CJbSsrD-kfr3Hg
[16:20:46] <ler_hydra> if you can make do with just one side (ie. not iambic) then it's just a matter of connecting it =)
[16:20:56] <tjtr33> andypugh, you could use a pivot bolt, take off the shoulder and threaded end, then thread the pivot length ( that really only gets you the socket head cylinder to work on )
[16:21:07] <Jymmm> ler_hydra: oh, ONLY NTC? No PTC?
[16:21:13] <Jymmm> ler_hydra: where's thefun in that?
[16:21:37] <Jymmm> ler_hydra: and what is making the noise? PC ?
[16:21:47] <ler_hydra> the load ;)
[16:21:52] <Jymmm> lol
[16:22:01] <andypugh> tjtr33: Hmm, I guess it might be worth looking at the socket size on shoulder bolts.
[16:22:21] <ler_hydra> hmm, well if PTC devices can be modeled as having a negative beta value then that'll work too
[16:22:32] <ler_hydra> so with some loads things can get... fun...
[16:22:48] <Jymmm> ler_hydra: no other feedback?
[16:23:05] <ler_hydra> not in the current firmware release at least
[16:23:16] <Jymmm> ler_hydra: I need to control some tanks
[16:23:20] <ler_hydra> there's a semi-free pin used for 4-pin PC fans that you could hack into
[16:23:22] <ler_hydra> ...
[16:23:30] <tjtr33> andypugh, very very nice shop and impressive step check
[16:23:44] <ler_hydra> as in a tank;
https://en.wikipedia.org/wiki/Tank ?
[16:24:01] <Jymmm> ler_hydra: as in liquid heating tanks
[16:24:14] <ler_hydra> haha, ok, right. that makes more sense
[16:24:19] <Jymmm> lol
[16:24:25] <ler_hydra> well, I'm guessing that you're talking about a high-power heater?
[16:24:30] <ler_hydra> IE. you'd need a relay anyway?
[16:24:53] <Jymmm> yeah, 120VAC heaters in most cases
[16:25:07] <Jymmm> relays I got, tempcontroller I dont =)
[16:25:14] <ler_hydra> in that case (IIRC) most relays will make enough noise
[16:25:16] <andypugh> I need to point out, that is not _my_ workshop.
[16:25:29] <ler_hydra> if not, you can just connect a random 3-pin PC fan during config
[16:25:35] <Jymmm> lol
[16:25:51] <Jymmm> then wth is the LEd for?
[16:26:22] <ler_hydra> uh I think that's just a power indicator
[16:26:33] <andypugh> It's this guy (who does the most exquisite stuff):
http://www.f1-2000.co.uk/index.php?f=conrods
[16:26:43] <ler_hydra> the entire device is I/O limited so there's not a whole lot of free GPIO
[16:26:49] <ler_hydra> (8-pin package)
[16:27:03] <Jymmm> ah
[16:27:34] <NickParker> PCW: ok thanks.
[16:39:19] <NickParker> ok I have a new problem, that's (hopefully) easier to fix. My PC seems to think the 6i25 is a graphics card right now, and defaults to it whenever it's in
[16:39:43] <NickParker> if i boot with the 6i25 installed it doesn't drive the display, and if i have the 6i25 out everything works fine
[16:40:13] <NickParker> this didn't happen before, but i've been uploading my obviously incorrect bitfiles to the 6i25, so i guess the last one has made it impersonate a display driver?
[16:40:27] <NickParker> there wouldn't happen to be a "restore to factory" button on it would there?
[16:50:40] <andypugh> That's quite funny.
[16:51:31] <andypugh> NickParker: You can run mesaflash through an ssh connection. (In fact that is the only way I have ever done it)
[16:52:06] <NickParker> easiest way to set up ssh on an unnetworked pc?
[16:52:13] <NickParker> actually i'll check google..
[16:52:23] <micges> NickParker: first try this 6i25 on different pc
[16:52:38] <NickParker> micges: I don't have any other PCs running linux
[16:52:44] <NickParker> let alone linuxcnc
[16:53:05] <andypugh> No idea, ssh is just there already on a mac
[16:53:36] <NickParker> oh i've got an ssh terminal, i just don't know how to make my linuxCNC box accessible through ssh
[16:53:36] <micges> NickParker: when this happended?
[16:53:52] <NickParker> I shut it down last night after loading a new bitfile.
[16:54:29] <micges> what bitfile?
[16:56:03] <NickParker> one i compiled with ISE myself
[16:56:11] <NickParker> I've still got it on a usb if it would be helpful
[16:56:29] <NickParker> also, I have to go in about 15 mins
[16:57:47] <andypugh> There are fallback bitfiles, I can't recall how to activate them. It might be a jumper.
[16:57:57] <NickParker> https://www.mediafire.com/?ak1k7t6uiq1g039
[16:58:07] <NickParker> andypugh: kk let me go look through the manual for that.
[17:00:02] <andypugh> Ah, it looks like the FPGA itself chooses to use a fallback if the other one is corrupted. A valid but useless bitfile probably won't trigger it.
[17:00:43] <NickParker> yep I don't see anything in the manual about jumpers..
[17:01:09] <andypugh> Pull out the 6i25, enable ssh on the Ubu box, put the 6i25 back, and ssh into it.
[17:01:51] <andypugh> sudo apt-get install openssh-server
[17:02:18] <andypugh> You are likely to want that anyway.
[17:02:43] <andypugh> It's just useful to have command-line access to all your machines.
[17:03:13] <NickParker> well I have to go for now, I guess when I get back I'll run ethernet from my router all the way to my linuxcnc box. I should see about getting a PCI wifi card for it...
[17:04:05] <andypugh> The other alternative is to use the Windows version of mesaflash to reflash the card
[17:04:11] <NickParker> ooh sweet that's actually a thing. Didn't expect that... I'll pick one up on the way home. ty everybody.
[17:04:31] <NickParker> andypugh: I don't dare plug this card into my main pc. its display driver situation is already shaky at best.
[17:04:41] <_DJ_> gn8
[17:04:45] <NickParker> anyway, back in a few hours.
[18:11:57] <andypugh> I need to assemble an extension on to a leadscrew. Access and other constraints mean that I can't fit a full-length ballscrew, bit need to add an extension piece in-situ. Furthermore I only have access to the joint in the axial direction.
[18:12:55] <cradek> yikes
[18:13:26] <andypugh> The solution I found was to put a taper (Morse angle) on the end of the screw, and a socket on the extension, then use a differential screw to draw the parts together.
[18:14:00] <andypugh> This works really well, and I am quite proud of it.
[18:14:03] <cradek> I first thought of heat
[18:14:22] <cradek> how did you cut the screw?
[18:14:38] <andypugh> But I have siince decided to upgrade the 12mm screw to a preloaded 16mm screw.
[18:15:08] <andypugh> I bought-in the screw. I am not in the ballscrew making business :-)
[18:15:44] <cradek> I meant the in-place screw you were extending
[18:16:36] <andypugh> The 12mm screw had a 6mm / 8mm differential screw, I think with the 16mm I can use 8mm / 12mm
http://imagebin.org/280271
[18:17:35] <andypugh> cradek: I machined it before fitting. In-place machining would have been too hard even for a man who likes a challenge
[18:19:18] <andypugh> My question is, should the differential screw be shorter than in the picture, or longer? Both options beat the arrangement shown, I feel sure.
[18:20:01] <andypugh> So, screw into the body of the ballscrew, or body of the ballscrew into the taper?
[18:24:40] <PCW> preemt_rt seems to have about 44 usec latency on my desktop at home so may be OK for Ethernet
[18:25:14] <skunkworks> PCW: are you testing micges stuff?
[18:25:23] <PCW> Yes
[18:25:27] <skunkworks> I was hoping to today.. but maybe tomorrow
[18:25:43] <andypugh> Does anyone need me to explain differential screws? They are really fun things
[18:25:52] <PCW> well i will when i figure out the magic ip and mac incantations
[18:26:02] <skunkworks> so - it is realtime? even though using normal ethernet drivers?
[18:26:07] <PCW> Yes
[18:26:12] <skunkworks> wow - pretty cool
[18:26:35] <PCW> looks pretty good just pinging around
[18:27:17] <PCW> (the config line needs the mac address so a static arp entry can be entered)
[18:30:23] <PCW> micges is not around or I would ask him the config format
[18:31:41] <andypugh> cradek: Please, opinionate, I am curious about how folk vsualise the situation. As a reference point the area of the 8mm screw is about 64, and the annular ring is about 80
[18:32:47] <andypugh> I think that a screw all the way into the meat of the ballscrew feels stronger.
[18:33:28] <andypugh> (Maybe make the taper shorter).
[18:33:53] <andypugh> But the screw will stretch, so how well is the load transferred?
[18:37:03] <andypugh> (OT: <giggle>
http://motocorsa.com/media/site/manigale-ducati-1199-wallpaper-12-comp.jpg
[18:42:57] <Tom_itx> PCW could you explain why linuxcnc report i have a 200k gate chip when i've got a 400k one when loading the bit file?
[18:43:32] <andypugh> Possibly because you have no comms at all
[18:43:47] <andypugh> (We should fix that)
[18:43:49] <PCW> 2 possibilities if the error is early it like any says no coms
[18:43:53] <Tom_itx> when i load a 200k gate compiled one it worked
[18:44:17] <PCW> chip selected is 200K?
[18:44:58] <Tom_itx> in xilinx yes
[18:45:06] <Tom_itx> i tried the 400k one first
[18:45:21] <andypugh> Yeah, but that checks the comms return against the (known) bitfile size. It's a really badly thought-out diagnistic, actuallly.
[18:45:21] <Tom_itx> i had to reload the zip file because of a problem then i started getting that
[18:45:35] <Tom_itx> it synthesized ok
[18:45:38] <Tom_itx> both
[18:45:53] <PCW> There are 2 things that need to be set: the GUI chip selection and the card type
[18:46:00] <Tom_itx> yes
[18:46:06] <Tom_itx> lemme fire that pc up
[18:46:22] <Tom_itx> and i'll tell you what's selected currently
[18:47:30] <PCW> a 200K bitfile for the 7I43 will be about 131K bytes a 400K bitfile will be about 212K bytes
[18:47:47] <Tom_itx> yeah it was around 209 iirc
[18:48:04] <PCW> you cannot program a 400K FPGA with a 200K bitfile or vice versa
[18:48:32] <Tom_itx> i just couldn't figure out why i got that error though since i was using a 400k seting
[18:48:59] <Tom_itx> hm2/hm2_7i43.0: board has FPGA '3s200tq144' but the firmware in hm2/7i43/SVST2_4_7I47SSS.BIT is for FPGA '3s400tq144'
[18:49:03] <Tom_itx> was the error
[18:49:05] <andypugh> Because a return value of zero can mean both 200k and no card at all. It's a bug
[18:49:49] <atom1> use work.i43_200card.all;
[18:49:53] <atom1> is currently selected
[18:50:08] <atom1> use work.PIN_SVST2_4_7I47SSS_48.all;
[18:50:11] <atom1> and the card
[18:50:12] <PCW> OK so thats wrong
[18:50:18] <atom1> yes
[18:50:52] <atom1> but i had this selected when i got the error: use work.i43_400card.all;
[18:50:52] <PCW> so the complaint is not surprising
[18:51:16] <atom1> so i switched _to_ the 200 to correct it
[18:51:20] <atom1> which seems wrong
[18:51:51] <atom1> i got the above error with the 400k setting
[18:52:47] <PCW> whatever is wrong it will not help to select the wrong card
[18:52:49] <PCW> if you have a 400K FPGA card you must select "use work.i43_400card.all;"
[18:53:03] <atom1> i agree
[18:53:10] <atom1> but why would i get that error?
[18:53:40] <atom1> those are the only 2 lines i changed in the file
[18:53:43] <PCW> no comms?
[18:54:05] <atom1> well i got the lcnc error screen
[18:54:32] <atom1> i will go thru it once more to triple check what i did
[18:54:37] <PCW> you can get that error if there's no EPP comms
[18:55:10] <atom1> i'll try again and see what happens with the 400 setting
[18:55:27] <PCW> a bad bifile can do this also (a bad bitfile that loads but doesnt work)
[18:55:43] <atom1> i'm gonna test it with a known good file
[18:56:14] <PCW> if you loaded a bad bitfile you may need to power cycle the 7I43
[18:56:24] <atom1> yeah i was each time
[18:56:38] <atom1> takes a while to churn on this pc
[18:57:00] <PCW> (since a bad bitfile will not respond to the "reset thyself" command)
[18:57:03] <atom1> i'm using the standard 7i47_48 vhd
[18:57:11] <atom1> yep
[18:57:59] <atom1> i'm not sure it likes remapping those pins like i was trying to do
[18:58:27] <atom1> using the spare B & index pins from the encoder for sserial
[18:59:04] <PCW> that cannot cause any problem
[18:59:12] <atom1> i did get sserial working, just trying to maximize pin use
[19:00:00] <atom1> if i don't use the full encoder function, i still count that as an instance in the header right?
[19:00:09] <atom1> even if it's just using channel A?
[19:00:46] <atom1> splitting the other 2 pins as an instance under sserial
[19:01:53] <atom1> i did notice while it was compiling it mentioned something about assigning zero to the unused pin functions
[19:02:04] <atom1> or such
[19:03:32] <PCW> Sure any part of an encoder still needs the whole encoder to be instantiated (but some unused logic will be trimmed)
[19:04:10] <PCW> none of these things can cause trouble at the interface level
[19:07:06] <atom1> ok
[19:12:54] <andypugh> Back to my screw, I thing this definitely looks better torsionally, though actually I am very much more concernerd about axial forces:
http://imagebin.org/280273
[19:13:12] <uw> whats that for
[19:13:55] <cradek> how do you tighten the screw?
[19:16:10] <andypugh> I screw the screw all the way into the ballscrew, then screw on the extension piece. Then backing out the screw (with a key through the axial bore) draws the taper up, because the two sections of the screw have a different pitch.
[19:18:29] <cradek> oh now I see the green part is hollow, of course
[19:18:44] <cradek> it's very clever
[19:18:46] <andypugh> This is not my invention, Wohlhaupter have been attaching the shanks to their boring heads this way for decades. It is _very_ clever,.
[19:19:43] <cradek> and you can unmount them!
[19:19:56] <cradek> without destroying everything with jacobs wedges...
[19:20:21] <cradek> those things suck
[19:20:29] <andypugh> Because the effective pitch of the screw is 0.25mm, but the actual teeth are 1.5mm, so the clamping force is huge, but the teeth are strong.
[19:21:44] <PCW> Dont micrometers use differential screws sometimes?
[19:22:11] <andypugh> Sometimes, yes. It's clever there too.
[19:23:32] <PCW> Thats very slick way to get a high pull-in/push out tapered connection
[19:23:34] <andypugh> What I like about the version I am using is that you sort of end up with a bolt with a very big virtual head, that could not possibly fit down the hole you have
[19:24:25] <andypugh> (And it also has an unfeasibly small effective pitch)
[19:25:57] <andypugh> The only drawback is that you have to remember what you did. Because it is clockwise to release :-)
[19:26:42] <cradek> fiendish
[19:26:44] <PCW> Yes you have so much leverage you could probably bust things
[19:27:14] <andypugh> Anyway... Shorter screw or longer screw? Equal length screw clearly combines the worst of all worlds.
[19:27:23] <cradek> in watches left-hand screws are usually but not always marked
[19:28:44] <andypugh> Nicks on the corners of the nuts are usual in welding (etc) gasses.
[19:29:40] <andypugh> In this case the screw is a whole allen-key length down a hole, you rather have to know it is there to even look for ir.
[19:31:02] <andypugh> I am not sure of the exact convention, but Oxygen regulators are right-hand thread and Acetylene are left-hand.
[19:31:21] <NickParker> hey PCI slots aren't supposed to take much pressure right?
[19:31:46] <andypugh> I am not sure if the convention is combustion safety or toxicity if inhaled.
[19:31:53] <NickParker> I've got a PCI wifi adaptor for my linuxcnc box here and it's ridiculously hard to get in. The profiles look like they match and nothing else is hitting though, so i don't get it.
[19:32:43] <andypugh> There are a lot of springy pins. Perhaps start at one end?
[19:32:45] <cradek> andypugh: isn't it the other way around?
[19:33:18] <NickParker> andypugh: wow that made a world of difference. Thanks
[19:33:30] <andypugh> Also make sure that there are no foreign objects in the slot.
[19:33:55] <PCW> like folded in pins :-(
[19:34:02] <andypugh> cradek: Could be, I am not a welder
[19:35:11] <andypugh> cradek: I could also believe that it differs horrribly from country to country.
[19:35:27] <cradek> oops google says you're right
[19:36:06] <cradek> but sure, could be different here, coriolis effect
[20:02:45] <andypugh> Goodnight all
[20:17:38] <Tom_itx> ok the bit file is working now
[20:17:49] <Tom_itx> not quite sure what i did wrong
[20:36:02] <Tom_itx> ok now with one channel the encoder velocity is inverted
[20:36:06] <Tom_itx> counting negative
[20:37:31] <pcw_home> Yes thats a consequence of unused pins being tried to logic '0'
[20:38:04] <pcw_home> so in up down mode with B at '0' the count is negative
[20:39:54] <Tom_itx> i tried inverting it by setting scale gain negative but that didn't work
[20:40:47] <pcw_home> setting the encoder scale negative should invert the count
[20:41:53] <Tom_itx> yep that did it
[20:41:54] <Tom_itx> thanks
[20:42:19] <Tom_itx> seems to all be working now
[20:43:02] <Tom_itx> funny mistakes happen when it's late and one's tired
[20:44:32] <pcw_home> Lots of weird things are possible with the Xilinx tools as well...
[20:45:24] <Tom_itx> kinda fun messin with these bit files
[20:46:26] <pcw_home> Ive been meaning to have the module ID part autogenerated but I haven't go to it yet
[20:46:56] <Tom_itx> it's not so bad once you know the rules
[20:47:18] <pcw_home> A tool to make the pin files would help as well
[20:47:21] <Tom_itx> i'm still not real sure about the sserial in's n outs
[20:47:52] <pcw_home> other then the resolver interface its probably the most complex module
[20:48:30] <Tom_itx> you've only implemented a single device per channel though right?
[20:48:37] <Tom_itx> up to 8 is allowed
[20:49:23] <pcw_home> no, 1 per channel 8 per port
[20:49:34] <Tom_itx> err that's what i meant
[20:49:44] <pcw_home> 4 ports max so 32 channels max
[20:50:30] <Tom_itx> but not 2 cards on the same pins
[20:51:11] <pcw_home> No, RS-485 mode is not supported (it may eventually be but not too likely)
[20:51:25] <Tom_itx> i'm not real sure what would need it
[20:52:02] <pcw_home> Yeah and it a fair bit more complicated (especially timeout calculations)
[20:52:10] <Tom_itx> i'm gonna wire up the 2nd channel in a bit to test it
[20:52:54] <pcw_home> in some cases there may be an advantage to more ports with fewer channels/port
[20:53:06] <pcw_home> (if you have the space)
[20:53:41] <pcw_home> (a bit faster)
[20:54:38] <pcw_home> But the CPU is fairly fast (75 MHz in the 7I43) so the per channel overhead is not much
[20:59:53] <skunkworks> http://www.youtube.com/watch?v=Hm3QlWTTcjo
[21:00:10] <Tom_itx> here's the .vhd file i ended up with:
http://tom-itx.dyndns.org:81/~webpage/cnc/configs/sherline/bitfiles/PIN_SVST2_4_7I47SSS_48.vhd
[21:01:22] <Tom_itx> sss.. sserial special
[21:03:03] <pcw_home> Nothing on the second connector?
[21:05:04] <Tom_itx> gonna try it now
[21:38:13] <Tom_itx> pcw_home, do i need 2 instances of sserial in the bit file to run cables on 2 pairs of tx rx pins?
[21:39:16] <Tom_itx> IO 12 & 20 are 1 and IO 14 & 21 are the other pair
[21:39:38] <Tom_itx> i think you told me only one
[21:42:16] <pcw_home> just one (up to 8 channels per sserial instance)
[21:43:31] <Tom_itx> i'm not getting anything from the 2nd one yet
[21:45:16] <pcw_home> config line correct?
[21:45:49] <Tom_itx> umm
[21:45:53] <Tom_itx> i bet not
[21:45:55] <Tom_itx> :)
[21:46:57] <Tom_itx> duh
[21:50:31] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/cnc/configs/sherline/bitfiles/showpin.txt
[22:08:16] <pcw_home> Probably a bit more I/O than you need
[22:08:51] <Tom_itx> it was a good exercise
[22:09:05] <Tom_itx> but yeah probably so
[22:09:56] <pcw_home> but you still haven't enabled the analog inputs and MPG encoders...
[22:10:25] <Tom_itx> i tried mode 2 for encoders but didn't see them show up in the list
[22:10:44] <Tom_itx> i need to read the docs more
[23:58:15] <Jymmm> This is awesome
https://www.youtube.com/watch?v=-RWzd3Q-Qtg
