#emc | Logs for 2010-07-07

[00:04:18] <theorb> theorb is now known as theorbtwo
[00:51:54] <Valen> got an odd problem with tuning on our mill
[00:52:22] <Valen> i do a G1 F100 and tune FF1 to zero out the error during cruise
[00:52:41] <Valen> if i then do a G1 F200 the error is back and bigger
[00:53:35] <Valen> basically i can tweak it to hover around 0 error at a specific feed rate with FF1, but as i change speeds the constant for FF1 seems like it should change
[00:53:47] <Valen> wondering what i'm doing wrong
[02:11:21] <PartScavenger> Hey guys, noob here, I'm trying to design my CNC machine. According to forums and other internet sites, I shouldn't run more than 1.75A per phase. The controller is rated 3A, 3.5A Max.
[02:11:51] <PartScavenger> My question is does my controller mean 3A per phase?
[02:12:15] <PartScavenger> I thought it meant 3A per phase,
[02:12:55] <PartScavenger> or is it supposed to add together and I just misunderstood?
[02:13:24] <PartScavenger> My controller is here:
[02:13:25] <PartScavenger> http://cgi.ebay.com/New-3-Axis-Stepper-Motor-Driver-Board-Controller-TB6560-/230494686336?cmd=ViewItem&pt=LH_DefaultDomain_0&hash=item35aa8dcc80
[02:13:34] <PartScavenger> It's a cheap chinese
[02:23:55] <alex_chally> PartScavenger, there are three of the actual driver chips on that, correct?
[02:24:04] <alex_chally> probably three chips under the big alum heatsink?
[02:24:11] <alex_chally> yeah, looks like it
[02:24:21] <alex_chally> so I would think that each axis will get 3A per axis
[02:25:11] <alex_chally> yeah, they call for a 8A PSU
[02:25:24] <PartScavenger> Hey!
[02:25:30] <PartScavenger> Yes three of the acutal chips
[02:26:12] <PartScavenger> I saw several people recommend that you don't run over 1.75A per phase,
[02:26:22] <alex_chally> so I would think that with that 8A PSU you are going to get ~2.6A out per axis, minus energy losses
[02:26:31] <alex_chally> PartScavenger, 1.75A per phase of your steppers?
[02:26:42] <PartScavenger> Yeah, apparently
[02:26:50] <PartScavenger> So that means 3.5A per axis right?
[02:27:14] <PartScavenger> I'm looking at a 3A stepper then?
[02:27:20] <alex_chally> you know, I am not entirely sure
[02:27:54] <alex_chally> I would think that you would split the current in half, but it could be pulsing one and then the other?
[02:28:05] <alex_chally> * alex_chally is not an EE
[02:28:08] <PartScavenger> http://www.kelinginc.net/KL23H286-20-8B.pdf
[02:28:13] <PartScavenger> Hehe, nor am I. ;)
[02:28:56] <PartScavenger> ...was the stepper I was planning to run with the said driver.
[02:31:54] <alex_chally> PartScavenger, take a look at page 5 of the datasheet
[02:32:00] <alex_chally> that is linked on the ebay page
[02:32:31] <alex_chally> it looks like there are two resistors, TQ1 and TQ2 that allow you to adjust the current output as a percentage of the whole
[02:33:01] <alex_chally> er, just pins
[02:33:04] <alex_chally> not resistors
[02:33:08] <PartScavenger> *looking...
[02:33:40] <alex_chally> I would be money that those DIP switches right next to the heatsink are the controls for those pins
[02:33:44] <alex_chally> bet money
[02:33:52] <alex_chally> I would be money too, coz I would have bet some money and won it
[02:36:11] <PartScavenger> LOL, You'd be right about the DIP switches I have seen them used on the internet to control the current.
[02:36:13] <alex_chally> PartScavenger, in fact, item 6 on that list "Calculation of the predefined output current" seems aplicable too
[02:36:44] <alex_chally> if only all of life's problems were so easily solved :D
[02:38:02] <alex_chally> PartScavenger, honestly the real answer is that you should totally go balls out on those
[02:38:11] <alex_chally> keep an eye on the temps of the motors
[02:38:53] <alex_chally> and you should be fine
[02:39:31] <alex_chally> but I am almost sure you can dump a bunch of current into those and have everything be fine
[02:39:44] <PartScavenger> It's not the motor's I'm worried about, it's this:
[02:39:45] <PartScavenger> http://www.mycncuk.com/forums/showthread.php?1373-eBay-TB6560-Stepper-Motor-Driver-Boards/page3
[02:39:50] <PartScavenger> Scroll mid page. =P
[02:40:45] <alex_chally> PartScavenger, so this is the part where he fucked up
[02:40:46] <alex_chally> This board is ment to be rated upto 36vdc, So I tried using my 36vdc 10A PSU.
[02:40:49] <alex_chally> is the quote
[02:41:05] <alex_chally> if you overvolt those chips you MUST have a current limited PSU
[02:41:11] <alex_chally> er
[02:42:10] <alex_chally> yeah, if you look at the datasheet for the drivers, they have a 3.5A max
[02:43:17] <PartScavenger> Oh ok
[02:43:23] <alex_chally> which means without a current limiting PSU
[02:43:41] <PartScavenger> So I guess it was unregulated and the voltage got too high,
[02:43:49] <PartScavenger> cause if you look at the spec sheets on some of the chips,
[02:43:53] <alex_chally> he could only put a load of 10.2 ohms on that before it would blow it up
[02:43:59] <PartScavenger> Voltage Max is 35V
[02:44:09] <alex_chally> remember V=I*R
[02:44:20] <alex_chally> so we have 36V=3.5A*R
[02:44:21] <PartScavenger> Yeah
[02:45:29] <alex_chally> in other words, having a current limited variable voltage bench PSU is awesome :D
[02:48:10] <alex_chally> PartScavenger, man, if I were designing that board I would have put the drivers in sockets
[02:48:34] <alex_chally> * alex_chally has a bag of dead easysteppers somewhere
[04:20:29] <elmo40> archivist: what was the picture for? you need to machine something? (IMG_0754.JPG)
[05:59:20] <archivist> elmo40, yup, may need to make a set of gears
[06:43:52] <bricofoy> hi
[06:45:21] <bricofoy> I've the bad feeling that I don't receive some of the ems-users emails. I sent one yesterday, don't received it. Is there a way to understat what happens ?
[06:53:53] <archivist> what was it about
[06:54:15] <bricofoy> classicladder / sequential editor
[06:54:35] <bricofoy> I'm looking in the list archive, it does not appear
[06:55:51] <bricofoy> I'm pretty sure there is a bug in the sequential editor that prevent using some functions
[07:37:55] <PartScavenger> @alex_chally Thanks so much for the help! Please forgive my manners, they were passing out chocolate cake, and I got distracted... :)
[07:38:46] <pjm> good morning
[07:45:11] <Valen> got an odd problem with tuning on our mill, i do a G1 F100 and tune FF1 to zero out the error during cruise, if i then do a G1 F200 the error is back and bigger, basically i can tweak it to hover around 0 error at a specific feed rate with FF1, but as i change speeds the constant for FF1 seems like it should change, wondering what i'm doing wrong
[08:01:51] <alex_chally> http://www.machineryforsale.com/images/TPC081276_Tahl%20Gap%20Bed%20Lathe.jpg
[08:01:52] <alex_chally> oh holy crap
[08:01:55] <alex_chally> I want a ride on toolpost
[08:02:14] <alex_chally> I wonder what Depth of cut they are taking....
[08:02:46] <bricofoy> x_x
[08:03:49] <alex_chally> that thing has a 230" swing
[08:03:54] <alex_chally> that is... 19 feet
[08:03:55] <alex_chally> and a bit
[09:11:22] <archivist> Valen, basicly tuning out at one speed takes little note of resonance, being critically damped at speed 1 and then running into overshoot at 2 so tune nearer to 2 or all over
[09:11:51] <archivist> tune the acceleration and stopping more than cruise
[09:55:16] <Valen> the acceleration etc is fine
[09:55:31] <Valen> I am just trying to null out the fixed offset
[10:09:13] <darren> darren is now known as redwizard-kde
[12:37:50] <jthornton> http://cgi.ebay.com/Hitachi-CNC-Ruffer-Mill-EMC2-Servo-Retrofit-/330446805833?cmd=ViewItem&pt=BI_Mills&hash=item4cf02a1b49
[12:38:44] <Valen> heh gotta like tux appearing on ebay
[12:38:52] <jthornton> yea
[12:39:02] <jthornton> you get your servos tuned up?
[12:39:05] <Valen> though a laptop
[12:39:14] <Valen> no, no idea what the problem is really
[12:39:38] <Valen> FF0 is position, FF1 is speed and FF2 is acceleration
[12:40:10] <Valen> its like I want FF1 to have an exponent or something
[12:41:15] <jthornton> I "think" I only used P I and FF1 to tune the lathe... I could check in a bit when I go outside
[12:41:31] <Valen> thats all your sposed to need
[12:41:41] <Valen> I added a little FF2 to help it start and stop
[16:49:17] <ChanServ> [#emc] "This is the #emc channel - talk related to the Enhanced Machine Controller and general machining. Website: http://www.linuxcnc.org/, wiki at http://wiki.linuxcnc.org/"
[16:51:04] <bricofoy> crash occurs on line 23 on my pastebin
[16:51:33] <bricofoy> the error on line 22 is at starting.
[16:52:53] <micges> bricofoy: you have emc from package?
[16:55:31] <bricofoy> micges, from mozmck experimental package
[16:55:42] <bricofoy> 2.4.0pre
[16:56:19] <micges> I see
[16:56:29] <cradek> weird, I don't even have a print button
[16:56:48] <bricofoy> hu !
[16:57:37] <cradek> I'm installing some packages and rebuilding (a guess)
[16:57:59] <cradek> #
[16:57:59] <cradek> inifile: warning: File contains DOS-style line endings.
[16:58:13] <cradek> you ought to fix your ini file and then never edit it on a windows machine again
[16:58:47] <bricofoy> I never edited it on a window machine, I don't understant where does this error comes from
[16:58:59] <bricofoy> I have only linux box in there
[16:59:12] <davidf> Hi
[16:59:18] <bricofoy> edited this file only with Gedit
[17:00:01] <cradek> did you do anything special to get print preview to crash? it works for me
[17:00:57] <bricofoy> http://imagebin.ca/view/kG0Siw.html this is what my classic ladder looks like
[17:01:34] <bricofoy> nope, just "print" then select "print preview" or "print" in the printer selection window.
[17:02:00] <bricofoy> but seems to work for smaller design, when I do the same on another section, it works
[17:02:02] <cradek> hm, that works here
[17:02:39] <bricofoy> I can pastebin my .clp file if you want to try with it ?
[17:03:15] <cradek> worth a try
[17:03:24] <cradek> filebin.ca
[17:03:31] <bricofoy> yep
[17:04:51] <bricofoy> http://filebin.ca/wpqzkr/custom.clp there it is
[17:07:39] <cradek> http://timeguy.com/cradek-files/emc/Screenshot-classicladder-preview.png
[17:08:25] <bricofoy> yeah for this section works ok for me
[17:08:35] <cradek> lunch
[17:08:42] <bricofoy> but try the same on the "sequential" section
[17:10:12] <bricofoy> ok, enjoy your meal
[18:12:38] <davidf> Got a question about a parport card -
[18:13:21] <elmo40> someone will reply ;)
[18:15:59] <davidf> It's a MCS9865 chipset card sold by Manhattan made in China as the name would imply. :(
[18:16:35] <davidf> Installation instructions don't work, not even close.
[18:16:54] <elmo40> fun.
[18:17:02] <elmo40> on a side note...
[18:17:22] <elmo40> i took a look at http://www.orchestracontrol.com/ Ran the liveCD they have. Not too impressed.
[18:17:31] <elmo40> i tried running things and they crashed
[18:17:43] <elmo40> or needed a login/passwrd to get in
[18:18:07] <elmo40> i couldnt even locate the CNC module. No clue what they are trying to do with this software.
[18:19:47] <davidf> back in a couple minutes
[18:25:32] <mrsun> yey it works =)
[18:25:37] <mrsun> all axis work perfectly =)
[18:25:44] <mrsun> so the mill is once again "whole" :P
[18:27:06] <davidf> OK The linux driver instructions don't work. My question is, do I really need the driver installed in order to use the card? Or does anyone know of a driver and installation instructions that will work? It should be pretty generic I think.
[18:27:36] <cradek> no, you probably don't need a driver
[18:27:53] <davidf> THanks cradek
[18:27:56] <cradek> if running emc2.4 it'll probably just show up in stepconf
[18:28:41] <bricofoy> cradek, did you try printing the sequential section of my clp file ?
[18:28:47] <cradek> no
[18:28:50] <davidf> cradek, when I do lspci -v I can see the card at something like $eob8 or some thing. But shouldn't the base address be $278 or $378?
[18:29:03] <cradek> davidf: no
[18:29:15] <bricofoy> cradek, because here the section you tried also work
[18:29:17] <davidf> OK. I have 2.3.3
[18:30:38] <cradek> bricofoy: I can't even switch to the sequential system to see it, without it crashing
[18:31:14] <cradek> davidf: you can use e0b8 in 2.3, or in 2.4 discovered parports get named like "1" and "2"
[18:31:42] <bricofoy> cradek, hu ! this is weird
[18:34:13] <davidf> Also I see a line with something like parport.1 OUT parport.2 IN
[18:34:40] <davidf> cradek, sorry I'm not at that computer - it's in the shop and not online.
[18:35:38] <davidf> Card doc says it handles epp modes etc. Can I set it for IN or OUT without driver software?
[18:35:51] <cradek> yes, hal_parport takes care of that for you
[18:35:59] <davidf> cool!
[18:36:17] <davidf> Awesome in fact. :)
[18:37:23] <davidf> It's getting harder to find PCI parallel cards. What happens when we can't get them?
[18:38:12] <cradek> I disagree with your premise, and also point out that there are PCI-E parallel port cards, and also point out that Mesa is currently making new PCI-E hardware in addition to all their other options
[18:38:34] <davidf> I stand corrected! Thanks
[18:38:50] <cradek> Search Terms: pci parallel port
[18:38:51] <cradek> (1-20 of 67 Results)
[18:39:01] <cradek> (a search on newegg.com)
[18:39:58] <davidf> OK I should check out Mesa.
[18:40:43] <cradek> mesa has a variety of awesome stuff that works nicely with emc
[18:40:59] <davidf> What's the url?
[18:41:25] <cradek> http://mesanet.com/
[18:42:09] <davidf> Thanks. When I said hard to find, I really meant at retail brick & mortar stores. It's pretty much all USB.
[18:42:27] <cradek> it's hard to find anything useful at those stores :-(
[18:43:12] <cradek> I'm much more likely to find what I need in my basement than a best buy store :-)
[18:44:15] <davidf> HA. I can see that.
[18:46:54] <bricofoy> I just find the words to explain why grafcet is interesting, and is complementary to ladder. Grafcet is a _simple_ way to implement state machines.
[18:48:48] <davidf> bbl
[18:57:56] <bricofoy> cradek, please (if you have time) try with this file : http://filebin.ca/wqbrzs/custom.clp
[18:58:23] <bricofoy> previous one had a design mistake, may be this is why it crashed for you
[19:02:23] <bricofoy> cradek, please just tell me if you have time to test this now. Not a problem if you don't, but please tell me because I have to go. But I can stay a while if needed.
[19:08:22] <bricofoy> cradek, ok, you're afk, so I assume you don't have time to test. So I go. see you tomorow :)
[19:08:30] <bricofoy> good night !
[20:33:24] <frallzor> what to mill in wood
[20:40:43] <andypugh> lepton: You were talking about a way to "numptify" the EMC2 interface? This is what my Lathe config looks like most of the time, running the same bit of adaptable G-code all the time in semi-manual mode.
[20:40:44] <andypugh> http://www.bodgesoc.org/lathe/lathe.html
[20:43:26] <lepton> andypugh: Yeah, that looks quite a lot like what we're invisioning
[20:43:51] <andypugh> Feel free to steal any of it you fancy
[20:43:58] <lepton> great! thanks
[20:44:43] <lepton> We're currently finishing up the wiring of all our I/O, and once we've got it electrically tested my last major task is to finish up the gui
[20:49:44] <cradek> andypugh: I've seen jmkasunich do a similar thing but he just edits numbers in the file
[20:50:31] <andypugh> He is probably more of a CLI geek to my GUIboi..
[20:51:25] <cradek> it'd be interesting to think of a way this could be done in touchy (with the wheel used to change the values)
[22:26:32] <andypugh> Well, my milling machine now has a drawbar, with one minor flaw.
[22:27:25] <andypugh> The only way to release it is to swivel the head all the way to upside-down and use the table to push the release plunger.
[22:40:17] <cradek> andypugh: ouch.
[22:41:27] <cradek> is the design defective or just not yet done?
[22:42:23] <andypugh> Not done. I need a pneumatic cylinder. Oh, and a compressor.
[22:43:21] <cradek> lever?
[22:43:55] <LawrenceG> funky new kind of tool changer.... easier to put a new tool in when its on top!
[22:44:41] <cradek> bbl
[22:46:10] <andypugh> A lever would work, but would need engineering itself, and I am not totally sure that I want to push against the top bearing. I would rather "pinch" to release. Though I am not sure how much it matters, it is a big bearing. However I notice that Tormach went to great trouble to avoid it.
[22:56:40] <JT-Hardinge> wow the forum has > 2000 members
[22:59:27] <andypugh> Do any answer questions rather than ask them?
[22:59:54] <JT-Hardinge> just a few and some demand an answer right away :)
[23:00:04] <JT-Hardinge> like they paid for the answer
[23:01:00] <JT-Hardinge> I'm running crippled atm with only 4 stations on the turret that work :/
[23:02:50] <andypugh> I really don't like forums, it is really difficult to read through everything in sequence.
[23:03:57] <JT-Hardinge> it's better if it is out of sequence?
[23:04:48] <andypugh> I am a completist, I hate to think I have missed something. That's why I like mailing lists. (I used to read all of rec,motorcycles every day too, which was a challenge)
[23:05:30] <JT-Hardinge> I guess I'm confused how could you miss anything on the forum?
[23:08:35] <JT-Hardinge> oh bother, the menus don't work after the last upgrade
[23:08:48] <andypugh> It isn't that, it is that you have to look for it. Search for new posts, click them one-by-one...
[23:09:43] <JT-Hardinge> yea, that is a real chore to scroll down a bit to see what sections have new posts :P
[23:09:43] <andypugh> It is manageable with the EMC forums, but anything half-way busy (like CNCzone for example) and there is no knowing what you are missing or how many parallel threads are covering the same issue.
[23:11:00] <JT-Hardinge> I have the same feelings about the mailing list and usually just delete them all as there is too much traffic to try and sort out the order and follow a topic because the date/time is wrong for all of them
[23:11:30] <JT-Hardinge> so you end up reading the end then message 22 then message 29 then 25...
[23:15:04] <andypugh> I find that a gmail account for mailing lists makes it all very pleasant.
[23:15:47] <JT-Hardinge> that is what I have but they still all come in out of order... can you remember what M1.. code I used to open my collet?
[23:17:30] <andypugh> Odd, they don't come in out of order for me. Maybe you ISP is playing games?
[23:19:24] <JT-Hardinge> dunno
[23:27:10] <andypugh> Right, night all.