#emc-devel | Logs for 2009-10-16

[03:00:09] <cradek> PCW: thank you! I will try to work on it this weekend. This would be very cool.
[11:09:56] <snegovick> hello, id like to get some help - im developing a usb cnc device, with g-code parser embedded in device, so i dont need rtlinux for work, but current build system doesnt support any non-rt devices, would you be interested in adding such a device to your source tree ?
[11:14:57] <micges_work1> micges_work1 is now known as micges
[12:49:42] <CIA-82> EMC: 03bigjohnt 07v2_3_branch * r445161ebf0d1 10/docs/src/config/ini_config.lyx: Add a bit to Geometry description
[12:51:05] <CIA-82> EMC: 03bigjohnt 07master * re2fd57b2510c 10/docs/src/config/ini_config.lyx: Add a bit to Geometry description
[14:42:52] <jepler> BJT-Work: thanks so much for continuing to work on the documentation
[14:53:21] <BJT-Work> np, jepler thanks for your work on the code
[14:53:45] <jepler> there hasn't been so much of that lately
[14:53:53] <jepler> one of the reasons I admire your dedication all the more
[15:17:50] <BJT-Work> thanks
[15:18:22] <BJT-Work> there should be some excitement in my shop soon, I'm hooking up the 5i20 with the THC to the plasma this weekend
[15:18:37] <cradek> cool
[15:19:12] <cradek> here too - I might work on the sub-period probing stuff - pcw made a firmware for me last night.
[15:20:52] <BJT-Work> I saw that
[15:23:15] <BJT-Work> and wondered what it was exactly :)
[15:24:02] <cradek> when probing, emc checks the probe input every servo cycle and if it has fired, it saves the position right then, stops the machine, and reports it to the interpreter so it ends up in #variables.
[15:24:29] <cradek> if you want the ultimate possible resolution, that means you can probe at maximum speed that gives one encoder count per servo cycle
[15:24:32] <cradek> brb
[15:24:42] <BJT-Work> cool
[15:29:02] <snegovick> hello again , im sorry for my persistence, but i wanted to contribute to your project, and i have a question about non-rt devices support by emc, maybe you are interested in adding support for such devices to the emc ?
[15:32:37] <micges> what kind of devices?
[15:33:58] <snegovick> im working on cnc device where all rt work is done on separate MCU, this device has USB interface for communication with pc
[15:35:30] <micges> what input data your device required?
[15:35:37] <snegovick> g-code
[15:35:58] <snegovick> it has its own scheduler
[15:36:16] <snegovick> but it doesnt have anything to monitor its state
[15:36:19] <jepler> It's impossible to say without seeing the code. If it turns out to fit nicely, then that makes it more likely to be acceptable.
[15:36:44] <micges> bbl
[15:38:16] <jepler> it is hard to see how a USB device like you describe would fit into emc2 and hal
[15:39:10] <snegovick> yes, thats what bothers me too, the idea of our device is to make it work on non-rt host machine
[15:39:55] <jepler> with git you can maintain your own modified software based on emc, and take new changes and improvements from us as long as possible.
[15:40:20] <jepler> git will also let us see your work and let us integrate it if there is a consensus that it should be
[15:40:53] <snegovick> thank you for answers :)
[16:40:38] <skunkworks411> http://www.cnczone.com/forums/showthread.php?p=678675#post678675
[16:41:23] <skunkworks411> http://www.cnczone.com/forums/showthread.php?t=91570
[16:41:40] <skunkworks411> ^what I meant to paste
[16:42:50] <skunkworks411> (small edit) http://www.cnczone.com/forums/showthread.php?t=91570
[17:03:43] <cradek> why would you need or want to measure 4 points to determine a rotation?
[17:07:45] <SWPadnos> so you can skew the output too?
[17:07:48] <cradek> edgefind front left, touch off Y=0, move over 10", edgefind, read new Y, rotate by Y/10 radians (but ugh, gcode uses degrees)
[17:08:18] <SWPadnos> it would be nice to have a version of G10L2 that could take X1Y1/X2Y2, and do the math for you
[17:08:36] <cradek> SWPadnos: I am baffled trying to think of a reason you'd want the final work distorted according to the original block's out-of-squareness
[17:08:36] <SWPadnos> XY = new center, IJ = another point
[17:08:42] <SWPadnos> yeah, you wouldn't
[17:08:55] <SWPadnos> you don't even want 3 points
[17:09:03] <cradek> no, obviously you want 2
[17:09:07] <SWPadnos> unless there's plane correction as well
[17:09:09] <SWPadnos> Z plane
[17:09:23] <SWPadnos> but I'd be really surprised if that's in there
[17:09:55] <cradek> maybe the person is asking about compensating for an out-of-perpendicular XY machine
[17:10:07] <cradek> but, he says "orient the workpart"
[17:10:19] <SWPadnos> no, I'd bet from the screenshot that they're talking about part orientation
[17:10:23] <SWPadnos> right, and he says so as wel :)
[17:10:25] <SWPadnos> l
[17:10:46] <cradek> I can imagine you could put a known-square thing on the table and by probing four points compensate for an out-of-square machine (but I doubt mach does this)
[17:11:08] <cradek> well it's ok - not the first time I've been baffled by something on a web bbs.
[17:11:32] <SWPadnos> it would be a bitch to do that in addition to coordinate system rotation about an arbitrary center
[17:12:54] <cradek> not sure what you mean by Z plane correction
[17:13:09] <cradek> compensate for the tool not being perpendicular to the table?
[17:13:18] <SWPadnos> well, it's kind of impossible, but correcting for a workpiece that isn't flat
[17:13:20] <SWPadnos> err
[17:13:21] <SWPadnos> not flat
[17:13:53] <SWPadnos> yes, but substitute workpiece for table in that sentence
[17:14:04] <SWPadnos> which is why it's not possible - holes wouldn't be perpendicular
[17:14:05] <cradek> but ... but ...
[17:14:10] <cradek> if you can't tip the tool
[17:15:26] <cradek> can a german speaker tell us what that screen does?
[17:15:52] <micges> alex?
[17:20:21] <cradek> drehen is rotate, abbrechen is cancel
[17:20:33] <jepler> you can either measure angle 1 or angle 2 or both
[17:20:44] <SWPadnos> I'm thinking that may be a wizard that issues G68/G69 commands for you
[17:21:00] <jepler> you choose the angle priority: angle 1, angle 2 or (I think) average
[17:21:22] <cradek> yes mittelwert = averaging
[17:21:24] <jepler> so you can use it when it's convenient to measure the angle from P1P2 or the angle P3P4
[17:22:56] <SWPadnos> I just looked through one Mach manual, and didn't see that screen anywhere, nor mention of any helper for G68/G69 (which is what Mach uses to enable and disable coordinate system rotation)
[17:23:20] <SWPadnos> would it make sense to add a modeto G10L2 that uses IJ instead of R?
[17:23:24] <cradek> it has a button you can press to get slapped
[17:23:39] <cradek> or be abducted by aliens in a flying saucer
[17:24:17] <jepler> I can't see what use the average would be
[17:24:42] <cradek> nonetheless it appears to be the default behavior
[17:25:12] <SWPadnos> if you have a non-square part, you can measure both edges and get something that might be slightly less inaccurate than either separately
[17:25:17] <cradek> SWPadnos: but a wizard or other gcode program can calculate it so easily...
[17:25:27] <SWPadnos> snark snark snark
[17:25:29] <SWPadnos> :)
[17:27:10] <cradek> it will be very obvious to any machinist who knows how to use a sine bar how to determine the rotation
[17:27:30] <cradek> if you can't use a sine bar, you probably can't use an edgefinder anyway, so you're stuffed
[17:27:55] <cradek> (and to think, you only typed "snark snark snark")
[17:28:19] <SWPadnos> heh
[17:28:38] <SWPadnos> I don't know how to use a sine bar (and don't have one), but I can certainly touch off two points easily enough
[17:28:56] <SWPadnos> I can also write a macro to do the ATAN2 function I'd need, but then again, I have a computer that can do that for me
[17:28:59] <cradek> but do you know rudimentary trig?
[17:29:10] <SWPadnos> depends on coffee intake :)
[17:29:59] <cradek> yes it's spelled ATAN[][] in gcode :-)
[17:30:19] <SWPadnos> is it really ATAN2 though?
[17:30:27] <cradek> seriously, if you want to add another G10 mode, have at it
[17:30:30] <cradek> yes it is
[17:30:36] <SWPadnos> ok, cool (on both)
[17:30:50] <cradek> er maybe it's spelled ATAN[]/[]
[17:30:59] <cradek> whatever (I know it is definitely atan2)
[17:31:04] <cradek> bbl, lunch!
[17:31:07] <SWPadnos> see you
[17:31:13] <alex_joni> * alex_joni is around now
[17:31:27] <alex_joni> is the translation still actual?
[18:10:40] <skunkworks411> wow - nice discussion :)
[18:12:32] <skunkworks411> and his reply 'as i understand your answer, there is no functionality comparable with the mach3 alignment possibilities integrated in ECM2. Thanks for your answer.'
[18:18:32] <SWPadnos> huh, that's a strange thing to take away from your comment and links
[18:18:51] <SWPadnos> the only thing missing is a screen that lets you touch off the points
[18:19:15] <micges> maybe he just knew before he asked ?;)
[18:30:01] <cradek> yes, what a very strange conclusion
[18:35:56] <skunkworks411> I like the fact that it looks like mach waves to you.
[18:36:16] <cradek> well I saw it as a slap
[18:36:28] <skunkworks411> heh
[18:36:40] <cradek> also I wondered why I'd want the flying saucer to abduct me
[18:36:54] <cradek> I don't know what the center one does, but considering the other two, I wouldn't push it
[18:40:16] <SWPadnos> it's kind of like a rebus
[18:40:22] <SWPadnos> "place the center punch in the hole"
[18:42:27] <jepler> rebus? you mean the monkey?
[18:43:21] <SWPadnos> no, the country singer
[18:43:36] <SWPadnos> or brother of rombulus
[19:01:26] <robh> i think that CNCzone post is trying to ask if there is a way to edge find two Y points, and have emc auto rotate your work offset for you. so you can put on a jig/vice etc not clock it, tough on two point have cnc realign your work, lazy mans way out of using a clock etc to align your work before you set the datums
[19:02:19] <cradek> you could easily write a gcode program to do this
[19:02:55] <robh> yes its like if u have renishaw software installed on your CNCmill, or have the matsura one, u get that function, along with millions of other probe cycles
[19:02:59] <cradek> find one point, save it in vars (say, use g28.1) find the other point, save it in vars (say, use g30.1), and do the very simple trig in gcode
[19:03:10] <cradek> sure if you have a probe, it can be all automatic
[19:03:25] <cradek> (in emc)
[19:03:43] <robh> yep
[19:05:30] <robh> i always clock my vice or fixure plate up so iv never realy had a bad need to use it
[19:06:11] <robh> i guess if u want to throw a vice on to do 1 or 2 parts, it might be handy put program in, and prob and go
[19:06:13] <cradek> my vise is keyed - I can indicate it in pretty fast, but why bother every time?
[19:07:18] <robh> only ever time u take it off put it on i mean
[19:07:40] <cradek> it's on and off constantly - small mill
[19:08:18] <cradek> my problem now is the new mill has bigger T slots than the bridgeport (that the keys fit)
[19:08:26] <robh> or if u have a odd part u make a fixure for and ur lazy and dont have end stop, u can probe fixure in vice and go i guess each time but if ur doing loads of parts ur going to end stop it or u have a long cycle time, theres numbers of things i guess
[19:08:59] <robh> all our vice arnt keyed not that u like to lug them on and off alot big hydrolic vices
[19:09:08] <cradek> * cradek laughs at "ur lazy"
[19:09:24] <cradek> it's hard to read your text with so many letters missing
[19:09:54] <robh> sorry
[19:10:08] <cradek> I bet the keys only work well if the vise is only used on one machine
[19:10:08] <robh> i will try type full english
[19:10:26] <cradek> cool, that seems to be the custom here
[19:10:43] <robh> my spellin aint to grate so you will have to excuse that though
[19:10:57] <cradek> if you had a bunch of machines, they'd probably all be different and the keys would only fit one
[19:11:16] <robh> yea all our mills have the same Tslot size and spacing, as same sub plate. if you take that off they are different then on one
[19:11:17] <cradek> I made them a tight fit - it just slides down into the T slot with a gentle push
[19:12:23] <robh> yea, we have one on our manual mill, it works grate. if we want to do something super critical i will clock it to check but its never a mile out.
[19:13:48] <cradek> yeah I bet it's always good to check, if it's important
[19:16:05] <robh> like when you use auto probe cycles to check bores or find the center of a part, you still have to tell the cnc where it is and a size to start with, by the time you have done all that you can probly do it manuly
[19:17:01] <cradek> a couple days ago I wrote a program for emc to probe a hole center (and show the diameter, which you get automatically)
[19:17:13] <cradek> you do not have to give it any information to start - only jog the probe tip into the hole.
[19:17:49] <robh> grate
[19:18:11] <robh> so if you could probe the bottom also then you can work out the taper also
[19:18:51] <cradek> I thought about that - would be neat if it would show you the taper after probing at two heights. just like the traditional two-ball-depth-micrometer technique
[19:19:07] <cradek> but the math is not real simple, so I didn't write it right away
[19:19:34] <cradek> so I will put it off until/unless I need it!
[19:20:07] <robh> would be nice to see them programs just to see how you did if possible
[19:20:11] <cradek> I have a project where I need to measure an outside taper very precisely to make a mating part... not sure how I will do it yet.
[19:20:19] <cradek> sure, one sec
[19:20:48] <cradek> http://timeguy.com/cradek-files/emc/P-hole.ngc
[19:21:15] <cradek> hole probe - but be careful if you try it - I have changed it since I last tested it
[19:21:47] <robh> abit like when people start telling me how they use there CNC to CMM on. but i they never belive you when you tell them there machine can not be that accurate they belive it is which is where alot of people fall over when get probe on there machine
[19:21:57] <cradek> the max radius is set to less than the max travel of my probe for testing/safety. for a bigger hole I would need to increase it.
[19:22:26] <cradek> robh: I probed a 4" gage block on mine and got 3.9999 with no tweaking
[19:22:43] <robh> did you get your renishaw working then
[19:22:48] <cradek> yes
[19:22:52] <robh> grate
[19:23:00] <robh> i have yet to do mine, on a long list of things
[19:24:48] <robh> find out all the low bat and turn on signals
[19:25:00] <cradek> yes
[19:25:44] <robh> which did you have again, the omm or omi pickup
[19:25:51] <cradek> omm I think
[19:26:08] <cradek> to MI12
[19:26:48] <robh> aah you have the interface unit
[19:27:05] <cradek> there is a LOT of stuff in the MI12
[19:27:16] <robh> im missing that unit
[19:27:25] <cradek> the OMM talks some kind of serial protocol I think
[19:27:35] <cradek> ouch, they are $1k+
[19:27:58] <robh> yea, OMI you can talk direct i know
[19:28:36] <robh> it would be £ for me
[19:29:35] <cradek> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=250434353344
[19:30:07] <cradek> you maybe could get it for $300 with make offer
[19:30:26] <robh> nice find
[19:30:34] <cradek> 2 available!
[19:31:26] <cradek> renishaw quoted me $931 for MI12 in June 08
[19:31:30] <robh> yea i did wunder what was in it, wether it was a 5v to 24v unit so you can interface to CNC as they are 24v or 12v
[19:31:49] <robh> no dout they gone up alot now like most things
[19:32:01] <cradek> it takes 24v and OMM wires in, gives some opto outputs
[19:32:06] <robh> ill get a nice coutoms import on that
[19:32:28] <cradek> I think MI12 has a lot of smarts in it
[19:32:37] <cradek> you will not be able to just hook up OMM without it - no way
[19:32:49] <robh> the board only version looks very complex
[19:33:01] <cradek> yep
[19:34:30] <robh> thirst thing to do is put mill back together i think
[19:35:08] <robh> did you add a brake resistor to your inverter you got for JR
[19:35:25] <cradek> no, it stops fast enough without
[19:36:01] <robh> untill we added one i was limited to 3.7sec from 6k to 0
[19:37:13] <robh> as we had to fit a new inverter , added brake resitor 10ohm 1.5Kw and stops in 1.6sec from 6K might do harder but cant see much point its as it was now
[19:38:49] <robh> i gtg now, nice chatting
[19:38:57] <cradek> bye
[19:48:12] <BJT-Work> ever hear of this one http://www.maxnc.com/3d%20Digitizing.htm
[19:49:00] <skunkworks411> you could make one probably as good...
[19:49:03] <cradek> I actually have one (somewhere)
[19:49:09] <skunkworks411> cheaper
[19:49:10] <cradek> yeah it is homemade quality
[19:49:21] <BJT-Work> ok, just wondering
[20:06:49] <BJT-Work> http://grandrapids.craigslist.org/tls/1396323832.html
[20:15:03] <cradek> great deal if it comes with the interfacing stuff
[20:15:10] <cradek> if you can see it work anyway
[20:20:21] <BJT-Work> and if you live in grand rapids
[20:20:43] <BJT-Work> time for me to head out talk to you later