#emc-devel | Logs for 2008-10-04

[00:47:29] <CIA-40> EMC: 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/drawing.c: Back to the thin blue line
[00:55:58] <CIA-40> EMC: 03cmorley 07v2_2_branch * 10emc2/src/hal/classicladder/drawing.c: Improve edit mode grid colour
[03:09:14] <seb_kuzminsky> hidee ho
[03:22:24] <cradek> hi seb
[03:22:31] <cradek> you always show up late
[03:38:06] <seb_kuzminsky> hi cradek
[03:38:14] <seb_kuzminsky> i come on after the kids go to bed ;-)
[03:38:18] <cradek> ahh
[12:14:47] <CIA-40> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/gui/axis.lyx: add a bit about the menu and a few spello's
[12:19:41] <CIA-40> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gui/axis.lyx: add a bit about the menu and a few spello's
[15:57:14] <rayh> fascinating. with today's emc2 version (2.2.6-1) I get both tkemc and axis when I start demo-step-cl
[15:57:37] <rayh> on my athalon 64 box
[15:59:27] <SWPadnos> that's odd
[16:19:19] <rayh> uh huh.
[16:20:42] <rayh> Ah. The first time I started it it failed. But it must have left the AXIS gui around and when I restarted it connected them both.
[16:21:01] <SWPadnos> that's odd
[16:21:03] <SWPadnos> :)
[16:21:06] <rayh> The failure was a config from a different version.
[16:22:23] <rayh> Let me see if I can reproduce the problem.
[16:22:49] <rayh> When I just stopped and restarted it it worked as expected.
[16:23:05] <rayh> But bot GUI's were connected and running.
[16:24:43] <rayh> There it is. Back with both displays.
[16:25:12] <SWPadnos> strange. how did you stop the first instance?
[16:25:36] <SWPadnos> (presumably, that's the one leaving AXIS around for the second instance)
[16:25:36] <rayh> Now I'll have to figure out what is going wrong with shutdown and restart when "emc is running, restart it?"
[16:25:46] <rayh> Yep.
[16:25:56] <SWPadnos> does the first config run two GUIs?
[16:26:05] <rayh> The odd thing is that it must be keeping and using the first NML.
[16:26:27] <rayh> It is a custom thing I made using stepconf.
[16:26:35] <SWPadnos> the NML stuff probably hangs around until everything using it has gone away, which never happens
[16:27:09] <rayh> Okay. And the second finds NML running and connects to it.
[16:27:21] <SWPadnos> right
[16:27:29] <SWPadnos> at least, that seems like a plausible reason ;)
[16:27:35] <SWPadnos> gotta run. good luck with it
[16:28:53] <rayh> Catch you later.
[16:29:23] <rayh> The stepconf config I was using had been modified outside of stepconf so I'll have to dig.
[17:05:58] <jmkasunich> hi rayh
[17:32:22] <skunkworks> wow - that email is the one I was talking about a week ago. (sent to the list by mistake)
[17:32:54] <jmkasunich> I think there is dirt in the tubes
[17:33:59] <skunkworks> I guess :)
[17:35:22] <jmkasunich> so how come no servo amp makers have prices listed?
[17:35:52] <skunkworks> because there are not enough of them.
[17:36:06] <skunkworks> (servo amp makers)
[17:37:06] <jmkasunich> even for plain ole brush motors?
[17:38:05] <jmkasunich> people like Advanced Motion Control and Servo Dynamics don't list prices - do they sell thru distributers or what?
[17:38:59] <skunkworks> I am sure that is part of it. When I was searching for a certain part number made by amc - I know I found a few re-sellers.
[17:40:49] <skunkworks> jmkasunich: what is the project?
[17:41:10] <jmkasunich> maybe nothing, but I've been contacted by someone with a 6" horizontal boring mill
[17:41:23] <jmkasunich> it's local to me
[17:41:34] <skunkworks> Neat. trying to get some prices togather?
[17:41:53] <jmkasunich> well, I won't have any concrete info till next week when I've seen the machine
[17:42:05] <jmkasunich> but I've been told that it is brush-DC motors
[17:42:24] <jmkasunich> and I'm just trying to find the ballpark price if we choose to replace the ancient GE HI-AK drives
[17:44:46] <jmkasunich> something like an AMC 50A20I is almost certainly big enough, maybe too big)
[17:44:58] <jmkasunich> it would be good to know if that is a $400 item or a $2000 item
[17:46:05] <skunkworks> http://www.electromate.com/products/series.php?&series_id=100309
[17:46:24] <jmkasunich> cool, thanks
[17:47:00] <jmkasunich> I was expecting prices in that neighborhood, but wanted something more concrete than my best guess
[17:47:19] <jmkasunich> oh, those are PWM input
[17:47:24] <jmkasunich> similar to what you are making
[17:47:30] <skunkworks> I guessed - I think they sell both
[17:47:33] <skunkworks> yes
[17:48:27] <jmkasunich> analog input is a little more: http://www.electromate.com/products/series.php?&series_id=100313
[17:49:23] <jmkasunich> heh, that seller is in canada
[17:49:51] <jmkasunich> seems bass-ackwards to buy a US made product from Canada for delivery into the US
[17:54:07] <skunkworks> :)
[17:54:36] <jmkasunich> if nothing else that link is a good starting point - if/when it is time to buy then contact AMC directly
[17:58:05] <skunkworks> exactly
[17:59:32] <skunkworks> http://www.servosystems.com/amc_pwm.htm
[17:59:41] <skunkworks> no pricing but US
[17:59:57] <jmkasunich> yeah, I found them already
[18:00:57] <jmkasunich> I think I'm set for now - there really isn't much more I can do till I see the machine and meet all the people involved
[18:01:11] <skunkworks> make sure to look at the brushes.
[18:01:20] <skunkworks> That might be a bit of money also
[18:01:34] <jmkasunich> yeah, but that will be the case no matter what they do
[18:01:49] <jmkasunich> I'm sure brushes are cheaper than replacing the motors with AC
[18:01:59] <jmkasunich> the machine works now, with an ancient AB control
[18:02:41] <jmkasunich> but it apparently is getting to be a wee bit unreliable :-)
[18:03:29] <jmkasunich> the way I see it there are three options
[18:04:08] <jmkasunich> 1) find a point where you can get at the +/-10V and encoder signals, keep the drives, and design a "quick swap" emc thing
[18:04:37] <jmkasunich> so you minimize downtime and can swap back if the retrofit has to work around production
[18:05:02] <jmkasunich> 2) replace the drives (which are as ancient as the control and probably as unreliable)
[18:05:12] <jmkasunich> 3) replace drives and motors with brushless or AC
[18:05:22] <jmkasunich> the last one would be extremely expensive
[18:05:36] <jmkasunich> if the motors are sound, that would be stupid IMO
[18:19:53] <rayh> Hi jmkasunich. Sorry was away for a bit.
[18:21:54] <rayh> Autumn leaves are so awesome this year. I have a difficult time being indoors.
[19:49:36] <alex_joni> skunkworks: lol
[20:13:23] <jmkasunich> rayh: I was gonna ask about the servodrives on the mazak and where you got them
[20:14:06] <jmkasunich> skunkworks pointed me to some answers, although I think the mazak used servo-dynamics, not AMC
[20:48:36] <issy> hi all
[21:11:17] <rayh> Ah. gone again. The drives are servo dynamics and Dave E got them for me.
[21:12:25] <rayh> The specific servo dynamics drives we used were nearing the end of their production run.
[21:12:46] <rayh> The replacement includes an AC power supply but costs a couple hundred more.
[21:13:00] <jmkasunich> I see
[21:13:14] <jmkasunich> you know anything about GE Hi-AK drives?
[21:13:28] <jmkasunich> other than "they're old" ;-)
[21:14:21] <jmkasunich> from what I've found, they are analog brush DC, tach feedback to the amps, and probably resolver feedback to the controls
[21:15:11] <rayh> I've worked on a few of those.
[21:15:33] <rayh> Sounds about right.
[21:15:57] <rayh> They did make an AC spindle drive under the HI-AK name also.
[21:16:11] <jmkasunich> if you had a machine with them, would you try to keep them running (good old robust stuff), or replace them (unobtainium parts)
[21:16:13] <rayh> They were tach feedback
[21:16:49] <rayh> Well if they happen to be the same as the drives Chris has in his Hardinge, I've got a couple left.
[21:17:03] <jmkasunich> probably not
[21:17:08] <jmkasunich> this is a much bigger machine
[21:17:29] <rayh> They are at the age when all the electrolytic caps ought to be replaced.
[21:17:45] <rayh> Otherwise tuning is a weekly thing.
[21:18:50] <jmkasunich> http://jmkasunich.com/pics/HBM-P1010003.JPG
[21:19:02] <jmkasunich> next weekend I should be able to look inside that box
[21:19:04] <rayh> I threw out some kollmorgen drives last year. Newer than the Hi-AC but they would drift so bad.
[21:19:21] <jmkasunich> "Transistorized Servo Drive"
[21:19:32] <jmkasunich> the machine: http://jmkasunich.com/pics/HBM-P1010004.JPG
[21:19:39] <jmkasunich> biam
[21:20:50] <rayh> I don't think that will fit in your basement!
[21:23:31] <jmkasunich> back
[21:23:36] <jmkasunich> its not mine
[21:23:45] <jmkasunich> a local shop is considering EMC for a retrofit
[21:27:25] <rayh> During the time when this machine was made, GE had a really nasty habit of stamping their part number over the real numbers.
[21:27:47] <jmkasunich> buy it from us at a 200% markup
[21:28:06] <rayh> It would take more time if you replace the Hi-AC for tuning but it would make for a much more dependable machine when done.
[21:28:23] <jmkasunich> that was my thought too
[21:28:44] <jmkasunich> a lot depends on how much downtime they can spare
[21:28:52] <rayh> Sure.
[21:29:14] <jmkasunich> in theory, if the retrofit is done at the +/-10V level, it can be done as a plug-in
[21:29:21] <rayh> Maybe it's a mini fest time again.
[21:29:37] <rayh> This is just 3 axis horizontal?
[21:29:47] <jmkasunich> so if the EMC retro isn't complete and a job comes in, you plug the old control back in
[21:30:05] <jmkasunich> I think its 4 - quill and table move parallel to each other
[21:30:14] <rayh> Ah okay.
[21:30:35] <jmkasunich> they also have a rotary table, but from what he told me, its more of an indexing table, not a rotary axis
[21:30:48] <jmkasunich> unlock, index (pneumatic), lock, cut
[21:31:11] <rayh> Many of these were. They sat down on a curvit(sp) coupline
[21:31:28] <rayh> Sorta like teeth meshing like 1 degree.
[21:31:45] <jmkasunich> I'll take a look at that while I'm there too
[21:32:08] <rayh> Many of these machines used a g45 to spin the indexer.
[21:34:15] <jmkasunich> sure will be hard to put a display in that pendant ;-)
[21:34:32] <jmkasunich> cut an LCD into strips and mount them end-to-end
[21:34:42] <alex_joni> 7" monitor?
[21:34:54] <jmkasunich> I bet its not even that wide
[21:35:12] <jmkasunich> 4-5" maybe
[21:35:16] <alex_joni> an LCD attached to the pendant?
[21:35:26] <alex_joni> maybe 2 to balance it :)
[21:35:54] <jmkasunich> or use the pendant only for "pendanty" things - jogwheel, FO, and some buttons
[21:35:56] <rayh> I've got a couple of wide screen automotive displays here. Been meaning to get one running for a machine I'm building.
[21:36:18] <jmkasunich> here is the main control screen: http://jmkasunich.com/pics/HBM-P1010002.JPG
[21:36:52] <jmkasunich> other than the keyboard, there are more buttons on the pendant than on the main control
[21:37:40] <alex_joni> what are those flip switches at the bottom?
[21:37:41] <rayh> Wow. That looks a lot like a 1050. I've got at least three matching displays out in the garage.
[21:38:08] <jmkasunich> alex_joni: so far all I have is these three pics
[21:38:18] <jmkasunich> I will see the machine and talk to the owners a week from today
[21:38:19] <alex_joni> rayh: the extra monitor up there might mean they need one :)
[21:38:44] <jmkasunich> he says the control is a 7320
[21:38:51] <rayh> Give 'em a real good deal.
[21:38:56] <jmkasunich> is a 1050 a control, or just a user interface panel?
[21:39:08] <rayh> 7320 sounds like AB.
[21:39:13] <jmkasunich> it is AB
[21:39:31] <rayh> Mine were 7360.
[21:40:11] <rayh> Hell to work on. But wow they had a nice bitwise computer interface.
[21:40:40] <jmkasunich> the diskette drive on the top looks a lot newer
[21:40:59] <rayh> You could set analog feed values right down to a couple millivolts
[21:41:04] <rayh> Got to be.
[21:41:16] <rayh> Nothing like it on the ones I had or serviced.
[21:41:32] <jmkasunich> paper tape only, or serial drip-feed?
[21:42:08] <rayh> They were paper tape fed but had enough memory to store several modest programs.
[21:42:38] <jmkasunich> I imagine this machine doesn't do 10,000 line contouring programs ;-)
[21:43:26] <rayh> Nope. It might with drip from a cassette.
[21:44:04] <jmkasunich> that box on top is labeled "CNC Minifile"
[21:45:29] <rayh> Finding out the machine's duty cycle and available back is one of the first things to do.
[21:45:40] <rayh> Then you can begin to plan a fit.
[21:46:07] <jmkasunich> yeah - there is a world of difference between "you get one week" and "it will be down for a couple months"
[21:49:22] <rayh> I can't quite read the legends on the pendant but most of them look like they are momentary.
[21:49:43] <jmkasunich> yeah
[21:49:55] <jmkasunich> a few toggles, some lights? and a bunch of buttons
[21:50:48] <rayh> We could develop a "crew" from guys on the lists and break it up into sections. Like pendant, ladder, machine IO, and such.
[21:50:57] <jmkasunich> heh
[21:51:25] <jmkasunich> actually, from what I've been told so far, the shop owners are very hands-on, and would be doing the majority of the work themselves
[21:51:44] <rayh> Sure. They just need a plan.
[21:51:48] <jmkasunich> they maintiain their machines now, to the extent of dragging out an oscilliscope when needed
[21:53:02] <jmkasunich> its a family owned shop that has been around since the 1930s
[21:53:44] <rayh> That is a good kind of shop to work with.
[21:53:52] <jmkasunich> yep
[21:54:08] <rayh> I don't exactly know how to handle the z axis kinematics.
[21:54:29] <jmkasunich> I think one would be Z and one would be W
[21:54:31] <rayh> But it would be a lot like a knee mill with both quill and knee.
[21:54:52] <jmkasunich> dunno which would be which, that is one thing I want to ask about - how do they program it now
[21:55:19] <rayh> Till F in Germany made a machine with two motors on the same Cartesian axis
[21:56:03] <rayh> So I know it can be done.
[21:56:28] <jmkasunich> I suspect they'd want independent control of quill and table, with two different letters in the G-code
[21:57:13] <rayh> There is a tool length advantage to know the distance between.
[21:57:44] <rayh> It would be a bit like the head on the mazak. Something we've not tried as yet.
[21:58:20] <rayh> But the table could just be a manual thing for now.
[21:58:41] <rayh> Or an extra axis under emc control but not coordinated.
[21:58:47] <jmkasunich> I need to talk to cradek about the w tool offset stuff
[21:59:00] <rayh> You're right about asking how they do it now.
[21:59:01] <jmkasunich> I bet EMC could do fully coordinated W
[21:59:13] <alex_joni> it doesn't do circles yet
[21:59:13] <rayh> No doubt.
[21:59:23] <jmkasunich> I think I'm gonna make a vismach model of an HBM, with a quill
[21:59:40] <alex_joni> jmkasunich: remember that you can model it in CAD now
[21:59:46] <jmkasunich> alex_joni: this machine doesn't have rotaries, so W is always parallel to Z
[21:59:48] <alex_joni> and only import the stl in vismach
[21:59:49] <rayh> I need to learn some of that uvw stuff also.
[22:00:04] <jmkasunich> alex_joni: I don't have cad that does STL
[22:00:04] <alex_joni> yeah, but what if you want to programm a G3 where a W moves
[22:00:11] <jmkasunich> and I can do it faster in vismach anyway
[22:00:38] <rayh> And not get black stuff under your fingernails!
[22:00:41] <jmkasunich> agreed - you can't do a G3 where W moves
[22:01:00] <jmkasunich> but arcs in that plan on this kind of machine are probably very rare
[22:01:05] <jmkasunich> that plane
[22:01:54] <alex_joni> maybe a helix
[22:02:03] <rayh> If the rotary table is an indexer only, you can't do a feedrate move with it anyway.
[22:02:18] <jmkasunich> the rotary isn't even normally mounted on the maching
[22:02:21] <jmkasunich> machine
[22:02:33] <rayh> Ah. I thought it was the table.
[22:02:42] <jmkasunich> no, its removable
[22:02:55] <jmkasunich> the four axes are X, Y, Z, W
[22:03:10] <rayh> Okay. I was imagining some boring machines I've worked on.
[22:04:00] <jmkasunich> they probably put it on when they are doing something that requires holes in the faces of a pentagon, or even just holes in all four faces of a block shaped part
[22:04:22] <rayh> Sure
[22:04:34] <jmkasunich> but he said that most times they leave it off, since they do a lot of large parts, and the combined weight of table and part is a bit much
[22:04:55] <fenn> alex_joni: are there any screenshots of vismach + stl?
[22:05:16] <alex_joni> fenn: I think jepler put some up
[22:05:25] <cradek> you can do g17/g18/g19 arcs while W also moves (g17 would be helical milling, the others would be stupid)
[22:05:46] <alex_joni> hi chris
[22:05:55] <jmkasunich> the arc would move Z in an arcy way, and W linearly?
[22:05:56] <cradek> hi
[22:06:07] <cradek> g18/g19 yes
[22:06:19] <cradek> would be useless I'm sure
[22:06:28] <cradek> g17 + w would be useful
[22:06:50] <jmkasunich> yeah, helical or thread milling using quill instead of table
[22:07:01] <cradek> yeah whichever it is
[22:07:14] <jmkasunich> rigid tapping always uses Z, or can it use W?
[22:07:35] <cradek> currently Z only. needs to be expanded to also allow W (or any direction)
[22:08:18] <Lerman___> Lerman___ is now known as Lerman
[22:08:23] <cradek> you'd want to tap with the highest accel axis, surely the quill
[22:08:31] <alex_joni> fenn: http://emergent.unpy.net/files/sandbox/base_stl.png
[22:08:58] <jmkasunich> cradek: dunno how far back you read - did you see the pics?
[22:09:02] <cradek> yes
[22:09:08] <cradek> looks fun
[22:09:29] <jmkasunich> I wouldn't be surprized if it has resolvers - I might want to talk to you about how Jon's resolver-encoder converters are working out
[22:09:34] <cradek> that machine looks either recently scraped, or not-much-used
[22:09:58] <jmkasunich> he said its in use for production
[22:10:00] <cradek> I have had exactly zero trouble with jon's boards
[22:10:20] <cradek> they only take minor fiddling for setup
[22:11:11] <cradek> you have to set the amplitude with some jumpers (just requires dual input scope or some other magic)
[22:11:48] <jmkasunich> dual channel scope is far from magic
[22:12:05] <alex_joni> but a lack of a dual scope can only be bridged by magic
[22:12:50] <cradek> I'm not sure how to do it without - that's what I meant
[22:14:10] <cradek> oh the machine is actually working? wow.
[22:15:15] <jmkasunich> from the first message I got: "The iron is good and the current control still works but I expect control maintenance issues to only increase over time. In fact, this machine was in production until just a few weeks ago when the operator left for another employer."
[22:16:17] <cradek> oh maybe it's idle because nobody knows how to run it - nice opportunity
[22:16:48] <jmkasunich> I got the impression that if they really need to run it, they can
[22:17:49] <cradek> nice. it will be much easier if it hasn't been sitting.
[22:18:19] <cradek> stuff on it will actually work
[22:18:24] <jmkasunich> heh
[22:18:36] <jmkasunich> unlike the vacuum pump on Stuart's G&L
[22:18:44] <cradek> yes and all the oil seals
[22:19:12] <cradek> and you can study it a bit and look for oddities/quirks
[22:19:17] <jmkasunich> from what he told me, there is far less hydraulic "stuff" on this machine - no hydrostatic bearings for example
[22:19:45] <jmkasunich> I wonder if the counterweight is hydraulic, or just a big old lump of heavy stuff
[22:20:06] <cradek> chain - must be a weight
[22:20:12] <cradek> hydraulics would push up, not pull down
[22:20:22] <jmkasunich> stuart's has a chain, but it is hydraulic
[22:20:34] <jmkasunich> the chain goes up and over, and the cylinder is in the column
[22:20:37] <cradek> oh, ok
[22:20:42] <cradek> strange
[22:20:48] <jmkasunich> you can't put the cylinder underneath unless you want to dig a deep hole
[22:20:56] <cradek> true
[22:22:30] <jmkasunich> I bet the cylinder travel is only half the head travel, and the chain runs over a moving pulley
[22:22:48] <jmkasunich> otherwise the extended cylinder would be twice as long as the travel
[22:24:24] <jmkasunich> almost 6:30, I should make dinner
[22:24:28] <jmkasunich> thanks for the input guys
[23:00:12] <alex_joni> good night all