#emc | Logs for 2011-05-05

Back
[00:03:20] <tom3p> check the 7i37 for the digital i/o and consider using the 7i37 with a software (HAL) encoder counter for the mpg, gotta go, bestOluck
[00:03:25] -!- theorb [theorb!~theorb@91.84.53.6] has joined #emc
[00:03:26] -!- tom3p [tom3p!~tomp@74-93-88-241-Illinois.hfc.comcastbusiness.net] has parted #emc
[00:03:41] -!- theorbtwo has quit [Ping timeout: 240 seconds]
[00:03:48] theorb is now known as theorbtwo
[00:06:39] <aggrav9d> hi guys
[00:06:57] <aggrav9d> anyone awake in here? I'm struggling with my CNC machine and I'd really love some help.
[00:09:52] <Jymmm> You would actually need to ask a question to see if someone can help you with the issue you're having.
[00:13:57] -!- mhaberler has quit [Quit: mhaberler]
[00:13:58] mhaberler_ is now known as mhaberler
[00:14:13] -!- mhaberler_ [mhaberler_!~mhaberler@macbook.stiwoll.mah.priv.at] has joined #emc
[00:14:14] -!- aggrav9d has quit [Read error: Connection reset by peer]
[00:14:35] -!- aggrav8d [aggrav8d!~aggrav8d@S0106000b6a15cda1.vc.shawcable.net] has joined #emc
[00:16:33] <PCW> One option is to use one of the 7I47s serial channels to connect to the 7I73 pendant interface (Via RS-422) but thats fairly bleeding edge
[00:17:17] -!- tris has quit [Quit: Leaving]
[00:17:24] -!- tris [tris!~tristan@173-164-188-122-SFBA.hfc.comcastbusiness.net] has joined #emc
[00:36:01] <Jymmm> PCW: Is it as bleeding edge as a drunk monkey with a machete?
[00:36:37] <ssi_> hm what happens if you command constant surface speed mode on a machine that doesn't have spindle speed control?
[01:03:02] <atom1> 7i73 isn't out yet is it?
[01:03:04] <atom1> iirc
[01:04:00] <atom1> being a noob i'm not sure i wanna bleed that much yet
[01:04:43] <atom1> what page is that displayed on?
[01:28:38] <jdhNC> the price list
[01:29:01] <atom1> well i saw it there
[01:44:54] <jdhNC> I looked at all the others the other night, never found it
[01:45:16] <atom1> same here
[01:45:29] <atom1> figured somebody might know
[01:48:35] <jdhNC> pcw would know
[01:49:56] <PCW> We have it working but it needs some updates to the sserial driver so you would have to use a patched 2.5 to make it work now
[01:53:43] -!- Techrat has quit [Quit: Leaving]
[01:59:46] -!- rooks has quit [Quit: So long, and thanks for all the fish.]
[02:25:42] <atom1> surplus cnc psu: http://tom-itx.dyndns.org:81/~webpage/temp/cnc/psu1.jpg
[02:42:50] -!- Techrat [Techrat!~TRat@c-24-2-169-60.hsd1.ct.comcast.net] has joined #emc
[02:42:50] -!- Techrat has quit [Changing host]
[02:42:50] -!- Techrat [Techrat!~TRat@unaffiliated/techrat] has joined #emc
[02:55:30] -!- Tom_L [Tom_L!~Tl@unaffiliated/toml/x-013812] has joined #emc
[02:55:42] -!- Tom_L has quit [Client Quit]
[02:57:46] <KimK> Jymmm: I saw a T-shirt for sale with the "10 Rules of Gun Safety". One of them was, "No matter how responsible he seems, never give your gun to a monkey."
[02:58:28] <KimK> aggrav8d: Do you have a specific question or problem?
[03:00:39] -!- WalterN [WalterN!~walter@tiwake.com] has joined #emc
[03:14:25] -!- warlock [warlock!warlock@cpc1-sgyl2-0-0-cust91.sgyl.cable.virginmedia.com] has joined #emc
[03:15:45] <Jymmm> KimK: lol
[03:22:06] <aggrav8d> so... i can jog the tool all over the envelope and return it to origin successfully. When I go to cut parts I get a gorram mess and I don't know why.
[03:27:29] <ssi_> my math, it is failing
[03:31:21] <KimK> aggrav8d: OK. What version of EMC2 are you running, what is the nature of your machine, and are you running steppers or servos?
[03:34:16] <ssi_> diameter mode is screwing me all up :P
[03:36:33] -!- aggrav9d [aggrav9d!~aggrav8d@S0106000b6a15cda1.vc.shawcable.net] has joined #emc
[03:36:46] -!- aggrav9d has quit [Client Quit]
[03:38:50] -!- aggrav8d has quit [Ping timeout: 276 seconds]
[03:40:36] -!- aggrav8d [aggrav8d!~aggrav8d@S0106000b6a15cda1.vc.shawcable.net] has joined #emc
[03:41:05] <aggrav8d> blacktoe 8x4. steppers. what was the rest of the question? emc2. uh... i don't know the version off hand. very recent.
[03:42:41] -!- Valen has quit [Quit: Leaving.]
[03:43:10] -!- Valen [Valen!~Valen@ppp121-44-231-147.lns20.syd7.internode.on.net] has joined #emc
[03:52:17] <KimK> Sorry, I didn't recognize your machine name, Googled, and found your videos too. What have you tried so far? writing with a pen appeared to go well.
[03:54:01] <toastydeath> what exactly is "a gorram mess"
[03:54:03] <KimK> What g-code file did you execute that did not go well?
[03:54:41] <aggrav8d> http://www.marginallyclever.com/samples/CNC/2011-05-04/small.ngc
[03:54:49] <aggrav8d> http://www.marginallyclever.com/samples/CNC/2011-05-04/capture.JPG
[03:55:37] <KimK> He's a Browncoat. http://en.wikipedia.org/wiki/Browncoat
[03:55:39] <toastydeath> image 404'ed
[03:55:39] <Valen> http://www.marginallyclever.com/samples/CNC/2011-05-04/capture.JPG = 404
[03:56:44] <aggrav8d> http://www.marginallyclever.com/samples/CNC/2011-05-04/Capture.JPG
[03:56:59] <toastydeath> 404
[03:57:02] <aggrav8d> i'm trying to get the results photo off my phone.
[03:57:13] <toastydeath> can you just describe what happened
[03:57:17] <aggrav8d> http://www.marginallyclever.com/samples/CNC/2011-05-04/Capture.jpg
[03:58:25] <toastydeath> those little divots in the gear?
[03:58:32] <toastydeath> gear teeth, that is
[03:59:16] <aggrav8d> divots?
[03:59:28] <KimK> aggrav8d: I notice you were trying to feed at 80ipm(?), what's your rapid(??).
[03:59:39] <toastydeath> what is the actual problem here
[03:59:50] <aggrav8d> rapid is maxed at 400 but I've gone to 700 no trouble.
[04:00:20] <aggrav8d> i don't trust my rapids or my ipm. i'm sure between them and the 23k rpm i'm doing it all wrong.
[04:00:38] <toastydeath> what is the problem?
[04:00:57] <KimK> What are we looking at in capture.jpg? How was this image made?
[04:01:06] <toastydeath> like, i see a on-machine toolpath and some g-code, what is happening that you are compliaining about?
[04:01:20] <aggrav8d> kimk - cambam. the program that makes the gcode. it's on an angle so you can see the paths.
[04:01:46] <KimK> So what are you saying, that the g-code had defects to begin with?
[04:01:48] <aggrav8d> anyways, you asked for a description. first time i was cutting 0.015 steps it made the inner circles pretty good but I could see some striations in the last hole and when it got to the gears, well...
[04:02:05] <aggrav8d> it made one pass around and when it started the second pass it was completely different from the first, just a total mess.
[04:02:33] <KimK> What does it appear to do if you cut air?
[04:02:56] <aggrav8d> appears to be fine (i think)
[04:03:07] <toastydeath> you are using .015" steps?
[04:03:11] <toastydeath> to cut gear teeth?
[04:03:47] <aggrav8d> i couldn't really tell it was getting the holes wrong until I paused the program and vacuumed very thoroughly. my vacuum setup isn't great.
[04:03:54] <toastydeath> try .0005" steps at 5 ipm
[04:03:57] <aggrav8d> but then I have a lot of issues with my setup.
[04:04:17] <aggrav8d> toastydeath - ah, well... see? how was I to know? these are decorative gears, not functional.
[04:04:28] <toastydeath> .0005" or less between move commands
[04:04:28] <aggrav8d> how am I supposed to know what rate to cut at?
[04:04:38] <toastydeath> experience, mostly
[04:04:45] <toastydeath> in time
[04:04:55] <toastydeath> cut as deep as your machine will handle
[04:05:08] <toastydeath> the final pass should be full-depth with a radial depth of cut of about .005"
[04:05:17] <aggrav8d> http://www.marginallyclever.com/samples/CNC/2011-05-04/photo.jpg -- results
[04:05:49] <toastydeath> those holes look like the part shifts substantially
[04:05:54] <toastydeath> during machining
[04:06:02] <KimK> I think toastydeath and I both suspect that you are losing steps. Both because you are overloading the steppers when cutting, and because you are trying to go too fast when rapiding, and maybe even fast feeding.
[04:06:08] <aggrav8d> it could be. I'm only clamping them down.
[04:06:25] <toastydeath> how you cut gears, even decorative ones, is at least a two step operation
[04:06:34] <toastydeath> you cut the blank out, and then you hold the blank very firmly.
[04:06:42] <toastydeath> using the internal holes.
[04:07:15] <toastydeath> but it looks at least as though the part is moving
[04:07:19] <toastydeath> so you need to deal with that first
[04:07:39] <toastydeath> second, when you are doing very detailed profiling in EMC, there is no "true" lookahead
[04:07:42] <KimK> toastydeath: It might look that way from losing steps also.
[04:07:43] <aggrav8d> ok, I can screw the stock onto the spoil sheet.
[04:08:04] <toastydeath> you must go very slowly so that the machine does not think it can go faster than it actually can
[04:08:31] <aggrav8d> i don't know what you mean by blank, and I don't know what you mean by lookahead. I'm not a machinist, I jumped in the deep end. I've been a programmer for too damn long.
[04:08:34] <toastydeath> a bunch of steps that have a gradual change in one axis then a dramatic change will fool the look ahead and cause the machine to drift substantially
[04:09:09] <Valen> if it lost steps wouln't it not join up? IE close the circle?
[04:09:24] <Valen> whats your acceleration btw?
[04:09:37] <KimK> aggrav8d: Try reducing the max velocity on your next run too. Maybe 25% of what it is now. And feedrates down to 25% too. Just to see.
[04:09:48] <KimK> Valen: An excellent question!
[04:10:31] <KimK> Try reducing your accel until you think it's way too slow, and try that, lol!
[04:10:44] <toastydeath> someone ought to eventually look into real look ahead
[04:10:48] <aggrav8d> I will try 20ipm. I've been impatient, it's true.
[04:10:52] <toastydeath> uh, 20 ipm?
[04:10:54] <toastydeath> no.
[04:11:00] <toastydeath> under 5.
[04:11:03] <aggrav8d> well i WAS doing 80.
[04:11:07] <toastydeath> jesus christ
[04:11:09] <Valen> cutting gears is hard to do, lots of detail
[04:11:20] <toastydeath> you are going WAY too fast for what you are working with.
[04:11:28] <Valen> wow thats pretty fast
[04:11:32] <aggrav8d> yeah... the guy who makes the machine says it can do 80 and i took that at face value.
[04:11:43] <Valen> it probably can "do 80"
[04:11:45] <toastydeath> I can do 80 on a full VMC with real look ahead, you cannot do it using vacuum hold down on a hobby machine using emc
[04:11:48] <KimK> What's your spindle/router(?) HP? and bit size?
[04:11:53] <Valen> like my car will do 240km/h
[04:11:59] <Valen> whats the machine?
[04:12:11] <KimK> It might be able to "do 80" while free-rapiding.
[04:12:16] <aggrav8d> 32000rpm porter cable.
[04:12:19] <Jymmm> Valen: downhill in a hurricane
[04:12:23] <aggrav8d> 1/8"
[04:12:28] <aggrav8d> 2 flute upturn.
[04:12:33] * atom1 puts on his hip waders
[04:12:34] <Valen> we ran ours cutting acrylic at 800mm (1/4 the speed of yours)
[04:12:49] <KimK> OK. So a countertop router or similar.
[04:13:01] <Valen> Jymmm: www.vapourforge.com/jake/general/hotness_touched.jpg uphill with a headwind ;-P
[04:13:33] <toastydeath> aggrav8d, you can only do very high speed cutting in a straight line
[04:13:39] <toastydeath> i.e. if you wanted to cut out a box
[04:13:48] <Valen> if you reduce the feed you will probably need to reduce the RPM as well so you don't burn the timber
[04:13:51] <aggrav8d> http://www.marginallyclever.com/samples/CNC/2011-05-04/photo2.jpg
[04:13:52] <toastydeath> you cannot do that with an interpolated move
[04:14:03] <toastydeath> or rather, a curve/path move
[04:14:25] <toastydeath> yeah, that looks like the part was moving
[04:14:25] <aggrav8d> ok, so i should be looking at something around 5rpm for all non-straight cutting?
[04:14:27] <Valen> actually yeah it looks like it lost steps on the top right corner of it all
[04:14:28] <KimK> aggrav8d: What's the deal on the cambam defects? Rounding errors or something?
[04:14:30] <aggrav8d> er, 5ipm
[04:14:40] <toastydeath> start at 5, you may want to experiment going faster
[04:14:52] <aggrav8d> nah, i'd rather go slow and have a success.
[04:14:54] <Valen> you need to find out what your acceleration maximum and conservative value is
[04:15:12] <aggrav8d> i need a success. I'm a little fragile these days.
[04:15:41] <Valen> I'd guess more like 10 -15 would probably do, but yeah start slow and work up
[04:15:57] <toastydeath> you also need to put an indicator on the edge of the board itself
[04:16:04] <toastydeath> connect the indicator to the table, somehow
[04:16:11] <toastydeath> and then the tip of the indicator against the board
[04:16:34] <toastydeath> this will tell you, while the machine is running, if the board is moving and you need a better vacuum setup
[04:16:41] <toastydeath> at 5 ipm it will be readable
[04:16:46] <Valen> or just bolt it down ;->
[04:17:03] <Valen> I'd cut the center holes out, put bolts through them, then cut the gears
[04:17:13] <toastydeath> Valen, that is what i would do as well
[04:17:23] <aggrav8d> to hold my stock down I've got some threaded inserts in the table top. I put the spoil down, then the stock, then clamp the stock to the table with a piece of 1/2" MDF that's got a hole in the middle. I put a bolt through the hole and into the threaded insert, plus something on the other end of the MDF to force pressure on the stock.
[04:17:46] <toastydeath> hm
[04:17:53] <toastydeath> that should be enough, you should be fine going slower
[04:18:28] <Valen> if the MDF starts burning/smoking reduce your spindle speed
[04:18:40] <Valen> too much spindle and you burn the wood
[04:18:52] <KimK> aggrav8d: There's another thing you might try before cutting material. EMC2 should come with an example program, tort.ngc or something like that. It's a random cloud of rapids, feeds, arcs, and helixes. What you want to do is run it and see if it returns to the same point. You may have to hand edit a new start and end point (same) somewhere out-of-the-way, so you can put a dial indicator (or homemade substitute reference) in a handy location.
[04:18:55] <Jymmm> and grabt eh extinguisher
[04:19:35] <KimK> Then run the program and see if you lose steps. If you do, adjust MAX_VEL, ACCEL, and so forth.
[04:20:32] <KimK> Oh, and you may have to scale it to fit your machine nicely.
[04:20:34] <Valen> you really need to get that sorted before your going to make detailed/accurate stuff
[04:22:27] <KimK> aggrav8d: Well, that should give you a bunch of stuff to try, right? Let us know what you find out.
[04:29:55] -!- ries has quit [Quit: ries]
[04:31:09] <aggrav8d> hmm...
[04:31:20] * aggrav8d absorbs twice as well as the next leading brand.
[04:32:03] <aggrav8d> my reference is an X drawn on the table top. :T
[04:34:16] <aggrav8d> you want even dumber? I read that I should never try to cut more than 1/2 the tool width. so in the last version of small.ngc the depth is 0.0625. only 4x more :P
[04:36:10] <aggrav8d> so i'm going to stick to 0.015 and 5ipm to start. my first pass will be cutting out the holes and the second pass will be profiling the exterior. my copy didn't come with a sample program. I'll have to find it.
[04:36:31] <aggrav8d> did i miss anything from your suggestions? screw the piece down to the table top?
[04:36:51] <aggrav8d> i'd rather not put holes in my table. again. :T
[04:37:10] <KimK> OK. If you don't have a dial indicator, or a caliper, or something more precise, then maybe you could just clamp a wooden cube near some convenient part of the machine, such that when it's in the starting position (nearest the block) you can set the block so you can slip, say, 2-3 slips of paper between the machine and the block, but not 4-5 slips of paper. It's a limited-precision measurement, but it's better than nothing.
[04:38:01] <aggrav8d> i have a digital caliper
[04:38:30] <aggrav8d> 0.005 repeatability, iirc?
[04:38:34] <KimK> Ah, excellent! Than use that (in the 0-6" range?) as convenient.
[04:38:40] <aggrav8d> yeah.
[04:38:55] <aggrav8d> explain again how i should use it? I don't grok.
[04:39:55] <KimK> It doesn't matter really. What you want to do is make sure that the machine ends up in the same place it started. (Your test g-code program must be set that way.)
[04:40:25] <aggrav8d> right. compare how it moves in the air and see if it comes back to the start.
[04:40:36] <KimK> You have to check one axis at a time, but you can record measurements for all three axes after a g-code run.
[04:40:44] <KimK> s/check/measure/
[04:41:25] <aggrav8d> from there i can experiment with speeds to check acceleration? rapids?
[04:41:59] <KimK> You just need to find some convenient spot to make your measurements. This is something like installing a digital readout (DRO) on a mill.
[04:42:07] <ssi_> grargh
[04:42:14] <KimK> Yes, exactly.
[04:42:16] <ssi_> I need to find a good way to simulate and debug gcode
[04:42:27] <ssi_> doing it on the lathe machine is too painful
[04:42:47] <KimK> ssi_: The simulators are not working for you?
[04:42:53] <ssi_> I haven't tried them
[04:42:57] <aggrav8d> aha.
[04:43:01] <ssi_> the only place I have emc running is on the machine that's running the lathe
[04:43:02] <ssi_> downstairs
[04:43:05] * aggrav8d didn't make that mistake. twice.
[04:43:09] <ssi_> with a nasty horrible waterproof keyboard
[04:43:27] <KimK> ssi_: Give the lathe simulator a try.
[04:43:57] <ssi_> is it one of the sample configs?
[04:44:05] <KimK> You might have to adjust it if it's "upside down" compared to your lathe, or some such.
[04:44:17] <KimK> Yes, under sim, I believe.
[04:45:34] <KimK> And you can also adjust it to fit your workspace size, offsets, etc., makes the sim more helpful, instead of just always saying, "run anyway".
[04:45:36] * Valen has one of those keyboards, pain in the ass lol
[04:47:33] <Valen> aggrav8d: for testing you could just do a g1 F100 X100 then g1 F100 X0 or something like that
[04:47:38] <Valen> just zip out and back
[04:48:27] <Valen> increasing acceleration until you loose steps then backing off (probably 2/3rds or so at a guess) then do it again and find what your maximum travel speed is
[04:48:35] <Valen> (ie increase F)
[04:48:52] <Valen> just zip away and back to your caliper
[04:49:08] <KimK> Valen: Yes, that's an excellent idea, especially when making your first adjustments.
[04:49:18] <Valen> your acceleration will set how detailed a part you can make at speed
[04:49:31] -!- ve7it has quit [Remote host closed the connection]
[04:49:46] <Valen> you will notice that when your cutting a gear like that the actual velocity will hardly ever reach the programmed velocity
[04:50:02] <Valen> when we were doing ours we had a feed rate of 800mm/min and were getting ~300 or so
[04:50:02] <ssi_> KimK: looks like the lathe sim is pretty close
[04:50:22] <ssi_> the only t hing that looks off is the tool profile that it renders on screen is backwards
[04:50:28] <KimK> aggrav8d: Also, when adjusting speeds and accels, etc., prepare to be disappointed. You might have to slow it down a lot more than you ever thought to make sure that it *never* misses steps. That's pretty common on the torture test.
[04:50:29] <ssi_> not the toolpath itself
[04:50:34] <ssi_> just the triangle that represents the tool
[04:50:45] <ssi_> it shows up as a left-hand tool
[04:51:09] <aggrav8d> well... it'll be really hard to tell if i missed juust one step.
[04:51:13] <ssi_> I guess that's the tool table
[04:51:35] * Valen needs to talk to psha
[04:51:58] <Valen> trying to get the HAL stuff working with his camera tab
[04:52:23] <Valen> but I'm having "issues" http://pastebin.ca/2053375
[04:52:30] <KimK> ssi_: It might be the tool table, or the lathe might be "upside down". You'll have to check. There's probably something you can do about that in ini, but you'll have to research it.
[04:52:42] <aggrav8d> KimK, thanks for the great tips. i'll be trying these out tomorrow. fingers crossed.
[04:53:14] <ssi_> looks like it's not the lathe that's upside down... I have one of my programs loaded, and the axes point the right direction, and the plot renders rightside up and forward
[04:54:04] <KimK> aggrav8d: No problem. Come back and tell us how it went. toastydeath, Valen, thanks for your help too.
[04:54:13] <ssi_> yea it's just the tool table
[04:54:31] <aggrav8d> will do.
[04:54:37] <KimK> ssi_: Great, an easy fix then? Enjoy.
[04:54:42] <ssi_> yep, thanks
[04:54:51] <aggrav8d> i like #emc more than cnczone or the forums where i got my machine.
[04:54:55] <ssi_> I need to dig into this and see how they did the little spindle speed panel
[04:55:06] <aggrav8d> g'night!
[04:55:12] <KimK> bye
[04:55:34] <KimK> ssi_: Yes, great fun, check out the PyVCP section.
[04:57:24] <ssi_> hm that's all there is to it, huh
[04:59:50] <ssi_> is there a way to somehow report or print out the values of variables in gcode while it's running for debugging purposes?
[05:00:08] <ssi_> I've been resorting to goofy crap like G92 Z#<someVar> M1 to be able to see what's in it
[05:02:44] <ssi_> like maybe... some user defined command that could assign a value to a hal signal, and then display that signal via pyvcp
[05:03:56] -!- i_tarzan has quit [Ping timeout: 260 seconds]
[05:06:23] -!- WesBaker has quit [Ping timeout: 276 seconds]
[05:07:22] <ssi_> aha
[05:07:34] <ssi_> (debug, some variable is set to #<someVariable>)
[05:07:45] <ssi_> my life is about to get a lot easier :D
[05:11:46] <KimK> ssi_: Wait, I could use that one too! That's a message function? Like (msg, blahblahblah)? Cool! I didn't know about that one. And I'm writing some g-code like that right now.
[05:12:37] <ssi_> (msg, blah) doesn't eval variables
[05:12:40] <ssi_> but (debug, blah) does
[05:13:08] <KimK> Exactly. Excellent, thanks!
[05:13:12] <ssi_> you bet :)
[05:13:43] <KimK> I'll go RTFM as soon as I'm done with what I'm on.
[05:14:36] <ssi_> TFM is surprisingly useful :D
[05:15:43] <KimK> I've written some g-code where everything (well, all the X & Y moves) are calculated from declared variables so as to be easily changeable, but now I can't tell if it's going to the right places without a chart, lol.
[05:15:52] <ssi_> I know what you mean
[05:15:53] -!- aggrav8d has quit [Ping timeout: 240 seconds]
[05:16:05] <ssi_> I'm writing a couple of subroutines that I can use as building blocks for hand-written code
[05:16:53] <ssi_> writing one right now that'll turn a chamfer into an edge at a given Z, you supply the diameter, edge Z, chamfer angle and depth, roughing and finishing tool, depth of cut, and feed
[05:17:11] <ssi_> I want to make a bunch of tormach holders, so I'm trying to get programs written to turn them :D
[05:18:07] <KimK> The many features of EMC2 make it easier to write by hand, or at least you can keep going longer before being forced to turn back and resort to CAM.
[05:18:17] <ssi_> yeah
[05:18:22] <ssi_> I've sat down and tried to get into cam
[05:18:30] <ssi_> and just the overwhelming nature of integration makes it hard
[05:18:45] <ssi_> I think it'll be easier for milling
[05:18:51] <ssi_> but lathe cam seems to be less polished
[05:19:17] <toastydeath> you have to remember that lathes have not needed cam for a very long time
[05:19:28] <toastydeath> because they have automatic roughing and finishing cycles
[05:19:53] <toastydeath> newer mill controls are just now catching up to that capability
[05:20:02] <toastydeath> ten or fifteen years behind lathes
[05:20:21] <ssi_> ah fair enough
[05:20:22] <toastydeath> so mill cam is much more advanced, for that same reason
[05:20:32] <KimK> HeeksCNC is supposed to be pretty usable now, but I'm no expert, I've only played with it just a little.
[05:20:39] <toastydeath> emc is an oddball for lathe control in that it doesn't have those canned cycles
[05:20:56] <toastydeath> nor some of the other lathe-centric things that make lathe programming go faster
[05:21:03] <KimK> Why is that?
[05:21:14] <toastydeath> i dunno, i don't develop emc
[05:21:26] <toastydeath> nobody has taken an interest in it yet, i guess is the answer
[05:21:28] <ssi_> I'm borrowing some of the routines that JT has for those kind of cycles and improving on them
[05:21:35] <ssi_> they're just subroutines, but it's a start
[05:22:08] <toastydeath> for lathes, the format is this:
[05:22:44] <toastydeath> g71 <bunch of commands>; g71 <even more commands> p100 q200; n100; <a toolpath>; n200; g70;
[05:23:11] <toastydeath> that would take, using the feed/stepdown/etc information given in g71, and rough the part down and finish it to the toolpath between n100 and n200.
[05:23:18] <KimK> And that's what, Fanuc?
[05:23:26] <toastydeath> fanuc and a bunch of other controls
[05:23:27] <ds3> Mmmmmm roughing cycles
[05:23:41] <ds3> be careful of where you start from ;)
[05:24:11] <ssi_> that would be handy
[05:24:35] <toastydeath> fanuc added similar milling cycles to their controls
[05:24:49] <toastydeath> so you have pocketing, contouring, facing, etc
[05:25:19] <toastydeath> and there's a at-tool g-code generator, you tell it what you want, and it will insert the appropriate g-code into your program
[05:25:33] <toastydeath> takes the bulk of the work out of it and all you have to do is minor tweaks, if you care about it that much
[05:26:41] <KimK> I understand <a_toolpath>, that would be your finished part profile, right? But what would be the <bunch of commands> and <even more commands> the initial roughing-as-needed?
[05:27:01] <toastydeath> the bunch of commands are things like feed rate, stepdown, how much stock to leave for the finishing pass
[05:27:12] <toastydeath> how many semifinishing passes to take, how much stock THOSE passes should have
[05:27:43] <ds3> but CAM obsoletes some of those cycles
[05:27:44] <KimK> Oh, so there are so many setup paramenters that it won't (or might not) fit on one line?
[05:27:52] <toastydeath> KimK, they split it for legibility
[05:27:58] <toastydeath> so you re-use the uvw xyz words
[05:28:07] <toastydeath> there aren't THAT many, it just makes it easier to understand with less to remember
[05:28:15] <toastydeath> ds3, cam does not obsolete those cycles
[05:28:24] <KimK> OK. I'll have to look that up sometime.
[05:28:25] <ssi_> good cam would just generate that
[05:28:31] <ds3> toastydeath: why would CAM generate those cycles?
[05:28:36] <toastydeath> those cycles make it so that I can touch a tool off and start cutting
[05:28:44] <toastydeath> while my co-worker is building a model
[05:29:03] <toastydeath> it depends on the part you are making and what your goal is
[05:29:08] <ds3> that's why you insist people give you a model first ;
[05:29:09] <ds3> )
[05:29:10] <toastydeath> cam wouldn't use those cycles, only programming by hand
[05:29:30] <ds3> how many people in industry hand program G code?
[05:29:30] <toastydeath> if someone is slow or bad at programming by hand, i could see why they would choose cam in every instance
[05:29:41] <toastydeath> a lot?
[05:29:46] <ds3> really?
[05:29:53] <toastydeath> is that a trick question?
[05:29:57] <ds3> no
[05:30:12] <ds3> that seems like a useless hobbist skill
[05:30:26] <toastydeath> then why are you asking me twice? just about every skilled cnc machinist i know of programs by hand as well as cam
[05:30:37] <toastydeath> some of the worst program just via cam.
[05:30:45] <toastydeath> i've only met two of those, however
[05:30:55] <ds3> thought about working as a machinist but figure there is no way I can keep up with others on a bridge port and most shops I know of go cam
[05:31:00] <toastydeath> could be a regional thing or industry thing, I imagine aerospace has very little need for hand programming
[05:31:15] <ds3> could be an industry thing
[05:31:37] <toastydeath> dunno, I found myself on a bridgeport fairly often
[05:31:50] <ds3> it seems the only thing I have seen in industry is a bridge port or a hybrid bridgeport with an intelligent controller (not full CNC)
[05:31:54] <toastydeath> it was nice to have a conversational control on it, that alleviated much of the work
[05:32:14] <toastydeath> but we had 4 full manual bridgeports, and two with hybrid
[05:32:19] <toastydeath> at my first shop
[05:32:27] <toastydeath> and the second just had a couple manual ones
[05:32:39] <ds3> no arguements on the bridgeport; it just seems the full CNC ones are all CAM programmed
[05:32:47] <ssi_> well my chamfer subroutine seems to be solid now
[05:32:48] <ssi_> :D
[05:33:00] <toastydeath> CAM is faster as part complexty increases
[05:33:08] <toastydeath> and the number of parts you are making
[05:33:25] <toastydeath> for some jobs, where you have to modify a bunch of parts
[05:33:35] <ssi_> seems to me like once you get into parts that have curves that aren't just constant radius arcs, cam starts to be more necessary
[05:33:47] <ds3> hand programming is easy - its all trig or so it seems
[05:33:51] <ssi_> I wouldn't want to hand-code chess pieces, for instance :P
[05:33:59] <toastydeath> ssi_, you'd be shocked how few parts are like that.
[05:34:06] <ssi_> no, I'm aware
[05:34:10] <ssi_> that's why I'm writing by hand :D
[05:34:21] <ssi_> I'm working on these TTS holders
[05:34:24] <toastydeath> most parts are stupid things with four walls, maybe a pocket, and a hole drilled in it
[05:34:24] <ds3> from a hobbist standpoint, cam is good when you have to surfacing
[05:34:28] -!- Dannyboy [Dannyboy!~Dan@ip68-101-102-133.oc.oc.cox.net] has joined #emc
[05:34:40] <ssi_> it's a facing op and two turning ops and a chamfer on the shank end, first setup
[05:34:50] <ssi_> and some way to relieve the shoulder
[05:34:52] <ssi_> still working on that one
[05:34:56] <toastydeath> chamfering is easy
[05:34:58] <toastydeath> don't write a routine for that
[05:35:03] <ssi_> well, it's not just a chamfer
[05:35:05] <ssi_> it's a taper
[05:35:07] <toastydeath> ah
[05:35:08] <ssi_> but it's a taper that's on an edge
[05:35:16] <ssi_> my routine takes an angle and a depth
[05:35:22] <ssi_> and the edge Z
[05:35:26] <ssi_> and does roughing and finishing
[05:35:58] <ssi_> I called it a chamfer because i'm not sure if the inputs are descriptive enough to work as a general purpose taper
[05:36:22] <ssi_> I suffer from analysis paralysis very frequently
[05:36:27] <toastydeath> lol
[05:36:41] <ssi_> trying to figure out how to cut this taper, I wanted to make a routine that would handle every single taper that could ever happen
[05:36:45] <ssi_> and my brain went OVC
[05:37:02] <toastydeath> i generally try not to solve problems i don't have yet
[05:37:10] <toastydeath> it's served me well in machining
[05:37:15] <ssi_> yeah, it's a bad habit of mine
[05:37:17] <ds3> lathe work is more fun when done by hand
[05:37:22] <ssi_> I'm a software architect, so I have to look ahead typically
[05:37:42] <toastydeath> 90% of the problems in machining are things you will not know to look for until after they've happened
[05:37:53] <toastydeath> so why worry overmuch?
[05:38:41] <ssi_> true enough
[05:41:08] <ssi_> http://www.stealthsystemsinc.com/chamfer.png
[05:41:11] <ssi_> there's my routine :D
[05:41:34] <toastydeath> nice
[05:41:37] <ssi_> setup to cut a 10 degree chamfer .125" deep, 0.020" doc roughing, with 4 .001" finish passes
[05:41:39] <toastydeath> wait, is this for lathe
[05:41:41] <ssi_> yes
[05:41:47] <toastydeath> you poor bastard
[05:41:50] <ssi_> haha what
[05:41:51] <toastydeath> do you have a tailstock
[05:41:54] <ssi_> sure
[05:42:09] <toastydeath> just knock it off in one shot
[05:42:27] <ssi_> that's hardly the point :)
[05:42:35] <ds3> use a file!
[05:42:43] <toastydeath> plasma cutter
[05:43:02] <ds3> file is the quickest way to put a chamfer on the lathe
[05:43:12] <toastydeath> he actually means a taper
[05:43:17] <ssi_> what about a jacobs taper? :D
[05:43:20] <ssi_> file me one 'em
[05:43:21] <ds3> Oh
[05:43:23] <toastydeath> also like
[05:43:31] <toastydeath> the traditional, fastest way to rough a taper
[05:43:41] <toastydeath> is to cut along z up to the surface
[05:43:44] <toastydeath> without moving x
[05:43:48] <ssi_> yeah, I know
[05:43:51] <toastydeath> then back out, drop x, and go in
[05:44:00] <toastydeath> that way your lathe is not piddling around for a half hour in rapid moves
[05:44:10] <KimK> ssi_: Are those your "(debug, blahblah)" messages showing in the corner there?
[05:44:10] <ds3> a jacobs might be short enough to use the compound
[05:44:18] <ssi_> KimK: indeed!
[05:44:23] <KimK> Great!
[05:44:28] <ssi_> ds3: this lathe doesn't have a compound anymore
[05:45:01] <ds3> for CNC, isn't a taper cut just a few G01's?
[05:45:10] <ssi_> yes, absolutely
[05:45:11] <ds3> (for shallow tapers)
[05:45:18] <ssi_> assuming you know the start and endpoint
[05:45:21] <ds3> then wtf a route? :D
[05:45:28] <toastydeath> he's problem solving, shh
[05:45:28] <ssi_> if you just know the angle, you need trig
[05:45:58] <ds3> so you grab your trusty calculator
[05:46:01] <ds3> anyways
[05:46:10] <ssi_> or I could just draw it in solidworks and cam it
[05:46:12] <ssi_> but again, not the point!
[05:46:25] <toastydeath> it's only one trig problem
[05:46:32] <toastydeath> you solve the z displacement for a given drop in x
[05:46:35] <toastydeath> boom
[05:46:54] <ssi_> yeah, that might be a better way to do it
[05:46:56] <ds3> it isn't even a trig problem if it is a stock taper.. the machinery handbook lists w/o it angle for a lot of the stock stuff
[05:47:17] <ssi_> mine is maybe overly complicated
[05:47:24] <toastydeath> you're problem solving.
[05:47:26] <ssi_> because I have it going out to a clearance plane on both ends
[05:47:51] <toastydeath> you have moved clearly into the territory of yak shaving
[05:47:58] <ssi_> haha
[05:49:39] <toastydeath> the only time I have seen someone work this hard at a taper is when my co-worker had to mill an internal, reverse taper on a part
[05:49:46] <toastydeath> with a 60 degree dovetail cutter
[05:49:53] <ssi_> sounds terrible
[05:50:03] <toastydeath> it took him five minutes because he has a calculator
[05:50:18] <toastydeath> and the problem was never seen again
[05:50:42] <ssi_> I'm not sure what you're implying
[05:50:54] <toastydeath> ?
[05:50:56] <ssi_> but I don't have a roughing cycle that I can feed a start and endpoint and have it do all the work for me
[05:51:04] <toastydeath> no, this was on a mill
[05:51:07] <toastydeath> no roughing cycles
[05:51:07] <ssi_> which is why I wrote one
[05:51:10] <ssi_> yeah, I understand that
[05:51:24] <ssi_> but the implication seems to be that I'm doing something wrong because I'm having to work hard at it
[05:51:29] <ssi_> and maybe I'm just missing the easy solution
[05:51:55] <ssi_> but when I flip this over
[05:52:08] <ssi_> and need to cut 1.625" down to JT6
[05:52:13] <ssi_> I don't think there's a button for that
[05:52:21] <toastydeath> not saying there is
[05:52:27] <ssi_> and I think it's more than plug x degrees into a calculator and make one G1
[05:52:30] <toastydeath> but you are approaching it in a very flourished way
[05:52:37] <ssi_> cause i don't have an infinite horsepower spindle
[05:52:44] <toastydeath> and also, it's one trig problem
[05:52:49] <toastydeath> it's as many g1 lines as you require
[05:53:21] <toastydeath> why don't you develop the subroutine as you need it, rather than thinking it out so far in advance?
[05:53:30] <ssi_> I need it now!
[05:53:42] <toastydeath> I suggest you do one without a subroutine.
[05:54:29] <ssi_> even without the subroutine, I'd do it with a loop
[05:54:56] <ssi_> the only thing I can see doing different is roughing along Z
[05:55:37] <toastydeath> what i am trying to do is get you to go through the steps that will convey the most domain specific knowledge as possible
[05:56:01] <toastydeath> and sometimes that's doing things the long way before jumping to a shortcut.
[05:56:08] <ssi_> I've been doing it the long way
[05:56:16] <ssi_> I'm just trying to clean it up and make it reusable
[05:56:35] <toastydeath> fair enough
[05:57:00] <ssi_> basically, I want to be able to make these holders in two setups
[05:57:11] <ssi_> one setup is raw 1.5" or 1.625" stock in a 3jaw
[05:57:39] <ssi_> faced, shank turned to .75" and tapered on the end 30 degrees, body turned to 1.5", shoulder faced and undercut
[05:57:59] <ssi_> then second op is shank held in a collet and the body machined, specific to whatever tool it will be
[05:58:19] <ssi_> I have all of the first op so far except for undercutting the shoulder
[06:00:31] -!- WalterN has quit [Read error: Connection reset by peer]
[06:34:23] -!- nullie [nullie!~nullie@dhcp249-250.yandex.net] has joined #emc
[06:50:19] -!- nullie has quit [Quit: Ex-Chat]
[06:58:24] -!- e-jones [e-jones!~jkastner@nat/redhat/x-cptacstfozwupvmm] has joined #emc
[07:00:03] -!- toastydeath has quit [Quit: Leaving]
[07:08:17] -!- pjm__ has quit [Quit: TTFO]
[07:12:05] -!- pjm [pjm!~pjm@uhfsatcom.plus.com] has joined #emc
[07:14:14] -!- nullie [nullie!~nullie@dhcp249-214.yandex.net] has joined #emc
[07:19:51] -!- capricorn_one has quit [Remote host closed the connection]
[07:30:21] -!- robh__ [robh__!~robert@5ace706f.bb.sky.com] has joined #emc
[07:33:00] -!- alex_joni [alex_joni!~alex_joni@emc/board-of-directors/alexjoni] has joined #emc
[07:51:57] -!- Dannyboy has quit [Remote host closed the connection]
[07:56:12] -!- Valen has quit [Quit: Leaving.]
[08:09:08] -!- acemi [acemi!~acemi@unaffiliated/acemi] has joined #emc
[08:14:10] -!- schwege [schwege!~veikko@dev.null.fi] has joined #emc
[08:15:36] <schwege> hi, anyone familiar with creating user-space HAL components?
[08:17:22] <schwege> I need to implement an EMC-to-VFD control, the VFD is hooked up over ethernet using modbus over tcp (thus no realtime)
[08:18:23] <schwege> don't worry about the possible lag that userspace creates, this has nothing to do with mills or lathes :-)
[08:19:47] <schwege> the problem is, I'd like to use comp for the convenience of it, but it doesn't allow me to link any external libraries
[08:21:01] <schwege> I'd guess that just preprocessing the .comp and then modifying the code would be the way to go, but what I'm lacking is instructions to replicate what 'comp' is doing on compile and install
[08:40:44] -!- awallin_ [awallin_!~quassel@2001:708:110:1020:224:7eff:feda:7c7d] has joined #emc
[08:53:46] -!- morficmobile has quit [Ping timeout: 246 seconds]
[08:55:22] -!- morfic has quit [Ping timeout: 240 seconds]
[09:19:41] -!- the_wench [the_wench!~the_wench@host81-149-189-98.in-addr.btopenworld.com] has joined #emc
[09:22:48] <schwege> okay, I got one step forward... it seems that somebody has thought of this (see the python source of comp, in function build_usr there are references to extra_compile_args and extra_link_args)
[09:23:08] <schwege> but the option parsing code doesn't like strings
[09:25:05] <schwege> with a bit of twiddling, though, I got it to pass them through to gcc, so not all hope is lost :)
[09:41:47] -!- mk0 [mk0!~geissler@fiztech.basnet.by] has joined #emc
[10:11:05] -!- rooks [rooks!~rooks@102-bem-18.acn.waw.pl] has joined #emc
[10:29:53] -!- mhaberler has quit [Quit: mhaberler]
[10:29:53] mhaberler_ is now known as mhaberler
[10:29:55] -!- mhaberler has quit [Quit: mhaberler]
[10:36:49] -!- Valen [Valen!~Valen@ppp121-44-231-147.lns20.syd7.internode.on.net] has joined #emc
[10:49:26] -!- mhaberler [mhaberler!~mhaberler@macbook.stiwoll.mah.priv.at] has joined #emc
[10:49:38] -!- mhaberler_ [mhaberler_!~mhaberler@imac.stiwoll.mah.priv.at] has joined #emc
[11:03:20] -!- WesBaker [WesBaker!~wes@68.67.115.73] has joined #emc
[11:18:50] -!- pingufan [pingufan!~rainer@goliath.hantsch.co.at] has joined #emc
[11:21:59] <mk0> for psha's log: +375 29 1396208
[11:30:23] -!- Paragon39 [Paragon39!~chatzilla@cpc13-bexl7-2-0-cust145.2-3.cable.virginmedia.com] has joined #emc
[11:41:43] <Paragon39> Hello All, I have just received an 8 x relay card that connects to the printer port for control. I wish to use this card with the Messa 7i43 IO card and am concerned about how much current it may draw. The inputs have 2K resistors which then appear feed into the base of BC639 transistors ( http://www.nxp.com/documents/data_sheet/BC639_BCP56_BCX56.pdf ) I am not sure what figure I need to use...
[11:41:45] <Paragon39> ...to calculate the current draw. Would it be Ibm (Peak Base Current) of 0.2A (200ma)?
[11:42:39] <archivist> just use ohms law on the resistors and voltages
[11:43:43] <Paragon39> OK so 5v/2k = .0025ma right?
[11:45:50] <archivist> more like a sensible number
[11:47:26] <Paragon39> archivist: So it would it be safe to assume that the 7i43 could drive this directly?
[11:51:49] -!- i_tarzan [i_tarzan!~i_tarzan@189.136.250.43] has joined #emc
[11:53:20] -!- ries [ries!~ries@200.125.129.54] has joined #emc
[12:06:30] -!- warlock has quit [Ping timeout: 252 seconds]
[12:07:47] -!- skunkworks has quit [Ping timeout: 240 seconds]
[12:11:13] <jthornton> Paragon39: did you get ngcgui up and going when you got home?
[12:32:12] <Paragon39> jthornton: ngcgui sure it was me JT? I was playing around with the pyvcp and we were discussing ladder logic yesterday but don't remember anything about ngcgui?
[12:33:06] -!- atom1 has quit []
[12:34:48] -!- atom1 [atom1!~Tl@unaffiliated/toml/x-013812] has joined #emc
[12:44:06] <JT-Shop> brain fart on my part it was ssi
[13:04:35] -!- pingufan has quit [Remote host closed the connection]
[13:15:52] -!- Valen has quit [Quit: Leaving.]
[13:22:54] <ssi_> JT-Shop: yea it works locally
[13:23:03] <JT-Shop> cool
[13:23:19] <ssi_> JT-Shop: I wrote a tapering routine that roughs and finishes
[13:24:23] <JT-Shop> nice, did you post it on the forum?
[13:24:27] <ssi_> not yet
[13:24:36] <ssi_> prolly need to tweak it a bit first
[13:26:12] -!- mk0 has quit [Quit: Ухожу я от вас (xchat 2.4.5 или старше)]
[13:33:03] -!- e-jones has quit [Quit: Ex-Chat]
[13:34:12] -!- mk0 [mk0!~x-ray@fiztech.basnet.by] has joined #emc
[13:37:40] -!- skunwkorks [skunwkorks!447329d2@gateway/web/freenode/ip.68.115.41.210] has joined #emc
[13:39:31] <bzzzz> gdb -tui
[13:51:26] <jdhNC> anyone tried the Linux DraftSight?
[14:03:35] -!- mhaberler_ has quit [Quit: mhaberler_]
[14:04:44] <Paragon39> Chaps, What the easiest way to connect a basic 2 wire probe. For example I wish to connect a croc clip to a toolbit and another to a pcb I wish to know when the probe touches the pcb. I have seen motion.probe-input but what is the easiest way to supply the +5v to the probe? Setting a spare output pin high maybe?
[14:04:46] -!- e-jones [e-jones!~jkastner@nat/redhat/x-iwrbmolklckjfjhy] has joined #emc
[14:05:12] <cradek> Paragon39: you can get 5v from a usb port.
[14:06:44] -!- mhaberler [mhaberler!~mhaberler@macbook.stiwoll.mah.priv.at] has parted #emc
[14:07:20] <Paragon39> cradek: True but that means mucking about with a usb cable which I have none spare etc. This is only quick temp solution as I will be adding 5v power supply eventually.
[14:11:22] <Paragon39> I had something similar setup for home switches though in this case the 3 switches were joined together and Emc would go through a sequence of homing each one. The 5v was taken from the paraport but I no longer have the configuration :-(
[14:12:32] <Paragon39> Is there a way of setting a pin to high within HAL?
[14:14:10] <cradek> sure, with setp
[14:15:14] <Paragon39> cradek: It's been a while... I think I need to invest more time in re-reading the EMC Manual ;-)
[14:19:44] -!- mk0 has quit [Quit: http://neo-technical.wikispaces.com/emc2-arch]
[14:21:52] -!- FinboySlick [FinboySlick!~shark@74.117.40.10] has joined #emc
[14:30:50] <Paragon39> cradek: Found it... setp parport.0.pin-08-out TRUE
[14:30:58] <bzzzz> aren't parports 3.3
[14:31:09] <bzzzz> v?
[14:32:21] <bzzzz> also, how do you guys avoid burrs on the copper traces when doing isolation routing?
[14:32:37] <cradek> sharper tool, spin faster
[14:34:14] <bzzzz> cradek: how fast? my 2hp bridgeport only goes up to 1700rpm. as for the tool, are fluted end mills inappropriate for this task?
[14:34:55] <ssi_> D shaped engraving bit spun way faster than that will work better
[14:34:59] <cradek> you want 20000+ rpm, using a tool made for isolation routing PCBs
[14:35:37] <cradek> mount an air powered die grinder or something
[14:35:49] <bzzzz> damn it, so i can't use my bridgeport, which is the only cnc i've got
[14:36:00] <cradek> you can't use your bridgeport's *spindle*
[14:36:06] <ssi_> yea just mount a secondary spindle
[14:36:08] <cradek> be creative and you can use your bridgeport
[14:36:23] <ssi_> people have had good luck with the proxxon grinders
[14:36:54] <bzzzz> that thing is huge
[14:37:17] -!- Birdman3131 has quit [Ping timeout: 276 seconds]
[14:37:19] <ssi_> compared to a bridgeport? :P
[14:37:26] <bzzzz> the disk, i mean...it's 2 inches!
[14:37:58] <bzzzz> cradek: "mount a secondary spindle" - so duct tape, say, a router to the bridgeport head?
[14:38:30] <ssi_> yea, just use machinist's duct tape
[14:38:31] <cradek> yes duct tape is exactly what I had envisioned
[14:38:31] <ssi_> it's more precise
[14:38:45] <bzzzz> christ, i can't tell if you two are being sarcastic now
[14:39:04] <ssi_> http://woodworker.com/fullpres.asp?PARTNUM=117-323&LARGEVIEW=ON
[14:39:17] <ssi_> see that little swivel fixture the one on the left is mounted in?
[14:39:25] <ssi_> make something like that which clamps on your quill
[14:39:30] <ssi_> without the swivel
[14:39:33] <ssi_> just the clampy bit
[14:39:45] <bzzzz> ssi_: i had no idea they make disks that tiny
[14:39:55] <ssi_> never used a dremel? :D
[14:40:11] <bzzzz> apparently not D:
[14:40:13] <ssi_> the proxxons are just dremels with better bearings
[14:40:38] <bzzzz> hm, lemme see how mountable my quill is
[14:40:39] <ssi_> whatever you end up with, you want something that'll hold small tools, spin very fast, and has very low runout
[14:41:49] <cradek> you need to hold 1/8 shank tools, nothing else
[14:41:50] <bzzzz> not sure how i'd mount it and get z-axis motion..
[14:41:55] <ssi_> I made a little high speed engraving spindle,.. bought a 1/2" shank ER16 chuck off ebay, turned a body for it and mounted it between bearings, added a timing pulley and a 140W RC motor
[14:42:34] <cradek> bzzzz: do you have QC30 or R8?
[14:43:53] <bzzzz> cradek: qc30, are you psychic? :o
[14:44:37] <cradek> can you make it attach to a tool holder? or just to the outside of the quill itself? that's a big enough quill.
[14:46:49] <bzzzz> the entire cylinder's ~4 inches when fully extended
[14:47:09] <bzzzz> ssi_'s grinder looks longer than 4
[14:47:40] -!- newbynobi [newbynobi!~norbert@p4FC9A831.dip.t-dialin.net] has joined #emc
[14:47:53] <bzzzz> problem with low rpms is that if i try to mill out smaller traces, i might tear up copper instead of cutting it, right?
[14:49:21] <cradek> you have 5" of travel on that thing
[14:49:43] <cradek> if you're at 1700 rpm you have to cut VERY slowly (like 1 ipm)
[14:50:00] <cradek> your spindle should be able to do 3000 rpm+. is something wrong with your varispeed?
[14:50:24] <bzzzz> um
[14:50:50] <bzzzz> lemme re-read the ratings engraved on the side plate
[14:51:09] <cradek> is it the varispeed head, or do you have belts to change?
[14:52:01] <bzzzz> varispeed
[14:52:29] <cradek> they'll go up to 4000, but the dial turns red (?) above 3000 if I remember right
[14:53:20] <bzzzz> um
[14:53:23] <bzzzz> it's 1700
[14:53:39] <ssi_> must be metric ;)
[14:53:53] <bzzzz> series 1 cnc
[14:54:00] <bzzzz> 1700rpm D:
[14:54:09] <bzzzz> lemme google a manual up
[14:54:43] <anonimasu> I have a secondary spindle to be mounted on my big mill
[14:55:37] <anonimasu> 1hp and 24krpm
[14:56:12] <bzzzz> wtf i don't get it
[14:56:24] <bzzzz> almost every series 1 i get from google goes up to 4krpm
[14:56:31] <ssi_> anonimasu: did you use a router, or what
[14:56:43] <bzzzz> cradek: you're right, varispeed problem
[14:56:51] <cradek> http://timeguy.com/cradek-files/emc/IMG_0948_1.JPG
[14:56:53] <anonimasu> ssi_: no
[14:56:55] <cradek> is it this head?
[14:57:07] <anonimasu> ssi: http://www.youtube.com/user/ahelin0?feature=mhum
[14:57:34] <ssi_> ah nice
[14:58:04] <anonimasu> 3000-22000rpm with nice torque curve
[14:58:10] <anonimasu> err 4000
[14:58:31] <bzzzz> cradek: it's this guy: http://images.google.com/imgres?imgurl=http://www.wotol.com/images/thumbs/800x800/152246_20f17acc5eac0e6bd9ab0fb6879a1775.jpg&imgrefurl=http://www.wotol.com/1-bridgeport-series-1-cnc-vertical-mill-vertical-m/second-hand-machinery/prod_id/152246&usg=__hJNCaXCnedAjMuCTf9h78mLJMUw=&h=640&w=480&sz=44&hl=en&start=0&zoom=1&tbnid=YZ7qn6rQdglW2M:&tbnh=146&tbnw=117&ei=6brCTZTcGOXY0QH2otDwAw&prev=/search%3Fq%3Dbridgeport%2Bseries%2B1%2Bcnc%26hl%3Den%26biw
[14:58:39] <bzzzz> holy crap
[14:58:44] <bzzzz> http://www.wotol.com/images/thumbs/800x800/152246_20f17acc5eac0e6bd9ab0fb6879a1775.jpg
[14:58:54] <bzzzz> sorry for the spam
[14:59:19] <skunwkorks> http://www.youtube.com/watch?v=n3wPBcmSb2U&feature=related
[14:59:42] <cradek> ah, that's a very different head
[14:59:50] <anonimasu> ssi_: tho it'l be really interesting to stick it into some aluminium ^_^
[15:00:00] <bzzzz> but according to http://www.wotol.com/1-bridgeport-series-1-cnc-vertical-mill-vertical-m/second-hand-machinery/prod_id/152246, that one runs up to 4.2krpm. so why is mine capped at 1.7k? D:
[15:00:15] <cradek> hm, no idea
[15:00:25] <ssi_> anonimasu: who makes that spindle?
[15:00:39] <bzzzz> cradek: thanks for trying anyway
[15:00:51] <anonimasu> isel
[15:00:58] <cradek> bzzzz: original control still working on yours?
[15:01:05] <anonimasu> but they seem to build them to order in germany
[15:01:26] <ssi_> http://www.milling-spindles.isel.com/milling_spindles.php#isa750
[15:01:28] <ssi_> something like that?
[15:01:29] <bzzzz> cradek: you mean the free-swinging panel at the right? yeah.
[15:01:34] <anonimasu> not cheap tho, but for the option of a UFM router spindle it was alot better
[15:01:44] <anonimasu> it's that 750 one
[15:02:08] <cradek> bzzzz: it's going to take a lot of paper tape to cut a pcb on that thing :-)
[15:02:14] <bzzzz> cradek: your pic says series 1 too, but it looks a lot lot newer :o
[15:02:28] <anonimasu> if you want a cheap spindle that works good they have a UFM series
[15:02:35] <anonimasu> which is like 1/10 of the price
[15:02:38] <cradek> yes series 1/2 is just the size of the main castings I think. there's a lot of variation.
[15:02:40] <anonimasu> err 3/10
[15:02:41] <ssi_> what's the cost of that 750?
[15:02:47] <anonimasu> like 1300eur
[15:02:52] <ssi_> yea that's spendy
[15:02:55] <anonimasu> with the inverter
[15:03:19] <bzzzz> oh well, guess i'll have to etch
[15:03:19] <anonimasu> 600 for the spindle without inverter
[15:03:37] <bzzzz> or build a cnc :>
[15:03:51] <anonimasu> what about a air driven spindle?
[15:04:16] <ssi_> can't you get some sort of gear hed that mounts in the spindle and gives you rpm multiplication?
[15:04:22] <anonimasu> if your cnc works you could build a pelton turbine
[15:04:38] <anonimasu> and drive a tool with that
[15:04:53] <cradek> yes, I've seen a QC30 speeder (once in my life)
[15:05:07] <cradek> I suspect they are very rare
[15:05:09] <anonimasu> they are very expensive
[15:05:17] <anonimasu> the speeders
[15:05:41] <cradek> yes even speeders for common spindles are expensive
[15:06:25] <anonimasu> a secondary spindle seemed better to me then buying a speeder off ebay unseen
[15:06:44] <ssi_> yeah cheapest easiest solution is a quill-mount secondary spindle
[15:06:56] <anonimasu> check isel ufm :)
[15:07:42] <ssi_> I need to find a deal on a series I :/
[15:08:02] <anonimasu> http://www.isel-germany.de/products/product.php?lang=en&ID=p110
[15:08:12] <bzzzz> mom says she'll kill me if i break her cnc
[15:08:25] <ssi_> hahahaha
[15:08:27] <anonimasu> I even saw someone make a iso something changer for one of thoose
[15:08:28] <anonimasu> :)
[15:08:37] <bzzzz> bz.courage_to_touch_spindle--
[15:08:52] <ssi_> sounds like what you need is your own little cnc engraver :D
[15:09:08] <anonimasu> http://www.usovo.de/shop/product_info.php?language=en&info=p528_Toolchangerwsystem-ISEL-UFM-500.html&XTCsid=01eacf9a7096fe8ff4397cb42528aefb
[15:09:18] <ssi_> anonimasu: how's the runout on the UFM ones
[15:10:02] <anonimasu> the seller of mine said it's very good for the price
[15:10:20] <ssi_> bzzzz: http://gallery.mac.com/tmbg/100089/IMG_1298/web.jpg?ver=13034445290001
[15:10:42] <bzzzz> ssi_: you built that :O
[15:10:45] <ssi_> yup
[15:10:50] <bzzzz> how long did it take you?
[15:11:04] <ssi_> two or three weeks of piddling with it evenings
[15:11:04] <bzzzz> is that...wood?
[15:11:07] <ssi_> aluminum
[15:11:15] <bzzzz> cool
[15:11:24] <anonimasu> ssi_: I dont own a ufm, it just looks like a better router spindle with better bearings
[15:11:30] <anonimasu> (better design)
[15:11:40] <ssi_> anonimasu: yeahh, that's what I'd figure
[15:11:54] <anonimasu> and they are made for cnc:ing
[15:12:04] -!- psha [psha!~psha@213.208.162.69] has joined #emc
[15:12:27] <ssi_> I need to get electronics and make that little machine run
[15:12:35] <ssi_> I want to redesign the Y axis a bit
[15:12:45] <ssi_> but my little lathe and mill are priority right now :P
[15:12:55] <bzzzz> x is the bottom one, right?
[15:13:02] <ssi_> bottom screw is Y
[15:13:10] <ssi_> gantry screw is X
[15:13:13] <anonimasu> seems like it would be miles better then the wood routers...
[15:13:14] <bzzzz> hm
[15:13:37] <ssi_> there's not enough width in the bearing block that the gantry's mounted on
[15:13:39] <ssi_> and it rocks a bit
[15:13:47] <ssi_> so I need to either widen it up a lot and lose some travel
[15:13:56] <ssi_> or make a new base where the gantry is fixed, and put the table itself on slides
[15:14:04] <ssi_> I'm leaning toward the latter
[15:14:21] <ssi_> cause I can make the base 24" long and get a full 12" Y travel if I want
[15:15:06] <bzzzz> is that cat hair? hehe
[15:15:11] <ssi_> ahhaha yep
[15:15:22] <ssi_> that's the cat's favorite chair
[15:15:54] <cpresser> what is wrong with this line of gcode? "#<x0> = [#<x0> + 86]"
[15:16:09] <ssi_> looks fine to me
[15:16:23] <cpresser> emc2 complains about a missing equal-sign
[15:17:11] <cpresser> complete code is here: http://nopaste.info/3bbac86ffb.html
[15:18:16] <cradek> weird error, but I see line 31 is an error
[15:18:28] <cradek> assume you meant to have a Z in there
[15:19:05] <cpresser> exactly
[15:19:07] <cradek> oh I bet it thinks 31 has the missing = sign, and you've got the wrong line.
[15:19:12] <cpresser> ty cradel
[15:19:33] <cpresser> ~cradek
[15:23:59] -!- nullie has quit [Quit: Ex-Chat]
[15:28:03] -!- WalterN [WalterN!~walter@tiwake.com] has joined #emc
[15:36:21] -!- capricorn_one [capricorn_one!~capricorn@zima.linwin.com] has joined #emc
[15:38:16] <cradek> cpresser: emc reads along the line from left to right. when it sees a variable with no word letter in front of it, the only thing following the variable that makes sense is an = to assign to it. that's why you got the error you did.
[15:41:22] -!- mhaberler [mhaberler!~mhaberler@imac.stiwoll.mah.priv.at] has joined #emc
[15:41:55] -!- mhaberler_ [mhaberler_!~mhaberler@macbook.stiwoll.mah.priv.at] has joined #emc
[15:47:10] -!- e-jones has quit [Quit: Ex-Chat]
[16:12:14] <skunwkorks> http://www.practicalmachinist.com/vb/cnc-machining/emc2-mach3-servo-conversion-221759/index7.html#post1569426
[16:18:30] -!- e-jones [e-jones!~jkastner@nat/redhat/x-vathbbbzxuyupbtl] has joined #emc
[16:28:15] -!- e-jones has quit [Quit: Ex-Chat]
[16:29:13] -!- morficmobile [morficmobile!~morfic@67.11.29.140] has joined #emc
[16:30:13] -!- morfic [morfic!~morfic@67.11.29.140] has joined #emc
[16:30:17] -!- mk0 [mk0!~Geissler@178.121.160.97] has joined #emc
[16:42:55] -!- ve7it [ve7it!~LawrenceG@S0106009027972e37.pk.shawcable.net] has joined #emc
[16:50:51] -!- aggrav8d [aggrav8d!~aggrav8d@S0106000b6a15cda1.vc.shawcable.net] has joined #emc
[16:51:05] <aggrav8d> kimk?
[17:22:18] -!- mhaberler_ has quit [Quit: mhaberler_]
[17:22:26] -!- mhaberler has quit [Quit: mhaberler]
[17:24:19] -!- awallin_ has quit [Remote host closed the connection]
[17:47:42] -!- vladimirek has quit [Ping timeout: 248 seconds]
[17:48:12] -!- vladimirek [vladimirek!~vladimire@bband-dyn166.178-41-220.t-com.sk] has joined #emc
[17:59:56] -!- new_by_nobi [new_by_nobi!~norbert@p4FC9B872.dip.t-dialin.net] has joined #emc
[18:00:50] -!- newbynobi has quit [Ping timeout: 264 seconds]
[18:03:30] <mrsunshine_> ahh casting pattern painted and ready for tomorrow =)
[18:05:20] <ssi_> :D\
[18:05:49] <mrsunshine_> new motor mount comming up, tho have not calculated amount of alu needed and i cant remelt my old motor mount as i need it to face of the surfaces of the new one :P
[18:06:00] <mrsunshine_> so i need to scavange some alu tomorrow also
[18:06:01] <mrsunshine_> sigh :P
[18:06:13] <ssi_> are you doing your own pours?
[18:06:16] <mrsunshine_> ofc, i have a huge block just laying around
[18:06:21] <mrsunshine_> ssi_, aye =)
[18:06:23] <ssi_> nice
[18:06:29] <ssi_> it's something I've been wanting to get into
[18:06:32] <ssi_> for a loooong time
[18:07:16] <mrsunshine_> its easy
[18:07:25] <jdhNC> does the Al have the same properties after melting/recasting?
[18:07:43] <mrsunshine_> some perlite or vermaculite ... fire mortar, a bucket, a couple of pipe couplings a mig welding tip, some work and a propane bottle :P
[18:07:54] <mrsunshine_> jdhNC, i guess ... if you dont overheat it
[18:07:59] <mrsunshine_> tho i do not heat treat my parts
[18:08:03] <mrsunshine_> havent figured that out yet
[18:08:13] <mrsunshine_> so i just let them naturaly age where they sit :P
[18:08:15] <ssi_> it's the refractory that I haven't been able to get my hands on
[18:08:21] <ssi_> not that I've tried THAT hard
[18:08:45] <mrsunshine_> ssi_, i just buy ordenary 1300 degree furnace mortar or whatever its called
[18:08:53] <mrsunshine_> 1300 as i celcius
[18:08:57] <ssi_> ah ok
[18:09:14] <ssi_> also, I would want to do the melting and pouring outdoors in a sandpit or something
[18:09:17] <mrsunshine_> holds up just fine for alu =)
[18:09:21] <ssi_> and I just don't know that I have the space for it right now
[18:09:32] <mrsunshine_> had my bucket for like 5 years now
[18:09:40] <mrsunshine_> done a bunch of pours with it =)
[18:09:52] <ssi_> are you using a homemade crucible or commercially made
[18:10:22] <mrsunshine_> sawed in half fire extinguisher... however that is spelled :P
[18:10:27] <ssi_> haha nice
[18:10:33] <mrsunshine_> talk about irony :P
[18:10:55] <mrsunshine_> made to put out fires, now its for meling metal ... :P
[18:10:57] <mrsunshine_> poor thing =)
[18:11:06] <ssi_> and you're using wood patterns?
[18:11:16] <mrsunshine_> for the most part
[18:11:21] <mrsunshine_> tho looking more and more on lost foam
[18:11:37] <anonimasu> wont you get inclusions by the steel into your cast Al
[18:11:41] <mrsunshine_> but when its a complex part i like wood patterns as i find it a pita to make the foam patterns :P
[18:11:51] <ssi_> cnc hot wire :D
[18:11:58] <mrsunshine_> anonimasu, from what i understand its not that bad :P
[18:12:18] <mrsunshine_> anonimasu, and for home use i dont realy care much, as long as it holds up its end :P
[18:13:12] <mrsunshine_> ssi, yeah, ive got myself some of those ball sliding bearings and going to build one when i get the time =)
[18:13:48] <mrsunshine_> but there also, complex parts is a pita as you would have to do it in many small pieces, then glue it together, then do fillets etc, havent figured the fillet part out :/
[18:13:58] <mrsunshine_> as sharp corners are your enemy when casting =)
[18:15:09] <mrsunshine_> ive still not had a casting fail on me =)
[18:15:14] <mrsunshine_> and its a dream to cut with cnc :P
[18:15:17] <mrsunshine_> compared to soft alu =)
[18:16:33] <jdhNC> any voids/etc from moisture boiling out of the wood molds?
[18:17:27] <mrsunshine_> are you talking about the patterns?
[18:17:33] <mrsunshine_> the wood things are not the problem there :P
[18:17:35] <jdhNC> I guess
[18:17:40] <mrsunshine_> as you remove them :P
[18:17:56] <mrsunshine_> the sand on the other hand, not had a problem as i see to it that the sand is just moist enough to hold together well
[18:18:38] <jdhNC> it's just sand? no binder? not permanent?
[18:18:54] <mrsunshine_> bentonite clay
[18:18:56] <mrsunshine_> as binder
[18:19:23] <mrsunshine_> and no voids so far, not big ones atleast, when lost foam casting ive caused some myself as im not used to it cooling alot slower then with green sand
[18:19:53] <mrsunshine_> have to make myself a molding bench and stuff tho, atm its very crude :P
[18:48:07] -!- robh__ has quit [Quit: Leaving]
[18:54:54] <KimK> aggrav8d: I'm here now
[18:56:50] <aggrav8d> so i'm told my machine cuts 80ipm and rapids 700ipm. if i really can cut 80ipm then at 32000rpm my 2-flute bit is shaving 0.00125" per turn. But I'm losing steps. So is this too much?
[18:57:40] -!- acemi has quit [Quit: WeeChat 0.3.2]
[18:58:02] -!- isssy [isssy!~isssy@78-83-51-185.spectrumnet.bg] has joined #emc
[18:58:34] -!- acemi [acemi!~acemi@unaffiliated/acemi] has joined #emc
[19:00:05] <aggrav8d> i mean, if i cut 5ipm then that's 0.00007812" per turn.
[19:00:20] <aggrav8d> that seems... well, i dunno. it seems pretty small.
[19:00:55] <aggrav8d> less than 1/5000 of a mm.
[19:01:40] <KimK> I don't know, it would be cutting quite a bit by volume, but it's wood (or was it MDF) and a router. It doesn't sound terrible for a, what, 1.5 HP router? As far as losing steps, you could be cutting air and still be losing steps so the alleged maximum speeds don't really prove too much.
[19:02:33] <KimK> Are you using a carbide tipped bit?
[19:02:45] <aggrav8d> 1.5hp, yes.
[19:02:56] <aggrav8d> mdf.
[19:03:01] <aggrav8d> hrm.
[19:03:14] -!- mhaberler [mhaberler!~mhaberler@imac.stiwoll.mah.priv.at] has joined #emc
[19:03:30] <KimK> hrm?
[19:04:14] <aggrav8d> thinking. it's got a z-increment of 0.015". 0.015*0.00125" is considered a lot of volume?
[19:04:28] <aggrav8d> i just don't know what is considered a lot.
[19:05:37] <KimK> Wait, the minimum Z step is 0.015"? That's about 1/64", or about 15-30 times most Z steps you normally run across.
[19:06:43] <Connor> Question guys.. what would be the best way to send Positional Data from a pendant to EMC.. (Over USB Serial).. Send a "tick" for each pulse from the MPG, or actual distance ?
[19:07:11] <cradek> usb serial is the wrong way to do it. you should hook the mpg up to something that can read it in realtime.
[19:07:22] <cradek> otherwise you'll have crappy mpg response.
[19:08:03] <KimK> aggrav8d: Oh, no, sorry. I wasn't thinking about it correctly, you're meaning it as if planing the edge of something, I was thinking of it as if making a "typical" (1/8" wide?) groove. Nevermind.
[19:08:35] <KimK> aggrav8d: But my question about the Z step size remains.
[19:09:18] <KimK> aggrav8d: What is your step size on the other two axes?
[19:09:39] <KimK> Hi Connor, cradek
[19:09:53] <cradek> hey
[19:10:34] <aggrav8d> really? 15-30 times more than most CNC machines will do?
[19:11:15] <aggrav8d> maybe I'm missing something. I've only been working with G-code i made myself. Do you have an example of g-code i can load in cambam that shows how something is SUPPOSED to be set up?
[19:12:31] <anonimasu> PCW: is it possible to interface a heidenhain linear scale with your boards?
[19:12:35] <Connor> cradek: okay, So, any ideas then? Wanting to use this based on a Arduino...
[19:12:48] <cradek> yes, the idea is don't use an arduino
[19:12:54] <anonimasu> 1�a pp.. with proper condisioning hardware ofcours... :)
[19:12:57] <anonimasu> conditioning
[19:13:02] <cradek> hook up the mpg to EMC using any realtime inputs, such as a parallel port
[19:13:22] <cradek> you can hook up any other switches/buttons that way too
[19:13:28] <ssi_> <3 mesa
[19:13:29] <ssi_> hehehh
[19:13:34] <KimK> aggrav8d: Well, not, "will do", but are designed for, I mean that the smaller step size gives you more cutting options, better cutter control (small stepper is more "geared down"), etc. Also some machines microstep their steppers which helps more, especially with mechanical resonance problems, which can lead to loss of steps and positional confusion.
[19:13:35] <PCW> anonimasu should be possible (isnt it 11 uA)
[19:13:43] <cradek> sure, or a mesa if you have that
[19:13:44] <anonimasu> err yeah
[19:13:55] <anonimasu> I have a spare one that would be very very nice for my lathe
[19:14:05] <anonimasu> :)
[19:14:23] <KimK> aggrav8d: What is your minimum step size on X & Y ?
[19:14:30] <anonimasu> should be significantly better then the chinese scales ^_^
[19:15:51] <aggrav8d> minimum isn't 0.015 on z. that's just the step size i use atm. let me go get the emc2 config file for you.
[19:16:16] <ssi_> what kind of machine is it?
[19:16:25] <Connor> cradek: Well.. That kind of not what I wanted.. I wanted a simple USB based pendant with a LCD in it for info and be able to toggle specific stuff on/off etc as well as have the DRO and stuff show on the LCD
[19:17:02] <KimK> ssi_: You're asking about anonimasu's lathe, right?
[19:17:07] <ssi_> aggrav8d's machine
[19:17:14] <anonimasu> the one for me is a easiturn
[19:17:21] <anonimasu> for further ref...
[19:17:27] <aggrav8d> kimk - http://pastebin.com/6vVi7cpy
[19:17:29] <cradek> Connor: that's a fine goal, and usb would work fine for most buttons and dro, but you want the mpg to be realtime
[19:18:15] <KimK> ssi_: Oh, OK. It's a gantry mill (router) made out of MDF and some metal bits (bearings, chains, sprockets, etc.) A Blacktoe, was it?
[19:18:36] <ssi_> ah ok
[19:18:44] <Connor> The only way to have the MPG be realtime would be to have a separate wires running to the parport or something.
[19:19:25] <cradek> yeah
[19:19:56] -!- toastydeath [toastydeath!~toast@c-69-140-223-139.hsd1.md.comcast.net] has joined #emc
[19:20:39] <aggrav8d> kimk - correct. actually it's made of MDO.
[19:20:51] <aggrav8d> everything else you said was right.
[19:21:54] <aggrav8d> i have it set for 1/16th? microstepping (it's been a while, I have to check)
[19:22:17] -!- acemi has quit [Quit: WeeChat 0.3.2]
[19:22:53] -!- skunwkorks has quit [Ping timeout: 252 seconds]
[19:24:26] <KimK> aggrav8d: Thanks for posting the file. Z=6400 steps inch, very fine pitch, nice. OK, microstepping, got it. And how did you arrive at the X & Y (calibrated?) values? It shows X is 1420 minus a smidge (to 7 decimal places), and Y is 1420 plus a smidge (to 7 decimal places)?
[19:24:54] <ssi_> I'm guessing he measured a move and applied a calibration factor
[19:24:58] <ssi_> I had to do that on my plasma table
[19:25:05] <aggrav8d> i marked 0,0 on the table, ran it to y=48, measured the difference, did the math.
[19:25:14] <aggrav8d> yeah, what ssi said.
[19:25:21] <aggrav8d> poor mans' method.
[19:25:52] <ssi_> I'm actually pretty surprised that moves on my lathe measure exactly as commanded with no scale adjustments
[19:25:56] <ssi_> my scales are right at 8000
[19:26:11] <aggrav8d> wow. nice construction! :)
[19:26:17] <KimK> ssi_: I'm guessing that too, but I wanted to hear it from him. aggrav8d: OK, sounds good, just curious.
[19:26:33] * aggrav8d still can't figure out why robotic arms are built by humans. wtf.
[19:26:36] <ssi_> heheh
[19:26:49] <ssi_> I'm not sure why human arms aren't built by robots!
[19:27:39] <KimK> aggrav8d: Have you run the cloud torture test experiment yet (tort.ngc)?
[19:28:06] <aggrav8d> robot porn: http://www.youtube.com/watch?v=0qM0znv70IA
[19:28:13] <aggrav8d> i haven't even found the file.
[19:28:19] <aggrav8d> brb, going to go look for it.
[19:28:58] -!- i_tarzan has quit [Ping timeout: 260 seconds]
[19:29:09] <KimK> Or, just to get started, what's the approximate size of your work envelope?
[19:29:18] <Connor> cradek: Do you think sending A/B encoder pulses over serial at 115k would be fast enough for most cases?
[19:30:10] <KimK> Connor: Easy. What's your encoder resolution and how fast are you spinning it?
[19:30:24] <Connor> 100 Click per
[19:30:27] <cradek> you were talking about usb before. now are you considering using a serial port?
[19:30:30] -!- psha has quit [Quit: leaving]
[19:30:46] <Connor> cradek: Serial OVER USB
[19:31:29] <cradek> you can't read usb in realtime, so whether it resembles serial somewhere along the line is moot
[19:31:48] <cradek> you'll be reading usb with a userland program, which adds more latency
[19:32:07] <cradek> a jog wheel will NOT work well in this configuration. that's why the motion controller in EMC reads jog wheels directly in realtime.
[19:32:49] * KimK wonders if KimK could send serial over usb, over ethernet, and send in the results by postal mail...
[19:33:36] <aggrav8d> kimk - found it.
[19:33:43] <cradek> before we had jogwheel handling in motion, I did it to halui (a userland program). It sucked, which is why we now have it in motion.
[19:33:47] <aggrav8d> will exceed minimum on x and y.
[19:33:49] <aggrav8d> boo!
[19:34:11] <schwege> KimK: just wait for power over ip ;)
[19:34:14] <Connor> cradek: What's the concern.. That it misses pulses, or just lags to far behind the machine movement?
[19:34:14] <aggrav8d> btw, thank you kimk for all the hand holding. I am doing my best to grok this stuff the first time around.
[19:34:33] <cradek> if you hook up the A,B to an actual serial port, you can read those bits in realtime, but it's not a serial protocol
[19:34:48] <KimK> aggrav8d: No problem. What's the size of your work envelope?
[19:35:33] <cradek> Connor: lagging and general unresponsiveness. but sure you could also drop pulses, depending on the details of your implementation. :-)
[19:35:57] <aggrav8d> 97x49x9. the limits are x=0,97 y=-0.5,48.5 z=-3,6
[19:36:57] <Connor> cradek: I guess I'll just have to play around with it and see how bad it is.. What do most people use pendants for these days? Real Machining, or just jogging the axis back and forth for setup and tool changes etc?
[19:37:21] -!- e3m [e3m!~IceChat7@217.96.27.167] has joined #emc
[19:37:21] <Connor> If I do it correctly.. I can give the person the option of using parport pins OR the serial method.
[19:38:15] <cradek> that I don't know - I don't think pendants are useful except on very large machines where you can't always reach the workpiece and control at the same time...
[19:38:31] <awallin> for 'manual-machining' arrow-buttons with constant speed are better than the jogwheel
[19:38:39] <cradek> IMO pendants just make things take two hands when there's no reason they can't be done with one hand
[19:38:53] -!- mk0 has quit [Quit: Ухожу я от вас (xchat 2.4.5 или старше)]
[19:39:05] <cradek> yes continuous jog is great for facing, etc
[19:39:14] <Connor> I use my joypad all the time.
[19:39:47] <Connor> Problem with joy pad is you can control X and Y at same time.. Cool thing is, you can control X and Y at the same time.. :)
[19:40:34] <KimK> aggrav8d: OK. Try writing your own exercise code. Try straight line runs, one axis at a time, as Valen suggested. Then work up to a basic helix, maybe. I'll try to get you a torture test that's more easily usable, I have one here somewhere. And you'll probably want a fixture offset, probably G54. Assuming you're keeping G59.3 for your machine offset (0,0,0) as many do.
[19:40:56] -!- i_tarzan [i_tarzan!~i_tarzan@189.136.250.43] has joined #emc
[19:40:57] <Connor> To be honest.. I use mine mostly when I'm setting up stuff.. to position the tool head, touch off and such..
[19:41:59] <Connor> It wouldn't be a big deal to run a extra wire siamesed with the USB cable..
[19:42:31] <Connor> have the MPG hooked to the Arudino so the MPG could also be used for menu selections etc...
[19:42:37] <aggrav8d> any chance you can take tort.nc and offset it to fit inside the envelope?
[19:43:04] <cradek> add a g21 and it'll fit...
[19:45:03] <KimK> cradek, Connor: I happened to be at a Haas dealer with some friends yesterday. He showed a trick on their (toolroom CNC) lathe control where you can turn on chamfering, so that moving the X handwheel is electronically geared to the Z handwheel. You can set it by rise/run or by angle. It looked like something that could be implemented in PyVCP.
[19:45:45] <KimK> OK, PyVCP/hal
[19:46:03] <cradek> if I ever get my lathe jogwheels installed, one will have a mode to do that - easy as pie
[19:47:11] <Connor> Yea.. I could add modes for Axis that would Do X+Y or something..
[19:47:21] <KimK> cradek: I think it has a G21, then it fits, but it's too small. I have a copy that I added scaling throughout to, so you can make it exactly fit. I'll find it and post it to git nc_files. Maybe as tort2.ngc?
[19:47:48] <cradek> if yours is an improvement, why keep the original?
[19:49:19] <KimK> Well, I didn't want to delete anything, but OK. What I'd really like is to recover (rewrite?) that lost "cloud generator" that I asked you about in the past, the one that you said was lost.
[19:49:59] <cradek> jepler wrote that - I bet it would be easier to rewrite than find the original.
[19:50:21] <KimK> Then I could add the scaling generation to the cloud generation, and all is well. Bah, another item on the "things that need doing" list.
[19:50:22] <cradek> making random valid gcode is not too hard a task...
[19:51:02] <KimK> Yes, that's why I thought it might ease me into Python or something.
[19:52:32] <KimK> aggrav8d: tort.ngc will fit inside your envelope (easily at G21!) but you'll have to use G54 to get it offset inside the envelope.
[19:53:24] -!- e3m has quit [Ping timeout: 260 seconds]
[19:59:50] <JT-Shop> KimK: most modern lathe controls can chamfer with the handwheel
[20:00:25] <atom1> so due to speed, mpg on usb is out. what about hooking one and some buttons switches and bells to a 7i43 or buffered thru a 7i47?
[20:00:47] <KimK> aggrav8d: But why don't you try making one-axis runs to check for step loss as Valen suggested. Something like (depending on where you start from, and which direction you go): F80 then G01 X50 then G01 X0. This is using your caliper to check for errors or lost steps.
[20:00:53] * JT-Shop e-mails cradek a jogwheel installing kit
[20:01:14] <atom1> seems to be considerable discussion on them lately
[20:01:56] <Connor> atom1: Not out, but, could be issues for some people..
[20:02:06] <awallin> atom1: reading mpg and button I/O at servo-tread rate (1ms) works fine
[20:03:40] <atom1> what about reading it thru the game port
[20:03:53] <KimK> aggrav8d: If you get errors or lost steps, then try again with different setup values (requires EMC2 restart). You can try G00 too.
[20:03:53] <atom1> or is that the same issue
[20:04:01] <Connor> atom1: How many machines have a game port now?
[20:04:10] <Connor> not mine.
[20:04:29] <atom1> i suppose
[20:04:31] <Connor> I would rather run a extra cable to hook to parport pins..
[20:04:45] <Connor> or have a option to try it usb if not, then us the parport..
[20:05:04] <atom1> so 4 to the wheel and another half dozen for the other buttons etc
[20:05:25] <Connor> 2
[20:05:32] <Connor> A & B.
[20:05:38] <atom1> or a binary switch
[20:05:50] <Connor> +5v and GND can be provided by USB
[20:06:00] <atom1> A&B will get you 4 selections
[20:06:20] <Connor> A&B will be for the MPG. which uses pulses..
[20:06:42] <Connor> If your using a different encoder.. then.. that's something else.
[20:06:50] <atom1> can hal read inputs that way?
[20:07:13] <Connor> It can read the A/B
[20:07:19] <atom1> i'm talking about a binary switch for the selector switch
[20:07:20] <Connor> not sure about a binary switch.
[20:07:33] <atom1> that would get you 4 axis
[20:07:50] <atom1> or 3 and 'off'
[20:08:15] <Connor> You talking about a selector switch ?
[20:08:18] <atom1> i'm considering sharing the data with an avr to send to a lcd
[20:08:23] <atom1> yes
[20:08:50] <atom1> the mpg isn't gonna be read at breakneck speed anyway
[20:08:54] <Connor> Just use a rotary switch with each leg tied to each other via 1k resistor. +5 on one side, gnd on the other and wiper to analog input
[20:09:30] <Connor> Cheap, Simple, Proven.
[20:09:34] <atom1> do you 'range' read them that way?
[20:09:47] <atom1> if it's between this and that it's position 1
[20:10:21] <atom1> or do you rely on a set value
[20:11:05] <Connor> You read the analog value.. 275 to 325 = 1, 375 to 425 = 2 475 to 525 = 3 etc etc..
[20:11:08] <ssi_> if you have a microcontroller in the pendant, just use an A/D... you could have a pretty high number of steps on the selector that way if you wanted
[20:11:21] <KimK> cradek: I'm currently helping a friend with a latency issue. We've changed video cards, helped a lot, runs 10,000-15,000ns generally, nice. (We're only running servos, no base thread, so, 1,000,000ns.) But sometimes (even if left alone) it will suddenly jump up to 24,000,000ns. Looking at RT & troubleshooting in wiki. Any other "usual suspects'?
[20:11:41] <Connor> I have a circuit and code setup like that on my bread board right now.. was testing it last night.
[20:11:58] <atom1> you're the one that wants to kit a pendant right?
[20:12:00] <Connor> FYI, a 128x64 LCD EATS up almost ALL I/O on a Arduino.
[20:12:11] <Connor> atom1: yea.
[20:12:18] <atom1> i'd be using a txt 4 x 20
[20:12:30] <atom1> since i have plenty already
[20:12:52] <atom1> and if i went graphic i'd go with an oled
[20:12:59] -!- vladimirek has quit [Remote host closed the connection]
[20:13:18] <Connor> yea. Well.. I found out how to save 1 pin already.. use a transistor as a inverter for the chip select, then only 1 pin needed.
[20:13:33] <Connor> need to free up 1 more.. I think I can use my I/O expander for the enable..
[20:13:34] <atom1> chip select for what?
[20:13:54] <Connor> on the 128x64, it has two chips, 1 for the left, 1 for the right (64x64)
[20:14:00] <atom1> oh
[20:14:18] <Connor> but, since we can only talk to one at a time.. no reason to burn a pin for the inverse of the other.
[20:15:20] <Connor> Only I/O I need for stuff except for LCD is 2 Analog for the Rotary switches, 2 other analog for I2C I/O and 2 digital on interrupts for the MPG
[20:15:45] <Connor> MIGHT need one more Digital for the I/O expander if I make use of the interrupt to tell the micro to check for a update.
[20:15:52] <Connor> vs polling.
[20:20:54] <atom1> have you found a nice small rotary switch yet?
[20:21:04] <Connor> How "Small" ?
[20:21:22] <Connor> http://www.mouser.com/ProductDetail/Alpha-Taiwan/SR2612F-0112-21R0B-D8-N/?qs=yA6kp8fx8Y7S2TB5GvzfgQ%3D%3D
[20:21:25] <atom1> handheld small
[20:21:49] <Connor> Pretty standard size. 1-12.. You can open them up and set the stop. 30 degrees.
[20:22:16] <Connor> I'm using a radioshack 1-6 double pole for testing ATM.
[20:24:42] -!- new_by_nobi has quit [Remote host closed the connection]
[20:25:06] <atom1> does it come with a knob?
[20:25:19] <Connor> No. You buy those seperate.
[20:25:29] -!- OoBIGeye has quit [Ping timeout: 260 seconds]
[20:25:50] -!- OoBIGeye [OoBIGeye!~kent@194-237-167-11-no48.tbcn.telia.com] has joined #emc
[20:26:36] <atom1> what are you reading the analog with?
[20:26:51] <Connor> Analog 0 and Analog 1 on the Arduino.
[20:27:02] <Connor> 1 each for each rotary switch.
[20:27:44] <Connor> The rotary switch acts like a standard POT with detent and fixed resistance values at each index.
[20:28:02] <atom1> yeah
[20:28:19] <Connor> you specify a range to account for variance, and de-bounce the input and your good to go.
[20:28:31] <atom1> so, one for the axis and one for the resolution
[20:28:38] <Connor> Yea.
[20:28:46] <Connor> and other stuff too.
[20:28:48] <aggrav8d> kimk - i used the offset to move the tort into the envelope, set up a reference point, ran tort.ngc, and then jogged it back to the numeric values of the reference point. I can't get my calipers between the reference and the bit, it's dead on.
[20:30:32] <atom1> Connor, i would suggest a series resistor before the first leg to avoid any direct shorts
[20:30:59] <atom1> or on the wiper
[20:31:18] <Connor> yea.. mine has a 2k, the 1k,1k,1k,1k,1k,1k,2k = 10k. that way I get .5v increments.
[20:31:21] <KimK> aggrav8d: OK, that's a good thing. You made one run so far? And you can add moves at the beginning and end (to/from same point) if that is more convenient for you, so you don't have to jog.
[20:33:37] <KimK> aggrav8d: What did the maximum travels look to be? (guessing is OK)
[20:35:17] <aggrav8d> don't understand the question.
[20:35:25] <aggrav8d> yeah, one run takes about 9 minutes.
[20:43:08] -!- isssy has quit [Quit: Visitor from www.linuxcnc.org]
[20:44:25] -!- MOGLI [MOGLI!~MOGLI@117.229.4.249] has joined #emc
[20:47:13] -!- JustinXJS [JustinXJS!~Justin@c-98-232-177-150.hsd1.or.comcast.net] has joined #emc
[20:49:06] <atom1> Connor i thought it was a 12 pos switch
[20:49:16] <atom1> or did you only wire part of them
[20:49:28] -!- JustinXJS3 has quit [Ping timeout: 260 seconds]
[20:49:45] <Connor> On my test, I only wired 6, since it was a 6 pos double from radio shack.. the 1-12's are on their way.
[20:51:37] <atom1> which mpg did you end up getting to test with?
[20:51:55] <Connor> http://www.pmdx.com/MPG-01
[20:52:53] <atom1> is this the same? http://homanndesigns.com/store/index.php?main_page=product_info&cPath=15&products_id=30&zenid=87985edee7234d38aa76bd9392a1f909
[20:52:57] <Connor> yes.
[20:53:02] <Connor> that one is in .au though.
[20:53:10] <Connor> and, I'm in the US.
[20:53:45] <atom1> what st?
[20:53:51] <Connor> Tenneessee
[20:53:57] <atom1> mmm
[20:54:03] <atom1> ks here
[20:54:39] <atom1> what's he charge for shipping?
[20:54:43] <Connor> 7
[20:54:50] <Connor> USPS Priority.
[21:09:23] -!- pjm has quit [Read error: Connection reset by peer]
[21:09:24] -!- pjm__ [pjm__!~pjm@uhfsatcom.plus.com] has joined #emc
[21:09:28] -!- robh__ [robh__!~robert@5ace706f.bb.sky.com] has joined #emc
[21:14:40] -!- FinboySlick has quit [Quit: Leaving.]
[21:22:22] <aggrav8d> kimk - maximum travels?
[21:23:33] -!- MOGLI has quit [Quit: Visitor from www.linuxcnc.org]
[21:38:52] -!- crazy_imp has quit [Ping timeout: 246 seconds]
[21:40:48] -!- crazy_imp [crazy_imp!~mj@a89-183-74-132.net-htp.de] has joined #emc
[21:47:41] <aggrav8d> damn. i still don't feel like I know what I'm doing. Ok, I'm going to try my test cut again at 10ipm instead of 80 and see how it goes.
[21:54:05] -!- Fox_Muldr has quit [Ping timeout: 276 seconds]
[21:55:33] -!- Fox_Muldr [Fox_Muldr!quakeman@frnk-4d01dafd.pool.mediaWays.net] has joined #emc
[21:56:55] -!- mhaberler has quit [Quit: mhaberler]
[22:07:22] <KimK> aggrav8d: Sorry, I was away for a bit. What did I miss, how's it going?
[22:09:07] <KimK> aggrav8d: Ah, on the maximum travels, I was just asking how far it appeared to move in each axis on the torture test. A guess is close enough.
[22:18:13] -!- stormlight [stormlight!~jasonandt@c-71-202-13-101.hsd1.ca.comcast.net] has joined #emc
[22:20:08] <KimK> aggrav8d: Oops, have to run an errand, back in a bit.
[22:36:06] * robin_sz wakes up
[22:36:10] <robin_sz> is it morning yet?
[22:44:07] <atom1> does the 7i43 have a male or female parport connector?
[22:44:59] <aggrav8d> kimk - i just got back. it took about 35 minutes at 11ipm to cut the gear i showed you in yesterday's gcode.
[22:45:23] <aggrav8d> it looks good, the 5 center holes are all 1" diameter and the teeth look great.
[22:45:49] <aggrav8d> i see a tiny bit of striation but that's probably due to the 0.0625 z step, not inaccuracy of the machine.
[22:46:10] <aggrav8d> incidentally, after I finished the cut I jogged back to origin and it looked to be right on again.
[22:46:25] <aggrav8d> one thing that concerns me is how much I can sense the machine's vibration.
[22:46:51] <aggrav8d> it doesn't do this when the mill is off so I'm guessing the bit isn't perfectly centered? maybe the collet needs cleaning?
[22:47:23] -!- ve7it has quit [Remote host closed the connection]
[22:48:29] <aggrav8d> one problem I've had a lot is cutting too deep or not cutting deep enough.
[22:48:35] -!- ve7it [ve7it!~LawrenceG@S0106009027972e37.pk.shawcable.net] has joined #emc
[22:49:13] -!- Valen [Valen!~Valen@ppp121-44-231-147.lns20.syd7.internode.on.net] has joined #emc
[22:49:22] <aggrav8d> in this case, the bit was touching the surface when I homed the z, but the first cutting pass didn't touch the material and after the last pass the gear was still very firmly attached to the rest of the wood.
[22:50:27] <aggrav8d> sometimes I think I've got it homed and then it cuts a little too deep. I'm sure this is a PEBKAC problem.
[22:52:38] -!- ve7it has quit [Remote host closed the connection]
[22:54:52] -!- ve7it [ve7it!~LawrenceG@S0106009027972e37.pk.shawcable.net] has joined #emc
[22:55:46] -!- ve7it has quit [Remote host closed the connection]
[22:56:01] <Valen> aggrav8d: your either loosing steps, or your gcode is wrong
[22:56:26] <Valen> keep in mind acceleration and velocity will likely be different for each axis
[22:59:03] <aggrav8d> valen - here's the gcode. see anything wrong? http://www.marginallyclever.com/samples/CNC/2011-05-04/small.ngc
[22:59:08] <aggrav8d> should cut 1/2" depth
[23:00:19] <Valen> your using a 5 axis mill?
[23:01:01] <Valen> (you have X Y Z I and J
[23:01:15] -!- factor [factor!~factor@r74-195-188-223.msk1cmtc01.mskgok.ok.dh.suddenlink.net] has joined #emc
[23:01:19] <Valen> also your feed rate is still 80, 20 for the plunge
[23:01:21] <aggrav8d> i and j are for curves, iirc.
[23:01:28] <aggrav8d> i only have a 3 axis.
[23:01:30] <Valen> ahh duh
[23:01:39] <Valen> dont mind me
[23:01:50] <aggrav8d> yeah, it's 80 but I force it in emc2 to be max 11ipm.
[23:01:52] <Valen> (I'm just used to our cam that spits it all out as G1's)
[23:02:13] <aggrav8d> note that the xyzij are g3.
[23:02:26] -!- warlock [warlock!warlock@cpc1-sgyl2-0-0-cust91.sgyl.cable.virginmedia.com] has joined #emc
[23:02:37] -!- stormlight has quit [Remote host closed the connection]
[23:02:43] <Valen> I see that *now*
[23:03:10] <Valen> what are your screws?
[23:03:16] <aggrav8d> screws?
[23:03:19] <Valen> like what type of screw are you using
[23:03:23] <Valen> leadscrew
[23:03:28] <Valen> thing to move stuff around
[23:03:35] <aggrav8d> acme 5 start 2tpi
[23:03:51] <aggrav8d> that's only on the z. the others are chain drive.
[23:03:58] <Valen> wow that is a fast screw
[23:04:06] <aggrav8d> ?
[23:04:15] <aggrav8d> oh.
[23:04:29] <Valen> ours are 15 tpi I think
[23:04:52] <aggrav8d> how many starts?
[23:04:55] <Valen> 1
[23:04:59] <aggrav8d> ah.
[23:04:59] <Valen> ballscrews
[23:05:18] <Valen> that won't affect the speed
[23:05:22] <aggrav8d> probably outlast mine my an order of magnitude.
[23:05:44] <Valen> pushing a 100kg head around though
[23:05:51] <aggrav8d> yikes!
[23:07:03] <Valen> have you looked at the accelerations to see if your loosing steps?
[23:07:40] <Valen> also if your loosing steps when plunging you might not have enough power
[23:13:34] -!- ve7it [ve7it!~LawrenceG@S0106009027972e37.pk.shawcable.net] has joined #emc
[23:14:42] <aggrav8d> i don't know how to look at acceleration and tell if i'm losing steps.
[23:17:56] -!- ve7it has quit [Remote host closed the connection]
[23:19:05] -!- ve7it [ve7it!~LawrenceG@S0106009027972e37.pk.shawcable.net] has joined #emc
[23:20:12] <aggrav8d> any tips?
[23:23:29] <Paragon39> Hello All, Just in from the workshop. I have finally had some really good results milling a pcb's. I setup basic probe within EMC which is used in conjuction with the following python script. It basically probes a grid across the pcb then uses the varied probe results to produce consistent tracks pcbprobe.py https://docs.google.com/leaf?id=0B53oJoJXbpkCMGRlZTYwMWItZjBjYy00OTczLTlmYjAtN2UxZTNiMTBkN
[23:23:31] <Paragon39> DBj&sort=name&layout=list&num=50
[23:23:54] <Paragon39> Ooops ... https://docs.google.com/leaf?id=0B53oJoJXbpkCMGRlZTYwMWItZjBjYy00OTczLTlmYjAtN2UxZTNiMTBkNDBj&sort=name&layout=list&num=50
[23:23:54] -!- ve7it has quit [Remote host closed the connection]
[23:25:52] <Paragon39> And heres a picture of a test pcb. I took no care in mounting it. In fact I had some carpet tape that was folded over from a previous pcb board. https://docs.google.com/leaf?id=0B53oJoJXbpkCODU5NDZiZjAtYTk0Yy00ZTBhLTk2NDctYzE0ZGUwZTFkMmQz&sort=name&layout=list&num=50
[23:26:41] <Paragon39> meaning the pcb was mounted unevenly.
[23:27:21] <JT-Shop> yuck, it want's me to log in :/
[23:27:47] <Paragon39> JT-Shop: those links do?
[23:27:55] <JT-Shop> yea
[23:28:15] <JT-Shop> the last one is the only one I clicked on
[23:28:31] -!- ve7it [ve7it!~LawrenceG@S0106009027972e37.pk.shawcable.net] has joined #emc
[23:28:46] -!- uwe_ has quit [Ping timeout: 246 seconds]
[23:28:51] <Paragon39> OK... Not sure what google docs is doing as they are shared to anyone with link. It keeps doing that. I open them to public..!
[23:29:05] -!- uwe_ [uwe_!~uwe_@dslb-088-064-208-151.pools.arcor-ip.net] has joined #emc
[23:29:24] <JT-Shop> might want to use pastebin.ca if it is up or pastebin.com
[23:30:11] <Paragon39> https://docs.google.com/leaf?id=0B53oJoJXbpkCMGRlZTYwMWItZjBjYy00OTczLTlmYjAtN2UxZTNiMTBkNDBj&hl=en
[23:30:43] <JT-Shop> that one works :)
[23:31:11] <Paragon39> And the picture... (iphone pic) https://docs.google.com/leaf?id=0B53oJoJXbpkCODU5NDZiZjAtYTk0Yy00ZTBhLTk2NDctYzE0ZGUwZTFkMmQz&hl=en
[23:31:31] <JT-Shop> that looks good
[23:32:31] <Paragon39> The tracks came out really well and issolated with a single pass running at 330mm/min from memory. The spindle is rather slow also about 2000rpm.
[23:32:48] <JT-Shop> the cut looks very even
[23:33:01] <JT-Shop> even width
[23:34:13] <Paragon39> I used cheap Chinese 60 degree cutters that where advertised as pcb cutters but where really router bits that had a flat so I filed them to a point with a diamond lap.
[23:34:52] <JT-Shop> # Remember in EMC you can probe up to 4000 points, in Mach3 up to 1000 :) I like this comment
[23:35:17] <JT-Shop> do they need to be pointy?
[23:36:41] -!- ve7it has quit [Remote host closed the connection]
[23:37:24] <Paragon39> I not really sure to be honest but the flat that was on the end was too wide as was so I just followed the contour of the bit so as to reduce. Guess I got lucky ;-)
[23:38:06] <Paragon39> If anyone is interested here is the original gcode before conversion https://docs.google.com/leaf?id=0B53oJoJXbpkCNjRlNTc1NjctZjY0Yi00N2YxLTkwNmMtY2I2ODhmOThjNWFh&hl=en
[23:38:41] <Paragon39> And after conversion :- https://docs.google.com/leaf?id=0B53oJoJXbpkCNTUwN2NlNDQtMTM5Yi00ZGUwLWIxNDQtNDU0ZDI3NjM4OTcx&hl=en
[23:39:34] <Paragon39> JT-Shop: 4000 vs 1000 :-)
[23:39:55] <JT-Shop> yea, it has to be at least a tad smaller than the gap wanted I guess
[23:40:27] <Paragon39> I must confess this is not my code I found it on the cnc forum but I cant find the link for it now.
[23:42:50] -!- ve7it [ve7it!~LawrenceG@S0106009027972e37.pk.shawcable.net] has joined #emc
[23:43:33] <Paragon39> The depth of cut on the above example is .1mm The great thing is it just worked first time, up to this point I just kept recking pcb boards due to the slight variations in height ect.
[23:43:41] -!- aggrav8d has quit [Ping timeout: 240 seconds]
[23:47:28] <ds3> why do you want to mill PCB traces?
[23:48:06] <TekniQue> because it's less work than etching
[23:48:33] <Paragon39> JT-Shop: The board that was milled on is as wavy as gravy. Some arse on ebay was advertising double sided copper clad boards. When I received them he had only gone and stuck a bunch of single sided boards back to back! The ingenuity of it ... lol
[23:48:59] -!- stormlight [stormlight!~jasonandt@c-71-202-13-101.hsd1.ca.comcast.net] has joined #emc
[23:49:34] <Paragon39> ds3: as TekniQue mentions a less work and mess etc. Plus it's quicker for proto typing.
[23:50:16] <Paragon39> The board I showed took around 10min
[23:50:21] <ds3> Paragon39: but it is so limiting...what kind of trace spacing can you do?
[23:50:28] <TekniQue> it's also handy
[23:50:35] <TekniQue> because you can do drilling and paneling
[23:50:40] <TekniQue> on the same machine
[23:50:46] <TekniQue> just change the tool
[23:50:53] <TekniQue> that's the real beauty of it
[23:50:58] <Paragon39> Well you can cut traces for surface mount chips.
[23:51:04] <TekniQue> stick the board down once
[23:51:12] <ds3> what's the min width of traces?
[23:51:20] <TekniQue> mill, drill and cut out
[23:51:52] <TekniQue> ds3: I've done 0.3mm
[23:52:11] <Paragon39> ds3: I'm not sure how far I can get down to as yet I will test that tomorrow with a template.
[23:52:35] <ds3> TekniQue: how close can you get the 0.3mm traces next to each other?
[23:52:49] <ds3> actually, what kind of tools are you guys using for it?
[23:55:06] <ds3> the reasonI am saying all this is - most of the interesting parts seems to be 0.5mm pitched which means I need about 0.2 to 0.15mm width traces with the same amount of space and the tools I tried didn't seem to be able to do it
[23:55:19] <Paragon39> Im using these but they are not the correct tool for the job although it mention they are for pcb's. I filed the point down with a diamond lap to take out the flat on the botom. http://cgi.ebay.co.uk/5x-Carbide-PCB-Engraving-Bits-CNC-Router-Tool-60D-0-1mm-/260779994233?pt=UK_Home_Garden_PowerTools_SM&hash=item3cb7b2ec79
[23:56:28] <Paragon39> ds3 check out pcb-gcode for eagle cad. (free) http://www.brusselsprout.org/PCB-Routing/
[23:56:36] <TekniQue> ds3: http://phk.freebsd.dk/CncPcb/
[23:58:09] <Paragon39> TekniQue: Arhhh thats the original author of the C based code that the above python code was based on.
[23:58:41] <ds3> that's the kind of bit I tried; the cut area expands as you go deeper making 0.2mm hard to do