Back
[00:51:25] <nlancaster> anyone have experience with beagle cnc?
[00:53:19] <CaptHindsight> nlancaster: you might get more useful info for the BBB on the mail list
[00:53:32] <nlancaster> didn't know there was a mailing list :D
[00:54:09] <CaptHindsight> yes, lots of discussion on the BBB with Linuxcnc and a few different IO boards
[00:54:53] <nlancaster> for that matter I have another question people here may be able to answer. anyone know of any other good low cost 4axis kits. like the xylotex?
[00:55:09] <CaptHindsight> servo or stepper?
[00:55:13] <nlancaster> stepper
[00:55:16] <nlancaster> low cost is the name of the game
[00:55:33] <CaptHindsight> reliability a concern?
[00:55:36] <nlancaster> want to build a system to sell locally for a small light duty cnc router. based around openbuilds designs
[00:55:54] <nlancaster> realiability in what away? I don't want it to die tomarrow
[00:56:32] <uw> sell for a?
[00:56:39] <CaptHindsight> there are several of those multi axis toshiba stepper driver boards
[00:57:09] <nlancaster> small light duty hobby cnc router uw
[00:57:24] <nlancaster> right now looking at total cost of parts about $700 for 30x30x4.5 inches
[00:57:55] <nlancaster> 225
[00:57:59] <uw> that sounds about right nlancaster
[00:58:17] <CaptHindsight> with plywood base and table?
[00:58:38] <uw> if you dont need super precision and just want to make things for cutting wood signs and such
[00:59:22] <CaptHindsight> http://www.aliexpress.com/item/4-axis-TB6560-Stepper-Motor-Driver-CNC-Controller-Board-V-Type-for-Engraving-Machine/1463283043.html
[00:59:37] <CaptHindsight> boards like this? ^^^
[00:59:45] <nlancaster> i have heard sad things about that board.
[00:59:53] <nlancaster> probabbly going to go with this.
[00:59:53] <nlancaster> http://xylotex.netfirms.com/OSCommerce/catalog/product_info.php?cPath=26&products_id=40
[01:00:07] <nlancaster> unless someone has a difffernt line on a decent setup that is cheaper
[01:00:21] <uw> lik what? the TB6560 drivers arent terrible for alot of uses
[01:00:37] <nlancaster> i have heard alot of blown drivers and capacitor issues
[01:01:24] <CaptHindsight> a few people use them here with some minor mods with pretty good success
[01:01:32] <nlancaster> don't want to have to mod them
[01:02:05] <nlancaster> I have heard good things about the xylotex product for several years. never used one
[01:02:07] <CaptHindsight> I don't see which driver chips they use
[01:02:21] <uw> i agree i dont see that
[01:03:22] <CaptHindsight> http://www.leadshine.com/producttypes.aspx?type=products&category=stepper-products&producttype=stepper-drives
[01:04:15] <uw> they look pricey
[01:04:45] <uw> but yea if you're not willing to mod/fix things yourself it's going to get harder to come under the $700 price point
[01:04:56] <nlancaster> thats my cost. not sell price :D
[01:05:04] <CaptHindsight> http://www.automationtechnologiesinc.com/products-page/stepper-motor-drivers/microstepping-driver-kl6050 are these expensive?
[01:05:22] <uw> oh you're going to sell this set up
[01:05:24] <nlancaster> yes
[01:05:35] <nlancaster> hope to sell aobut 10 of them locally
[01:05:43] <nlancaster> not going to make a career out of it.
[01:06:21] <nlancaster> mostly to raise funds to either 1, get a cnc routerparts 4848, or build my own around chinese 20mm supported rail and ballscrews.
[01:06:38] <nlancaster> on top of selling other items made on my current router
[01:06:42] <nlancaster> joescnc2006
[01:06:45] <CaptHindsight> http://smoothieware.org/smoothieboard how about this junk?
[01:06:58] <nlancaster> i have heard so so on thatone
[01:07:28] <uw> never heard of that
[01:07:37] <uw> kickstarters are lame though
[01:07:41] <CaptHindsight> http://www.dhgate.com/product/cnc-4-axis-stepper-motor-driver-board-tb6560/151239010.html
[01:07:45] <nlancaster> also was looking at this.
[01:07:45] <nlancaster> http://zapmaker.org/projects/grbl-controller-3-0/
[01:08:56] <nlancaster> https://www.synthetos.com/project/grblshield/
[01:09:32] <uw> too bad that smoothie board doesnt run linux
[01:09:52] <uw> it's not a bad idea
[01:10:05] <nlancaster> no, and I have heard good and bad things about it.
[01:10:10] <nlancaster> but I think its mostly directed at 3d printers
[01:10:20] <uw> but i kinda like having the non-integratedness of a separate computer and drives
[01:10:26] <CaptHindsight> grblShield uses TI DRV8818 stepper drivers
[01:10:39] <nlancaster> is that good or bad?
[01:11:45] <CaptHindsight> 2.5 amps per winding (bipolar steppers)
[01:12:08] <uw> thats not bad
[01:12:15] <uw> for most stuff right?
[01:12:21] <CaptHindsight> http://synthetos.myshopify.com/products/gshield-v5
[01:12:25] <uw> i dont even know what current mine draw typically
[01:12:30] <nlancaster> i have seen a few people working with it. and I just want it for a light duty belt drive cnc router
[01:12:38] <nlancaster> my joes cnc runs 36v at 3amps
[01:12:43] <nlancaster> thru kl4030 driers
[01:12:45] <nlancaster> drivers
[01:13:51] <CaptHindsight> somebody built >50 routers for cheap with the 4-axis Chinese drivers and epson printer steppers
[01:14:00] <nlancaster> lol
[01:14:26] <nlancaster> this is my insperation.
http://www.openbuilds.com/builds/openbuilds-ox-cnc-machine.341/
[01:14:28] <uw> ugh epson printer stepper?
[01:14:34] <uw> do you have any pics for that?
[01:15:06] <uw> must have either been real small ass cnc, or really shitty
[01:16:02] <CaptHindsight> how do the open builds routers hold the parts to the table?
[01:16:28] <uw> looks like tape
[01:16:48] <nlancaster> tape, screws, add track if you want.
[01:16:48] <nlancaster> this is supposed to be a cheap machine
[01:16:58] <uw> i'll say LOL
[01:17:30] <CaptHindsight> wouldn't it just be cheaper to just use a router at a hackerspace?
[01:17:44] <nlancaster> we don't have any hackerspaces in portland
[01:17:51] <nlancaster> we have a few "maker clubs" but they charge a fortune
[01:18:03] <Mr_Mayhem> Very clean looking design. Wonder about rigidity and whatnot.
[01:18:04] <nlancaster> and the only routers they have are farmed out to help pay for the maker club
[01:18:18] <nlancaster> light duty Mr_Mayhem , it is belt drive after all
[01:18:29] <Mr_Mayhem> I see, not ballscrew.
[01:18:32] <CaptHindsight> lots of people around that don't value their time that would router for you for next to nothing
[01:20:03] <Mr_Mayhem> Chain would be an interesting upgrade for that design maybe.
[01:20:28] <Mr_Mayhem> Does chain have backlash?
[01:20:47] <uw> yes i think so
[01:21:04] <nlancaster> no not ballscrew, and I have seen a chain design running, done right they are ok. but pick the rock sprocket and the can "jump" at each tooth
[01:21:45] <Mr_Mayhem> Hmm, seems like there are too few options on that stepper to linear aspect.
[01:21:47] <nlancaster> wrong sprokect
[01:22:17] <nlancaster> what we really need is a linear stepper :D with high enough steps per inch/mm
[01:22:29] <Mr_Mayhem> Ahmen.
[01:23:01] <Mr_Mayhem> Make a coil winder for linear drive. Then bolt together sections to make any size.
[01:23:18] <Mr_Mayhem> need encoder too of course.
[01:23:45] <Mr_Mayhem> Would be very neat.
[01:24:33] <Mr_Mayhem> They exist now, and have good speed and force, but expensive isn't the word. Crazy expensive. Hmmm.
[01:24:35] <nlancaster> got discoed did i miss anything?
[01:24:47] <nlancaster> yes, no more rotary to linear conversion issues: D
[01:24:48] <nlancaster> :D
[01:25:15] <Mr_Mayhem> Would be nice to see an open design pop up.
[01:26:32] <Mr_Mayhem> Maybe put the coil in the track, and the magnets in the moving part, to save on magnet costs. I am trying to learn what I can about the various designs to see. Iron, vs open core, and some others.
[01:27:12] <Mr_Mayhem> It's basicly a servo unfolded flat.
[01:27:57] <nlancaster> yep
[01:28:23] <nlancaster> ideal would probably be eletrocmagets in each face of the linear motor
[01:28:28] <nlancaster> i think that is how they do it now
[01:29:53] <nlancaster> but anyway, back to xylotex good choice, bad choice?
[01:34:46] <Mr_Mayhem> Probably fine for sheet plastics and wood, but not aluminum.
[01:35:06] <nlancaster> the xylotex or the machine. because the machine I don't think would do it for aluminum either :D
[01:35:42] <Mr_Mayhem> It comes down to how fast you can cut a given job without exceeding the limits.
[01:35:47] <nlancaster> sure
[01:36:02] <nlancaster> light duty, low DOC at relativly low IPM
[01:36:09] <Mr_Mayhem> yeah
[01:36:18] <nlancaster> only going to put a 1 - 1.25 HP router on it
[01:36:56] <Mr_Mayhem> How much is the xylotex package?
[01:36:57] <nlancaster> just came in to see if anyone knew of any other low cost 4axis controlls that don't have to be "modded" or fixed
[01:36:59] <nlancaster> 300
[01:37:07] <nlancaster> 4 axis, 4 motors, PSU
[01:37:24] <nlancaster> http://xylotex.netfirms.com/OSCommerce/catalog/product_info.php?cPath=26&products_id=40
[01:37:32] <Mr_Mayhem> maybe the gecko with drives?
[01:37:38] <Mr_Mayhem> g540?
[01:37:41] <nlancaster> alot more
[01:37:46] <nlancaster> almost that much for just the g540
[01:37:58] <Mr_Mayhem> yeah, kinda pricey.
[01:38:02] <nlancaster> http://www.automationtechnologiesinc.com/products-page/featured-cnc-products/g540-4-axis-nema23-381ozin-48v7-3a-psu
[01:38:14] <nlancaster> i mean its a great system. but this router isn't for me, and its supposed to be low cost
[01:38:28] <nlancaster> my entire bom is only about $700 including the control and steppers
[01:39:28] <nlancaster> I will be happy if the machine is reliable at 40IPM and .125 DOC with 1/4inch bit
[01:39:34] <nlancaster> that would be phenominal for the price
[01:40:38] <Mr_Mayhem> Looks like the one xylotex comes with is decent, what's wrong with that one?
[01:41:11] <Mr_Mayhem> Like it's made for it.
[01:41:14] <nlancaster> nothing, just seeing if anyone knew of a cheaper one :D
[01:41:37] <Mr_Mayhem> I think you are already scraping the bottom at 2.5 amp per winding.
[01:41:41] <nlancaster> did you see the machine I was talking about building, using for insperation?
[01:42:16] <nlancaster> my josecnc2006 runs at 200ipm and cuts fine at 120ipm on only 3amp 36v
[01:42:23] <nlancaster> keling kl4030 drivers
[01:42:55] <Mr_Mayhem> yeah, the openbuild one.
[01:43:40] <Mr_Mayhem> Again, I think the design is one of the cleanest I have seen for a low end router.
[01:43:47] <nlancaster> i know, it looks great
[01:44:15] <nlancaster> i think they are just going ot use the grbl shield
[01:44:23] <Mr_Mayhem> The 4030 boards are good, but you need 4?
[01:44:37] <nlancaster> well i gues stechinically I could chain off one drive to 2 motors.
[01:45:32] <Mr_Mayhem> Oh, because of the double y axis steppers.
[01:45:35] <nlancaster> yes
[01:46:14] <nlancaster> alghouth, i could probably just get some drillshaft and a few bearings and make a single stepper y-axis drive
[01:46:14] <Mr_Mayhem> Here is the 4030 from Keiling, is a leadshine rebranded. Good baseline, 3amp.
[01:46:16] <Mr_Mayhem> http://www.automationtechnologiesinc.com/products-page/kl-stepper-drivers/kl-4030-24-40vdc-3-0a-microstepping-driver
[01:46:19] <nlancaster> but would cost almost as much
[01:46:27] <nlancaster> yes i have the kl4030 on my big joes cnc
[01:46:50] <nlancaster> but for these inexpensive machines, I don't want ot have to wireup the control box to that extent, and the keling 4axis kit is over 100 more
[01:48:25] <nlancaster> http://www.youtube.com/watch?v=W2GxLMvkjQw
[01:48:30] <nlancaster> thats the drive system, very ingeinous
[01:48:53] <Mr_Mayhem> Given your goals, I'd be leaning towards the xylotex, looks like a good fit. Don't know about others like that but the layout looks clean on the pcb.
[01:49:13] <nlancaster> and I have heard people driving machines as big as a joes cnc2006 or even a joes evolution with it.
[01:49:37] <Mr_Mayhem> hehe
[01:49:54] <nlancaster> so if someone bought my machine, they could use it to build a bigger better machine in the future :D
[01:50:32] <Mr_Mayhem> Yeah, true. Nice to start up a little overbuilt, rather than have to redo everything later.
[01:51:22] <nlancaster> yep, thats why I am not thinking of nema 17s even though it would still work at that low duty level
[01:52:32] <nlancaster> http://youtu.be/3gpeDRt4V6I
[01:52:38] <nlancaster> there is the speeds it is capable of
[01:55:02] <Mr_Mayhem> Watching the video.
[02:01:41] <Mr_Mayhem> The moving part is so light, I bet it is fast and zippy.
[02:02:06] <Mr_Mayhem> Didn't see a motion example hooked up.
[02:03:13] <Mr_Mayhem> I guess you have the router spindle, but still, very minimal bracket and carrage weight. Would be great for a laser maybe.
[02:06:41] <Deejay> moin
[02:09:37] <Mr_Mayhem> Oh, ok there is the movement example. Yeah, pretty fast.
[02:13:12] <nlancaster> wouldnt want to put alot of load by doing large DOC but otherwise looks great for the price
[02:14:43] <Mr_Mayhem> I wonder what the physical limit is before the belt skips teeth. And maybe do a two belt design or something for added force.
[02:15:39] <nlancaster> i know my nema17 inch 3d printer will stall if I grab a axis while moving, instead of skipping teeth
[02:17:00] <s1dev> what's a breakdown of things I need to convert a machine to CNC? steppers, stepper mounts, electronics right?
[02:17:21] <nlancaster> steppers, mounts, couplers to attach the stepper shaft to the leadscrew, electronics
[02:17:48] <s1dev> what size stepper would I be looking for and how much are electronics
[02:17:58] <Mr_Mayhem> software-pc-breakout board-driver boards for each axis-wires-limit switches, for starters.
[02:18:23] <s1dev> limits switches are a few cents
[02:18:43] <Mr_Mayhem> Depends on the type.
[02:19:03] <s1dev> would I want anything fancier than a standard microswitch?
[02:19:10] <Mr_Mayhem> I use radio shack cheapies too. But eyeballing the inductive ones hehe.
[02:19:44] <Mr_Mayhem> In a word, repeat-ability. OK two words.
[02:20:02] <s1dev> hall effect limit switches
[02:21:29] <Mr_Mayhem> The steppers you need depend on the size of the machine and what it will be cutting.
[02:21:57] <nlancaster> what machine are you going to convert?
[02:23:49] <Mr_Mayhem> Also, the nicer limit switches or sensors have de-bounce built in or don't bounce in the first place. Bounce is two or more signal changes in short space of time when it clicks. Makes it less precise sometimes.
[02:23:57] <Jymmm> Does anyone recall the NAME of those crimp on pins (for bare wire) typically used for screw terminals?
[02:25:04] <s1dev> nlancaster a logan 920
[02:25:10] <Jymmm> Ah.. Ferrules
[02:25:39] <nlancaster> keeping the stock acme screws?
[02:25:42] <s1dev> yeah
[02:25:45] <Mr_Mayhem> Best to also solder them after crimping.
[02:25:48] <s1dev> I'd like to anyway
[02:26:14] <nlancaster> probably need minimum nema23, maybe nema34 motors then
[02:26:20] <Jymmm> Mr_Mayhem: That is completely WRONG. Yo NEVER crimp AND solder, it's one or the other.
[02:26:28] <Jymmm> You*
[02:26:28] <nlancaster> why?
[02:26:45] <Jymmm> If you solder, it makes the wire brittle
[02:27:06] <Jymmm> Then compressing the wire with a crimp makes a thing/weak spot.
[02:27:15] <Jymmm> thin/
[02:27:22] <Mr_Mayhem> Yeah, I could see that. You put the solder near the tip, not the bending end. But I see your argument.
[02:27:24] <nlancaster> crimp then solder?
[02:27:42] <Jymmm> nlancaster: No, one OR the other, but NEVER both.
[02:27:54] <nlancaster> hmm, my EE mate likes both. lol
[02:28:01] <nlancaster> i always thought it was over kill
[02:28:10] <Jymmm> nlancaster: If you crimp first, then you have no place for the solder to flow.
[02:28:32] <Jymmm> solder wicks into the strands of the wire.
[02:28:39] <nlancaster> he did as Mr_Mayhem said, soldered from the crimp end
[02:28:44] <nlancaster> yes i know, just saying
[02:28:51] <s1dev> how much would stepper controllers + boards, etc be?
[02:29:05] <nlancaster> we were just having that convo earlier :D
[02:29:17] <nlancaster> there is a kit that many people have had good luck with from xylotex
[02:29:18] <Jymmm> Well, ppl think doing both is better, but don't know better is all.
[02:29:22] <Mr_Mayhem> Maybe on some connectors. My cnc came with crimped and soldered molex connectors for the steppers. But what do the chinese know? A lot of their stuff is lame wiring design anyway.
[02:29:40] <nlancaster> http://xylotex.netfirms.com/OSCommerce/catalog/product_info.php?cPath=26&products_id=39
[02:30:12] <Jymmm> Just get a pair of good crimpers and you'll break the wire before it ever comes out ofthe connector.
[02:30:21] <nlancaster> that has been my experience
[02:30:36] <Jymmm> T&B are EXCELLENT and lifetime warranted
[02:30:42] <Mr_Mayhem> Yeah, having the correct tool is very important. Not to wing it with pliers.
[02:31:44] <Mr_Mayhem> I used to install cable tv for a living, and I can appreciate the right stripper and crimper vs the "lets see what I can use" method.
[02:31:58] <Jymmm> These are the BEST DAMN crimpers ever, Not even the snap on/kline and other clones come close....
http://www.amazon.com/Thomas-Betts-WT111M-Insulated-Terminals/dp/B0018LD2PU/ref=sr_1_4?ie=UTF8&qid=1387700094&sr=8-4&keywords=t%26b+crimpers
[02:32:25] <Jymmm> Well, one does, but it's been discontinued form the mfr.
[02:33:39] <Jymmm> These have been discontinued, but I hear good things about them, though have never tried them
http://www.amazon.com/Channellock-909-Crimping-Tool-Cutter/dp/B00004SBDI/ref=pd_sim_hi_3
[02:34:36] <Mr_Mayhem> Nice straightfoward design. I will save that for future buy.
[02:35:06] <Jymmm> And I do have a bunch of fancy ratching crimping tools too, but I will always grab the T&B ones first.
[02:35:41] <Jymmm> The ratcheting ones are mostly for funky connectors.
[02:36:27] <Mr_Mayhem> I suppose you do a bit of wiring of things and you have had the time to compare.
[02:36:47] <Jymmm> LOL, slight understatement there =)
[02:37:02] <Jymmm> My T&B are 20+ years old
[02:37:39] <Jymmm> I've had the same pair THAT long and they still are great
[02:38:17] <Mr_Mayhem> hehe. Now you have me wondering about the sea of wiring jobs you have done. What is your typical application?
[02:38:21] <Jymmm> They are worht the $50 just in preventing headaches alone
[02:38:54] <Mr_Mayhem> Like wiring up cnc machines?
[02:38:59] <Jymmm> Having a crimp come lose, then having to go back and trace it out... fuck that.
[02:39:12] <Mr_Mayhem> Yeah, tracing can suck.
[02:39:46] <Jymmm> No, I'm good at it, but it's it's an intermittant issue, it can be a helll of a bitch to diagnose and isolate the issue.
[02:40:10] <Mr_Mayhem> Right, the intermittent problem is the hard animal to catch.
[02:40:21] <Jymmm> And I'll be damned if I'm going to let a fucking $0.12 connector be the source of the problem.
[02:40:33] <Mr_Mayhem> Ha! I feel you.
[02:41:07] <Jymmm> I've wired everythign from automaotive, electronics, industrial electrical
[02:42:07] <Jymmm> You gotta realize that when a tool hasn't changed it's deign in 20+ years and is still sellling, that says something about it.
[02:42:38] <Jymmm> Hell, even snap-on and others have cloned the T&B ones.
[02:42:52] <Jymmm> But not that close =)
[02:43:57] <Mr_Mayhem> Well done wiring is an art form, esp when there are hundreds or thousands of wires in a system. People show off their work in pictures and all that.
[02:44:52] <Jymmm> And they are all shit too! I saw one harness wiring that used only wax string.
[02:44:58] <Jymmm> PERFECT bends
[02:45:25] <Jymmm> You could VISISUALLY trace each wire in the harness all thru the machine.
[02:45:44] <Jymmm> There were to cable clams either, the entire harness was self-supporting
[02:45:50] <Jymmm> ...in mid air.
[02:46:00] <Jymmm> It was a beautiful thing to see.
[02:46:05] <Mr_Mayhem> Yeah, that's old school wax string method, very nice.
[02:46:26] <Mr_Mayhem> Wire ninja.
[02:47:07] <Mr_Mayhem> The ones with tie wraps make me wonder if they overstress the wires.
[02:47:24] <Mr_Mayhem> Long term.
[02:47:54] <Jymmm> Not really, in the person does it properly.
[02:48:52] <Mr_Mayhem> I have seen those "wire combs" used for network cable harnesses. Are they any good?
[02:49:13] <Jymmm> combs?
[02:49:37] <Jymmm> You mean panduit?
http://sigma.octopart.com/5387266/image/Panduit-G1X1LG6.jpeg
[02:49:47] <Mr_Mayhem> https://www.youtube.com/watch?v=-cFToZt_KzI
[02:53:32] <Jymmm> LOL, fuck that thing... More of a PITA and I can see that it actually kinked a couple fo the cables in the bunch.
[02:53:59] <Jymmm> Just do a clean job.
[02:54:21] <Mr_Mayhem> hehe
[02:56:24] <Mr_Mayhem> Yeah. Seemed kinda interesting, but if you already have methods, maybe what it does is redundant, or gets in the way.
[02:58:49] <s1dev> nlancaster so for the xylotex, what else needs to be purchased besides limit switches? serial port card?
[02:59:08] <Jymmm> Dont buy/use xylotex.
[02:59:33] <Mr_Mayhem> What is the issue with it?
[02:59:38] <Jymmm> serial? you mean parallel ?
[03:00:00] <Jymmm> Has known thermal issues.
[03:00:14] <Jymmm> underrated for most applications.
[03:00:24] <Jymmm> I have one, wish I never bught it.
[03:00:30] <Mr_Mayhem> Maybe you could put heatsink bar across the chips with thermal grease between?
[03:00:39] <Jymmm> I was trying to be cheap and it bit me in the ass
[03:00:48] <Mr_Mayhem> Oh.
[03:00:53] <Jymmm> HA, good luck with that
[03:01:11] <Jymmm> I have SIX fans on mine.
[03:01:38] <Jymmm> And, it's not at it's MAX rating either.
[03:01:59] <Jymmm> It also doens't have any mid-band stability
[03:02:08] <Jymmm> aka, stall prevention.
[03:02:31] <Mr_Mayhem> What would you choose if you could go back in time?
[03:02:38] <Jymmm> G540
[03:02:56] <Jymmm> http://www.geckodrive.com/geckodrive-step-motor-drives/g540.html
[03:03:41] <Jymmm> It includes a BOB, higher ratings, mid-band compensation,
[03:04:14] <Mr_Mayhem> It's over the budget of s1dev.
[03:04:42] <Jymmm> So will having to buy a new xylotex when it burns out
[03:05:03] <Jymmm> Plus the xylotex doesn't have a BOB, you still need to buy that.
[03:05:14] <Mr_Mayhem> Maybe back to a set of
http://www.automationtechnologiesinc.com/products-page/kl-stepper-drivers/kl-4030-24-40vdc-3-0a-microstepping-driver
[03:05:46] <Jymmm> The G540 also has estop and heartbeat so your computer does fuck anything up during power on.
[03:06:29] <Jymmm> Mr_Mayhem: no mid-band compensation.
[03:06:34] <Jymmm> that I can see
[03:07:28] <s1dev> so, for the gecko
[03:07:32] <Jymmm> Also, that 3A I suspect INCLUDES BEMF too
[03:07:57] <Jymmm> so you'll have to lower it a bit to combensate
[03:08:41] <jesseg> One old stepper driver I saw used like 48v to drive the 5v steppers, but it used giant power resistors to limit the current. Thus, during higher speed operation, the voltage was available but during low speed or stop, the motors didn't get overheated.
[03:08:43] <jesseg> Night folks
[03:09:02] <Mr_Mayhem> I had the el-cheapo toshiba version of that animal, and it was ok, but I overheated it by not connecting my fan properly one day. So I am looking myself. But the dsp ones are way over the budget our friend is trying to follow.
[03:09:04] <Mr_Mayhem> night.
[03:09:27] <Mr_Mayhem> So s1dev, what are your thoughts so far?
[03:10:42] <s1dev> the gecko G540 I just get a parallel port board for my controller computer and 3-4 steppers and I'm set?
[03:11:10] <Jymmm> and a PS
[03:11:46] <Jymmm> s1dev: I'd HIGHLY RECOMMEND you read this beofr buying anything
http://www.geckodrive.com/step-motor-basics
[03:11:48] <Mr_Mayhem> Yeah, it plugs right in to the parallel port cable and power supply. Plus the name Gecko carries weight for people buying from you- it says quality to most people.
[03:12:31] <s1dev> Jymmm I'm familiar with steppers
[03:12:35] <Connor> http://www.automationtechnologiesinc.com/products-page/digital-stepper-motor-driver/3-axis-dsp-based-digital-stepper-drive-max-60-vdc-6-0a
[03:12:41] <Mr_Mayhem> That G540 is hugely popular in the hobby cnc and maker crowd.
[03:13:10] <Jymmm> s1dev: Not well enough that you had to ask what you needed. That guide is VERY WELL WRITTEN
[03:13:44] <Mr_Mayhem> I was gonna mention the 3 axis model from leadshine, but your machine needs 4. I would not put 2 stepper motors on one drive.
[03:14:37] <Mr_Mayhem> Maybe 3 + one extra drive, but that would be kinda weird. Just keep the 2 y axis on the 3-model.
[03:15:06] <Mr_Mayhem> So they have the same behavior.
[03:15:17] <Connor> I like Kelling Stepper Drivers..
[03:15:52] <Connor> and Automation Tech has that new 3 axis similar to a G540, but more voltage and amperage capability.
[03:16:07] <Connor> I tend to use single driver units, easier to replace if one goes bad.
[03:16:12] <Mr_Mayhem> I hear nothing but good things, and wiring individual drives is not as big a deal as you are making it out to be.
[03:16:40] <Mr_Mayhem> Keiling and Leadshine are the same, and both good.
[03:18:07] <Mr_Mayhem> You can even buy a card like this:
http://www.cnc4pc.com/Store/osc/product_info.php?cPath=33&products_id=270
[03:18:41] <Jymmm> You dont need that BOB if you have the G540
[03:18:46] <Connor> Nahh.. that just adds more stuff to break.
[03:18:55] <Jymmm> Plus, it's not optially isolated iirc
[03:19:02] <Connor> Those stupid Rj45 break out boards cause issues.
[03:19:07] <Jymmm> optically*
[03:19:41] <Mr_Mayhem> Yeah, they are not as fast as they could be sometimes. And the leds are one more thing to get crummy.
[03:19:41] <Connor> What was the budget ?
[03:19:51] <Jymmm> $0
[03:20:18] <Connor> Well then, crappy a$$ 4 Axis board from ebay with a few mods. :)
[03:20:49] <Jymmm> Eh, pay now in cash, or later in blood and cash, their choice.
[03:20:55] <Mr_Mayhem> He was aiming quite low.
[03:21:25] <Connor> Frankly MESA 5i25 with break out board, with Keiling drivers..
[03:22:05] <Mr_Mayhem> I am in that camp. Already have the Mesa 5i25 and 7i76 for $210 + shipping from CA.
[03:22:36] <Mr_Mayhem> Very nice also to have speed control for the spindle built in, and a shitload of inputs and outputs.
[03:22:58] <Mr_Mayhem> I gotta get the drives next.
[03:23:04] <Connor> Yup. Frankly.. the 5i25 is my next upgrade.. I'm using a cheap a$$ BOB with mine via parport..
[03:23:31] <Jymmm> Those Keiling drives don't have mid-band compensation
[03:24:05] <Connor> with a C41 PWM Board for the spindle and a chage pump board.
[03:24:06] <Mr_Mayhem> Do you need dsp type of drive to get that feature, or do the lower class basic designs offer it?
[03:24:29] <Connor> Mid-Band compensation ?
[03:25:15] <Mr_Mayhem> For solving stepper resonance issues.
[03:25:20] <Jymmm> Ah, they offer it on their "digital drives"
[03:25:20] <Mr_Mayhem> Yes
[03:25:57] <Jymmm> Connor:
http://www.geckodrive.com/step-motor-basics
[03:26:15] <Mr_Mayhem> That's what I thought. Digital drives have an fpga chip. New models offer a fancy bancy way to adjust all aspects via serial port on pc.
[03:27:13] <Mr_Mayhem> The FPGA runs the fast math for DSP functions to smooth the steps, anti-resonance, etc.
[03:27:52] <Connor> Yea, If I could afford it, I would upgrade to digital drivers to..
[03:28:45] <Connor> I'm not sure where the mid-band is with my steppers.. or if I have issues with it..
[03:28:52] <Mr_Mayhem> I saw a video of the difference on steppers side by side on a table, and the dsp is silent at low speeds, no clicking sound. Pretty impresssive difference at first glance.
[03:29:32] <Connor> I think I've tuned it so that I'm probably just below it on my top speed.
[03:30:10] <Jymmm> DAMN, they re expensive now...
http://www.ebay.com/itm/Parker-stepping-motor-driver-OEM750X-/281186690075?pt=LH_DefaultDomain_0&hash=item417808641b
[03:30:34] <Jymmm> But 7 for $445, ins't too bad at all
[03:30:53] <Jymmm> and MAKE OFFER too, but from china =(
[03:31:25] <Jymmm> Oh, and you doent need 750X, just 750
[03:31:31] <Jymmm> don't*
[03:32:07] <Connor> Jymmm: Is that for a single one? Or all 7. Says they have 4 available, 1 sold..
[03:33:00] <Jymmm> Connor: Auction doesn't say, but it should be for all 7 but who knows.
[03:33:28] <Jymmm> Connor: everyone else is selling them for $200 each
[03:33:29] <Jymmm> http://www.ebay.com/sch/i.html?_nkw=parker+oem+750&_frs=1
[03:33:53] <Connor> so, maybe it's just 1 each..
[03:34:03] <Connor> Time for bed.
[03:38:56] <Mr_Mayhem> Never saw that before. What machine is it typically used with
[03:39:00] <Mr_Mayhem> ?
[03:41:07] <Mr_Mayhem> Seems like more of a professional motion control item.
[04:30:11] <Loetmichel> Mr_Mayhem: i bend the sheets by hand, sometimes with the help of a vice
[04:30:49] <Loetmichel> but usually you can bend a 1,5mm sheet aluminium sored to 075mm by hand easily
[04:51:45] <Mr_Mayhem> Thanks, I was wondering. Again, very nice sheetmetal work. I will try it soon.
[05:02:52] <archivist> smooth jaw vice is very useful for small bending work
[08:16:07] <dosas> hello i installed the rtai kernel as described here
http://wiki.linuxcnc.org/cgi-bin/wiki.pl?LinuxCNC_On_Ubuntu_Precise
[08:16:27] <dosas> now i would like to compile lcnc from source which git branch should i use?
[08:17:56] <dosas> i cannot find branch-rt in the repository
[08:30:12] <Mr_Mayhem> Give it a few, someone who knows will answer. A bit early in the morning.
[08:32:09] <archivist> dosas, I think what you want may be on the buldbot
http://buildbot.linuxcnc.org/
[08:33:18] <archivist> but that seems sim only and I have seen talk of it being tested
[08:34:20] <archivist> http://buildbot.linuxcnc.org/buildbot/waterfall
[08:46:32] <Tom_itx> colorful
[09:00:00] <dosas> okay but there mus be a git repo for these builds ?
[09:00:52] <Jymmm> http://git.linuxcnc.org/
[09:01:07] <dosas> yes but which branch
[09:01:27] <Jymmm> Um, read? Search?
[09:01:57] <dosas> git branch -r | grep rt
[09:01:58] <dosas> origin/dynload-rtapi
[09:01:58] <dosas> origin/jepler/rtapi-more-headers
[09:01:58] <dosas> origin/jepler/rtos-master-v0-linux_rt
[09:01:58] <dosas> origin/nort_testing
[09:01:58] <dosas> origin/rtapi-bitops-gcc-intrinsics-rc
[09:02:00] <dosas> origin/rtos-integration-preview3-merged-into-master
[09:02:02] <dosas> origin/rtos-master-v0
[09:02:04] <dosas> origin/rtos-testmerge-into-master
[09:02:10] <Jymmm> DONT FLOOD
[09:25:10] <eric_unterhausen> there were two branches that were packaged and mentioned on the wiki page, I would use one of them
[09:34:30] <dosas> you probably mean master-rt or v2.5_branch_rt but these names do not occur
[09:34:58] <eric_unterhausen> leave off the rt
[09:35:07] <dosas> kk thanks
[09:35:40] <eric_unterhausen> I don't think there are any non-rt branches, but there are non-rt packages
[09:35:58] <dosas> :D
[09:53:31] <dosas> anyone from the rtai-modules-3.4.55-rtai-2 developer here?
[09:54:01] <dosas> or people involved with this page
http://wiki.linuxcnc.org/cgi-bin/wiki.pl?LinuxCNC_On_Ubuntu_Precise
[10:02:04] <eric_unterhausen> I installed it, why?
[10:02:40] <dosas> i installd it too from the repos but my realtime seems to fail to load
[10:02:56] <dosas> NOTE: 'rtapi' kernel module must be loaded
[10:03:16] <eric_unterhausen> are you sure you are running the realtime kernel?
[10:03:16] <dosas> can you do a lsmod
[10:03:30] <dosas> uname -r says3.4.55-rtai-2
[10:03:33] <eric_unterhausen> you can do an lsmod or modprobe
[10:03:41] <dosas> but i'm using debian not ubuntu
[10:03:53] <eric_unterhausen> don't think it should matter
[10:04:06] <dosas> probably some modules not laoded
[10:04:44] <dosas> http://pastebin.com/cLjis4hF
[10:04:54] <dosas> you know the modules names
[10:04:55] <eric_unterhausen> realtime components are run later
[10:06:22] <eric_unterhausen> if you run a latency test, that would tell you if it loads rt modules
[10:06:40] <dosas> i did that and got this error
[10:07:09] <dosas> http://pastebin.com/1p1FAPv2
[10:07:20] <dosas> then i checked
http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?LockedMemory
[10:07:27] <dosas> but the line is in
[10:08:57] <eric_unterhausen> did you try that solution?
[10:09:04] <dosas> as is said
[10:09:08] <dosas> the line was already in there
[10:09:37] <dosas> maybe i try a reboot
[10:10:00] <dosas> wait
[10:10:02] <dosas> strange
[10:13:36] <dosas> okay i needed to restart in order to get the mem limit right
[10:13:44] <dosas> latency teste works now
[10:23:37] <Guest65040> just trying out IRC
[10:24:15] <archivist> you need to pick a nick someone does not already own :)
[10:24:46] <eric_unterhausen> it turns out I have the most common name ever
[10:25:08] <Guest65040> nick scotty
[10:25:19] <Guest65040> NICK scotty
[10:25:22] <Guest65040> haha
[10:25:22] <eric_unterhausen> ok, so not as common as that
[10:25:39] <Guest65040> is my syntax correcy
[10:25:41] <Guest65040> correct...
[10:25:44] <eric_unterhausen> no
[10:26:07] <archivist> /nick name that is not in use
[10:26:09] <Guest65040> nick <scotty_cnc>
[10:26:14] <Guest65040> ahhh
[10:27:09] <eric_keller_is_t> oh, nicks can't be long enough to say eric_keller_is_the_most_common_name_ever
[10:27:51] <eric_keller_is_t> having a common name provides a certain amount of anonymity
[10:28:20] <scotty_cnc_123> i was going to ask some question on the forum
[10:28:31] <scotty_cnc_123> my name on there is "scotti"
[10:28:37] <scotty_cnc_123> figured id try this out
[10:28:50] <archivist> irc can be faster
[10:28:53] <eric_keller_is_t> scotti is taken on freenode?
[10:29:04] <eric_keller_is_t> depends on who is on
[10:29:30] <archivist> nicks can be protected, they dont need to be on
[10:29:48] <eric_unterhausen> ya, if I got a nick I liked, I would protect
[10:30:07] <eric_unterhausen> but every variation on my name is currently in use and nobody uses unterhausen
[10:36:28] <eric_unterhausen> I should go ride my bike, heat death of the universe scheduled for tomorrow
[10:39:10] <ChuangTzu> i have a stepper that no longer has any holding torque, but steps fine
[10:39:14] <ChuangTzu> i wonder if the driver is fried
[10:39:33] <scotty_cnc_123> did you try it on another axis
[10:39:53] <eric_unterhausen> sometimes holding torque is a setting
[10:40:05] <scotty_cnc_123> anyone there can tell whether i need or should care about having full closed loop all the way back to linuxcnc
[10:40:20] <eric_unterhausen> many people use open loop
[10:40:33] <scotty_cnc_123> i was going to play around with some ac servos
[10:40:35] <ChuangTzu> the only settings this driver has are enable and not enable
[10:40:41] <ChuangTzu> and enable is default
[10:40:53] <ChuangTzu> i'd try another axis but i only have one of these motors
[10:40:58] <scotty_cnc_123> as i understand it the ac servo driver does all the heavy lifting...
[10:41:01] <eric_unterhausen> doesn't sound good
[10:41:04] <ChuangTzu> got another driver on order though
[10:41:10] <scotty_cnc_123> nothing gets back to linxcnc
[10:41:29] <eric_unterhausen> scotty, what kind of ac drives, most have encoder out
[10:41:38] <eric_unterhausen> or is it a step/dir drive?
[10:42:57] <scotty_cnc_123> its one of those dmm tech kits... i think they have some sort of a magnetic encoder or something
[10:43:07] <Jymmm> I like #3
http://jdgroover.wordpress.com/2013/08/20/the-2013-darwin-awards-are-out/
[10:43:12] <scotty_cnc_123> the drives themselves can be configured for step/dir
[10:44:05] <scotty_cnc_123> if the drive has encoder out back to linuxcnc it would only be for informational purposes though... correct?
[10:46:46] <eric_unterhausen> I don't know about those drives, checking
[10:48:00] <eric_unterhausen> do you have a link?
[10:48:05] <eric_unterhausen> http://dmm-tech.com/Dyn3.html
[10:48:17] <eric_unterhausen> has either step/dir or full feedback
[10:50:52] <scotty_cnc_123> sorry screaming baby
[10:51:07] <scotty_cnc_123> http://www.dmm-tech.com/Dyn2.html
[10:51:32] <eric_unterhausen> have to stop pinching the baby
[10:52:14] <eric_unterhausen> Serial interface [UART,SPI], Pulse/Direction or CW/CCW, and Analog communication modes
[10:52:40] <eric_unterhausen> so you could use a, for example, mesa card with analog control
[10:52:51] <eric_unterhausen> or use step/direction open loop control
[10:54:49] <scotty_cnc_123> so with analog control there is no step/dir? its +-10 volts to the drive
[10:55:15] <scotty_cnc_123> i guess what is no clear to me is what is the drive actually outputting back to the mesa card
[10:55:16] <eric_unterhausen> yes, then you need an encoder output on the drive for feedback
[10:55:30] <scotty_cnc_123> there is a position command port
[10:55:42] <eric_unterhausen> it might be bi-directional
[10:55:46] <scotty_cnc_123> JP3 which has Pin 5: On position signal (low active) for CW/CCW mode, serial data output for SPI mode (Optional) .
[10:56:44] <scotty_cnc_123> and JP4: Sensor port Pin 4: Serial data to sensor. Pin 5: Serial data from sensor.
[10:57:05] <scotty_cnc_123> ill have to ask the dmm guys about this...
[10:57:40] <scotty_cnc_123> but my main question was .. what is the advantage to having the encoder output come all the way back to linuxcnc
[10:58:26] <eric_unterhausen> never mind, dyn 2 does not have a usable analog control mode. Unless you can tap into the motor encoder lines directly
[10:59:21] <eric_unterhausen> running open loop requires some faith that your drive will detect a following error properly
[11:00:08] <eric_unterhausen> and tuning and other things where you would like to have a computer in the control loop
[11:01:36] <scotty_cnc_123> my crude machine has small servos and gecko 320s
[11:01:49] <scotty_cnc_123> i never did any tuing at all... got lucky and it just worked.
[11:01:56] <eric_unterhausen> sometimes that happens
[11:02:22] <eric_unterhausen> but if you got tired of slow performance, you could run into problems
[11:02:24] <scotty_cnc_123> i do see people using the scope/meter which seems to show encoder feedback... which would probably be useful
[11:02:55] <scotty_cnc_123> if i was trying to tune the system using linuxcnc and not the vendor provided windows app
[11:03:26] <scotty_cnc_123> the other option is a used yaskawa ac servo setup from overseas
[11:03:35] <scotty_cnc_123> less money
[11:03:50] <scotty_cnc_123> but no support... i guess the dmm guys are pretty good with support
[11:04:15] <scotty_cnc_123> thanks for the info
[11:04:38] <scotty_cnc_123> i think im going to end up try out one of the mesa 5i25 cards
[11:04:51] <Tom_itx> can't go wrong there
[11:36:46] <s1dev> gecko g540 or mesa card?
[11:38:49] <archivist> s1dev, you are not comparing like with like
[11:39:15] <archivist> are you asking stepper or servo or what
[11:39:24] <s1dev> stepper
[11:40:08] <s1dev> and g540 is mostly self contained, but the mesa cards require stepper controllers right? both require a power supply
[11:43:08] <eric_unterhausen> does g540 have stepper drivers?
[11:44:05] <eric_unterhausen> it looks like it's a bob + 4 drivers?
[11:44:19] <Tom_itx> it is a stepper driver
[11:44:22] <Tom_itx> 3 axis
[11:44:49] <s1dev> it says it contains 4 stepper drives + BOB
[11:47:57] <eric_unterhausen> the comparison is to mesa + 4 drives or g540
[11:48:16] <eric_unterhausen> I'm sure the mesa would be better, but by how much I don't know
[11:48:28] <eric_unterhausen> depends on the drives you use with it
[11:48:42] <eric_unterhausen> mesa could be better depending on drives
[11:56:53] <pcw_home> you can drive a G540 directly from a 5I25/6I25
[12:07:22] <s1dev> what is the advantage of a FPGA vs microcontroller
[12:09:13] <archivist> microcontroller is a red flag in here, microcontroller for what use
[12:09:53] <IchGuckLive> hi all B)
[12:10:10] <s1dev> controlling the drives, its really G540 vs Mesa and why?
[12:10:46] <IchGuckLive> s1dev: better G540 vs Leadshine
[12:10:48] <archivist> you can have G540 AND mesa card
[12:10:53] <pcw_home> G540 is a quad step motor driver
[12:10:54] <pcw_home> Mesa makes interfaces
[12:11:27] <IchGuckLive> s1dev: mesa makes the pulses for the drives
[12:11:38] <IchGuckLive> better then the parport
[12:11:42] <s1dev> and what mesa card is suggested?
[12:11:45] <pcw_home> LinuxCNC --> interface ---> MotorDrive
[12:11:51] <archivist> mesa card allows higher pulse rates
[12:11:52] <Tom_itx> s1dev, for one thing an FPGA can run process in parallel
[12:12:01] <Tom_itx> where a microcontroller usually can't
[12:12:25] <IchGuckLive> s1 are you in the USA
[12:12:29] <s1dev> yes
[12:12:31] <pcw_home> interface can be parallel port, GPIO bits on Arm processor, FPGA etc
[12:12:45] <IchGuckLive> so im in germany and therfor out of the helpers phalanc
[12:13:11] <pcw_home> Motor Drive can be Gecko G540, Leadshine m542, analog servo etc etc
[12:14:09] <Tom_itx> pcw_home i was wandering around the directories, what's the USBIOPR8 driver for?
[12:14:31] <Tom_itx> and USBSDMC
[12:15:10] <s1dev> would a 5i25 have enough ports?
[12:15:17] <pcw_home> USBIOPR8 is a simple USB 6X8 bit GPIO interface
[12:15:30] <Tom_itx> not used by lcnc is it?
[12:15:34] <pcw_home> USBSDMC is USB interfaced SoftDMC motion controller
[12:15:39] <pcw_home> No
[12:16:21] <pcw_home> (No not used by linuxcnc)
[12:16:30] <Tom_itx> gotcha
[12:17:45] <pcw_home> The 5I25 will directly connect to a G540 for example and then the
[12:17:47] <pcw_home> other port can be used for extra I/O, depending on how much I/O is needed
[12:18:27] <pcw_home> Or if you have a lot of I/O and individual drives like the leadshine you could use a 7I76
[12:19:04] <pcw_home> (which adds spindle control and 48 24V I/O points)
[12:19:42] <pcw_home> Or you can just use a parallel port if its fast enough for you
[12:19:58] <pcw_home> (Or 2 parallel ports if you need more I/O)
[12:28:29] <grimpirate> Would it be alright to ask a question about a cnc build concept as opposed to questions relating directly to linuxcnc software?
[12:29:17] <skunkworks> grimpirate: go for it
[12:29:22] <CaptHindsight> grimpirate: it happens all the time
[12:29:54] <grimpirate> Thanks, just wanted to make sure I wasn't violating a room rule or something, my question relates to an H Belt setup
[12:30:05] <grimpirate> Specifically I understand how to generate motion along the x-axis and y-axis
[12:30:45] <grimpirate> What I can't discern is how the motors would work together to generate angular lines? i.e. not perpendicular or parallel to x/y axes
[12:33:04] <IchGuckLive> H Belt setup is used in most 3D printers
[12:33:39] <IchGuckLive> http://www.triomotion.com/tmt3/imgs/prog_examples/kin4.gif
[12:34:09] <IchGuckLive> this is only good for less then 2,5m
[12:34:15] <CaptHindsight> grimpirate: ever play with an etch-a-sketch? turn both knobs at once
[12:34:26] <IchGuckLive> 100inch
[12:35:00] <grimpirate> Right, what I'm saying is if I turn both motors cw or ccw I get motion in one axes
[12:35:09] <grimpirate> If I turn them in opposing directions I obtain motion in the other axes
[12:35:19] <grimpirate> This concept is understood
[12:35:24] <IchGuckLive> grimpirate:
http://www.youtube.com/watch?v=V6kllGspehQ herei is my mashine drawing for this
[12:35:48] <grimpirate> So let's say I want to make a 45 degree line
[12:36:10] <grimpirate> Would the rpm of each motor be what affects the angular movement?
[12:36:13] <IchGuckLive> both motors will run together
[12:36:33] <IchGuckLive> controled on given parameters
[12:36:44] <IchGuckLive> the interpreter will generate the pulses
[12:37:00] <IchGuckLive> on 45deg it might be the same on both
[12:37:27] <IchGuckLive> on less or more one motor will turn less or more rmp
[12:37:36] <IchGuckLive> ;-) rpm
[12:38:26] <grimpirate> ok so when the RPMs perfectly match each other, that's when you obtain parallel or perpendicular motion to the major axes
[12:38:40] <grimpirate> And when RPMs differ you achieve an angular movement
[12:39:18] <IchGuckLive> on a pulsed mashine StepDIR it will be the pullse space not the Rpnm
[12:40:11] <grimpirate> Right but pulse space/interval basically dictates the RPM of the motor?
[12:50:33] <pcw_home> yes and pulse rate is determined by:
[12:50:34] <pcw_home> Gcode program --> motion /Kinematics --> joint pos --> stepgenerator --> pulse rate and direction
[12:51:28] <grimpirate> Understood
[12:52:04] <grimpirate> I suppose the difficulty was just in visualizing how the different motor speeds would affect motion as it's one belt and should only exhibit one linear speed throughout
[12:52:53] <pcw_home> So you need to chose the kinematics that match your mechanics
[12:53:35] <grimpirate> Well I planned on writing the software myself
[12:54:12] <grimpirate> My issue was just understanding how motor rpm affected x/y movement
[12:56:53] <grimpirate> thank you all for your help, definitely much less confusing now
[12:58:49] <pcw_home> http://www.youtube.com/watch?v=_e7yy9u6QmA
[12:59:07] <Mr_Mayhem> Hey pcm_home, what part of the stepgen aspect is offloaded to fpga chip, like what gets replaced with what, signal wise? Do moves get turned into move instructions or something?
[13:00:57] <Mr_Mayhem> Maybe that's a dumb way of phrasing it. I suppose high frequency stuff gets replaced with low frequency signals.
[13:01:15] <Mr_Mayhem> or instructions to make the steps.
[13:03:20] <grimpirate> Yeah I've seen that video pcw_home, sadly, it only demonstrates parallel/perpendicular motion
[13:05:34] <grimpirate> The only one I've seen that shows an angular motion is here
http://www.macrondynamics.com/videos/hbot-gantry
[13:06:05] <grimpirate> Where the operator stops the motion of one of the "motors" and spins the other, demonstrating that the gantry moves angularly
[13:07:57] <pcw_home> You may need to built the hardware and twist the shafts by hand to really get a feel for whats happening
[13:09:22] <pcw_home> Mr_Mayhem: for the step generator, the hardware is just a
[13:09:23] <pcw_home> rate /pulse generator (a DCO) with position feedback to linuxcnc
[13:09:30] <grimpirate> Yeah that is the end goal, I just wanted to be sure before investing the funds
[13:10:04] <grimpirate> So I can sketch it on CAD first and see where I can reduce costs and build complexity
[13:10:39] <archivist> what is the machine for
[13:11:14] <grimpirate> Laser cutter
[13:11:18] <grimpirate> But specifically for paper
[13:11:27] <grimpirate> So a small 200 mW laser
[13:11:51] <pcw_home> (They seem to be not terribly accurate but can be very high speed)
[13:12:28] <pcw_home> you are not going to be very high speed with 200 mW...
[13:13:32] <grimpirate> Not at all, but speed isn't a top notch concern
[13:14:15] <grimpirate> Reliable positioning, from what I'm given to understand is achieved by having a rigid enough frame/carriage and proper belt tension
[13:15:18] <IchGuckLive> at that site of a router better to go balbelt for pricise
[13:15:32] <IchGuckLive> or even with 2 belts to get full speed
[13:15:48] <IchGuckLive> ballbearing is cheep 1205
[13:15:58] <IchGuckLive> and retches 20m/min
[13:16:16] <IchGuckLive> im off BYE have a good sunday
[13:20:15] <Mr_Mayhem> It seems to me many folks could use a relatively cheap linear motor system that can be bolted together in sections to get different sizes. No such animal apparently. Even a pretty weak one would be nice for that guy's application, instead of a belt. But either you need a pile of expensive magnets, or a pile of driver semiconductors, to drive each coil independently.
[13:21:01] <Mr_Mayhem> So it looks like no way to make it cheap, really. But they shouldn't be thousands or tens of thousands.
[13:21:15] <tangentaudio> i have used some if this stuff for a lightweight application:
http://openbuildspartstore.com/openrail/
[13:22:05] <tangentaudio> not really suitable for a machine with any large cutting forces
[13:22:22] <Mr_Mayhem> That was a useful idea for light stuff, yeah.
[13:23:02] <tangentaudio> i used their openrail and guide wheels bolted onto a piece of 80/20, and made my own belt drive arrangement to work with it
[13:25:54] <Mr_Mayhem> The laser and knife folks are lucky in that respect, minumal side loads. But they have other challenges, like fumes and fire hazards, hehe.
[13:27:07] <Mr_Mayhem> Hey tangent, by the way, I am working on a desktop electrostatic loudspeaker set. I will keep you updated when I have it together. I have done a few prototypes so far, and this should be a neat improvement. Wire stators with resistor eq network, big toroids. Acrylic and aluminum.
[13:28:03] <Mr_Mayhem> Saw your site, very nice.
[13:28:46] <tangentaudio> sounds cool
[13:30:08] <tangentaudio> homemade diaphragm or did you source it from somewhere?
[13:30:11] <Mr_Mayhem> The wires go to a pcb front and back to distribute the signals to the individual wire pairs.
[13:30:39] <Connor> Aluminum? Why not hardened steel..
[13:31:16] <Mr_Mayhem> I just got the thin mylar and stretch it in the usual way with tape. Next I will try tabs with pulleys and weights, and a mic to see the resonant freq to match the pair.
[13:31:49] <tangentaudio> how low do you expect the frequency range to extend to?
[13:31:59] <grimpirate> Well the cheapest one I've seen so far is at corexy.com
[13:32:22] <Mr_Mayhem> I use wires for the electrodes, so I can alter their frequency response to fade off the highs towards the edges. Kills the "beaming" effect and widens the pattern.
[13:32:55] <Mr_Mayhem> Also flattens the response.
[13:33:39] <tangentaudio> sounds neat... i've never really done much speaker building, save for a pair of subwoofers about 15 years ago... certainly nothing esoteric like electrostatics
[13:34:17] <Mr_Mayhem> Well this is a small desktop design, so maybe 300 hz or maybe 250? I use a sub to fill in the lows.
[13:34:36] <tangentaudio> i think i've only ever even *heard* a pair of electrostatics once... martin logans. i've heard ribbon or planar magnetic speakers much more
[13:34:57] <Mr_Mayhem> The mids and highs are magical in the last one, but too bright without that resistor network to soften the highs.
[13:36:57] <Mr_Mayhem> I saw the electrostatic is one speaker design someone can diy and get respectable if not amazing audio from it, so I went in that direction.
[13:37:44] <Mr_Mayhem> Anyway, not to hijack any topic here, but look foward to having something to show when I fix my cnc's driver cards.
[13:38:23] <Mr_Mayhem> Just gonna get a new set.
[14:30:18] <scotty_cnc_123> m
[14:34:25] <scotty_cnc_123> hello, is there anybody out there
[14:34:48] <asah> I am...
[14:34:50] <scotty_cnc_123> just nod if you can here me
[14:34:53] <asah> skunkworks around?
[14:34:57] <scotty_cnc_123> is there anyone home
[14:35:07] <asah> I am working on gearbox code today.
[14:35:24] <scotty_cnc_123> anyone out there done a nut driven screw
[14:40:12] <skunkworks> asah: I have not done any comps in a couple of years...
[14:41:32] <scotty_cnc_123> comps?
[14:42:34] <skunkworks> realtime compoants written in a C like languate
[14:42:37] <asah> ok, I am good to do the comp. I lost the reference you gave me.
[14:42:43] <skunkworks> oh
[14:42:47] <asah> sorry .=(
[14:43:14] <asah> my main problem right now is just making sure the io is mapped right, and then figuring out what the logic is.
[14:43:29] <skunkworks> http://electronicsam.com/images/KandT/conversion/testing/config/
[14:43:46] <asah> there are 5 switches per shifting gear, and they all don't show up on my io.
[14:43:51] <asah> thanks.
[14:44:15] <asah> do you have a maho sam?
[14:44:43] <skunkworks> no - it is a Kearney and trecker
[14:45:12] <scotty_cnc_123> i see
[14:45:17] <asah> right, I thought you might also have a maho.
[14:45:20] <pcw_home> scotty_cnc_123: Andy pugh has driven ball-nuts in his mill conversion (for space reasons)
[14:45:30] <scotty_cnc_123> thanks
[14:45:31] <skunkworks> what do you mean they don't show up in your i/o?
[14:45:39] <scotty_cnc_123> i got a couple of BIG screws
[14:45:51] <scotty_cnc_123> 50mm diameter
[14:45:58] <scotty_cnc_123> 60+ inches long
[14:46:02] <asah> most of the switches for detecting state of the gear changing show up in my io table.
[14:46:14] <asah> one of them (at least) doesn't.
[14:46:25] <asah> it must be used internally. might be for mid gear finding.
[14:46:55] <asah> http://www.practicalmachinist.com/vb/dmg-mori-gildemeister-maho-cnc/anyone-here-got-cnc-maho-gear-drive-spindle-159398/ is what I am looking at
[14:47:14] <asah> those three gear selectors
[14:47:59] <asah> I made myself a little control panel and I can swtich the gears forwards and back etc. now I just need to make a comp that can detect the different states and stop the changing.
[14:48:19] <asah> I found the mapping table between colored dots and speeds, so I should be good.
[14:48:25] <asah> a nice slog.
[14:48:37] <asah> but the good news is that seems to be the last major hurdle on my conversion.
[14:48:47] <asah> (before I change the servos out =) )
[14:48:59] <asah> I got spindle to spin yesterday!
[14:49:04] <skunkworks> yay
[14:49:08] <skunkworks> why change the servos?
[14:49:25] <asah> I have to move the thing to a single phase shop in a month or so.
[14:49:42] <asah> plus my y axis brake is shot.
[14:49:58] <asah> plus I like making more work for myself.
[14:49:59] <asah> =)
[15:01:00] <skunkworks> heh - what kind of servos are currently on it?
[15:02:08] <skunkworks> The K&T was converted from hydaulic servos...
[15:54:03] <mrsun_> gah, how to color in the text of my carvings? :/
[15:56:24] <awallin> paint before v-carving, then carve, then face-mill off the paint everywhere else but the carved bits as the final operation
[15:56:40] <awallin> or something like that :)
[15:58:08] <mrsun_> but now its all done already ... i just want the text to show better :P
[15:58:47] <JT-Shop> paint, sand
[15:59:45] <mrsun_> paint will be sucked by the fibers in the wood wont it? :/
[16:03:20] <JT-Shop> wood sealer
[16:05:12] <asah> anyone dealt with a contactor "chirping" in a start delta starter?
[16:06:03] <asah> my motor (380 ac induction spindle) has issues getting into delta it seems when it is "cold" perhaps.
[16:06:15] <asah> it might be the contactor is dirty? anyone else dealt with this?
[16:06:45] <asah> the problem will likely go away when I change it over to a vfd, but I thought I might ask.
[16:50:07] <asah> anyone want to talk tuning PID loops?
[16:50:37] <asah> I am tuning a position -> velocity loop.
[16:51:09] <asah> I am using peters suggestions to keep the main component FF1 and use P gain to tune.
[16:51:16] <asah> still taking a long time to settle.
[16:51:35] <JT-Shop> have you seen my tutorial on tuning a velocity drive?
[16:51:53] <asah> probably not.
[16:52:00] <asah> where can I find it?
[16:52:13] <JT-Shop> http://gnipsel.com/linuxcnc/tuning/index.html
[16:52:38] <asah> perfect, thanks!
[16:53:46] <asah> what is the connection here though? it looks like you are using velocity-cmd as the input to the loop?
[16:54:11] <mrsun_> JT-Shop, hmm wood sealer .. what would that be ? polyurethane stuff ? shellac ?
[16:54:31] <mrsun_> i sprayed it all with shellac now .. looks quite nice so will see if that will do or if i will do more to it tomorrow =)
[16:55:59] <asah> my servos are velocity mode.
[16:57:01] <JT-Shop> I'm not sure what it is made of but it says wood sealer on the can
[16:58:15] <asah> why is it so hard to find example hal files?
[16:59:14] <asah> I cannot find one single example of a hal file setup of a velocity mode amp with pid setup in it.
[17:02:48] <JT-Shop> because they are all different, that's the beauty of HAL!
[17:06:18] <asah> yeah. well. I am trying to setup the simplest PID example and I cannot set it up correctly.
[17:06:49] <asah> axis.1.joint-vel-cmd does not seem to be doing the correct trapezoidal thing in a jog move of .1 in
[17:09:27] <cradek> asah: you need to use position as the input to pid, just like ALL these servo config sample hal files
[17:09:49] <asah> I don't see ANY config sample hal files.
[17:10:12] <asah> and the tuning tutorial JT just sent me has velocity as the command.
[17:10:24] <asah> but I can't tell cause I can't see the setup.
[17:10:45] <asah> cradek, where are ALL those servo config sample files. I am an idiot and cannot find them.
[17:10:46] <cradek> hm2-servo is a good sample
[17:11:26] <cradek> net emcmot.00.pos-cmd axis.0.motor-pos-cmd => pid.0.command
[17:11:28] <asah> ok. looking in the source.
[17:11:31] <cradek> net motor.00.pos-fb hm2_[HOSTMOT2](BOARD).0.encoder.00.position => pid.0.feedback
[17:11:43] <asah> I have grepped myself a hm2-servo.hal
[17:11:58] <cradek> not sure what you're working from. if you pick hm2-servo on the config picker, it'll offer to copy it to your home directory for you to customize.
[17:12:23] <asah> thanks cradek... Ill dig through this.
[17:14:50] <JT-Shop> asah, do you know where the config picker is?
[17:17:42] <asah> pccnf?
[17:18:06] <JT-Shop> no
[17:18:23] <JT-Shop> the main Ubuntu menu CNC > LinuxCNC
[17:18:25] <asah> stepconf wizard?
[17:18:52] <JT-Shop> Applications > CNC > LinuxCNC
[17:19:12] <JT-Shop> open up Sample Configurations
[17:20:33] <asah> I don't see that option.
[17:20:52] <asah> ah. ok. sorry.
[17:21:17] <asah> Im lame. I have been using the linuxcnc commandline for so long to launch my .ini file, I didn't see those samples.
[17:21:19] <asah> sorry!
[17:21:24] <JT-Shop> no problem, we all start at the beginning and know nothing
[17:21:47] <asah> =)
[17:33:44] <Tom_itx> wb JT-Shop
[17:34:06] <Tom_itx> snow finally hit here
[17:35:14] <Tom_itx> JT-Shop, i ran your facing routine last night
[17:35:28] <JT-Shop> Tom_itx, I can tell that no one has used the hole ops yet, it lacks the F word
[17:35:51] <JT-Shop> we got a light dusting over by Springfield when we left there this morning
[17:35:59] <JT-Shop> how did the facing come out?
[17:36:09] <Tom_itx> i had to tweak the code a bit
[17:36:30] <Tom_itx> i dunno if i entered something wrong or what but the offsets were off a bit
[17:36:48] <Tom_itx> and i wanted it to exit past the material
[17:36:56] <JT-Shop> the start position?
[17:37:00] <Tom_itx> yes
[17:37:48] <Tom_itx> i used zig zag and changed the exit path to stop when the tool reached the centerline of the cutter
[17:41:19] <JT-Shop> I see I still have some debugging output on zig zag
[17:41:58] <Tom_itx> yeah i saw that too
[17:42:33] <JT-Shop> I just made a test one and it seems right to me
[17:42:58] <Tom_itx> the engraving is certainly better;
http://tom-itx.dyndns.org:81/~webpage/cnc/pendant8.jpg
[17:43:00] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/cnc/pendant/pendant2.jpg
[17:43:06] <Tom_itx> compared to the old one
[17:43:39] <Tom_itx> i should have gone a tad deeper on the bottom part
[17:44:12] <JT-Shop> did you paint the letters then sand the surface?
[17:44:23] <Tom_itx> yes
[17:44:36] <JT-Shop> looks good
[17:45:00] <Tom_itx> left a spot for another button but i couldn't think of anything to add there yet
[17:45:19] <Tom_itx> if it were a big mill i'd have a bigger estop
[17:45:35] <JT-Shop> did you specify the start position in the G code generator?
[17:45:45] <JT-Shop> lol
[17:45:56] <Tom_itx> i don't remember, i'd have to fire it up and reenter the numbers
[17:46:04] <Tom_itx> i don't think i did
[17:46:57] <Tom_itx> just did some rewiring, fixin to make another bit file to try
[17:47:10] <JT-Shop> ok
[17:47:48] <Tom_itx> on the facing, i left the x y reference at 0 and left the start blank
[17:48:06] <Tom_itx> i forgot how the two were related
[17:48:30] <JT-Shop> it will start with the tool in the correct position in Y and off the material in X
[17:48:53] <Tom_itx> it acted like the xy 0 was the left front
[17:48:56] <Tom_itx> instead of rear
[17:49:03] <Tom_itx> i just offset it a bit more and ran it
[17:49:09] <JT-Shop> unless you specify somewhere else to start
[17:49:29] <JT-Shop> that's odd
[17:50:07] <Tom_itx> i don't think i have any axis wired backwards, my cad cam works fine
[17:51:17] <Tom_itx> i guestimated the size of the cutter too since it's a flycutter
[17:51:27] <Tom_itx> should have been fairly close though
[17:51:34] <JT-Shop> http://imagebin.org/283412
[17:52:01] <JT-Shop> that is a screen shot I just ran and it looks correct to me based on the numbers I entered
[18:01:10] * JT-Shop heads inside
[18:01:28] <Tom_itx> i'll mess with it some more
[18:01:42] <Tom_itx> once i get these bit files straightened out
[18:01:59] <JT-Shop> ok, thanks for testing it
[18:02:11] <Tom_itx> it did save some time
[18:02:16] <JT-Shop> I see another problem with the metric tap pitch conversion
[18:02:21] <Tom_itx> i had to mill that plate down to thickness a bit
[18:02:40] <Tom_itx> that little mill just takes a long time to get anything done
[18:04:04] <JT-Shop> at least you can get it done with the little mill even if it takes longer
[18:04:47] <Tom_itx> nothing like what i was used to
[18:04:55] <Tom_itx> but it's still fun
[18:05:58] <Tom_itx> lemme know when you post a new rev
[18:06:25] <Tom_itx> this is 1.5.7
[18:09:02] <jthornton> ok
[20:43:14] <Tom_itx> pcw_home, i ran a couple .vhd files to check the timing as you suggested. first i took a stock SVST2_4_7I47.vhd and it did ok, then i did the custom one with the sserial and it failed to meet the timing constraints
[20:43:27] <Tom_itx> any ideas to get it to meet them?
[20:44:09] <Tom_itx> possibly remove one of the other unused functions?
[20:45:13] <Tom_itx> and assign the pins as Nulltag?
[20:45:57] <Tom_itx> i'm not using the PWMTag
[20:49:17] <scotty_cnc_123> non linux cnc related...
[20:49:38] <scotty_cnc_123> am running a 8020 bolt together router...
[20:49:58] <scotty_cnc_123> is it even possible to build a fast/rigid frame out of 8020
[20:53:02] <toastyde1th> what are you calling "rigid"
[20:53:20] <Tom_itx> build it on a granite slab
[21:01:36] <scotty_cnc_123> rigid enough that i cant see the dam n thing flexing and moving about when i do quick reversals
[21:02:13] <scotty_cnc_123> my issue is that its going into a basement... otherwise id weld up a frame with square steel tubing
[21:13:33] <pcw_home> Tom_itx: setting optimization to speed, removing unused modules, having the complier try harder etc
[21:14:11] <Tom_itx> it is set for speed
[21:14:18] <Tom_itx> i removed the pwm and trying again
[21:14:35] <Tom_itx> what do you mean have the compiler try harder?
[21:16:16] <pcw_home> its has a lot of options for optimization some include trying harder at getting timing closure
[21:17:06] <Tom_itx> if this doesn't get it, i'll take a look and see if i can find those settings
[21:17:24] <Tom_itx> i suspect this was the problem the first time
[21:17:31] <Tom_itx> with the unusual behavior
[21:18:15] <pcw_home> also some timing errors are meaningless, can you check that the 7i43.ucf file has TIG statements for
[21:18:16] <pcw_home> clockmed <--> clocklow cross clock domain checking
[21:22:52] <atom1> well i didn't get the timing error this time through
[21:23:10] <atom1> generating pin file...
[21:33:44] <atom1> TIMESPEC "TSLowToMed" = FROM "async_low" to "async_med" TIG;
[21:33:54] <atom1> TIMESPEC "TSMedtoLow" = FROM "async_med" to "async_low" TIG;
[21:34:15] <atom1> that?
[22:07:41] <Tom_itx> ok things seem a little better so far
[22:07:55] <Tom_itx> sserial works, spindle encoder seems ok so far
[22:08:32] <Tom_itx> moved the sserial off the spare encoder pins so they're separate from anything else
[22:08:38] <Tom_itx> not that it mattered
[22:09:20] <Tom_itx> removing the unused pwm seemed to help
[22:10:14] <Tom_itx> pcw_home, i'm sure each module varies on it's timing constraints but have you tested to see how many can be active in a bit file?
[22:11:11] <Tom_itx> it wouldn't make practical sense to have certain ones active with others but i could see a need for pwm for some things
[22:11:34] <Tom_itx> i'll run it like this for a while and see how it does
[23:53:37] <tjb1> 🚣
