# #emc | Logs for 2007-07-27

Back
[00:00:15] <ds2> you are refering to the area near PCC?
[00:00:30] <cradek> R0BH: ask either way - sometimes someone else can jump in and help
[00:00:30] <Jymmm> yes
[00:00:37] <R0BH> iv set g53, its in center spindle now, and g28 54mm to get to home on edge of x travel where i like it to be, but if i g28 home, and turn off, when turn back on it says machine is in middle of x when it is not
[00:01:15] <R0BH> so in effect its 54mm out of postion each time
[00:01:32] <Jymmm> I also miss hitting in-person auctions.... that is really a lot of fun.
[00:01:32] <cradek> yes you have to home every time
[00:01:43] <R0BH> on 53
[00:01:45] <cradek> can you add a home switch at least on X?
[00:02:29] <R0BH> there is a limit switch on x as it is, which is triggered on ether over travel
[00:02:35] <ds2> eh? onlything on there that is really surplus is C&H and they are worse then Halted here
[00:02:38] <cradek> home on that sucker
[00:03:21] <ds2> hmm they claim to be open sat...
[00:03:53] <R0BH> ok, ill have to wire him up then
[00:14:32] <ds2> for a small desktop machine with a 1hp spindle, is 300W a good ballpark power budget for the steppers, drivers, and control?
[00:14:43] <ds2> or would that be overkill?
[00:20:28] <SWPadnos> does "control" include the PC?
[00:20:53] <ds2> yes, PC + a small (<9inch) LCD + any LEDs, buttons, etc
[00:22:33] <ds2> nope, it will probally be something like an 800MHz VIA C3 unless RT/EMC fails on it...
[00:22:45] <SWPadnos> ok, that'll be a low power unit then
[00:24:12] <ds2> didn't think the choice of MB's would be that significant for a ball park estimate...esp. since anything more modern then a PIII is not even considered
[00:44:47] <R0BH> think i have found a bug in css
[00:47:40] <R0BH> when contoring in css it seems to rapid on the G02 and G03 moves willing to send program to someone to test out
[00:48:47] <cradek> is it CSS or FPR that causes it? are you using both?
[00:49:06] <R0BH> let me show u the program
[00:49:43] <R0BH> http://innovative-rc.co.uk/EMC/TEST1-5CUT-R.ngc
[00:49:47] <cradek> my lathe is not set up right now so the more precisely you can describe the problem you see, the better
[00:51:49] <R0BH> where it would retract on the arc on ruthing it will rapid out along it not fpr
[00:52:08] <jmkasunich> so you are using FPR
[00:52:10] <R0BH> same on finish finish, on g02 and g03 lines it will rapid around the moves and go back into fpr on g01
[00:52:18] <jmkasunich> (you said "bug in css")
[00:53:02] <R0BH> sorry getting late here :s
[00:53:19] <cradek> oh! I see it here too
[00:54:03] <cradek> I think I see the bug - testing a fix
[00:55:18] <R0BH> no problem
[00:55:25] <SWPadnos> come on - this is taking too long!
[00:55:36] <R0BH> lol
[00:56:16] <R0BH> nearly 2am here
[00:56:33] <CIA-8> 03cradek 07TRUNK * 10emc2/src/emc/kinematics/tp.c: fix FPR arcs
[00:56:35] <R0BH> can u do us some autocorner radius and chamfering please ;)
[00:57:13] <R0BH> many thanks will update and retest
[01:00:20] <cradek> R0BH: you should get your cam to put a G21 at the beginning of your program - when I loaded it here it was a BIG part!
[01:03:08] <cradek> (adding G21 was the only change I had to make to run it on my inch configuration)
[01:03:58] <R0BH> aah yea, here we are allways in metric and all our full size CNCs are pre G21 anyway
[01:04:09] <R0BH> it does put it in but we normlay take it out to save on NC code space
[01:04:21] <cradek> ahh, not an issue with emc!
[01:04:24] <R0BH> as 17k on CNC does not go too fair these days
[01:04:49] <R0BH> yes we have a leadwell CNC we would like to EMC onto when we have time
[01:05:09] <cradek> wow 17k doesn't seem like much - how old is that control?
[01:06:12] <R0BH> one we want to convert has a mitsubishi control, where bigger one we have has fanuc
[01:06:55] <R0BH> it has resolvers not encoders so we have got to buy converters or new encoders for it also to convert it
[01:07:15] <R0BH> control is about 1980 something i belive
[01:07:58] <R0BH> cvs updated and is fully working
[01:08:07] <cradek> cool thanks for testing so fast
[01:08:57] <R0BH> so didt u want to make a chess set for your garden with out G21 ;)
[01:09:27] <R0BH> but thank you for fixing the problem so fast to
[01:09:29] <cradek> I think it would be about as tall as me!
[01:09:42] <cradek> you're welcome, I knew right were to find this one.
[01:09:51] <R0BH> ye about 30inch high lol
[01:09:55] <Jymmm> 1/3 human height chess pieces are huge.
[01:10:24] <Jymmm> It looks funny when they're moving them too
[01:10:37] <R0BH> ye even hubals mirrors where wrong when they could not get conversion right metric to imperial
[01:11:23] <R0BH> do you guys have any plans for diamiter programming soon?
[01:11:41] <Jymmm> Plan, what plan?
[01:12:06] <cradek> in diameter mode, arcs aren't round anymore - it's a big can of worms
[01:12:25] <ds2> when is G71 ready for testing? ;)
[01:12:45] <Jymmm> ds2: It's in the plan
[01:12:59] <Jymmm> ds2: didn't you get the memo?
[01:13:13] <R0BH> best one we have used here is one written by philips for there control
[01:13:40] <ds2> the one written in hyroglphics carved on the stone tablet? no... it cracked in transit
[01:14:01] <cradek> hmm, time for dinner II - hungry again
[01:14:02] <Jymmm> ds2: well, glue the pieces back together
[01:14:19] <R0BH> good night
[01:14:23] <R0BH> speak again soon
[01:14:42] <ds2> Jymmmm: I sent the messenger back to get a new one
[01:17:01] <R0BH> going to land of nod now, cya all soon
[01:17:14] <R0BH> does g83 work on lathe
[04:07:21] <GuySoft> hello all, i want to build a controllable camera mount (2 axis) for this i need a 2 stepper motor controller, did any of you get something like this working under linux?: http://tinyurl.com/3b8qmw .
[04:08:28] <GuySoft> anyone alive today?]
[04:08:30] <SWPadnos> I don't know of anyone doing that
[04:08:37] <SWPadnos> err - using that
[04:09:38] <GuySoft> dont you work with CNC devices? its the same principle.. i need 2 stepper motors controllable in linux..
[04:09:44] <GuySoft> who are the geckodrive?
[04:10:01] <SWPadnos> nevermind - someone recently asked a similar question on a mailing list
[04:10:30] <SWPadnos> yes, I work with CNC devices, but I don't know of anyone using that device with Linux
[04:12:16] <SWPadnos> they don't appear to publish the command set for their controller, so it's unlikely to be usable under Linux
[04:13:31] <GuySoft> SWPadnos, id be glad if you recommend a simular device.. that was more or less my budget..
[04:14:08] <GuySoft> LPT... thats what i have on this laptop
[04:14:24] <SWPadnos> err - so USB yea but LPT no?
[04:15:33] <GuySoft> SWPadnos, i would like this to be as flexible as possible.. however if LPT is much simpler, ill go for it.. the best i am familiar with LPT, is turning LEDs on and off it
[04:15:58] <SWPadnos> driving small steppers basically only requires beefier transistors
[04:16:16] <SWPadnos> but it needs to be a real parport, not something connected via USB
[04:16:33] <Jymmm> SWPadnos: I was gonna ask you a question since yesterday, but I forgot! LOL
[04:16:59] <SWPadnos> the other option, if you're comfortable writing some microcontroller code, would be to get an inexpensive microcontroller development kit and write the stepper code yourself. it's not too hard
[04:17:28] <GuySoft> SWPadnos, well, i want it to interface with INDI : http://indi.sourceforge.net/index.php/Main_Page
[04:17:27] <SWPadnos> Jymmm, good. I'm in the middle of some work right now, but decided to help out a newcomer :)
[04:18:07] <GuySoft> SWPadnos, can you direct me to such a kit, or something.. i am just new and looking for where to start
[04:18:53] <ds2> SWPadnos: did you comment any more about my power q during the netsplit?
[04:19:04] <SWPadnos> ds2, I don't think so
[04:19:06] <GuySoft> also, if anyone can give me hands-on experience advice.. this is what i am looking for.. i am new to this and trying to figure out what to use for what..
[04:19:21] <ds2> okie
[04:19:29] <GuySoft> SWPadnos, well i took the motors of a 5.4 harddrive.. thats a start..
[04:19:44] <ds2> i'll go with the 1KW figure and verify later
[04:19:53] <SWPadnos> hmmm. how many wires do they have?
[04:20:02] <GuySoft> i found something that does more than what i want here: http://fred.unis.no/StampII3Mx/Default.htm
[04:20:10] <GuySoft> 5 pins
[04:20:41] <GuySoft> here is a diagram of it: http://de.wikipedia.org/wiki/Bild:Schrittmotor.PNG .. i was lucky to find this.. only exists on the german wiki! (and i dont know german)
[04:22:01] <GuySoft> so there you have it, the plan is to connect these two on one side.. and kstars on the other :)
[04:22:01] <SWPadnos> ok - that's just a standard unipolar 4-phase stepper diagram
[04:23:02] <GuySoft> SWPadnos, if its simple to you, then you are wiser than me.. i figured out how they worked with zaping the pins around with 3 watss battary
[04:23:23] <GuySoft> still, i need to figure out what controller i am using on the electric side
[04:23:34] <SWPadnos> well, I'm an electrical engineer, and I've been involved with EMC for several years (and motor control starting about 20 years ago :) )
[04:23:48] <SWPadnos> hmmm. maybe only 18 years ago
[04:23:53] <GuySoft> i started last winter..
[04:24:25] <GuySoft> and it wasnt really motors for most of the time.. i never even soldered before then.. now i have a desk for it here..
[04:24:31] <SWPadnos> anyway, that step driver you saw - they're just using the ULN5803 driver chip to get high current output, and they use a microcontroller to control that
[04:24:40] <SWPadnos> well, you're doing fine so far, it seems
[04:24:58] <SWPadnos> sorry - ULN2803 - typo
[04:25:08] <GuySoft> 2003 it says there
[04:25:20] <ds2> or a UN2003 + 8bit latch on a USB printer cable?
[04:25:31] <SWPadnos> you can make a PC board with a 2803 that plugs into the parallel port, and drive it directly with EMC
[04:26:03] <SWPadnos> or at least HAL - you probably don't need all the G-code wizardry - just the step generator and a means of feeding it positions
[04:26:13] <GuySoft> carnt i even use for that ua741s? i have 2 of them here...
[04:26:42] <ds2> 741 as in the opamp?
[04:26:55] <SWPadnos> so no, you can't use those
[04:27:04] <GuySoft> SWPadnos, i lost you now.. whats an 8bit latch?
[04:27:22] <SWPadnos> ds2 was mentioning another possibility for that board, or for you to use
[04:27:31] <SWPadnos> I'd say, don't bother with that :)
[04:27:35] <GuySoft> SWPadnos, ok.. about the 741 it was just an idea because i mucked about with them last week
[04:27:45] <SWPadnos> USB takes the "real" out of "real-time"
[04:27:58] <ds2> but it isn't like the camera will "crash"
[04:28:12] <SWPadnos> but you'll only get one step every 2 ms
[04:28:19] <GuySoft> ah, the 8-bit latch question was directed at ds2 ..
[04:28:25] <ds2> only reason I even mention it is to have a quick and dirt way of driving a camera
[04:28:33] <SWPadnos> and the torque on those motors means there'll probably need to be gearing
[04:28:55] <SWPadnos> sure, but a parport should be easier, no?
[04:29:07] <ds2> if you can find the paraport on a modern computer :(
[04:29:09] <SWPadnos> 8 data lines - connect one to each input of the driver ...
[04:29:13] <GuySoft> SWPadnos, well i am using this for digital astro-photography, 2ms is more than what i need.
[04:29:15] <SWPadnos> I'm pretty sure GuySoft said he has one
[04:29:26] <ds2> GuySoft: something like a 74LS373
[04:29:40] <GuySoft> SWPadnos, yes i have one.. r52 IBM laptops come with an LPT port :)
[04:29:54] <SWPadnos> 2 ms per step, and you'll want very very precise angular positioning, I suspect
[04:29:57] <ds2> Oh if you have a real parallel port, go with that
[04:30:16] <GuySoft> i also rather use USB through because i am scared to fry it.. dont have money for a new one
[04:30:56] <GuySoft> USB have in their standard things to cope with overloads.. lpt dont..
[04:31:20] <SWPadnos> LPT has no 5V supply
[04:31:32] <ds2> not really
[04:31:33] <SWPadnos> which can be a pain in many cases, but may be a benefit here
[04:32:04] <ds2> you send 10KV down the USB line, it'll release smoke just as well as anything else
[04:32:31] <ds2> main thing USB has is current limit on the power lines if the followedthe spec
[04:32:42] <GuySoft> ds2, i can also send a hammer to it.. still, the safer the better...
[04:33:09] <ds2> GuySoft: eh? driving a motor can generate 10KV back EMF during the switching!
[04:33:30] <SWPadnos> ideally the 2803 would catch that
[04:33:31] <GuySoft> anyway. if it will work with LPT, then LPT it is.. now how do i find a controller for that.. and also drivers would be handy.. coding it from scratch would be really long and painful
[04:33:49] <GuySoft> EMF/
[04:33:51] <SWPadnos> there's a mode in the EMC2 step generator that will generate 4-phase waveforms
[04:33:50] <GuySoft> ?
[04:33:51] <ds2> "ideally" heheh
[04:34:09] <SWPadnos> EMF=ElectroMotive Force = voltage
[04:34:09] <jmkasunich> GuySoft: have you seen the Mel Bartels software for driving Dobsonian telescopes with steppers?
[04:34:30] <GuySoft> jmkasunich, i dont think so.. do you have a url?
[04:35:08] <jmkasunich> http://www.bbastrodesigns.com/BBAstroDesigns.html
[04:35:11] <jmkasunich> heh
[04:35:25] <jmkasunich> yep, I didn't know the URL either, but google found it
[04:35:28] <GuySoft> oh, and mine is a reflector.. however i might use this first on a camera mount.. lets energy needed
[04:35:33] <GuySoft> googling now..
[04:35:53] <GuySoft> ah, nm just saw you posted it, thanks :)
[04:36:35] <jmkasunich> because scopes need low speeds, a simple 4 transistor unipolar drive works fine
[04:36:45] <jmkasunich> two axes = 8 bits = easy to do on a printer port
[04:37:02] <jmkasunich> he uses PWM to microstep the motors for smoothness
[04:39:13] <GuySoft> PWM?
[04:39:27] <jmkasunich> pulse width modulation
[04:39:35] <GuySoft> also, i think i saw this before.. its just i am more in to the basics now of this
[04:39:58] <SWPadnos> gotta run, good luck.
[04:40:37] <GuySoft> thanks, SWPadnos i think i might be back here.. get other peoples ideas..
[04:41:28] <GuySoft> jmkasunich, i think splitting that acronym didn't do me any good at :-S ..
[04:41:49] <GuySoft> jmkasunich, i am trying to find stuff that would be more easy to make thaugh..
[04:42:23] <jmkasunich> PWM means instead of just turning a bit on continuously, or off continuously, you turn it rapidly on and off, to get an average value somewhere between 0 and 1
[04:42:43] <jmkasunich> GuySoft: I don't think you can get much easier than the BBastro stuff
[04:43:46] <jmkasunich> you have a choice between totally free (schematic and bill of materials are published, build it yourself), totally turnkey (assembled board ready to use) for 145, or multiple steps in between [04:43:59] <jmkasunich> the software is free in both price and freedom (GPL'ed) [04:44:07] <GuySoft> jmkasunich, maybe, just keep in mind that the part lists a hard for me to understand, also the circuit diagrams.. usualy i need someone that knows electronics to explain lines to me [04:44:43] <GuySoft> jmkasunich, i am looking at it.. i am just not sure how the hardware even works yet. let alone if i can build it [04:44:55] <GuySoft> jmkasunich, i am looking at it as we are talking :) [04:45:10] <jmkasunich> if you can't build that stuff (which is in kit form) you won't be able to design and build something from scratch [04:45:36] <jmkasunich> that software has been evolving for about 10 years [04:46:14] <jmkasunich> I used to be interested in telescopes, before I got into machine tools instead... Bartels has a very good reputation as one of the leaders in computer controlled scopes for hobbiests [04:47:11] <GuySoft> jmkasunich, of course i wont.. but if you saw what i posted before, you could see what i was looking at.. i thought of appending different devices to something working [04:47:45] <jmkasunich> running some motors with EMC/HAL isn't hard [04:47:48] <jmkasunich> but thats just the start [04:48:20] <GuySoft> jmkasunich, i am scheming along that data on this page.. there is a lot to take in. should i look at semething spesific? i want to figure out what chips ill need for this.. i dont know if i could get them around here [04:48:32] <GuySoft> jmkasunich, why is it the start? [04:48:41] <jmkasunich> the bartels stuff does polar alignment, lets you go to specific stars, etc - lots of astronomy specific features that you won't get from EMC [04:49:04] <GuySoft> when i interface with indi i am done... it will get the commands directly from kstars [04:49:55] <jmkasunich> I don't know what indi is [04:50:11] <jmkasunich> I didn't read the entire conversation (and I'm not going to, its almost 1am here) [04:50:38] <jmkasunich> I just wanted to let you know that somebody has already invented what you are looking for, and has been refining it for about 10 years [04:50:47] <jmkasunich> and its available under the GPL [04:50:50] <GuySoft> http://indi.sourceforge.net/index.php/Main_Page here it is , they say its a: INDI is a distributed control protocol designed to operate astronomical instrumentation. INDI is small, flexible, easy to parse, and scalable. It supports common DCS functions such as remote control, data acquisition, monitoring, and a lot more. [04:51:23] <jmkasunich> * jmkasunich looks [04:51:24] <GuySoft> jmkasunich, i know its invented. its the implementation thats the hard part now [04:51:58] <jmkasunich> ? [04:52:03] <jmkasunich> the bartels stuff is already implemented [04:52:41] <GuySoft> jmkasunich, yes, but i need to repeat the implantation ;-) [04:52:49] <jmkasunich> why? [04:52:57] <SWPadnos> building another unit [04:53:11] <jmkasunich> "unit" = what? [04:53:13] <jmkasunich> a scope mount? [04:53:16] <SWPadnos> assembly [04:53:21] <SWPadnos> not implementation, assembly [04:53:33] <jmkasunich> please clarify [04:53:36] <SWPadnos> GuySoft, you're in Italy? [04:54:07] <SWPadnos> jmkasunich, all the parts are designed, but the assembly for him isn't yet [04:54:08] <GuySoft> no, israel.. [04:54:14] <jmkasunich> parts for what? [04:54:19] <GuySoft> why, you saw il ? [04:54:22] <jmkasunich> are we talking about a scope mount? [04:54:35] <GuySoft> SWPadnos, yes sorry, the word slipped from my mind [04:54:34] <SWPadnos> building up kits (low budget), figuring out all the rest of it [04:54:37] <SWPadnos> yes [04:55:14] <jmkasunich> so GuySoft wants to build and sell scope mounts, not just build one for his own use? [04:55:18] <SWPadnos> no [04:55:23] <GuySoft> jmkasunich, a scope mount/camera mount.. the motors i have dont have sufficient power at the moment.. but i should be able to solve that [04:55:25] <SWPadnos> well, I don't know :) [04:55:58] <GuySoft> what was the question just now about me? [04:55:58] <jmkasunich> indi is NOT a scope mount controller, it is a high level interface to a controller [04:56:19] <jmkasunich> you have to provide the controller, and if there isn't an indi driver for it, you have to provide the driver [04:56:28] <jmkasunich> the bartels stuff IS a scope mount controller [04:56:47] <GuySoft> jmkasunich, i know.. its a long way up from the physical level [04:57:30] <SWPadnos> the idea is to make the mount controllable from some other Linux software, like the indi thing or kstars [04:57:48] <jmkasunich> http://indi.sourceforge.net/index.php/Devices [04:58:16] <jmkasunich> I notice the bartels controller is one of the target devices for indi, although there is no page behind that link (which makes me think there is no support yet) [04:58:50] <GuySoft> jmkasunich, they also have a manual for writing new drivers.. it shouldn't be hard to write if i have the right existing functions already on a driver [04:58:50] <SWPadnos> right [04:59:03] <jmkasunich> GuySoft: if you want/need to build a controller, I can't think of any better source of info than bbastro and the links from that page [04:59:31] <GuySoft> jmkasunich, right, i just noticed you said they are using it in the indi project [04:59:33] <jmkasunich> the amateur telescope making list (bartels is an active member) has been refining the design for years, you can learn lots there [04:59:56] <jmkasunich> theres also a scope-drive group at yahoo-groups [05:00:00] <jmkasunich> (link on the bartels page) [05:00:03] <GuySoft> jmkasunich, i am sure they are, much better than me [05:00:38] <jmkasunich> most of that work is completley public - reading those pages will teach you what you need to know far better than we can [05:01:08] <GuySoft> jmkasunich, where do you see that they are working on bbastro integration? [05:01:22] <jmkasunich> I didn't say they were working on it [05:01:38] <GuySoft> jmkasunich, well you are more understandable than the pages :) [05:01:51] <jmkasunich> under "Telescopes" on the indi Devices page there is a link for Mel Bartels Controllers [05:02:10] <jmkasunich> GuySoft: the problem is that I cannot spend weeks interpreting that info for you [05:02:32] <GuySoft> jmkasunich, ah yes i see. hang on, it might be supported already.. let me check kstars [05:02:36] <jmkasunich> that page and the pages linked to it are as close as you will get to an encyclopedia of amateur scope drive design [05:02:41] <jmkasunich> you can educate yourself there [05:02:56] <GuySoft> jmkasunich, thanks for all the help [05:03:13] <jmkasunich> are you subscribed to the indi mailing list? ask there.... [05:03:23] <GuySoft> just can we go back to the first problem? [05:03:40] <jmkasunich> I have to go to bed [05:03:58] <GuySoft> jmkasunich, i will. they just came back online.. a few hours ago they were down for maintenance [05:04:21] <GuySoft> jmkasunich, me 2, its 8am here.. i slept only 3 hours.. and i did that because there was a power cut here.. [05:04:35] <jmkasunich> you must be young [05:04:57] <jmkasunich> I have to get up for work in the morning [05:05:06] <GuySoft> YAY, mel bartels is allready integrated to kstars.. if i use that, it will save me alot of work [05:05:23] <jmkasunich> ;-) [05:06:19] <GuySoft> jmkasunich, i am, however i am on vacation from the army, this is the time i can muck about with hardware.. and not be surrounded by semi-evolved simians [05:06:52] <jmkasunich> heh [05:07:49] <GuySoft> its mandatory service here.. i am counting the days down.. i hope by the time i get out ill have better home-built astronomical equipment .. [05:08:05] <GuySoft> well, i wont keep you, sleep well, thank you for the link.. i have alot to read now [05:08:18] <jmkasunich> you should really subscribe to the scope-drives yahoo list, and/or the amateur telescope makers mailing list [05:08:32] <jmkasunich> great groups of people on both lists, very helpfull [05:10:08] <jmkasunich> goodnight [05:13:33] <GuySoft> i find IRC more responsive. goodnight [06:07:33] <fenn> grumph [06:44:23] <Jymmm> grumf? [10:18:45] <Jymmm> alex_joni: you got ballscrews? [10:43:43] <Jymmm> Well, I inspected each of the 68 balls in the ballnut.. nice and round (no ovals) and all the same size. [10:45:55] <Jymmm> Inside the ballnut itself, I did see a shiny track along the ball thread groove, looks like wear. Checked out the return tube too, seems look... balls travel thru it unobstructed. [10:46:27] <lerneaen_hydra> Jymmm: how did you inspect hte balls? [10:46:32] <lerneaen_hydra> just look at them? [10:46:38] <lerneaen_hydra> or something more advanced? [10:47:00] <Jymmm> Not sure what to do at this point. Before I took it apart, I turned the leadscrew vertical, and I'd notice when I let the nut free-spin, it would drag about every 1.5" [10:47:34] <Jymmm> when going from A to B, the other way B to A, no drag. [10:47:49] <Jymmm> Under a magnifying glass [10:48:46] <Jymmm> lerneaen_hydra: Under a magnifying glass [10:50:22] <lerneaen_hydra> oh ok [10:51:29] <Jymmm> No micrometers, so I used my dial calipers to measure each ball, move it around, measure it again. All were consistant size. [10:52:03] <Jymmm> and I didn't drop one =) [10:52:44] <toast> measuring balls? [10:53:10] <Jymmm> (03:43:41) Jymmm: Well, I inspected each of the 68 balls in the ballnut.. nice and round (no ovals) and all the same size. [10:53:10] <Jymmm> (03:45:54) Jymmm: Inside the ballnut itself, I did see a shiny track along the ball thread groove, looks like wear. Checked out the return tube too, seems look... balls travel thru it unobstructed. [10:53:10] <Jymmm> (03:46:27) lerneaen_hydra: Jymmm: how did you inspect hte balls? [10:53:10] <Jymmm> (03:46:30) lerneaen_hydra: just look at them? [10:53:11] <Jymmm> (03:46:36) lerneaen_hydra: or something more advanced? [10:53:15] <Jymmm> (03:47:00) Jymmm: Not sure what to do at this point. Before I took it apart, I turned the leadscrew vertical, and I'd notice when I let the nut free-spin, it would drag about every 1.5" [10:53:16] <Jymmm> (03:47:33) Jymmm: when going from A to B, the other way B to A, no drag. [10:53:18] <Jymmm> (03:47:47) Jymmm: Under a magnifying glass [10:53:20] <Jymmm> toast: ^^^ [10:53:38] <toast> ty sir! [10:53:46] <Jymmm> np [10:53:54] <toast> also there's no way to measure bearings with a micrometer [10:53:58] <toast> they're too accurate [10:54:36] <toast> verrry tiny doodads in the surface geometry will cause problems [10:54:48] <Jymmm> I wonder if I should NOT be using grease to lube them [10:55:01] <toast> dunno, i see grease on a lot of ballscrews, and not on others [10:55:22] <toast> i think some manufactuers reccomend grease, some tell you to hit it with waylube [10:55:47] <Jymmm> It seems to place drag on them, and lot build up inside the ballnut even after I cleaned it up. [10:56:50] <Jymmm> I have to use a grease gun for my rails, lithium grease is available easily, so I've been using that on the ballscrews as well. [10:56:57] <toast> oooh. [10:57:21] <toast> yeah, grease is only a good idea if there's a grease joint thingy [10:57:31] <toast> otherwise, it is a bad idea [10:57:38] <Jymmm> zirk fitting on the rail trucks [10:57:46] <toast> oh, neat [10:58:28] <toast> i'd probably use waylube over grease on screws [10:58:38] <Jymmm> maybe not a good thing for the ballsscrews though, maybe I'll call thompson tomorrow when wake up. [10:58:42] <toast> you might want to ask the manufacturer [10:58:42] <Jymmm> what's waylube? [10:58:42] <toast> yar [10:58:55] <toast> waylube is a specific kind of oil meant for machine ways [10:59:15] <toast> big machines are really particular about what kind of oil your hit the ways with [10:59:59] <Jymmm> it's an oil? [11:00:03] <toast> yeah [11:00:07] <toast> almost like syrup or honey [11:00:08] <Jymmm> not grease? [11:00:11] <toast> yes [11:00:17] <Jymmm> spray can? [11:00:20] <toast> loool [11:00:25] <toast> you aren't gonna get waylube to spray [11:00:30] <toast> you might be able to brush it on [11:00:30] <Jymmm> k [11:00:41] <toast> it's for automatic oilers [11:00:45] <Jymmm> I have an oil can here [11:00:47] <toast> both single shot and full auto [11:00:52] <toast> yeah, that will work [11:01:06] <toast> one of the real popular ones is Mobil Vatra 2 [11:01:10] <Jymmm> just not easy since I have to apply it upside down [11:01:24] <toast> also you have to be careful with waylube [11:01:35] <toast> is your ballscrew shielded? [11:01:47] <toast> because waylube will turn into jeweler's rouge very fast if it gets dirty [11:01:50] <Jymmm> like wipers? [11:01:57] <Jymmm> no wipers [11:02:01] <toast> wipers or a sheath [11:02:07] <Jymmm> nothing [11:02:07] <toast> or just generally out of the way of chips [11:02:14] <toast> hmm [11:02:23] <toast> you might want to come up with some sort of like, cloth sheath for it [11:02:39] <toast> i'd sew an accordian thing for it [11:02:53] <toast> run some strings on self winding crap [11:02:59] <toast> so that the sheild bellows are held off the screw [11:03:09] <Jymmm> sawdust [11:03:13] <toast> ooooooh [11:03:15] <Jymmm> and plastic partical [11:03:24] <toast> do not put waylube on your ballscrews [11:03:24] <Jymmm> goes everywhere [11:03:26] <toast> until you shiled your screws [11:03:32] <toast> which i suggest you do anyway [11:04:30] <toast> if you have access to a sewing machine, protecive bellows are just doughnuts or squares [11:04:36] <toast> of material [11:04:36] <Jymmm> theres no direct drop on the screws, just flying dust, so cover wont do much good [11:04:47] <toast> i'm saying a full 360 degree cover [11:05:14] <lerneaen_hydra> toast: how do you seal the seams? tight stitches and PCV "cloth"? [11:05:32] <toast> kind of, yeah [11:05:38] <toast> you can glue the seams after sewing [11:05:41] <lerneaen_hydra> tight enough? [11:05:46] <toast> most of the time the seam is good enough [11:05:50] <toast> by itself [11:05:55] <lerneaen_hydra> oh ok [11:06:08] <toast> any heavy cloth will work [11:06:12] <lerneaen_hydra> so no need to seal it with heat/adhesive [11:06:14] <toast> right [11:06:19] <Jymmm> I'm outta here.... Zzzzzzzzzzz [11:06:21] <toast> you're not like, making it water/airtight [11:06:26] <toast> just trying to keep crap off it [11:07:09] <lerneaen_hydra> yeah [11:07:44] <toast> it increases the life of the machine by a very good factor [11:07:57] <toast> and allows you to run better lubricants on the ways and screws [11:08:09] <toast> i.e. waylube, which almost eliminates stick/slip [11:08:25] <lerneaen_hydra> oh, nice [11:08:39] <lerneaen_hydra> how are those felt scrapers compared to bellows? [11:08:46] <toast> big machines use both [11:08:51] <lerneaen_hydra> oh ok [11:08:54] <toast> felt scrapers are like, necessary [11:09:10] <lerneaen_hydra> bellows to get rid of most of the crap and scrapers to clean the stuff that gets through? [11:09:16] <toast> indeed [11:09:28] <toast> and also wipes the old oil off the way [11:09:57] <toast> the stuff that's turned into lapping compound =) [11:10:14] <lerneaen_hydra> heh yeah [11:10:30] <toast> a big problem on smaller machines is the use of compressed air [11:10:41] <toast> especially bridgeport sized machines [11:10:51] <toast> because nothing is protected and compressed air will blow chips right under the wipers [11:11:04] <toast> into the soft, unhardened bearing surface [11:11:13] <toast> and also onto the screws [11:11:32] <lerneaen_hydra> yeah [11:11:52] <lerneaen_hydra> therefore people using air on machines should be executed [11:11:53] <toast> but a well-designed bellow will allow a person to use air [11:12:05] <toast> people using air on machines that are not designed to handle compressed air [11:12:07] <toast> should be executed [11:12:13] <toast> and air should just be minimized [11:12:17] <lerneaen_hydra> yes, naturally [11:54:25] <fenn> i like les' way of shielding the ways, he has this roll-up window blind thing [11:54:30] <fenn> and its under constant tension [11:55:30] <fenn> http://www.lmwatts.com/cnc.html [13:04:42] <maddash> hm [13:05:01] <maddash> someone wanted to know about lostirc? [13:09:39] <fenn> lost on irc island... [13:11:29] <toast> i've always wondered if you could make a tv show [13:11:38] <toast> that is just essentially irc [13:12:05] <fenn> its called 'talk radio' [13:12:19] <fenn> picture rush limbaugh with some headphones and a mic [13:12:26] <toast> except this would be irc people talking about wangs [13:12:30] <toast> instead of like, stuff nobody cares about [13:12:35] <fenn> um, right [13:12:55] <fenn> 1) talk about wangs 2) ??? 3) profit! [13:13:28] <toast> that is essentially how irc works right now [13:13:34] <toast> "hey guys, wangs" [13:13:37] <toast> "holy shit three hundred dollars" [13:13:41] <lerneaen_hydra> utter boredom with the occasional, rare moment of utter terror [13:17:50] <archivist> trawl the logs for a good strory line [13:18:44] <toast> yeah man [13:18:46] <toast> and just play it back [13:18:50] <toast> according to the timestamps [13:19:28] <archivist> but there have been some choice statements worth saving for posterity (I save some in the #mysql bot) [13:20:52] <lerneaen_hydra> or better yet, print qdb.us [13:20:52] <toast> ? [13:21:04] <toast> re: mysql bot [13:21:07] <lerneaen_hydra> you've never been to qdb.us? [13:21:26] <lerneaen_hydra> http://www.qdb.us/top [13:21:30] <lerneaen_hydra> worlds best IRC quotes [13:21:38] <lerneaen_hydra> what bash.org was [13:21:46] <toast> i know the guy who runs qdb [13:21:53] <lerneaen_hydra> oh cool [13:22:00] <toast> well, know/sit in the same chatroom [13:22:02] <toast> digdug [13:22:08] <lerneaen_hydra> ah ok [13:22:20] <toast> never met the guy irl [13:22:24] <toast> HOTHACON never happened [13:22:34] <toast> so sad [13:32:47] <skunkworks> <TeamsterX> man watching 6 MSCE's around a sun box, looks alot like the opening scene's of 2001:space odyssey and the monkey's with the monolith [13:33:31] <cradek> TeamsterX: But, maybe they know not to form plurals with an apostrophe. [13:33:59] <jepler> zing! [13:34:14] <skunkworks> :) [13:34:35] <fenn> and rely on capitalization alone? sounds dangeroso [13:34:39] <skunkworks> atleast he is consitant [13:35:22] <skunkworks> consistent [13:35:43] <fenn> Did you know, Alt+Z makes your text go backwards in most IRC clients? [13:35:54] <skunkworks> ;) [13:36:36] <skunkworks> * skunkworks was going to exit for the fun of it.. ;) [13:41:33] <cradek> did you know you can cause someone else to exit irc by typing /quit and then their name? [13:42:33] <skunkworks> * skunkworks has quit ("leaving") [13:43:58] <cradek> is alt-Z something special in winders? [13:44:56] <jepler> if it said ctrl-z you might have gotten me :-P [13:57:33] <maddash> sheesh [13:58:23] <lerneaen_hydra> O_o [13:59:02] <lerneaen_hydra> * lerneaen_hydra see's a +o cradek, a maddash: oh shit, and a maddash was kicked ("better luck next time") [14:00:43] <maddash> sees*? [14:01:03] <lerneaen_hydra> oh kill me now [14:01:21] <lerneaen_hydra> I need to move to an english speaking country and fix my broken linguistics [14:01:45] <lerneaen_hydra> who the hell makes mistakes like that? 6th graders? [14:01:52] <lerneaen_hydra> bleh... [14:02:59] <jepler> lerneaen_hydra: almost any american might type "see's" [14:03:46] <cradek> * cradek shivers [14:04:34] <skunkworks> * skunkworks whistles to himself.. [14:04:38] <maddash> jepler: ahem. [14:07:19] <tomp> Cutting a Dash: a BBC radio series on how punctuation is murdered, this episode is about the (mis) use of the apostrophe try this cmd "mplayer rtsp://rmv8.bbc.net.uk/bbc7/1345_mon_thu.ra" [14:09:06] <skunkworks> npr has a show called 'a way with words.' I try to catch evey weekend. Doesn't seem to help much ;) [14:09:33] <lerneaen_hydra> jepler: still, it's very ugly IMO [14:09:47] <lerneaen_hydra> lack of common grammar skills [14:10:40] <cradek> nobody learning english should aspire to the average american's skills [14:11:40] <lerneaen_hydra> cradek: nobody learning anything should aspire to the average american's skills [14:11:47] <lerneaen_hydra> ;) [14:12:13] <maddash> what's with this growing trend of bashing Americans? [14:12:43] <jepler> maddash: r u kiddin'g? [14:13:24] <cradek> the average american has poor language skills: not bashing, just a simple statement of fact [14:17:13] <lerneaen_hydra> http://www.wellingtongrey.net/miscellanea/archive/2007-03-18--of-angels-monkeys-and-fat-men.html [14:17:53] <jepler> ba ha ha haha [14:27:19] <jepler> welcome, cnc_engineer, barum, and Vq^ [14:28:23] <cnc_engineer> i am unable to get lenearity in PWM output please help to solve the problem [14:30:11] <jepler> cnc_engineer: last time we talked about this, I suggested you use a component to map from requested RPM to PWM duty cycle: http://emergent.unpy.net/files/sandbox/linear8.comp [14:30:14] <jepler> cnc_engineer: did you ever try this approach? [14:30:50] <jepler> 17:38:57 <jepler> cnc_engineer123_: the is component, which I've just written, might be helpful. It performs linear interpolation with up to 8 data points: http://emergent.unpy.net/files/sandbox/linear8.comp [14:30:54] <jepler> 17:43:24 <jepler> you would find several pairs of (pwmgen input):(measured spindle speed); use (pwmgen input) as the yi values and (measured spindle speed) as the xi values [14:30:57] <jepler> 17:46:24 <jepler> if you use less than 8 value pairs, then set the 'n' parameter to the number of pairs you chose to use [14:32:06] <jepler> (from IRC log http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-06-28.txt) [14:35:25] <cnc_engineer> Jepler: i could not try this approach, i think its first time we are discussing this matter [14:37:13] <maddash> rofl [14:38:51] <cnc_engineer> JeplerL: i need to convert Pwm to Analouge voltage command to control my servo moters but when i need to open 50% duty cycle actual thers 75% and when there should be 60% actual duty cycle goes 100% [14:40:30] <cradek> are you using pwmgen on the parallel port to control full size servos? [14:41:05] <jepler> cnc_engineer: when you say there is a difference between the 50% duty cycle (desired?) and 75% (measured?), how are you making the measurement? [14:44:09] <cnc_engineer> i am using osciloscope to check the duty cycle and i give position commands whithout powerup the motor [14:49:11] <SWPadnos> cnc_engineer, test the output pin by hooking a charge_pump component to it [14:49:27] <SWPadnos> the output you're using for PWM may have asymmetric drive strength [14:50:11] <SWPadnos> hmmm. nevermind. the jitter in interrupt timing will be far worse than any pin drive differences [14:50:36] <x3rox> I have a question on the L297+L298 IC's. Is it better to set the "control" input high or low? High=Chopper acts on phase lines ABCD, Low=Chopper acts on INH1+2. Currently I use "Low". What is better to lower heating of the L298? [14:52:01] <jepler> in halscope I can see that the pwmgen output is exactly what I would expect [14:53:09] <SWPadnos> halscope follows the software jitter [14:53:13] <jepler> indeed [14:53:27] <SWPadnos> you should get a real scope :) [14:53:31] <jepler> it's hard to see how jitter could systematically lengthen all duty cycles [14:54:04] <SWPadnos> that's true [14:54:33] <SWPadnos> there is probably some systematic issue from drive strength, but it should be some fraction of a percent [14:55:05] <SWPadnos> though if it's using the PDM mode (or other highest-frequency method), it could be more significant [14:59:01] <jepler> x3rox: the l297 datasheet says the effect of setting control high is to reduce power dissipation on the sense resistor .. I don't know that either setting affects the heating of the L298 itself [15:01:26] <x3rox> Ah, yes. I totally overread that. I always had in mind that this belongs to the drivers in the L298. So, actually, it doesn't matter as long as the resistors are cold enough... - Thanks. [15:02:29] <x3rox> jepler: Do you think that a PII-266 is fast enough for my mill? I want to mill with tools of 0.5 ... 2.0 mm ø, so I must move slow... [15:04:02] <x3rox> jepler: (This is an old industrial PC in 19" enclosure. Would be great to have a usage for it. [15:04:04] <jepler> x3rox: you can only find out by trying it.. [15:04:13] <x3rox> What happens in worst case? [15:04:18] <skunkworks> The slowest computer I have runn is a 400mhz. It ran fine. [15:04:51] <jepler> x3rox: worst case you can't get acceptable response time from the GUIs while running a BASE_PERIOD fast enough to get the step rate you want [15:04:57] <skunkworks> pentium II [15:04:57] <lerneaen_hydra> x3rox: you'll have to run slowly with a large base period [15:05:17] <lerneaen_hydra> you may have to have a remote graphical control [15:05:42] <fenn> reducing current thru the sense resistor would reduce power dissipation everywhere else too [15:05:48] <x3rox> Oh - So it is actually no real fun to work with? [15:06:07] <lerneaen_hydra> btw, if you SSH into a box, and start X over SSH, is there a lot of graphical load on the host machine or is most of it on the client that displays the graphics? [15:06:49] <x3rox> I assume that in addition the LAN will be a bottle neck... [15:07:28] <lerneaen_hydra> surely a 10 mb/s connection is fast enough? [15:07:46] <fenn> its prolly 100Mbit anyway [15:07:54] <x3rox> I will try to take a faster CPU and run everything on one PC. surely better. ;-D [15:07:58] <lerneaen_hydra> 200mhz, was 100base common then? [15:08:13] <fenn> eth cards werent common at all [15:08:20] <lerneaen_hydra> oh, right [15:08:33] <lerneaen_hydra> they weren't even integrated in the motherboard then [15:08:40] <x3rox> I think so. 100 MBit is not such new... [15:08:59] <x3rox> It's a Slot1 board :D [15:09:38] <x3rox> With some ISA slots! So I could add my old ISA LPT cards, if I need... :D [15:09:51] <fenn> cool beans [15:10:02] <x3rox> :D [15:10:45] <x3rox> As my father sais: Don't throw old things away, you never know if they will be useful sometimes... [15:11:21] <fenn> exercise moderation in all endeavors [15:12:10] <skunkworks> I come from a long line of horders.. [15:13:41] <skunkworks> hoarders? [15:14:03] <fenn> horde = yurt on wheels [15:15:00] <x3rox> How should I start testing the stepper hardware? Currently I fed a square-wave signal in, generated by an oscillator, but this is no real test. The stepper turns and I can change the direction, but I don't see if steps get lost. Would like to run a loop which moves some thousand times left/right a longer distance and then should stop exactly at the staring point again. How can I tell this EMC? Can it run loops? [15:15:21] <fenn> put a flag on the stepper so you can watch its rotational position [15:15:26] <fenn> line up the lines [15:16:07] <fenn> emc does loops with the O routines but i'm lazy so you'll have to look it up yourself [15:16:30] <fenn> note that it can lose steps in both directions and end up at the same place [15:16:51] <fenn> so a better test would be with gage blocks and a dial indicator or something [15:16:56] <skunkworks> now that is just crazy talk.. ;) [15:16:58] <fenn> ya [15:17:18] <fenn> easier to look at where the flag is pointing and calculate the "nominal" position [15:18:17] <x3rox> That's clear. Does EMC start immediately with full speed defined with F command, or does it use an acceleration ramp? [15:18:23] <fenn> btw skunkworks you can lose steps in both directions if your direction pulse isnt held long enough before the step pulse [15:18:36] <fenn> so you lose the first step in a new direction [15:18:53] <skunkworks> I can see that.. I hate those kind of problems. [15:18:57] <x3rox> And does EMC control the lines correctly? [15:19:16] <x3rox> (including the timing)? [15:19:22] <skunkworks> emc does everthing ;) - it accelerates and controls the lines the way you set it up.. [15:19:24] <fenn> x3rox: g0 does full speed, g1 goes F, always with acceleration ramp [15:19:45] <fenn> you define the accel per axis and also overall [15:20:44] <x3rox> Can I configure EMC to know up to which speed it can immediately start with 1 step, and from which speed on it must use acceleration? [15:21:37] <fenn> no [15:22:04] <skunkworks> it starte the accelleration from 0. [15:22:23] <fenn> you cant step less than a pulse anyway [15:23:24] <maddash> fenn: g0 obeys the acceleration limits, too [15:23:38] <fenn> that's what i meant if it wasnt clear [15:23:49] <x3rox> Don't understand. So if I mill a polygon, EMC accelerates from 0 to F, mills the vector, then breakes down to 0, then does the same with next vector? [15:24:08] <maddash> x3rox: only if you disable blending [15:25:12] <x3rox> Ok. "blending" remembers current speed in every axis and accelerates only to a new speed? [15:25:17] <fenn> x3rox: if you want to mill an exact polygon you must stop at the corners, otherwise you will have infinite acceleration at the corner [15:25:48] <fenn> in practice it's not a big deal unless you're doing high speed machining [15:26:22] <x3rox> You are right. With a tool-diameter below 1mm I will not have much speed at all. :-D [15:26:50] <fenn> actually blending is still important at that scale [15:27:07] <x3rox> What does blending in detail? [15:27:15] <fenn> say that again [15:27:32] <x3rox> ? [15:27:41] <fenn> do you mean what part of the code calculates blending? [15:27:54] <skunkworks> x3rox: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TrajectoryControl [15:28:17] <x3rox> No. I mean: What is the result of blending=enabled. [15:28:31] <fenn> the yellow curves on the first pic on that page [15:29:06] <x3rox> Huh... [15:29:11] <maddash> damn, I hope my brother doesn't eat all the spaghetti [15:29:28] <x3rox> Grab some before he can... ;-) [15:30:50] <x3rox> Ok - I browsed over the text in the link. Seems to be better to disable blending at all. [15:31:31] <fenn> try it out both ways and see what you like [15:31:41] <fenn> 99% of the time a rounded corner doesnt matter [15:32:00] <fenn> and you can program a stop in regular g-code easily enough [15:32:05] <fenn> for just one corner [15:32:24] <x3rox> Are the differences noticable, or are we talking about such minor differences that one has to search them? [15:32:46] <x3rox> (Is this G61, G64, ...) [15:32:47] <fenn> it would be hard to tell the difference in the final part, but the machine will go a lot smoother and faster with blending on [15:32:52] <lerneaen_hydra> usually your acceleration is so high and speed so low that's it's irrelevant [15:33:02] <lerneaen_hydra> (the rounded corners) [15:33:41] <x3rox> I see. So we are actually talking about 0.0x millimeters of a radius? [15:33:58] <x3rox> Or is it more? [15:34:02] <maddash> x3rox: depends. are you making nun-chucks? [15:34:15] <x3rox> What is a nun-chuck? [15:34:20] <fenn> ya .05mm sounds about right [15:34:47] <x3rox> So it is really no problem. [15:35:12] <maddash> not if you want to break skin [15:35:27] <fenn> hey arent nun-chucks usually round? [15:35:35] <archivist> where do you get hex nuns from [15:35:38] <SWPadnos> x3rox, it'blending is a physics problem. it's physically impossible to get a non-rounded corner unless you (a) use exact stop mode or (b) put big loops in the G-code that overshoot the endpoint and then come back on the new tangent [15:36:31] <maddash> fenn: concave polygons work quite well, actually [15:36:58] <SWPadnos> the path blending is controllable using G61 and G64. I think G64Pxx gives a tolerance of xx units before it slows the feed rate [15:37:10] <SWPadnos> but I could be wrong there - the docs are your friend :) [15:37:35] <x3rox> Yes, because if I want to mill a 90° corner, the second axis would need an infinite accel in such cases. I thought that blending sees such things and deccelerates before the endpoint to a level that the other axis stays inside its limits. [15:38:23] <SWPadnos> it does, and the Pxx tolerance option allows it to go a bit faster, as long as the resulting path is "close enough" [15:39:00] <maddash> omg the rs274ngc code thinks in metric [15:39:04] <maddash> atrocity! [15:39:05] <SWPadnos> hmmm. I may have mis-spoken there. I'm not sure that the normal blending mode will slow down to take corners [15:39:18] <SWPadnos> I know the tolerance mode does [15:39:18] <fenn> geez maddash what would you prefer, furlongs? [15:40:07] <x3rox> I see, the fine-tuning will somewhat difficile... [15:40:08] <fenn> SWPadnos: afaik the Pxx just lets you specify the tolerance, rather than the hard-coded 0.0001 or whatever [15:40:28] <SWPadnos> fenn, that could be [15:40:47] <fenn> and its not uniform, like, the tolerance is .0001 inches but its .001 mm [15:41:06] <fenn> unless you use g64 P.. [15:41:07] <lerneaen_hydra> what's the current default? [15:41:09] <lerneaen_hydra> keep speed constant, no limit on blending? [15:41:29] <fenn> default is g64 p.0001 inches or .001 mm [15:42:20] <fenn> i love this picture http://wiki.linuxcnc.org/uploads/planner-test.jpg [15:42:48] <x3rox> It is really impressive (and amazing) what EMC can do. I am only afraid that all this withh need a lot of time to get configured correctly. Or is a good average config for a 3-axes-mill with steppers included, which only must be fine-tuned? [15:43:13] <fenn> yep [15:43:20] <jepler> x3rox: start with sample-configs/stepper/stepper_inch.ini [15:43:29] <maddash> x3rox: stepper/stepper_*.ini [15:43:34] <maddash> oh, heh. [15:43:54] <x3rox> fenn: This is really a nice picture. ... so - abstract...? [15:43:59] <lerneaen_hydra> maddash: and having rs274 in metric is bad because...? [15:44:23] <jepler> x3rox: select your pinout ("standard" or "xylotex", or make your own by modifying the .hal file), and enter stuff like the steps per inch, max speed in inches per second, max accel in inches per second squared, etc [15:44:56] <jepler> x3rox: planner-test.jpg was a randomly generated program that included linear and circular moves in 4 axes (XYZA) [15:45:26] <jepler> x3rox: it was run with a special setup that checked 1000 times per second that the commanded acceleration and velocity were within the limits specified in the .ini file [15:45:57] <jepler> x3rox: also it could be visually inspected (actual path vs preview plot) to see if there were any gross errors in blending [15:46:09] <jepler> a few bugs of both kinds were found, but then we (cradek) fixed them [15:46:08] <x3rox> jepler: This is exactly what my mill will be able to do in final state. I can mount a turning axis on the machine desk. [15:46:24] <maddash> lerneaen_hydra: because i'm orthodox english and i lose accuracy from the conversion [15:47:04] <fenn> maddash: i believe it also uses floating point math, horror of horrors [15:47:24] <fenn> how will you ever get your 1/3 inch widget exactly perfect? [15:47:24] <lerneaen_hydra> yes you might loose 1e-23 somewhere there! [15:47:56] <x3rox> Because I am from Austria (and we use here the metric system), I would prefer to calibrate the mill in millimeters. Is this a problem? [15:48:19] <lerneaen_hydra> x3rox: no, it can be configured either way [15:48:41] <x3rox> And what's that mentioned above? [15:49:16] <jepler> x3rox: if you want millimeters, then start with stepper_mm.ini [15:49:35] <lerneaen_hydra> you loose an infitesemal amount of accuracy when running in inch mode becuase the conversion can't be completely accurate, due to the length of one inch [15:49:47] <maddash> x3rox: it would seem that you're better off with metric [15:49:49] <fenn> actually an inch is 25.4mm [15:49:51] <lerneaen_hydra> 1/25.4 is an irrational number [15:49:52] <maddash> lose* [15:49:54] <jepler> then you enter steps per mm, max speed in mm per second, max accel in mm per second squared. in gcode files, specify g21 so that coordinates are also in millimeters [15:49:57] <fenn> no it isnt [15:50:01] <lerneaen_hydra> isn't it? [15:50:12] <lerneaen_hydra> I always thought it was [15:50:12] <maddash> lerneaen_hydra: no, rational. [15:50:12] <fenn> .0393700787401574803149606299212598425196850393700787401574803149606\ [15:50:25] <cradek> irrational means it can't be expressed as a ratio [15:50:26] <jepler> lerneaen_hydra: the ratio of two integers is always rational [15:50:31] <maddash> lerneaen_hydra: 1/25.4 = 10/254 = a/b [15:50:40] <fenn> see the repeating er.. mantissa/ [15:50:43] <lerneaen_hydra> oh, right, transcendantal? [15:50:47] <maddash> cradek: irrational means that it's crazy. [15:50:49] <lerneaen_hydra> or whatever it was called [15:51:01] <fenn> 039370078740157480314960629921259842519685 repeats [15:52:06] <x3rox> Ok. Thanks. -- How do I find out the right calibration of my mill? It's a shame, but I don't know the pitch of the leading spindles, I only know that the steppers are 200 steps/turn (1.8°). [15:52:24] <cradek> you have to measure it then [15:52:31] <x3rox> I think that counting 10 turns and measuring is in best case an estimation? [15:52:33] <fenn> 42 digits [15:52:35] <cradek> first try using a thread gauge [15:52:48] <cradek> or measure the axis travel with something [15:52:48] <jepler> For all multiples of .0001 inch from -10.0000 to 9.9999, the maximum error introduced by converting from inches to mm to inches is 1.7763568394e-15. test program: http://pastebin.ca/636041 [15:52:54] <lerneaen_hydra> fenn: so the sequence repeats ad infinitum? [15:52:59] <fenn> ya [15:53:03] <jepler> that value is so much smaller than the accuracy of physical positioning that it can be ignored [15:53:14] <lerneaen_hydra> jepler: hence that the error is nearly infinitessimal [15:53:25] <fenn> lerneaen_hydra: check this out: 123456789/999999999 [15:54:27] <fenn> oh i guess i should mention you can get an arbitrary number of decimal points if you type 'bc' and then 'scale=999' [15:54:41] <lerneaen_hydra> fenn: type where? [15:54:48] <fenn> in the bc shell [15:54:48] <maddash> jepler: clever. [15:54:50] <jepler> interestingly, the maximum error in representation of a multiple of .0001 inch is larger: 1.45e-11, since .0001 is not exactly representable [15:54:59] <lerneaen_hydra> oh, right [15:55:17] <maddash> fenn: it only seems that way because your denominator is so close to 1000000000 [15:55:47] <fenn> i'm pretty sure there's a proof it goes on forever, but i'm no mathematician [15:55:51] <lerneaen_hydra> fenn: neat [15:55:53] <maddash> fenn: it doesn't. [15:56:10] <fenn> * fenn shrugs [15:56:14] <fenn> it's a religious thing [15:56:42] <x3rox> cradek: It's an acme thread, ø12.67mm. [15:56:56] <maddash> fenn: it does, then 1123456789/(10^9*(10^9-1)) [15:57:03] <maddash> fenn: if it does, then 123456789/(10^9*(10^9-1)) [15:57:09] <maddash> =0 [15:57:13] <lerneaen_hydra> x3rox: that's 1/2 inch diameter [15:57:38] <lerneaen_hydra> might be inch-based pitch [15:57:39] <maddash> ^^ equals zero, not some stupid emoticon [15:57:50] <fenn> uh, no [15:57:50] <lerneaen_hydra> * lerneaen_hydra pouts [15:58:26] <fenn> that thread is probably an inch thread [15:58:32] <cradek> x3rox: it's probably 1/2 inch then, it'll be an integer number of turns per inch [15:58:35] <fenn> and 1/0 is infinity and that's final [15:58:44] <x3rox> cradek: So I can be nearely sure that the thread will also have an "inch-pitch"? [15:58:55] <lerneaen_hydra> fenn: *coughnot*really*cough* [15:59:06] <cradek> x3rox: yes [15:59:30] <fenn> * fenn guesses 10 threads per inch [15:59:43] <jepler> if you had an algebra textbook, you could find the page where they prove that 1/3 = \sum_{i=1}^\inf 3/(10^i) [16:00:02] <maddash> jepler: but this isn't an infinite sum [16:00:15] <maddash> fenn: 1/0 is undefined [16:00:18] <jepler> the same property of sums of geometric sequences will show that the pattern in 123456789/999999999 repeats indefinitely [16:00:44] <cradek> fenn is right - it'll probably be 10tpi [16:00:53] <maddash> jepler: if it does, then 999999... = 99999...+1 [16:01:02] <maddash> jepler: if it does, then 999999...9 = 99999...9+1 [16:01:10] <fenn> you mean infinity = infinity+1? [16:01:23] <maddash> fenn: the second one, not the first one. [16:02:02] <fenn> infinity isnt in the set of addable numbers [16:02:25] <fenn> that's god giving you a type error [16:02:30] <maddash> infinity isn't even a number [16:02:33] <x3rox> This will become funny to convert/calibrate. Which pitch will usually be if the pitch I measure is ~3.12mm? (I pressed a sheet of paper on the oily thread and measured 6 turns) [16:02:48] <maddash> rofl novel [16:03:08] <jepler> x3rox: 1/8 inch is 3.175mm [16:03:12] <cradek> x3rox: if it's 10tpi, your pitch in mm is .3937 [16:03:20] <cradek> true it might be 8tpi [16:03:44] <cradek> if you have an inch scale you will be able to see easily which one it is [16:03:54] <fenn> oo there's that number again [16:04:06] <fenn> 3149606299 [16:05:31] <x3rox> yes, 8tpi looks good. Accidently I have exactly 8 stamps on the sheet of paper and they are 1 inch. So I hane 1/8 inc per turn.? [16:05:41] <cradek> yes [16:06:07] <x3rox> Will get a strange correction numer... [16:06:20] <jepler> You have: (1|8 in / rev) / (200 steps / rev) [16:06:18] <jepler> You want: steps / mm [16:06:18] <jepler> reciprocal conversion [16:06:18] <jepler> * 62.992126 [16:06:20] <cradek> so you could either configure the machine in inches, or use .31496 [16:06:21] <maddash> in class EMC-TRAJ-CIRCULAR-MOVE, how is the 'turn' member used? [16:06:49] <jepler> (or use 400 if half-stepping) [16:07:13] <x3rox> I'll use half-steps because this can L297+L298 do. [16:08:09] <maddash> no, that came out wrong. my problem is with the line of code in emccanon.cc: "circularMoveMsg.turn = rotation - 1;" So if I want the usual arc (eg, from 45 deg to 47 deg), then would 'turn' = 0? [16:08:09] <x3rox> So I will have to set (.31496 * 2) as correction value? [16:08:27] <cradek> maddash: http://www.isd.mel.nist.gov/personnel/kramer/pubs/RS274NGC_3.web/RS274NGC_34a.html#1000406 [16:08:28] <x3rox> (For metric setup) [16:09:04] <SWPadnos> x3rox, do you have any belt reduction? [16:09:20] <x3rox> No. [16:09:25] <SWPadnos> ok [16:09:27] <jepler> x3rox: [AXIS_0]INPUT_SCALE gives the number of steps to take to move one machine unit. I get 125.98425 steps per mm based on what you describe [16:09:52] <maddash> cradek: whoa, cool [16:10:15] <cradek> maddash: now there is also UVW but otherwise I think that documentation is still right [16:10:35] <x3rox> Thanks. I will try that out soon. (First I must finish the stepper drives.) Thanks a lot for your help! [16:10:45] <SWPadnos> those functions are actually in the user manual PDF, near the end [16:10:53] <cradek> maddash: for arbitrary arcs I think you want a new canon call ARBITRARY_ARC_FEED (or some such name) [16:11:17] <cradek> I wouldn't change this part of the API (I guess you could make it call ARBITRARY_ARC_FEED) [16:11:33] <x3rox> I'll re-read this part. Now I know (thanks to your help) which thread I have, so it is easier now. [16:11:40] <x3rox> Have a nice weekend! [16:11:56] <SWPadnos> have fun x3rox [16:12:01] <x3rox> Bye! [16:12:44] <maddash> cradek: i'm only adding to the api, and taking nothing away. [16:13:54] <cradek> * cradek cringes at the "mapping of rotation to turns" comment [16:14:41] <maddash> what the hell does it mean? [16:14:52] <maddash> "1 0 2 1 -1 -1 -2 -2" [16:14:56] <cradek> at least it seems to match the code [16:15:35] <maddash> the code's pretty screwed up, actually [16:15:41] <cradek> maddash: you're working in cvs trunk aren't you? don't tell me you're not [16:15:48] <cradek> maddash: I fixed that comment in trunk [16:15:57] <maddash> cradek: er...2.1.6? [16:15:59] <maddash> * maddash runs [16:16:11] <cradek> maddash: making substantial improvements to a release version means you will have to rewrite them for trunk if you want them incorporated [16:16:48] <SWPadnos> is it just me, or does that description leave out what a, b, and c are? [16:17:01] <SWPadnos> nevermind [16:17:12] <SWPadnos> * SWPadnos goes to make a pot of coffee [16:17:48] <cradek> SWPadnos: I agree, but it's "obviously" the endpoints [16:19:24] <maddash> cradek: 'center,' 'end,' and 'normal' aren't necessary [16:22:23] <maddash> s/it's/they're/ [16:22:27] <maddash> :P [16:25:10] <SWPadnos> I got that they're the endpoints, though it's interesting that the function doesn't take a pointer to the current position, which raises the "global variables are bad" issue [16:25:22] <SWPadnos> I'm sure that's all over the place though, so no biggie [16:25:36] <cradek> all of the canon calls specify the "new" position only [16:25:55] <SWPadnos> right - the globals are part of the spec, so it shouldn't be an issue [16:26:19] <SWPadnos> "but what if we want one canon to plan for 2 machines?" :) [16:26:31] <cradek> meh [16:26:36] <SWPadnos> heh [16:26:49] <maddash> in an XY arc, why is only the x tool offset accounted? [16:27:14] <maddash> 'case CANON_PLANE_XY:' [16:27:24] <SWPadnos> it's assumed that the tool is spinning, and therefore round (I think) [16:27:35] <cradek> there are only two tool offsets, X and Z [16:28:26] <maddash> why would you only account for the X and not the Y? it's an XY arc, so Y offset matters as well [16:28:45] <cradek> there is no Y tool offset [16:28:51] <maddash> oh shit, these are *tool* offsets, not work offsets [16:28:56] <cradek> right [16:28:56] <SWPadnos> the Y offset is the same as X, since the tool is round -you only need one number to offset the radius [16:29:03] <cradek> SWPadnos: nope [16:29:12] <SWPadnos> really? [16:29:16] <maddash> lathe. and mill. [16:29:27] <SWPadnos> sure, for a lathe it's totally different :) [16:29:41] <cradek> canon knows nothing about the tool shape, these calls move "the controlled point" only [16:29:54] <maddash> and why are the additions done in sequence and not all at once? some arcane optimization procedure? [16:30:30] <SWPadnos> ah, so the compensated arc is passed to canon [16:30:44] <jepler> most likely it's just a matter of programming style [16:30:45] <cradek> yes radius/shape compensation is an interpreter function [16:31:17] <SWPadnos> ah yes - I think there was some discussion about putting compensation into canon [16:31:23] <SWPadnos> I like it where it is, personally [16:31:49] <barum> Where to find smaller emc2 Live image than 660MB. Found old Puppy iso (51MB), but if it exist with newer version that whill be better. [16:32:17] <cradek> barum: I think those are your two choices fora Live CD [16:32:29] <maddash> third choice: make your own. [16:32:29] <barum> Thanks [16:32:45] <cradek> barum: I think the Puppy version is very old [16:33:19] <fenn> that xemet ever write up how he did it? [16:33:27] <fenn> s/that/did/ [16:33:31] <cradek> fenn: there's some stuff in the wiki I think [16:36:02] <fenn> all i see is emc-puppy [16:37:23] <jepler> I thought xemet's was also based around the emc2/ubuntu live CD [16:38:30] <fenn> yeah [16:38:48] <fenn> presumably once you get it on a usb stick you can start deleting stuff like mad [16:38:59] <fenn> i forget if i ever got emc to run on DSL [16:45:20] <maddash> feck [16:45:48] <maddash> can anyone else confirm that pmCirclePoint in _posemath.c handles arbitrary helices as well? [16:46:51] <cradek> looks like it to me [16:46:53] <maddash> or can I just forget about helices for now? [16:47:01] <maddash> ah, ok. [16:47:09] <cradek> (and it looks pretty well documented) [16:48:30] <maddash> hm, we need to discuss the syntax of G02.4 [16:48:45] <cradek> did you find docs saying how another system does it? [16:48:56] <maddash> yes, I'm mimicing fanuc as closely as I can [16:49:07] <cradek> are those docs on the web? [16:49:44] <maddash> (googling) [16:50:36] <maddash> www.compumachine.com/Support/Downloads/Fadal/GE%20Fanuc/0010__GE_FANUC_User_Manual.pdf [16:50:56] <maddash> search terms: "G02.4" [16:51:28] <maddash> fanuc actually implements a g03.4, but it's redundant [16:51:48] <maddash> page 719 of 1257 in that doc [16:53:09] <maddash> anyway, we've no support for X_1/2 Y_1/2, etc, so the next best thing I could come up with was using I,J,K [16:53:59] <maddash> right now, the syntax I have in mind is " G02.4 X- Y- Z- I- J- K- ," where xyz is the endpt and ijk is the midpt [16:54:36] <cradek> you could do this two block thing, I think [16:54:43] <cradek> but I don't think it supports helixes [16:55:01] <cradek> (I don't follow the alpha/beta stuff yet either) [16:55:24] <maddash> the problem arises when you want an arbitrary helix [16:55:40] <cradek> yes [16:56:17] <maddash> * maddash should stop staring at the keyboard when typing [16:56:19] <cradek> oh alpha/beta means you can specify more axes there (like our abcuvw) [16:56:27] <jepler> cradek: I think that it means you could command a move on up to two other axes (ABCUVW in the case of an adaptation to emc) [16:56:35] <jepler> right [16:56:39] <cradek> must be right if we both think that [16:56:41] <maddash> * maddash is contagious [16:57:01] <jepler> ("only two additional axes" seems like a stupid limitation..) [16:56:59] <cradek> up to two is a strange limitation [16:57:04] <cradek> hahaha [16:57:16] <cradek> one of us is redundant here [16:57:20] <jepler> so leave [16:57:26] <cradek> wanna go to lunch? [16:57:28] <maddash> the fanuc standard was designed specially for the fanuc machine, which only has 5-axes [16:57:58] <maddash> five axes* [16:59:21] <maddash> at any rate, i'm bringing up the syntax because of the helix issue. frankly, it would seem simpler to leave out helices entirely (because I'm a lazy-ass). [17:05:18] <lerneaen_hydra> one thing that IMO warrants thinking about: when would you need to code a helix manually? is there ever a case where you need a helix *and* need to code by hand? [17:05:43] <lerneaen_hydra> 3d arcs I can understand, smaller programs, faster loading [17:06:07] <lerneaen_hydra> helices can be appoximated with a couple 3d arcs [17:07:39] <SWPadnos> a helix is a good pattern for milling circular holes, and that can't be done with arcs if they're constrained to a plane [17:07:53] <SWPadnos> err - I should go drink some of that coffee [17:08:44] <maddash> a helix is also needed for threading [17:08:48] <maddash> useful* [17:11:26] <fenn> you cant thread without helices [17:11:30] <archivist> clockmakers need a special helix [17:12:04] <SWPadnos> you only need helixes for threadmilling, no? [17:12:09] <SWPadnos> (relative to threading) [17:12:51] <archivist> Im thinking of the Fusee [17:13:38] <SWPadnos> sure - you need a decreasing arc helix for that (or increasing) [17:15:43] <archivist> often a turned item on a fusee lathe, or could be milled with a rotary axis on a vertical mill [17:20:27] <maddash> hm, can I default circularmovemsg.turn to 0, since helices are left out? [17:20:52] <maddash> why am I even asking? [17:22:50] <lerneaen_hydra> what I meant is when you do stuff like that you're not going to code by hand, you'll use a cam app [17:23:27] <lerneaen_hydra> and with with the exception of threading that can be done with multiple G1 movements [17:23:43] <lerneaen_hydra> as I take it the threading needs spindle syns [17:23:59] <lerneaen_hydra> *sync [17:24:10] <lerneaen_hydra> though that depends on the type of cutter of course [18:05:30] <Roguish> hey all, what's the 'best' cad program for linux open source? [18:06:00] <Roguish> something just to sketch with.... [18:06:08] <jepler> Roguish: I don't know of any really good ones. I have put together a few drawings for emc2 documentation with qcad, however [18:06:56] <Roguish> how 'bout : http://brlcad.org/ [18:06:59] <jepler> http://linuxcnc.org/docs/devel/html/hal/drivers/img4.png and http://linuxcnc.org/docs/devel/html/hal/drivers/img6.png are two of them -- however they've suffered in quality by the conversion from dxf -> eps -> png [18:07:03] <jepler> I haven't used brlcad [18:08:08] <SWPadnos> if you can get your head around it, brlcad can do everything. getting past the UI isn't easy - I haven't managed it yet [18:08:20] <SWPadnos> you can join #brlcad if you have questions [18:09:15] <archivist> the UI sucks absolutely [18:09:41] <jepler> "varkon - A CAD-system with parametric modelling [18:09:42] <jepler> " [18:09:53] <jepler> has anyone use this? I see it in 'apt-cache search' but I have never heard of it [18:10:01] <archivist> nor me [18:10:50] <SWPadnos> I did install it on a machine, but I'm not sure I was able to do anything with it [18:11:06] <SWPadnos> I don't remember any results :) [18:12:04] <jepler> must be pretty old -- http://www.tech.oru.se/cad/varkon/scrdmp4.htm -- looks like fvwm + tk or motif [18:12:22] <archivist> tis old 85-86 [18:12:48] <archivist> see cant do at bottom http://www.tech.oru.se/cad/varkon/whatis.htm [18:13:07] <anonimasu> heh..i [18:13:09] <anonimasu> it's horrid [18:14:16] <archivist> about the same cant do as brlcad [18:15:49] <fenn> varkon isnt exactly a cad program [18:16:08] <fenn> its more of a software library [18:17:03] <archivist> as brlcad is [18:17:12] <fenn> fair enough :) [18:17:48] <archivist> although brlcad is more of a solid modeller [18:18:02] <fenn> brlcad is more of a solid modeller than most anything out there [18:18:31] <fenn> parallelepiped [18:18:52] <fenn> i cant remember the wacky geometries you'll never use [18:18:55] <archivist> I lurk in there hoping for some sense in the UI and real world dimensioning being added [18:21:09] <archivist> a solidworks UI on top of brlcad would be heaven probably [18:22:34] <archivist> but as a solidworks user I just cant go back to the Ark [18:30:17] <Roguish> how about qcad? [18:31:43] <fenn> qcad is good for simple drafting [18:32:10] <Roguish> that's all i am looking for. got Proe for design. [18:33:58] <fenn> skencil might be worth playing with too [18:36:02] <Roguish> just put in qcad. looks good. thanks. [18:37:25] <fenn> wow skencil is really similar to inkscape [18:38:02] <bill2or3> * bill2or3 googles [18:38:56] <fenn> it has this 'blend' function which is really neat though [18:39:40] <Jymmm> Does EPS, but no mention of PS [18:41:00] <fenn> meh, they're similar enough [18:41:20] <Jymmm> close, but no cigar [18:41:38] <fenn> skencil outputs .ps not .eps, but ps2eps works [18:41:52] <Jymmm> it says EPS, not PS [18:42:22] <fenn> well i just saved a .ps file so i dont care what it says [18:42:34] <Jymmm> Exports Encapsulated PostScript [18:42:48] <Jymmm> look at the file in a text editor [18:43:12] <Jymmm> then pastebin it [18:43:15] <fenn> %!PS-Adobe-3.0 EPSF-3.0 [18:43:24] <jepler> eps is postscript with additional restrictions [18:44:13] <jepler> "encapsulated" means there are special comments that can be read by compatible software, and restrictions on the operations that can be done inside so that a program can embed it in a larger document without worrying about the details [18:44:43] <fenn> apparently ps is a rather powerful language [18:44:53] <Jymmm> fenn: yep [18:45:11] <Jymmm> fenn: actually, very cool lang [18:47:30] <Jymmm> fenn: This was created using PS lang with randomly generated patterns... http://farm1.static.flickr.com/43/105089849_07aeb24900.jpg [18:49:26] <Jymmm> And yes, there is a solution. It's not just a optical illusion [18:50:24] <fenn> reminds me of mac-plus days [18:51:20] <Jymmm> fenn: you should see the EXTREME puzzles I've created... [18:51:37] <fenn> you gonna cut that maze on your router? [18:52:41] <Jymmm> I haven't figured out how to cut the labyrinths, as they're layers. But I have of the mazes. [18:56:02] <fenn> http://www.dollins.org/mazes/rt-25-25-4-closeup2.png [18:56:38] <Jymmm> yep, like that [19:00:30] <fenn> he's got some nice celtic knots too [19:00:48] <fenn> or, something like that [19:04:00] <fenn> http://davidbau.com/downloads/pymaze.py [20:24:07] <David36_LA> Hi, does the TRAJ_PERIOD variable affect how many GCODE blocks per second EMC can process ? [20:25:44] <awallin> I think so [20:25:48] <cradek> are you having trouble? [20:26:08] <cradek> (a lot of things can affect that) [20:26:36] <David36_LA> No, no problem. Bu assuming all others factors are not an issue (machine speed, etc) [20:27:06] <David36_LA> I was curious if the EMC code that reads the gcode file deals with only one block per traj_period [20:27:24] <cradek> if the userland code (interpreter etc) can keep it fed, the realtime motion controller can use up to one motion per traj/servo cycle [20:27:31] <cradek> often that's 1000/sec [20:27:54] <cradek> whether the interpreter can keep it fed is a matter of your machine speed and how busy it is doing "other" things like running the GUI [20:28:10] <David36_LA> In most configs file I saw servo is often 1ms but traj is more like 10ms [20:28:34] <jepler> however, the realtime motion controller begins to give poor performance (lower velocity than requested) when an individual segment is not long enough for the accel/decel phase, which is almost always well before 1 segment per ms [20:28:46] <cradek> I don't actually think the traj cycle is used, the planner runs every servo cycle [20:29:15] <cradek> yes jepler is right, that's why I asked if you were having a problem - to figure out exactly what question you were asking [20:30:21] <David36_LA> ok, a customer doesn lots of 3D with short segments [20:30:44] <David36_LA> he is asking me if EMC2 is good at fast 3D work in wax and such [20:30:45] <cradek> with tiny segments, be sure to use tolerance mode (G64 Px.xx) [20:30:53] <David36_LA> on a mill [20:31:05] <cradek> with P = an appropriate machining tolerance [20:32:39] <awallin> cradek: sounds interesting if traj_cycle is never actually used !? shouldn't that be documented and traj_cycle removed from the ini? [20:32:41] <David36_LA> For example, If I run sprial.ngc from EMC2 files [20:33:30] <cradek> whether it's "good" at "fast" is hard to answer without details about what "good" and "fast" mean [20:33:37] <David36_LA> and set feed rate to 115 ipm (my mill can take it and I am not actualy cutting material), the curves seemed to move well bellow 115ipm [20:33:57] <David36_LA> it seems to slow down alot towards the center of the spiral [20:34:14] <cradek> do you have the machine's acceleration set as high as possible? [20:34:26] <cradek> making any curvy path requires acceleration, so that may be limiting the motion [20:34:34] <cradek> also, try G64 Px.xx (tolerance) like I suggested [20:34:58] <David36_LA> ok [20:35:00] <David36_LA> will try [20:35:15] <jepler> tolerance mode probably won't make much difference on the spiral until near the end [20:35:22] <cradek> right, the acceleration setting is the most important for making small moves and direction changes [20:36:54] <David36_LA> thanks [20:36:57] <jepler> For the emc2 realtime trajectory planner to contour well, each segment should be long enough to have a "cruise phase": assume the mill is stopped when you begin the segment, accelerates to the requested speed, "cruises" for a time at the top speed, then decelerates to stop exactly at the end -- trapezoidal velocity profile [20:37:18] <jepler> if the segment becomes short enough relative to the acceleration and velocity, the profile becomes triangular and there is no "cruise phase" [20:39:25] <awallin> jepler: but in G64 mode you start out with some velocity, and can end the segment also at non-zero velocity, so the full acc-ramps at the beginning and end should in principle not be required ? [20:40:02] <cradek> awallin: yes but that's how the algorithm works [20:40:15] <davidf> hi.. [20:40:20] <cradek> hi davidf [20:40:22] <jepler> awallin: it's true that a different method altogether would not necessarily have this particular limitation [20:40:26] <jepler> but the one that is currently in emc does [20:40:41] <davidf> cradek, Hey :) [20:41:41] <davidf> Just wanted to stop in and pass this on. Probably not anything new to you guys but just in case: [20:42:23] <davidf> I was using windows for ACE converter dxf to g-code, then putting on floppy. [20:42:51] <davidf> Then I tried installing wine, and ace on ubuntu. Works! :) [20:43:09] <cradek> nice, no windows license required! [20:43:29] <awallin> jepler, cradek: ok, I understand. coming up with an alternative algorithm is not trivial... but that's still something I'd like to learn about when I have time. The problem when I read about this last time is how you react to changes in feed-override. It seems that for better lookahead you have to sacrifice a bit of response-time to feed-override... [20:43:56] <davidf> But you just have to use todos to convert the resulting g-code file for linux linefeed style. [20:45:21] <cradek> actually I think EMC2 will load CRLF gcode without complaining [20:45:23] <davidf> Oops. That's not right. I meant to say you need to convert your linux style dxf file with todos so ace can read it properly. [20:45:34] <cradek> oh ok! that's different [20:45:47] <davidf> Yes. very. sorry. [20:46:17] <cradek> certainly no apology necessary [20:46:33] <davidf> Anyway, I am finding that on ubuntu, with QCad for making dxf and ace. there is really great possibility. [20:46:38] <cradek> I'm going to suggest wine/ace to a friend who is wanting to work with dxfs, thanks for letting us know [20:46:51] <cradek> right, that's exactly what he might want to do [20:47:06] <cradek> a gcode export in qcad would be better ... all free/open source software then [20:47:12] <cradek> but maybe this is a good stopgap [20:47:14] <davidf> Can't tell you what a relief it is not to have to work on two boxes anymore. [20:49:49] <davidf> On my emc machine, I'm still running ubuntu 5.1 or something like that. Can I install eme 2.1.6 without a linux upgrade to 6.06? [20:50:03] <davidf> emc 2.2.6 [20:50:07] <cradek> yes you can update just emc2 to 2.1.6 [20:50:28] <davidf> Great. There is a ubuntu binary, right? [20:50:31] <cradek> ubuntu have stopped supporting version 5.10 - you should consider updating soon [20:50:35] <cradek> yes [20:50:48] <ds2> if someone got the time, the ACE source code is available [20:51:36] <davidf> Yeah, but I have a real tiny 4 gig HD on that machine, and it is kinda slow too, so I thought I'd probably wait to upgrade the whole thing soon. [20:51:42] <cradek> ah [20:51:57] <cradek> (I think 4G is still enough for ubuntu 6.06) [20:52:17] <davidf> Yes, the ace download has all the source code as well as the windows exe. [20:52:31] <cradek> oh, interesting - is it a free license? [20:52:34] <davidf> cradek, Oh that's great. [20:52:44] <davidf> Yup! [20:52:59] <davidf> GPL I think. [20:53:07] <cradek> very interesting [20:53:10] <davidf> I'll check... [20:53:16] <cradek> should be pretty easy to build a native unix binary then [20:53:18] <ds2> it is not a straight forward compile on LInux, unfortunately [20:53:23] <cradek> er, possible, maybe not easy [20:54:18] <davidf> Yes, it is GPL. [20:55:30] <davidf> I had a look at the source yesterday. Looks doable, but ace as is works great anyway. Does prioritized layers, arcs, optimization, etc. [20:56:50] <archivist> hmm looks interesting [20:56:58] <davidf> I was thinking of just trying to use the source as a guide to rewrite it for linux, then thought, duh. What if it will work with wine? Glad I tried it! [20:58:55] <davidf> cradek, You remember that old bug that gives an axis following errer sometimes, and you have to restart emc or wait a while? Is that problem totally fixed now? [21:01:40] <davidf> Also, I'm using emc to cut flutes on the reamer bits I make. Sometimes I'll be too deep by a couple thou, out of the blue. I think I saw a ref to some kind of 'overshoot' bug in older versions. Could that be it? Haven't been able to identify what does it. It is intermittent. [21:02:14] <cradek> davidf: I'm not familiar with either of these bugs... I suspect you have some kind of misconfiguration if you are having random problems [21:02:49] <cradek> if you are using servos, you may not have them tuned very well [21:03:31] <cradek> I have to run ... be back later [21:03:59] <davidf> Well, I remember bringing it up many moons ago, and either you or jepler I think, said 'someone was supposed to have fixed that, but then you realized he hadn't. I think swapdnos was in that chat as well. It had something to do with accel I think. [21:04:05] <davidf> OK. Bye. [21:04:37] <davidf> I'm suing steppers btw. [21:04:55] <davidf> ha. using steppers, not suing. [21:05:42] <davidf> Anyway, Thanks for the info. hope the ace thing helps some folks. I really like it. Bye... [21:09:33] <ds2> wow suing... hehe [21:33:48] <Roguish> hey, what's the latest status on M5i20 fpga code? 2.2.0? [22:08:01] <edmoore> hi chaps [22:08:22] <edmoore> completely newb question that would be better served be reading the website most probably, but, can emc do 5 axis cnc work? [22:08:29] <SWPadnos> yes [22:09:23] <SWPadnos> you'll need an appropriate CAM program to generate 5-axis G-code that will work with your machine [22:10:14] <edmoore> yes sure [22:10:29] <edmoore> i appreciate that's probably more likely to be the week link for the hobbyist [22:10:38] <edmoore> weak* [22:11:17] <anonimasu> SWPadnos: does that mean 5 axis interpolated work? [22:13:01] <floppy1234> hello [22:17:32] <SWPadnos> anonimasu, XYZABC (and now +UVW) interpolation is pretty well described in the manual [22:17:58] <SWPadnos> you can't do tooltip velocity with ABC included, or if you use any of XYZ along with any of UVW [22:18:15] <anonimasu> SWPadnos: I havent looked for quite some time.. I dont care as I dont have a 6 axis mill.. [22:18:43] <anonimasu> im not even sure about uvw [22:19:26] <floppy1234> i want to control a pneumatic system with emc. i've seen "etch-servo" which configures a closed loop system. how fast is this way to control the servopneumatic linear axies? [22:19:45] <anonimasu> floppy1234: are you building a new machine? [22:19:52] <SWPadnos> I think UVW are treated the same as ABC if there is any XYZ motion (ie, the UVW move completes in the time taken for XYZ) [22:19:57] <floppy1234> anonimasu yes [22:20:03] <anonimasu> floppy1234: dont go there please.. [22:20:08] <SWPadnos> and if there's no XYZ, then UVW are trerated as the cartesian group for tooltip velocity [22:20:10] <anonimasu> r:) [22:20:10] <floppy1234> anonimasu why? [22:20:15] <SWPadnos> treated [22:20:34] <anonimasu> floppy1234: because pneumatics arent too great at positioning [22:20:52] <floppy1234> anonimasu i have the axies [22:21:15] <floppy1234> anonimasu and they should cut with a hot wire [22:21:34] <floppy1234> do you know "styropor"? [22:21:49] <floppy1234> it is a light wight foam [22:22:27] <anonimasu> well, pneumatics have stiction and friction that makes them bad for positioning [22:23:07] <floppy1234> i drive with controled exit air [22:23:16] <floppy1234> so it works.. [22:23:28] <floppy1234> i mean, it is not the best way to positioning [22:23:51] <anonimasu> it's probably one of the worst ones.. [22:23:56] <anonimasu> but sure you could use emc to drive it.. [22:24:01] <floppy1234> but conside: the axies are a gift ;) [22:24:24] <floppy1234> and we want to drive 1mm or so [22:24:49] <anonimasu> that small? [22:24:51] <floppy1234> the incremental encoder's resolution is 5µm [22:25:16] <floppy1234> i mean we need a accuracy about 1mm [22:25:26] <anonimasu> I think that's hard to acheive.. [22:25:29] <floppy1234> our driveway is about 1200mm [22:26:25] <floppy1234> well, i think, if we can drive with a accuracy about 5mm it is also okay [22:27:08] <floppy1234> but i know the axies can be controlled more exacting [22:29:09] <floppy1234> hm... if the pneumatic control doesnt work, i changeover to spindles, but i first try with the existing systenm [22:29:12] <floppy1234> system [22:29:28] <skunkworks> what are you using for a valve? [22:29:34] <anonimasu> you probably need some hardware for controlling your valves.. [22:29:46] <floppy1234> hmmm [22:29:54] <floppy1234> the valves are from bosch rexroth [22:30:20] <floppy1234> they can be controlled between 0 to 100mA and let flow 0 to 6 bar [22:30:29] <skunkworks> I think it would be a cool experiment. :) [22:30:58] <floppy1234> so i made a operationamplifier circuit which control with 0 to 5V the current from 0 to 100mA [22:31:36] <floppy1234> this ciruit works... i have a adjustable power suply and can control very good [22:31:47] <floppy1234> wanna see some pictures? [22:32:20] <skunkworks> sure. what are your plans to interface emc to your control? [22:32:30] <floppy1234> http://www.rsp-design.de/plotter/Foto113.jpg [22:32:58] <floppy1234> skunkworks first i want to drive with PWM like the EMC's demofile "etch-servo" [22:33:07] <skunkworks> ah - cool [22:33:31] <floppy1234> if this is to slow, i wanna change to a microcontroller for each axis [22:33:39] <floppy1234> skunkworks what is cool? ;) [22:35:26] <skunkworks> using pwm out of the printer port - and the slides. you said the encoders res is .000005M? or or .005mm? [22:35:44] <floppy1234> the res is 5µm [22:35:57] <floppy1234> also 0.005mm [22:36:22] <skunkworks> that may be pretty hard to count with the printer port.. speed wise [22:36:39] <floppy1234> i think it will go very sloooooow [22:36:48] <floppy1234> but it must drive ;) [22:38:30] <skunkworks> The pluto would be a cool/cheap 4 axis servo controler you might want to look at. 4pwm out 4 encoder in and some i/o [22:38:50] <floppy1234> hmmm [22:39:06] <floppy1234> exists there a open source device? [22:39:09] <skunkworks> yes [22:39:39] <floppy1234> is this the controller called "pluto"? [22:40:17] <skunkworks> http://emergent.unpy.net/01166412010 [22:40:39] <skunkworks> if you want more power - the mesa card is open source also. [22:41:09] <skunkworks> the pluto is printer port - the mesa is pci [22:42:02] <floppy1234> well... 60 for the plutoboard isnt expensive
[22:42:15] <floppy1234> i think for this price i dont build a board
[22:44:41] <floppy1234> i cant find something about the mesa card
[22:44:53] <anonimasu> did anyone ever try making anything with fluid/pneumatic bearings+
[22:44:56] <anonimasu> ?
[22:45:37] <floppy1234> no
[22:45:41] <anonimasu> probably not..
[22:45:49] <anonimasu> :)
[22:45:50] <floppy1234> it is my first project ;)
[22:46:44] <anonimasu> I need a grinding attachment for my lathe
[22:47:31] <floppy1234> uuuuh
[22:47:39] <ds2> isn't a foredome handle in a boring bar holder equivalent?
[22:47:57] <anonimasu> ds2: to a grinder?
[22:47:58] <anonimasu> heh
[22:47:58] <skunkworks> http://www.mesanet.com/motioncardinfo.html
[22:47:59] <floppy1234> whar is a foredom?
[22:48:05] <anonimasu> or what do you mean
[22:48:10] <skunkworks> 5I20
[22:48:15] <ds2> yes, to a grinding attachment
[22:48:27] <anonimasu> I dont know what a "foredom"
[22:48:28] <anonimasu> is
[22:48:41] <ds2> think big dremel on flex shaft
[22:48:49] <anonimasu> hm, _flex_ shaft.. no
[22:48:50] <anonimasu> not really
[22:49:15] <anonimasu> http://www.golmatic.de/GOLmatic_EN/zubehoer_en/supportschleifer_en.htm
[22:49:41] <floppy1234> anonimasu are u german?
[22:49:43] <anonimasu> no
[22:49:51] <floppy1234> okay ;)
[22:50:08] <floppy1234> is there a german emc chat?
[22:50:17] <floppy1234> cause my english is very bad :p0
[22:50:18] <floppy1234> :p
[22:50:24] <ds2> http://www.fandfwoodcarving.com/images/TX&HP.jpg
[22:50:26] <ds2> much better picture
[22:50:39] <floppy1234> i tortue "dict.leo.org" all the time :p
[22:50:44] <anonimasu> _woodcarving_
[22:50:58] <anonimasu> ds2: I want to grind steel stuff.
[22:51:02] <ds2> yes, but you can put a mini grinding stone on it
[22:51:38] <floppy1234> hmmm
[22:51:56] <floppy1234> ive yet another question
[22:52:06] <archivist> anonimasu, thats lathe abuse, but Ive done it with a dremel as well
[22:52:12] <ds2> like this guy -
[22:52:13] <ds2> http://www.5bears.com/tshaft_1.htm
[22:52:17] <ds2> see bottom of the page
[22:52:40] <anonimasu> he crude homemade toolpost grinder shown in Part 1 was a failure. The bearings in the Foredom handle were simply not precise enough, and the finish was poor. So I scrapped the grinder concept and went forward with the traditional home-shop method of producing the shaft.
[22:52:51] <floppy1234> what is this like? can i display a 4 axis hot wire cut machine?
[22:53:07] <ds2> DOH
[22:55:48] <anonimasu> ds2: I want a spindle that'll go to 10krpm..
[22:55:50] <anonimasu> ;)
[22:55:58] <anonimasu> ds2: hence fluid bearings and stuff..
[22:56:44] <ds2> just did :/
[22:56:48] <floppy1234> so... i open a german beer ;)
[22:57:09] <floppy1234> are you all amarican?
[22:57:16] <anonimasu> no
[22:57:18] <anonimasu> im swedish
[22:57:35] <floppy1234> archivist is german?
[22:57:41] <archivist> no
[22:58:32] <floppy1234> hmm i think so cause you are in mysql.de
[22:59:28] <archivist> no I run the bot for mysql mysqlde and mysqles and thats why im there
[23:00:00] <floppy1234> mmm kay :)
[23:00:02] <anonimasu> http://pergatory.mit.edu/rcortesi/portf/spindle/mighty_spindlel_files/frame.htm
[23:00:07] <archivist> Im english
[23:00:25] <floppy1234> pity...
[23:00:56] <floppy1234> so i must break down in english :p
[23:02:50] <ds2> * ds2 refrains from complaining about machines using a mix of whitworth and metric screws
[23:03:15] <anonimasu> :p
[23:03:27] <archivist> its worse here, we still use BA as well
[23:03:35] <floppy1234> ds2 i think everytime so, when i must repair my car
[23:03:53] <ds2> serious? thought you guys outlawed anything nonmetric
[23:05:07] <anonimasu> ds2: http://pergatory.mit.edu/perg/awards/TURBOTOO.html
[23:05:10] <anonimasu> why does that make me drool :p
[23:05:48] <ds2> looks expensive to make ;)
[23:05:52] <floppy1234> yeah
[23:06:19] <floppy1234> ds2 looks depressing when it breaks ;)
[23:06:31] <anonimasu> :)
[23:06:44] <ds2> should ask for quotes at the local shops to scare them ;)
[23:06:51] <anonimasu> haha
[23:06:53] <archivist> ds2 actually BA is a british standard that used metric dimensions
[23:06:56] <floppy1234> hehe
[23:06:57] <anonimasu> 200 000rpm..
[23:07:07] <floppy1234> anonimasu so slow? *G*
[23:07:16] <anonimasu> or 100 000 rpm and 100kw
[23:07:17] <anonimasu> :p
[23:07:34] <anonimasu> have you ever seen a 100kw cut being made?
[23:07:40] <ds2> archivist: Hmmmm I am looking at a machine from the 80's and the plans call out something like 1/4-20BA so...
[23:07:49] <floppy1234> uuh.. there is my grandma faster with a gimlet
[23:07:50] <floppy1234> *GGG*
[23:07:57] <anonimasu> I'd love to see that :p
[23:08:27] <ds2> it doesn't look that horrible now that i study it a bit
[23:08:43] <anonimasu> :)
[23:08:45] <floppy1234> anonimasu i belive ;)
[23:08:47] <ds2> 4 axis mill can do most of it
[23:08:53] <anonimasu> yep
[23:09:02] <anonimasu> you need to grind the outer surface..
[23:09:08] <anonimasu> that's the complex part..
[23:09:11] <anonimasu> and the housing..
[23:09:11] <ds2> Oh
[23:09:13] <anonimasu> I guess..
[23:09:20] <anonimasu> hone it or something
[23:09:34] <anonimasu> and ofcourse it needs to be balanced..
[23:09:35] <floppy1234> hm mm
[23:09:52] <ds2> wonder if you put a die grinder stone in a collet and have the same CNC machine grind it
[23:09:59] <ds2> (assuming you have a 10K+ spindle)
[23:10:07] <anonimasu> no
[23:10:14] <anonimasu> I dont think you do :)
[23:10:25] <anonimasu> or do you mean to sharpen the flutes?
[23:10:29] <floppy1234> well my question again: can i show the 4 axis erode machine on my screen?
[23:10:41] <anonimasu> yes
[23:10:42] <ds2> 10K CNC spindles are an option on the Haas
[23:11:26] <anonimasu> dosent haas have everything as option?
[23:11:33] <ds2> heh
[23:11:39] <anonimasu> though they are cheap..
[23:11:41] <ds2> I know the tiny taig has a 10K spindle default
[23:11:49] <anonimasu> matsuura has 40krpm standard..
[23:11:53] <anonimasu> and 60k as option
[23:12:07] <anonimasu> or well 8-15krpm for other machines :)
[23:12:19] <ds2> nice, and how thick is the shelding plastic/glass? =)
[23:12:19] <floppy1234> anonimasu you mean yes to my question?
[23:12:28] <anonimasu> floppy1234: Yes
[23:12:36] <floppy1234> anonimasu how can i do?
[23:12:37] <ds2> a 1/8" bit breaking on a 60K spindle can hurt!
[23:12:42] <anonimasu> floppy1234: you do know you only seewhere the machines are going..
[23:12:47] <anonimasu> and not the actual machine
[23:12:49] <anonimasu> just the toolpath..
[23:12:57] <floppy1234> yes i konw
[23:13:01] <floppy1234> know
[23:13:02] <floppy1234> thats i wantr
[23:13:04] <floppy1234> want
[23:13:12] <anonimasu> how does that differ from a 3 axis mill?
[23:13:20] <anonimasu> :)
[23:13:38] <floppy1234> hmmm
[23:13:41] <anonimasu> except one axis less :)
[23:13:49] <anonimasu> ds2: yep
[23:13:53] <floppy1234> ok...
[23:14:03] <floppy1234> i think about when the machine run
[23:15:11] <anonimasu> ds2: I'd love to try that kind of setup
[23:15:25] <ds2> anonimasu: engraving or grinding?
[23:15:30] <anonimasu> making a tool like that
[23:15:33] <floppy1234> i wanna se the machine run ;)
[23:15:55] <floppy1234> how fast is the parallelport?
[23:15:57] <anonimasu> floppy1234: I dont get it.. you can see how it moves in realtime..
[23:16:02] <anonimasu> with axis :)
[23:21:30] <anonimasu> ds2: milling
[23:22:35] <floppy1234> kk
[23:24:09] <anonimasu> ds2: tell me 100krpm isnt a cool thing ;)
[23:27:15] <archivist> pcb drilling machines work at insane speeds
[23:27:44] <anonimasu> archivist: they dont cut aluminium..
[23:27:49] <anonimasu> :)
[23:27:51] <floppy1234> hmmm
[23:27:51] <archivist> I know
[23:28:53] <floppy1234> which expirience do you have with threaded bars from the do-it-yourself store to position a machine?
[23:29:43] <archivist> they are rolled threads to a price dont expect accuracy
[23:30:54] <floppy1234> hmmmm
[23:31:11] <floppy1234> you mean, i should not take them?
[23:31:22] <floppy1234> i will build a seccond machine
[23:31:39] <floppy1234> to drill pcbs and mill little things
[23:31:46] <anonimasu> they work.
[23:31:53] <floppy1234> okay
[23:32:19] <anonimasu> the bottom line is how mich are you willing to pay for accuracy
[23:33:49] <floppy1234> well... i think special spindles give not more accuracy but hey durable
[23:33:53] <floppy1234> or not?
[23:34:07] <archivist> has EMC got leadscrew compensation
[23:34:08] <anonimasu> floppy1234: 01:31 < floppy1234> which expirience do you have with threaded bars from the do-it-yourself store to position a machine?
[23:34:11] <anonimasu> 01:32 < archivist> they are rolled threads to a price dont expect accuracy
[23:34:50] <floppy1234> oh... i dont understand right, sorry
[23:36:38] <floppy1234> hmmm you mean, i cant await accuracy when i dont buy recirculating ballscrew?
[23:37:43] <anonimasu> floppy1234: yes.
[23:38:00] <floppy1234> okay
[23:38:02] <anonimasu> floppy1234: "How much should accuracy cost?" do you get that?
[23:38:48] <archivist> accuracy depends on quality, or it can be calibrated or movement can be measured so it does not matter so much
[23:38:55] <floppy1234> but is the gain of accuracy drastic for a homemade lowcost pcb mill?
[23:39:04] <archivist> no
[23:39:05] <anonimasu> no
[23:39:06] <floppy1234> okay
[23:39:10] <floppy1234> thank you
[23:39:15] <floppy1234> thats all i wanna kow
[23:39:21] <floppy1234> but stop....
[23:39:24] <floppy1234> not all :p
[23:39:45] <floppy1234> how is it about the galling?
[23:40:12] <anonimasu> the more stuff I do with cnc's the more I realize that what I want is a machine I can depend on to work, and not break..
[23:40:50] <archivist> yes, the toy here is not reliable
[23:41:14] <anonimasu> I hate fscking around when I need to facemill something
[23:41:34] <floppy1234> well
[23:41:36] <floppy1234> i go to bed
[23:41:47] <floppy1234> when i have a problem tomorrow, ill come back...
[23:42:06] <floppy1234> otherwise ill come also back
[23:42:28] <toast> fff
[23:42:29] <archivist> face milling here is on a turn of the century machine 1899-1900
[23:44:43] <anonimasu> :)
[23:46:40] <floppy1234> good night
[23:47:44] <archivist> www.archivist.info/skeleton
[23:48:34] <archivist> down in the basement
[23:53:18] <Ziegler> is that machine belt driven?
[23:53:29] <Ziegler> the spindle?
[23:53:29] <archivist> yes
[23:53:32] <Ziegler> cool
[23:53:48] <archivist> we think pre ww1
[23:53:57] <Ziegler> bet that has been fun to play with
[23:54:21] <archivist> and googling just now confirms that sort of date for the company
[23:54:31] <Ziegler> what company?
[23:54:37] <archivist> its a bit odd to use
[23:54:54] <archivist> Garvin machine co new york
[23:55:10] <Ziegler> ah... are the ways in good shape?