#emc | Logs for 2008-02-12

[00:04:18] <fenn> oo i like that - detect torque via differential measurement of shaft position
[00:25:43] <BigJohnT> anyone know of an ADC chip/circuit that I can use to input binary into the parallel port?
[00:26:34] <BigJohnT> for about $225 I can get a PLC with an analog input card and output binary to the parallel port
[00:27:24] <BigJohnT> any electronic wizz's about
[00:30:26] <SWPadnos> digikey has many choices for parallel output ADCs
[00:31:07] <BigJohnT> so many that it's mind boggling...
[00:31:53] <BigJohnT> I should have a digikey and mouser catalog around here somewhere
[00:32:49] <BigJohnT> 10,721 items in the digikey catalog under ADC...
[00:33:54] <SWPadnos> heh
[00:34:29] <SWPadnos> how many bits are you looking for, and at what conversion speed / update rate?
[00:34:54] <SWPadnos> incidentally, if you check "in stock", the list is drastically reduced
[00:35:22] <BigJohnT> hmmm, I want to convert my arc voltage signal to I'm guessing 8 bits
[00:35:42] <SWPadnos> ok, the ADC is the least of your worries then, I think
[00:35:50] <BigJohnT> I know that I have to use a voltage divider circuit to scale it down to 5 or 10 vdc max
[00:35:55] <SWPadnos> isolation and level shifting will be much more cahllenging, as well as noise reduction
[00:36:02] <SWPadnos> challenging
[00:37:00] <BigJohnT> the source is not from the arc circuit but from a arc signal circuit from the controller board
[00:37:09] <SWPadnos> ok, that's a good thing :)
[00:37:20] <SWPadnos> do you know how clean it is, or whether you need isolation?
[00:37:22] <BigJohnT> would it not be kinda not noisy?
[00:37:37] <BigJohnT> not really, I don't have a scope anymore
[00:37:54] <SWPadnos> not necessarily, if it's near a high voltage/high power switching supply for the plasma
[00:38:12] <BigJohnT> it is designed to plug into a torch height control
[00:38:21] <SWPadnos> what kind of sample rate are you looking for?
[00:38:55] <BigJohnT> I have not done the math on that but fast enough to update the torch height while cutting
[00:39:21] <SWPadnos> I'm not sure that you'll want to do that in software.
[00:39:25] <SWPadnos> I don't know though
[00:39:51] <BigJohnT> why is that?
[00:40:02] <SWPadnos> response time
[00:40:11] <SWPadnos> I don't know how touchy a torch controller has to be
[00:40:33] <BigJohnT> what I would be adding is one thing to Dallur's hal file
[00:40:50] <SWPadnos> I'm betting faster than 1ms though, and you'll have to do some testing if you want to use FP in the base thread
[00:41:06] <SWPadnos> (due to latency, not any specific technical problem)
[00:41:19] <BigJohnT> instead of up down, I would input a binary number
[00:42:17] <BigJohnT> what is FP?
[00:42:36] <SWPadnos> I may be overthinking it - I just came back from a customer who needed an extra HAL parameter and error checking for "load resistance too low" :)
[00:42:39] <SWPadnos> floating point
[00:42:54] <BigJohnT> ok
[00:43:31] <BigJohnT> in Dallur's case he is doing everything but comparing the input number to a preset number
[00:43:44] <SWPadnos> ok
[00:44:13] <BigJohnT> using pyVCP to input a voltage number then compare that to the binary number on port 1 and adjust up or down as needed
[00:44:19] <SWPadnos> if it's more or less about keeping the plasma a certain distance above the wrok, and it's controlled by a Z motor, then it could work
[00:44:24] <BigJohnT> that is my thought process at the moment
[00:44:35] <BigJohnT> yes, that is exactly it
[00:44:38] <SWPadnos> pyvcp isn't even close to realtime, it all runs in user space
[00:44:48] <SWPadnos> think about that a bit before continuing :)
[00:45:11] <BigJohnT> I just set the preset in pyvcp
[00:45:17] <SWPadnos> ok, phew
[00:45:28] <BigJohnT> hal does the compare
[00:45:51] <BigJohnT> I've actually got all the hal part worked out and tested
[00:46:06] <BigJohnT> somewhat
[00:46:27] <SWPadnos> heh
[00:47:11] <BigJohnT> my next step as I see it is to use my lab power supply to feed an ADC circuit to test that part but I need to come up with an ADC circuit/chip
[00:48:07] <SWPadnos> I think the next step is to find out what kind of interface and filtering might be necessary to connect to that output
[00:48:31] <BigJohnT> the output of the plasma torch?
[00:48:46] <SWPadnos> if you see that you need isolation, then you may need to use a serial A/D or a microcontroller
[00:48:48] <SWPadnos> yes
[00:49:03] <SWPadnos> a THC may be expected to have plenty of protection on its input - I don't know for sure
[00:49:31] <BigJohnT> how can I find out what I need?
[00:49:49] <SWPadnos> ideally, from manufacturer spec sheets
[00:50:01] <SWPadnos> for the plasma and /or THC products
[00:50:03] <BigJohnT> do I need to scrounge a scope?
[00:50:06] <SWPadnos> you might
[00:50:28] <SWPadnos> a meter might also do it, or looking at the components that drive the output
[00:50:32] <BigJohnT> I didn't see any thing on my plasma but I'll look at some of the THC products and see what I can dig up
[00:50:52] <SWPadnos> isolated analog isn't too easy, so I'm suspicious that the output isn't isolated
[00:51:17] <SWPadnos> step 1: define the problem. step 2: figure out the solution :)
[00:51:29] <SWPadnos> step 3: profit!
[00:51:50] <BigJohnT> profit what is that?
[00:52:05] <SWPadnos> dunno - I keep forgetting that one
[00:53:16] <BigJohnT> so worst case I have to buy a PLC and use that to convert from analog to digital
[00:53:35] <BigJohnT> that I know I can make work
[00:54:02] <SWPadnos> heh - that would be the worst case, I think
[00:54:03] <BigJohnT> but if I only need $50 worth of chips and parts I'd rather do it that way
[00:54:08] <fenn> SWPadnos: i'm having trouble going from 2 to 3
[00:54:21] <SWPadnos> $50 should be enough, even if you need isolation
[00:54:34] <SWPadnos> fenn, yeah, I know the problem. haven't figured out a good way around it though
[00:56:12] <BigJohnT> on the emc side I'll use wsum to convert the binary input from the parallel port to an integer
[00:56:30] <SWPadnos> yep
[00:57:08] <BigJohnT> then convert that to a float and do a comp on the setpoint vs the actual
[00:57:55] <BigJohnT> * BigJohnT hears the dinner bell ringing and I'm never late for dinner bbl
[00:58:00] <SWPadnos> heh
[01:11:52] <BigJohnT> mmmm, chicken n dumplings perfect on a nasty winter night
[04:15:58] <gezar> i totally do not understand the chain rule, derivatives of exponents, derivatives of e and log functions sigh
[04:30:22] <tomp> SWPadnos: do you design ECM generators/controls?
[04:30:45] <SWPadnos> nope
[04:30:53] <SWPadnos> well, sort of, indirectly
[04:31:16] <SWPadnos> but not the hardware or "design" - I implement the software to do whatever the deisgners think they need
[04:31:59] <tomp> ok, i was looking for an opinion on pulsed dc current sensors
[04:32:25] <SWPadnos> hmmm. I'm not sure what they used
[04:32:38] <SWPadnos> the scale may be off by a few orders of magnitude as well ;)
[04:32:47] <tomp> My application is 28-32Vdc 300Apeak. Max cycle 300uS 'on' (discharge) with 50uS 'off' (pause), so near 3Khz. Shunt voltages too noisy, so i looked at http://www.vektrex.com/Products/sensors.html. I want to stop the process when a threshold is exceeded.
[04:33:01] <tomp> )i had that ready for someone to lok at ;)
[04:33:04] <SWPadnos> heh
[04:33:20] <SWPadnos> ok, only ~2 orders of magnitude low ;)
[04:34:27] <SWPadnos> this is what was used, I don't know what brand: http://en.wikipedia.org/wiki/Rogowski_coil
[04:36:31] <tomp> thx, new to me ( was looking in the magnetic sensor area)
[04:37:04] <SWPadnos> yeah, the name was new to me as well, though there was a paper about them from about 1912 (according to wikipedia)
[04:38:36] <SWPadnos> google told me about this company: http://www.pemuk.com/index.html
[04:41:39] <tomp> yeh, that flexiprobe looks expensive ( hp always were )
[04:41:46] <SWPadnos> heh
[04:42:32] <tomp> that advert moved away from any dc claims ( most CTs do )
[04:50:10] <jmkasunich> tomp: have you looked at hall effect sensors?
[04:50:30] <tomp> these are hall http://www.vektrex.com/Products/sensors.html
[04:50:32] <tomp> thx
[04:51:07] <tomp> and they seem to be the magnitude needed, a laser stack user quoted success at 32V 300A
[04:52:02] <jmkasunich> http://web4.lem.com/docs/products/LA%20255-S%20E.pdf
[04:52:09] <tomp> i'd hope that a comparator with a setpoint would handle this ( a stop gap bandaid till the generator design is fixed )
[04:53:01] <jmkasunich> the one I posted is one of many - list of product ranges they make: http://www.lemusa.com/hq/en/content/view/269/206/
[04:53:26] <tomp> thx, damn fast, the 32ohms worries me, will look thru thier offerings :)
[04:53:53] <jmkasunich> 32 ohms?
[04:56:14] <tomp> oops, that was secondary coil, not seen by process
[04:57:01] <jmkasunich> LF205-S looks like it would work too - cheaper
[04:57:26] <jmkasunich> we use LEM sensors in drives, I think I've used the LF205 a while ago
[04:58:02] <jmkasunich> we mostly use LF505, LF1005 and LF2005 for the drives I've been doing lately
[05:15:11] <tomp> thanks, those look good, i just gotta find one with a bigger hole :) 4 flat cables gotta pass thru. looks great
[05:15:35] <tomp> hmm, no , can use a busbar , even better
[05:16:13] <maddash> the surfaces that I mill come out horribly rough -- is there anyway to fix this automatically?
[05:17:52] <toastydeath> rough how
[05:17:57] <toastydeath> do you have pictures
[05:18:40] <maddash> darn, should've thought of taking pics
[05:19:09] <toastydeath> can you describe the roughness
[05:19:43] <toastydeath> does it look torn, or pulled apart?
[05:19:49] <toastydeath> scalloped?
[05:19:56] <maddash> I have a cylindrically shaped mill tool, whose bottom is flat, and I use the tool to carve out a small rounded rectangle well (pit) in a block of acrylic
[05:20:09] <maddash> scalloped may be the right word
[05:20:28] <toastydeath> what's your feed, speed, and size of the endmill
[05:20:29] <SWPadnos> unless you're really careful, acrylic will always look like shit
[05:20:44] <maddash> because at the bottom of the well that I mill, I see circular lines indicating the path that the tool bottom took
[05:20:58] <toastydeath> you will always see those
[05:21:00] <SWPadnos> your millhead may not be square (trammed)
[05:21:21] <toastydeath> you can slow your feed down a lot, and speed your RPM up
[05:21:32] <toastydeath> but i doubt you're going to eliminate all the tool marks
[05:21:49] <toastydeath> lube helps
[05:21:50] <toastydeath> a lot
[05:21:53] <maddash> so in the world of milling, one will always see the tool marks?
[05:21:57] <toastydeath> yep
[05:22:18] <maddash> SWPadnos: 'trammed' meaning a tool defect?
[05:22:31] <toastydeath> trammed meaning head perpendicular to the table
[05:22:35] <SWPadnos> no, "trammed" meaning that the spindle is perfectly perpendicular to the table movement
[05:22:36] <tomp> squareness of spindle axis to x y travel
[05:22:57] <toastydeath> if you have some kerosene you can brush some of that on
[05:22:59] <toastydeath> that will help
[05:23:03] <toastydeath> wd-40 also works
[05:23:22] <tomp> mount and indicator in the spindle, swing over table, adjust till reading same n/s/e/w
[05:23:29] <tomp> and/an
[05:23:30] <maddash> are there specific materials that those two work on? or can I use this on acrylic?
[05:23:43] <toastydeath> give it a test on a scrap piece to make sure it doesn't discolor
[05:23:45] <SWPadnos> you can tell when the machine is more or less trammed corerctly when you have about the same number of circular marks in one direction as the other (ie, there's no real bias toward marks from the right side of the tool vs the left)
[05:23:46] <toastydeath> but it should work perfect
[05:23:59] <toastydeath> swpadnos: he wants the tool marks to go away
[05:24:17] <SWPadnos> understood - tramming is important for that - you have gouges with really bad tramming
[05:24:19] <toastydeath> tramming it will make it worse
[05:24:26] <toastydeath> unless it's way out
[05:24:33] <SWPadnos> err - not in my limited experience :)
[05:24:40] <toastydeath> in my less limited experience it does
[05:24:45] <SWPadnos> heh
[05:24:47] <toastydeath> the back drag tears the material and leaves a crappy finish
[05:24:48] <tomp> http://www.jjjtrain.com/vms/mill_movements_head/mill_movement_head_00.html
[05:25:00] <toastydeath> if you bump the tram out just a bit, .002" or so
[05:25:08] <toastydeath> leaning into the cutting direction
[05:25:11] <SWPadnos> hmmm. maybe I always use perfect machining practice, so I don't have that problem
[05:25:12] <toastydeath> you will get a better finish
[05:25:15] <SWPadnos> naaaahhhh :)
[05:25:54] <toastydeath> but wd-40 or some other similar thing will work great
[05:25:57] <maddash> SWPadnos: meaning that you never have to polish up the part that comes out of the mill?
[05:26:03] <toastydeath> and it should work for acrylic
[05:26:09] <SWPadnos> no, I still have marks :)
[05:26:11] <toastydeath> a very sharp hss tool
[05:26:22] <maddash> 'hss'?
[05:26:23] <toastydeath> i can make a part on a mill where you can't see the marks
[05:26:29] <toastydeath> hss, high speed steel
[05:26:32] <toastydeath> versus a carbide tool
[05:26:45] <maddash> maybe acrylic is a bad choice of material
[05:26:57] <toastydeath> well, not really
[05:27:07] <toastydeath> anything you cut is going to have tool marks
[05:27:11] <toastydeath> especially in a pocketing operation
[05:27:28] <toastydeath> the only chance you ever have on a mill to remove obvious tool marks is on a flycut
[05:27:39] <toastydeath> and that only does large surfaces
[05:28:36] <toastydeath> the sharpness of your tool, the speed it spins at, how slow you feed it, and how lubricated it is
[05:28:43] <toastydeath> are going to be major factors in your control over surface finish
[05:29:00] <toastydeath> the spindle bearings are something that also has a very large contribution but isn't in your control
[05:29:07] <ds2> can't see as in producing an optically clear surface in acrylic?
[05:29:07] <maddash> the last three I can control, but I'm not sure about the first
[05:29:18] <toastydeath> maddash: use a new tool
[05:29:21] <maddash> ds2: if it's possible, yes
[05:29:33] <toastydeath> you aren't going to get an optical surface from milling it
[05:29:58] <maddash> hm, what process would be required, then?
[05:30:04] <toastydeath> polishing the darn thing
[05:30:09] <Jymmmm> mill a mold, and use cast acrylic
[05:30:20] <toastydeath> you'd have to polish the mold
[05:30:30] <Jymmmm> not really
[05:30:31] <ds2> how would you polish the inside of a pocket?
[05:30:38] <maddash> does that mean using sandpaper? ugh.
[05:30:43] <tomp> maddash: try skim cutting a piece of junk material, got scallops? you'll have same in pocket bottom, problem seen on skim is not depth related, check head square to table motion
[05:30:46] <toastydeath> ds2: carefully, with sandpaper/very fine grain files
[05:30:53] <toastydeath> riffler/die sinking files
[05:30:59] <tomp> torch
[05:31:04] <ds2> oh Hmmm
[05:31:09] <Jymmmm> y'all are assuming the mold is out of metal
[05:31:29] <ds2> Jymmm: silicone molds work on acrylic?
[05:31:30] <toastydeath> i am assuming he's casting plastic and it's going to take on whatever surface finish the mold has no matter what material?
[05:32:32] <toastydeath> unless you've got an idea to create a mold with a great finish from the get go, which i would not pee on insofar as ideas go
[05:32:34] <maddash> Jymmmm: omg, "cast acrylic"? that's waaaaaay out of my league
[05:32:47] <toastydeath> i just don't know any methods to produce an optically smooth mold without polishing
[05:32:49] <toastydeath> in any material
[05:33:14] <toastydeath> i'd get a bunch of sandpaper and a lot of wd-40
[05:33:16] <toastydeath> and wet sand it.
[05:33:21] <maddash> maybe I could rig my mill to do the sanding too
[05:33:23] <toastydeath> and eventually work my way up to polishing cloth/diamond paper
[05:34:01] <ds2> guess laminating it is easier then polishing acrylic pockets
[05:34:08] <Jymmmm> ds2: Create an account http://cyro.custhelp.com/cgi-bin/cyro.cfg/php/enduser/acct_login.php?p_sid=Uf4MM5Yi&p_accessibility=&p_lva=&p_sp=&p_li=&p_next_page=std_alp.php
[05:34:14] <toastydeath> lamanating is a good idea
[05:35:16] <toastydeath> why does this have to be see through
[05:35:43] <maddash> toastydeath: it's going to be a fancy birthday present for my ex-ex
[05:36:02] <toastydeath> can you make it in two parts, so you can use a "frame" with translucent pocket borders
[05:36:05] <toastydeath> from sanding
[05:36:09] <toastydeath> and then use a stock piece with the optical finish as a faceplate
[05:36:36] <toastydeath> and use like chrome bolts or something to hold the two together and still look fancy
[05:36:55] <maddash> haha
[05:37:18] <maddash> come to think of it, it doesn't have to be transparent, translucent is good too. maybe opaque could work, as long as those horrid tool marks don't show up
[05:37:31] <toastydeath> sand it, it'll go translucent
[05:37:35] <toastydeath> and no tool marks.
[05:37:46] <maddash> toastydeath: I'll try your lube/rpm/hss advice tomorrow, thanks
[05:37:50] <toastydeath> np
[05:37:57] <toastydeath> those marks ain't going away, it will just be less sanding
[05:38:06] <toastydeath> (full disclosure)
[05:41:05] <ds2> toastydeath: going to try to make some keychains and I want to protect the engraving
[05:41:14] <toastydeath> ..?
[05:41:20] <toastydeath> oh, for laminating?
[05:42:16] <ds2> I want to create a clear pocket to glue over the engraving
[05:42:33] <toastydeath> oh
[05:42:37] <ds2> so instead of doing it as one pocket, I'll get flat stock with a center cut out and a thin clear peice
[05:42:42] <ds2> and glue the 3 peices together
[05:42:52] <toastydeath> yeah, that's a much saner plan
[05:48:02] <ds2> btw, those roper/whitney clone punches from HF are amazing things for sheetmetal work
[11:52:00] <micges> hi all
[13:39:18] <rayh> Question. What is EMC_TASK_PLAN_RESUME and how do I get rid of it if I don't want to resume?
[13:44:42] <jepler> I think that's the NML message sent (by the user interface) to continue after a M0 (in gcode) or EMC_TASK_PLAN_PAUSE (via NML)
[13:45:56] <rayh> Thanks jepler
[13:46:28] <rayh> I'm seeing now an EMC_TASK_EXEC_WAITING_FOR_MOTION_AND_IO
[13:47:18] <rayh> The confusion is that I don't think I asked for it to sit and wait for some external signal.
[13:48:03] <rayh> Is there a description of motion and IO around?
[14:03:27] <jepler> at this point I'd be forced to start searching around in the source files..
[14:04:19] <jepler> looks like the only places that EMC_TASK_EXEC_WAITING_FOR_MOTION_AND_IO is set is in emctaskmain.cc
[14:06:58] <rayh> What the ...
[14:07:33] <rayh> So I seem to be sitting waiting for some IO to complete so that motion can take off again.
[14:07:51] <rayh> This is with a tool change and a spindle start.
[14:08:07] <rayh> But I commented both of them out and still get it.
[14:08:25] <rayh> Thanks again Jeff.
[14:09:00] <fenn> is it waiting for tool-changed or spindle-at-speed?
[14:09:14] <fenn> in hal
[14:10:12] <rayh> I was wondering if spindle at speed might do it but didn't see a problem.
[14:10:20] <rayh> * rayh looks again
[14:11:29] <fenn> which tool is active, the new or old tool?
[14:11:59] <fenn> could give some hint where you're stuck in the tool change cycle
[14:13:02] <rayh> Well spindle is on
[14:13:12] <rayh> spindle speed out is 5200
[14:13:34] <rayh> brake is off
[14:15:16] <rayh> and and halui says tool #1 so the change has completed.
[14:15:41] <alex_joni> hey rayh
[14:16:17] <alex_joni> rayh: might be a HAL thingie which sets feedhold until at speed
[14:16:46] <alex_joni> or a foobar in the NML comm
[14:18:14] <rayh> I don't see a spindle at speed here. This is 2.1.7
[14:21:53] <fenn> maybe spindle-at-speed was never implemented
[14:26:15] <rayh> This behavior is really strange. Perhaps I'll recompile a 2.1.7 and try again.
[14:26:28] <rayh> Thanks for the help guys.
[14:26:39] <skunkworks> Hey rah - good to hear from you again
[14:28:28] <skunkworks> * rayh
[14:28:37] <rayh> Hey. you as cold there as we are here.
[14:28:41] <skunkworks> * skunkworks is not hearing from the sun god
[14:28:51] <skunkworks> no - finally
[14:28:54] <rayh> yesterday <-30f
[14:29:07] <rayh> today <-18f
[14:29:11] <skunkworks> yikes - +5 here heat wave
[14:29:31] <skunkworks> It was cold this weekend.. but not that cold
[14:29:33] <rayh> Wow. Hope it gets here before long.
[14:29:43] <skunkworks> snowing again :0
[14:29:59] <rayh> And a 40 mph wind on top of those.
[14:30:16] <skunkworks> yep -sounds like what we had sunday
[14:30:18] <rayh> You have had more than enough snow.
[14:30:34] <skunkworks> yes - I don't mind though.
[14:30:40] <skunkworks> an acutal real winter
[14:31:04] <rayh> Keeps the riff-raff out?
[14:31:28] <skunkworks> heh - and the crazies on the ski hills
[14:56:47] <Roguish> good morning. what is 'teleop' mode as compared to 'free' and 'coord' modes?
[15:07:17] <cradek> teleop lets you jog axes instead of joints on a kinematics machine
[15:15:02] <Roguish> ah, hello cradek. thanks for the explanaition. i just read a brief description in motion.h
[15:27:38] <eta> 'lo, anyone here?
[15:27:41] <eta> I can
[15:27:43] <eta> '
[15:27:53] <SWPadnos> that's tthe spirit!
[15:28:17] <eta> t seem to find the documentation for the auto-tool measurement that cradek did
[15:28:35] <eta> gah, damn keyboard with a messed up layout and huge return key ;)
[15:28:36] <SWPadnos> tool-length-probe.ngc I believe, in the nc_files dir
[15:28:45] <eta> oh, a sample .ngc file
[15:28:47] <Roguish> if an axis can be jogged in teleop mode, why can't it be homed in teleop mode also?
[15:29:02] <SWPadnos> Roguish, because homing uses the free mode planner ;)
[15:29:27] <gezar_> howdy, the video about the rotary broach tool, that does require significant horse power to perform, im talking lots
[15:29:37] <SWPadnos> probably not
[15:30:04] <SWPadnos> broaching uses very small changes in tooth size, so you're shaving off ~0.010 thickness per tooth
[15:30:48] <gezar_> ive used that particular hole to make square holes before, its a lot of power required
[15:30:48] <SWPadnos> I did broaching of aluminum keyways on my bridgeport by sticking the broach in my spindle and crankingthe knee up
[15:31:10] <gezar_> oh yeah, but thats a lot of power there as well
[15:31:28] <SWPadnos> no, not lots of power, good force multiplication
[15:31:39] <gezar_> its more then say a sherline lathe can do is what im saying
[15:31:43] <SWPadnos> it was still the 0.3 or so HP I can generate with my arms (on a good day)
[15:32:23] <gezar_> well, yeah with aluminum, but once you get into steel the requirement goes up
[15:32:31] <SWPadnos> possibly. the canted angle of those broaches may act like a chip-breaker, so you're only pushing one tooth at at time
[15:32:34] <SWPadnos> yep
[15:33:09] <skunkworks> There is no reason you can't do rigid tapping on a lathe - right?
[15:33:15] <gezar_> yeah, thats to help break the chip, when it roates at a wobble its compressing then shearing off a chip
[15:33:37] <skunkworks> (with the tap mounted in the tool post)
[15:33:57] <skunkworks> (as long as emc can reverse the spindle)
[15:34:20] <SWPadnos> bbiab - I have to go commune with my morning coffee
[15:35:04] <gezar_> SWPadnos: imm havine mine now
[15:35:27] <gezar_> skunkworks: its still a good practice in a lathe to use a floating holder to make up for any slight missalignments
[15:38:03] <SWPadnos> the stuff smells so good when it's brewing :)
[15:38:53] <gezar_> I guess my experience on machines is different for a ridgid tapping on a lathe v. mill application because most lathe spindles have a lot more mass, designed to spin a large dia work piece, so the resulting slow/reverse gets tricky, especially at higher rpms but I guess at home a lathe spindle is going to be much like a mills spindle
[15:40:14] <dave_1> much of a lathe spindle's inertia in in the chuck
[15:40:19] <rayh> I saw a small Mazak rigid tapping at 4500 rpm in plastic.
[15:40:37] <rayh> I'll bet it barely got to speed before reverse.
[15:41:25] <dave_1> that is why small collets systems are nice
[15:41:25] <skunkworks> heh
[15:42:16] <rayh> This one had a small chuck and was bar feed.
[15:42:41] <gezar_> no floating holder?
[15:46:28] <rayh> nope.
[15:46:40] <rayh> but plastic bar
[15:48:00] <gezar_> nice, im not saying it cant be done, all things are possiable, im just saying that its more dificult normally
[15:48:28] <rayh> Sure. And a ten hp servo spindle makes a difference
[15:55:06] <dave_1> just a bit different than my 2 hp lathe spindle ... 240 v single phase ;-)
[15:55:38] <dave_1> still done carefully it should thread decently
[15:56:20] <dave_1> be back in a bit
[15:56:21] <skunkworks> we where rigid tapping aluminum with the mazak
[15:56:26] <skunkworks> at the workshop
[15:56:51] <skunkworks> *we - as in the emc developers.. ;)
[15:58:13] <SWPadnos> and plastic
[15:59:18] <rayh> That was great work.
[15:59:41] <skunkworks> http://axis.unpythonic.net/01181942178
[15:59:42] <SWPadnos> yeah -how did you like the stepped screws? :)
[16:00:02] <SWPadnos> (useless but cool :) )
[16:00:30] <rayh> Right. How about pipe thread with a variable pitch?
[16:00:48] <SWPadnos> hmmm. it would have to be small pipe to fit on the Sherline ;)
[16:00:51] <rayh> Hey did that cold weather get to you yet. I sent it your way.
[16:01:01] <SWPadnos> thanks, I think we have a little of it
[16:01:08] <SWPadnos> but not as bad as you were describing
[16:01:30] <Roguish> cradek: thanks for adding the ';' comment line!!!!!!!!
[16:02:03] <fenn> Roguish: joints and axes are different things
[16:02:18] <fenn> you can't home an axis, period.
[16:02:47] <SWPadnos> except when you have trivkins, and everyone calls axes joints and vise versa
[16:03:02] <fenn> sure, when anything can be everything for everyone, anything goes
[16:04:24] <Roguish> fenn: ok. i see that. i am getting wierd behavoir with the 5 axis configuration when using tkemc and axis. the behavior are different and i am trying to figure out what is going on so i can ask
[16:04:36] <gezar_> im going back inside now to work on cooking and some math, well lots of math
[16:04:39] <Roguish> more intelligent (less stupid) questions.
[16:04:51] <SWPadnos> there are no stupid questions
[16:05:00] <SWPadnos> only stupid people asking questions ;)
[16:05:04] <fenn> gezar_: rotary broaching doesnt require any particular amount of power, just enough torque to cut off a decent sized chip (as with any metal cutting)
[16:05:10] <SWPadnos> </8th grade jokes>
[16:05:23] <Roguish> oh yeah there are!!! color me stupid sometimes then.
[16:05:29] <SWPadnos> heh
[16:06:05] <fenn> teleop is poorly tested and definitely buggy
[16:06:07] <gezar_> fenn : power torque but true, in 4140 a 3/4" square hole broach can pull 40% on a 25hp spindle
[16:06:36] <fenn> gezar_: ?
[16:07:02] <cradek> Roguish: you're welcome
[16:07:04] <gezar_> fenn thats what the load was when broaching 36rhc 4140 was with a 3/4" broach made of m2
[16:07:34] <Roguish> like i said, there are problems going from joint to world mode and back. it seems that some of the numerous 'home' flags may not be getting set or cleared.
[16:07:38] <gezar_> machine didnt like doing it at all
[16:07:54] <fenn> ok, 10hp seems high
[16:08:13] <gezar_> fenn I dont know what its like to do in softer metals, but yeah, probably not much work at all
[16:08:27] <cradek> Roguish: G0 X1 (MSG,It was interesting; this can't be a comment); but this is
[16:09:09] <SWPadnos> heh
[16:09:12] <fenn> gezar_: generally i think of HP as how much a certain tool can handle before it liquifies
[16:09:37] <Roguish> sure. ';' must be in the front of the line not mid-line. that's ok fine.
[16:09:42] <gezar_> you mean surfacefootage?
[16:09:46] <SWPadnos> no, it can be anywhere in the line
[16:09:55] <gezar_> gezar_ is now known as gezar
[16:09:56] <cradek> no I allowed it anywhere
[16:09:56] <SWPadnos> but a semicolon in a () comment doesn't count as a comment char
[16:10:06] <SWPadnos> that was the "fun" part
[16:10:32] <fenn> gezar: what i'm saying is there's not really a minimum power level for any particular cut type
[16:10:52] <gezar> eh?
[16:11:10] <fenn> because you can gear it down
[16:11:15] <SWPadnos> there's a minimum force, which when integrated over time gives you a minimum power (so for a given cut, to do it in some time, you need some minimum power)
[16:11:25] <Roguish> oh. ok. that's ok also. ';' anywhere on the line except in a '( )' comment.
[16:11:35] <gezar> fenn yeah, thats true for sure
[16:11:48] <cradek> no, haha, they're allowed anywhere, including in () comments
[16:12:07] <SWPadnos> gezar, you can derive all this stuff with that calculus homework ;)
[16:12:07] <cradek> otherwise the change would break existing programs
[16:12:10] <Roguish> except in message?
[16:12:21] <gezar> im not saying its not possiable, im just saying that to do it on a small home made lathe its not going to be easy
[16:12:33] <SWPadnos> Roguish, you can put a ; in a () comment, but it doesn't make the rest of the line a comment, the ' is just part of the message
[16:12:46] <Roguish> ok.
[16:12:48] <gezar> SWPadnos: its killing me man, practiable application != formulas a lot of the time :)
[16:12:49] <SWPadnos> so you could (msg, change tools; then press start)
[16:12:59] <SWPadnos> without an error
[16:13:06] <cradek> except from the punctuation police
[16:13:11] <SWPadnos> otherwise the ending parens would be nmissing and stop the program
[16:13:15] <SWPadnos> heh
[16:13:19] <fenn> gezar: rigidity is the issue then, or maybe torque but i doubt it
[16:13:44] <fenn> unless you're broaching huge forms
[16:13:50] <gezar> fenn yeah
[16:13:51] <skunkworks> http://www.cnczone.com/forums/showpost.php?p=408985&postcount=101
[16:14:12] <gezar> we ended up going with a single point tool to do the broaching
[16:14:15] <SWPadnos> rigidity is resistance to deforming under force (/ pressure)
[16:14:24] <Roguish> i just wanted the front of the line, but anywhere on the line is way more better.
[16:14:34] <gezar> during the rotary broaching, it would make the turret twist
[16:14:38] <cradek> is it just me or does this tap have a surprising twist to it? http://axis.unpythonic.net/files/01181942178/img_6417.jpg
[16:15:15] <gezar> but if you can index your spindle, then a single point tool taking .003 or so a wack makes broaching in a lathe easy
[16:15:19] <bill20r3> that's pretty twisty.
[16:15:25] <Roguish> who made the '5axis' utube video and what software was used to catch the images?
[16:15:33] <skunkworks> jepler
[16:15:35] <cradek> I wonder if our hole was undersize
[16:16:08] <skunkworks> Roguish: http://emergent.unpy.net/01196105360
[16:16:40] <skunkworks> that picture isn't getting here. I think we are overloading poors jeplers site
[16:18:09] <Roguish> i got to that link. thanks. i will try that for my 5axisgantry.
[16:18:13] <gezar> okay, off to do my homework
[16:18:56] <skunkworks> cradek: I really doubt if that tap is twisting..
[16:19:14] <cradek> yeah but it sure looks twisty
[16:19:53] <skunkworks> you can follow the flute up - I think that is what it is.. (the flute is exiting the tap)
[16:21:07] <skunkworks> (if that made sense)
[16:21:29] <cradek> you might be right
[16:22:18] <jlmjvm> you can actually but a spiral flute tap,has much more twist than the 1 in the pic
[16:22:34] <jlmjvm> osg has them
[16:22:46] <cradek> yeah they're good for blind holes
[16:23:06] <jlmjvm> yes,they sure are
[16:26:16] <skunkworks> it is cool seeing the shavings come strait out the flutes of the tap.. (of the spiral taps)
[16:26:51] <acemi> I can't find bitlbee in lenny repo. was bitlbee removed from lenny?
[16:26:57] <fenn> boy konqueror sure didnt like showing that progressive jpg on my little computer
[16:27:07] <acemi> wrong channel
[16:27:20] <skunkworks> heh :)
[16:29:22] <jlmjvm> i learned something last night from the mazak demo ini on homing,instead of my sequence being z,y,then x it now homes z,then x and y simultaniously
[16:31:02] <jlmjvm> looks like a store bought machine homing out
[16:31:36] <cradek> wish my store-bought machine would do that :-)
[16:31:53] <cradek> not enough computing power available I guess
[16:32:08] <fenn> wish i could buy a machine in a store
[16:32:23] <fenn> they dont seem to have any around here
[16:32:34] <cradek> fenn: well more like "someone bought it from a store when I was in grade school"
[16:42:35] <eta> ah, cradek is here, excelent
[16:42:44] <eta> eta is now known as lern_hydra
[16:43:46] <lern_hydra> I'm trying to implement a tool length probe function in my CAM app
[16:44:05] <lern_hydra> I've been looking at your tool-length-probe.ngc file a bit
[16:44:41] <lern_hydra> does touch-off use g54 and/or will it work with g38.2?
[16:45:49] <cradek> my scheme is this: with no TLO in effect (G49) touch off with the first tool. This one now considered zero length
[16:45:57] <cradek> is now
[16:46:22] <cradek> at tool changes the difference between the original tool and new tool is calculated and TLO is given that number
[16:47:01] <lern_hydra> Ah, I see
[16:49:21] <lern_hydra> hmm, where are the G30 coords defined?
[16:49:31] <cradek> what is G30
[16:50:19] <lern_hydra> some type of predefined coordinate (you used it in your ngc file ;) )
[16:50:43] <cradek> oh G30, right
[16:50:50] <cradek> go to alternate reference point position
[16:51:05] <cradek> just set it over the switch
[16:55:01] <dave_1> anyone qualify as a 5i20 expert?
[16:55:31] <cradek> dave_1: just go ahead and ask your question
[16:55:54] <cradek> when asked point-blank, nobody who's smart will claim to know anything!
[16:56:17] <dave_1> long sad story ... my motenc card just went down ... it holds the bus down so computer won't boot. :-(
[16:56:19] <cradek> unless he's trying to sell you something maybe
[16:56:21] <lern_hydra> cradek, how do I set it over the switch?
[16:56:59] <dave_1> so ... I need to fire up my 5i20/7i33/7i37 which I've never used
[16:57:03] <cradek> lern_hydra: don't make me URL you
[16:57:20] <dave_1> does anyone have a suggestion on a good breakout board ...
[16:57:41] <jlmjvm> campbell designs
[16:57:48] <dave_1> and are there pics/instructions someplace to make cabling easy
[16:58:03] <cradek> lern_hydra: http://linuxcnc.org/docs/2.2/html/gcode_main.html#sub:G28,-G30:-Return
[16:58:14] <dave_1> brb
[16:58:17] <lern_hydra> cradek, O_o it doesn't say in the manual, it says how to config in a gcode file, but not globally
[16:59:01] <cradek> those parameters are saved, so once you set it (even in MDI) it will be set forever until you change it
[16:59:18] <lern_hydra> ah, ok
[16:59:21] <lern_hydra> right, thanks
[16:59:32] <cradek> in trunk, but not 2.2, I've made G30.1 (and G28.1) which set those parameters to the current position
[16:59:43] <cradek> otherwise I think they're a pain to set
[17:00:35] <lern_hydra> oh I see
[17:00:37] <lern_hydra> bbl
[17:00:42] <lern_hydra> thanks :)
[17:13:31] <skunkworks> dave_1: yeck
[17:13:57] <dave_1> indeed
[17:14:52] <tom1> dave_1: the pinouts of the m5i20 are in the emc pdf manuals, the cheap<er> breakouts are from http://www.daqstuff.com/
[17:15:37] <dave_1> tom1 ok, since I don't need opto isolation
[17:16:59] <dave_1> tom1 I just like the instructions for dummies ... P2 on "xxx" goes to P2 on "yyy" , etc.
[17:18:21] <skunkworks> dave_1: any one thing that took it down do you think?
[17:18:54] <tom1> dave_1: anders wallin has diy opto iso brkouts ( i spose you could jmp across the opto pads )
[17:20:19] <dave_1> don't see that I want the opto ....
[17:20:36] <dave_1> the '33 talks directly to the encoders and the amps
[17:20:55] <dave_1> and the '37 is opto already so I think I'm set
[17:22:08] <dave_1> skunkworks ... no hint what took it down .. I was shutting down and firing up ... emc ... working on setting for homing and all of a sudden it wouldnt boot
[17:22:35] <dave_1> has to be the pci chip on the board ... tried two different motherboards
[17:23:49] <dave_1> tom1 are all pins individual or do they gnd all even pins
[17:24:26] <tom1> dunno dave, its in the docs & i gotta fly
[17:24:35] <dave_1> have fun
[17:25:33] <dave_1> gonna run also .... have fun ...
[18:10:52] <alex_joni> http://businessclass.org/truthaboutinternet.jpg
[18:12:25] <skunkworks> heh
[18:14:05] <eta> eta is now known as lern_hydra
[18:14:45] <lern_hydra> back again, can't seem to find (neither in the wiki nor in the user manual) any way of doing an incremental move, anyone know the syntax for that?
[18:15:16] <lern_hydra> ie. incremental_g0 Z5 moves the Z axis 5 units
[18:15:43] <Vq^> G0 Z5 G91
[18:16:13] <lern_hydra> oh, G91
[18:16:18] <Vq^> and G90 to restore to absolute positioning
[18:16:49] <lern_hydra> oh, so it's modal
[18:16:49] <Vq^> if im not completely mistaken...
[18:17:25] <lern_hydra> yep it appears to be modal
[18:17:28] <lern_hydra> thansl
[19:26:36] <fwhite> anyone familiar with working with .dxf files and can assist me? I am not a cad user and disabled and was told the drawings I had made from pdf need to be corrected so the parts fit together
[19:31:42] <archivist> I was just about to say something!!
[19:34:47] <skunkworks> archivist: anything new?
[19:35:34] <archivist> like what?
[19:36:05] <archivist> I had to change my gear prog to match the extra axis today
[19:36:33] <skunkworks> that good
[19:36:35] <skunkworks> :)
[19:37:06] <archivist> I must fit some homing switches and measuring to improve the cycle
[19:37:43] <skunkworks> what is the link to your sit again? (cnc pictures and such)
[19:37:56] <skunkworks> site
[19:38:07] <archivist> will be noo vid later when uploaded
[19:38:21] <archivist> www.archivist.info/cnc
[19:38:43] <archivist> not added this weekends work yet
[19:41:48] <archivist> won some useful parts today http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&rd=1&item=180213396882&ssPageName=STRK:MEWA:IT&ih=008
[19:42:31] <skunkworks> Nice
[19:55:25] <gezar> buahahahahah http://www.c505.com/vvvvvv/projects/ascii_rock/index.html
[19:55:32] <gezar> thats really close to the end of the internet
[19:56:47] <archivist> and this to get ideas at this week http://www.devicelink.com/expo/macbuild08/fairataglance.html
[21:02:14] <SkinnypuppY34> Anyone have a link that has small parts for Dumore TP grinders? One of my dumore spindles came with a 1/8 thread on type collet that's been spun out and overclamped :o( to death.
[21:13:14] <SWPadnos> http://www.dumorecorp.com/ ??
[21:18:52] <SkinnypuppY34> I'd been looking on the dumore site and hadn't received an e-mail back. Tried travers and mcmastercarr sites..
[21:30:28] <gezar> homemade tamales are so good
[21:31:24] <seb_kuzminsky> gezar: what speeds & feeds do you use for tamales?
[21:32:17] <SkinnypuppY34> mmmmm
[21:33:01] <gezar> 10-20sfm with a feed rate of about 1 corn husk covered and wrapped /min
[21:33:32] <gezar> they are not nearly as hard as I had originally thought they were going to be to make
[21:34:04] <quiteBIGeye> gezar: you have to promise to paste the g-code for tamales :O)
[21:34:10] <gezar> and I cant understand how the last 10 places ive ordered them at have managed to screw them up so badly
[21:34:33] <gezar> I will, this was my trial run, so ill post a beef one how is that?
[21:34:46] <gezar> I used some leftover chicken that I boiled down
[21:35:04] <gezar> it wont be iso standard Gcode is that okay?
[21:35:26] <SkinnypuppY34> *(&^ just got outbid again on a ballscrew.... anyone here just win an NSK basslcrew auction?
[21:35:32] <quiteBIGeye> as long as you use proper mcodes, i'll be okay with it...
[21:35:34] <tomp> SkinnypuppY34: dumore 1/8 collet http://www.jlindustrial.com/DUM-70066A/SEARCH:CATEGORY/product.html
[21:35:53] <gezar> SkinnypuppY34: what size screws you after?
[21:36:12] <SkinnypuppY34> Woah tomp good dig! thanks
[21:36:26] <tomp> be sure b4 buying
[21:37:25] <SkinnypuppY34> gezar something to drive the quill on my bp clone http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&rd=1&item=150211343552&ssPageName=STRK:MEDW:IT&ih=005
[21:38:00] <gezar> oh wow, your thinking about using a screw and some how attatching an arm to the quill?
[21:38:35] <fenn> how else would you do it
[21:38:55] <gezar> directly into the existing rotary gizmo
[21:38:59] <SkinnypuppY34> Yep, hopefully attaching to the adjustable stop's location I've seen others with a clampon foot around the spindle
[21:39:16] <gezar> I guess my way isnt very accurate though
[21:39:20] <SkinnypuppY34> I though of it but there is a little play in the handfeed
[21:39:22] <fenn> lots of backlash
[21:39:54] <gezar> clamp on foot would give you some mounting freedom
[21:40:42] <skunkworks> then you could spin the nut...\
[21:41:03] <gezar> i was thinking about that, one of those through hole steppers?
[21:41:33] <gezar> or servos, who knows
[21:41:49] <tomp> lock the depth nut& jam nut on left of BP, then see your blacklash by feeling the hand feed lever on right
[21:47:19] <tomp> no, that'll show the slop in the feed gear...
[21:49:10] <tomp> i think the depth nut is against a boss on the quill cylinder, then there wouldnt be any Z lash there
[21:50:00] <fenn> jmk has a drawing of Z with a spinning nut
[21:50:23] <tomp> there's a spinning nut design for mini mills...
[21:51:06] <fenn> http://jmkasunich.com/cgi-bin/blosxom/shoptask/Z-axis-12-16-06.html
[21:51:10] <SkinnypuppY34> I was just looking at the turn to Z travel , one turn gives me 3 1/2 " quill travel
[21:51:39] <tomp> another http://www.truetex.com/dcdrv.htm
[21:52:17] <fenn> uh, that's a spindle conversion?
[21:54:47] <SkinnypuppY34> Looks pretty nice
[21:55:42] <tomp> spinning nut http://www.stirlingsteele.com/millplans.html
[21:55:54] <tomp> uh, right fenn
[21:58:59] <tomp> SkinnypuppY34: the thingy that gives you 3 1/2" is sloppy, jmk's design is cnc-worthy
[22:00:50] <tomp> maybe put one of those boyscout collapsable cups over the exposed screw ;)
[22:01:29] <SkinnypuppY34> OIC just looked at the JMK design
[22:04:52] <SkinnypuppY34> Here's one of the conversion threads I had ideas from though I think I want to mount onto the machined pad the depth rule is attached to instead of the four clamping bolts http://cnczone.com/forums/showthread.php?t=8749&highlight=j-head+conversion
[22:08:13] <SWPadnos> that's how some of the Anilam quill drives mount
[22:08:36] <SkinnypuppY34> On that machined pad?
[22:09:17] <SWPadnos> yep
[22:09:41] <SWPadnos> the bolts are used to hold it down, but the mating surface is that lip at the top/bottom of the scale drive nut
[22:10:05] <SWPadnos> in fact, the drive is a peg that goes into a hole on that bushing (that the depth stop ring goes to)
[22:10:57] <SWPadnos> I can take some photos of the Anilam quill drive I have if you like (but probably not for a few days at least)
[22:11:01] <SkinnypuppY34> Good to know looked to be the best datum to work from , jeez an expanding plug what what I had in mind
[22:11:26] <SkinnypuppY34> Sure thing I wouldn't mind seeing
[22:11:33] <SWPadnos> it's not the best way, but it is good, considering what's there
[22:11:58] <SWPadnos> I think I have to drill / tap some holes in the millhead though, which is part of the reason I haven't mounted that yet
[22:14:48] <SkinnypuppY34> What all do you have on your mill? I've been enjoying my 2 axis steppers and need to get in gear on the quill
[22:15:12] <SWPadnos> I have handles ;)
[22:15:53] <SkinnypuppY34> I had handles nothing wrong there ;)
[22:15:58] <SWPadnos> I'm going for the "machine retrofit procrastination" record, so it may be a while before I put the servos on :)
[22:17:52] <skunkworks> you have that problem too?
[22:18:12] <SWPadnos> what problem? :)
[22:18:45] <SkinnypuppY34> I sure do , that and cheapness has led me to months with no Z ..
[22:25:11] <fenn> swp what were you going to make that got you into cnc anyway
[22:25:41] <fenn> some heat exchanger or something?
[22:29:09] <SkinnypuppY34> I want to build one of these http://www.trevormarshall.com/waveguides.htm
[22:35:45] <archivist> hmm "Really folks - plus or minus 1 mm will not kill your antenna!" er depends /me is into microwaves a bit
[22:39:39] <SkinnypuppY34> +- 1mm sounds like a mile to me
[22:39:58] <SWPadnos> mm wavelength is ~300GHz, so it shouldn't matter much on an antenna
[22:46:02] <archivist> Ide like to test one on my network analyser to check
[22:50:00] <archivist> a mm error in width of the slots will change the impedence match, dunno by how much though
[22:50:37] <SWPadnos> ok, in waveguides it could be a problem ;)
[22:52:14] <archivist> and iirc some slot antennas use a phased refection from the ends, he has loads to dump it
[22:53:13] <archivist> if it all goes out the slots none will get to the ends
[22:54:40] <SkinnypuppY34> Anyone familiar with roton ballscrews? http://www.roton.com/Mating_Components.aspx?family=7059321
[22:56:29] <SWPadnos> heard of them, never used them
[22:58:16] <seb_kuzminsky> there was a recent thread on CNC Zone about them: http://www.cnczone.com/forums/showthread.php?t=51159
[22:58:29] <seb_kuzminsky> gotta go, back later
[22:58:30] <archivist> SWPadnos, do you recognise the center bottom testgear item on that guys page looks a bit like a scalar network analyser to me
[22:58:43] <SWPadnos> which guy?
[22:59:02] <archivist> http://www.trevormarshall.com/waveguides.htm
[22:59:29] <SWPadnos> or a vector network analyzer
[23:02:56] <SkinnypuppY34> Roton maybe not "Roton ballscrews are C10 or T for transport grade,lead accuracy .009."
[23:09:36] <micges> hello\
[23:59:10] <alex_joni> good night all