#emc | Logs for 2007-12-15

[00:00:11] <BigJohnT> just a little tired and a lot excited to see it move
[00:08:50] <jessica> evenin, all
[00:10:32] <BigJohnT> hi
[00:14:45] <BigJohnT> I got RTAPI: ERROR: Unexpected realtime delay on task 1
[00:14:56] <BigJohnT> when I started axis
[00:14:59] <SWPadnos> hmmm. your BASE_PREIOD may be too low
[00:15:01] <eric_U> that's because you are a genuinely bad person
[00:15:02] <SWPadnos> PERIOD
[00:15:20] <BigJohnT> I got it when the base period was higher too
[00:16:15] <eric_U> did you run the latency test?
[00:16:20] <BigJohnT> yes
[00:16:39] <BigJohnT> my base period is slower than needed by a bit
[00:16:48] <eric_U> I think there is something strange about that rtapi error, but they tell me I'm full of it
[00:17:06] <BigJohnT> It started with 2.2
[00:17:51] <eric_U> of course, people were telling me I was full of it before I started getting rtapi errors
[00:18:04] <BigJohnT> lol
[00:27:22] <jessica> All Your Base Are Belong to Us
[00:27:31] <SWPadnos> AYBABTU
[00:27:55] <jessica> great minds, SWP, great minds...
[00:28:01] <SWPadnos> heh
[00:28:18] <SWPadnos> I forgot what game that was, but it is a funny story
[00:28:30] <jessica> folks, I just picked up 4 NEMA-23's from probotix.
[00:28:40] <jessica> thing is, they don't have keyways.
[00:28:48] <SWPadnos> do they have flats?
[00:29:07] <jessica> my couplers have set screws. Should I grind flats into the shaft?
[00:29:11] <jessica> (no flats)
[00:29:22] <SWPadnos> probably not necessary for the amount of torque a NEMA23 can output
[00:29:26] <fenn> the easy way to couple a stepper to a piece of threaded rod is to use hose clamps and a short length of rubber hose
[00:29:45] <fenn> a solid coupler will inevitably bind
[00:30:07] <jessica> All my rubber hose is in use. I flog all my slaves with it...
[00:30:11] <jessica> *grins*
[00:30:21] <SWPadnos> you can get more at Home Slave Depot
[00:30:49] <fenn> * fenn goes back to reading about US metrication
[00:31:12] <fenn> apparently the navy measures distance in kiloyards
[00:31:15] <jessica> these 23's have hella torque. 'bout want to take my arm off.
[00:31:24] <jessica> kilo yards?
[00:31:30] <jessica> lordy
[00:31:38] <SWPadnos> which ones did you get? (I see 4 motors, all NEMA23)
[00:32:33] <jessica> can't 'member the specs off the top of my head. the ones with the highest torque curve
[00:32:42] <SWPadnos> the big green ones?
[00:32:46] <jessica> on the probotix site
[00:32:48] <BigJohnT> I'm up to 480 IPM with max acceleration 75
[00:32:49] <jessica> blue
[00:33:00] <SWPadnos> hmmm. did you get the unipolar ones?
[00:33:06] <jessica> yeah
[00:33:09] <SWPadnos> bummer
[00:33:12] <jessica> y?
[00:33:34] <SWPadnos> well, they're easier to drive with do-it-yourself electronics, but they can't be driven by high end drivers like geckos
[00:33:51] <archivist> er 6 wire?
[00:33:59] <jessica> i'm cheap -and- easy. plus I'm broke
[00:34:08] <BigJohnT> acceleration is 100 now
[00:34:18] <SWPadnos> I'm not sure if they're worse than bipolar per se, but I don't know of any machines (that you can't lift easily) that use unipolar motors
[00:34:21] <BigJohnT> might need to wire in my limit switches
[00:34:22] <fenn> says they can be driven bipolar
[00:34:23] <jessica> arc, yeah. 6 wire, so I could use them as either uni or bi, right?
[00:34:31] <archivist> yes
[00:34:38] <SWPadnos> ok, good :)
[00:34:48] <archivist> 5 no 6 yes
[00:34:49] <fenn> 2000oz*in is decent torque
[00:34:56] <SWPadnos> or 200
[00:35:12] <archivist> * archivist just got 180's
[00:35:26] <jessica> so, the question is, to grind a flat or not. What's the vote?
[00:35:42] <BigJohnT> is there any practical limit to max acceleration?
[00:35:53] <archivist> yes inertia
[00:35:56] <BigJohnT> bring them by my shop and I'll machine them
[00:36:00] <fenn> you're driving a 1/4-20 rod with this? i wonder if it will take the punishment
[00:36:15] <SWPadnos> try it without, and if the pulleys/couplers slip, then make flats
[00:36:17] <eric_U> why do they say 4 wire bipolar on the page for the unipolar steppers?
[00:36:18] <BigJohnT> keep going till I start to loose steps?
[00:36:20] <jessica> fenn, me?
[00:36:24] <fenn> yes
[00:36:30] <jessica> yeah, a 1/4-20
[00:36:33] <SWPadnos> because they mistitled the motors
[00:36:44] <SWPadnos> they should say "6-wire steppers"
[00:36:49] <fenn> jessica how long is the maximum distance between the nut and leadscrew bearing?
[00:36:52] <archivist> eric_U, just dont connect the center tap for bipolar
[00:37:00] <SWPadnos> just make to tape the center taps (white and yellow wires, I think)
[00:37:05] <SWPadnos> make sure to ...
[00:37:20] <jessica> i've got a bearing on each end. max length is 26"
[00:38:00] <fenn> eh, well, um.. just dont put too much force on it
[00:38:08] <jessica> heheh
[00:38:28] <jessica> I went 1/4-20 cuz the shaft was 1/4
[00:38:58] <jessica> bad logic?
[00:39:16] <fenn> if you don't have a lathe it's good thinking
[00:39:29] <jessica> btw, fenn, did you have your meeting last night?
[00:39:47] <jessica> I don't have a lathe, but i have access to one. why?
[00:40:02] <fenn> yeah we talked about C and how foreign investors are buying up all of the US's assets
[00:40:24] <jessica> whould you have put a 1/4 pocket in a 3/8 threaded rod?
[00:40:28] <fenn> and i showed off one of my lathe's axes
[00:40:34] <fenn> axisees
[00:40:58] <jessica> how many ppl showed up?
[00:41:14] <fenn> five
[00:41:49] <BigJohnT> up to 510 IPM with 500 as max accel
[00:41:57] <jessica> all good. perhaps after the holidays I'll swing by. I'm in b'ton often enuf.
[00:42:14] <jessica> I'm taking a baguazhang class in town
[00:42:15] <SWPadnos> if you have access to a lathe/grinder, then you're better off using a screw that has the inner thread diameter about equal to the motor shaft diameter
[00:42:19] <fenn> BigJohnT: what does max vel have to do with max accel?
[00:42:35] <SWPadnos> you get rid of the threads on the end, and you can make a nice bearing shoulder, etc.
[00:42:47] <BigJohnT> dunno but my limit seems to be 510 on velocity
[00:42:52] <jessica> swp, that makes sense
[00:42:59] <BigJohnT> I'm just trying to figure this out
[00:43:01] <SWPadnos> also, when you turn it down, you can make the end the same diameter as the motor shaft, so a coupler is a little easier (maybe)
[00:44:13] <jessica> does the rubber hose solution add any backlash? it seems it would.
[00:44:21] <fenn> not really
[00:44:39] <fenn> the reinforced braided hose is pretty stiff actually
[00:44:50] <archivist> use reinforced hose
[00:45:01] <fenn> anyway steppers act like a spring within a half step
[00:45:09] <SWPadnos> it's not like you use a 1-foot length- more like 1/16 or 1/8" between shaft ends
[00:45:21] <BigJohnT> jessica: use a hydraulic hose
[00:45:22] <SWPadnos> maybe a little more, but not a lot
[00:45:24] <BigJohnT> no wind up
[00:45:49] <jessica> *smiles*
[00:46:33] <jessica> shame I only made a 5/8 clearance hole for the coupling spot.
[00:46:40] <jessica> tight fit for a hoseclamp
[00:46:57] <SWPadnos> yeh, you might have screwed the pooch there
[00:47:11] <BigJohnT> ok I found out what max acc is limited by changing directions
[00:47:27] <SWPadnos> do you have backlash set?
[00:47:34] <BigJohnT> me?
[00:47:39] <SWPadnos> yes
[00:47:50] <jessica> i like to build things tight. unfortunately, that bites back if you make a mistake.
[00:47:50] <BigJohnT> no backlash on this machine that I can tell
[00:47:55] <SWPadnos> ie, is there a non-zero BACKLASH setting in EMC?
[00:48:05] <SWPadnos> that will also affect max usable accel, I think
[00:48:12] <BigJohnT> ok
[00:48:25] <fenn> jessica: a good motor mount is simply a square tube turned sideways to give access to the coupling, and some holes drilled through
[00:49:07] <jessica> Now he tells me...
[00:49:36] <fenn> or you can use a standoff on each corner
[00:50:04] <jessica> I used blocks of wood...
[00:50:43] <fenn> that should work.. i tend to have a distaste for wood in machinery
[00:51:22] <jessica> If I could afford aluminium, i'd use that.
[00:51:43] <jessica> My machine is a rockcliff variant
[00:52:07] <jessica> by this time, it's more variant than anything
[00:52:16] <fenn> what or who is rockliff?
[00:52:39] <jessica> rockcliff is a company that sells plans for homebuilt mills
[00:52:57] <jessica> cnczone has a forum for them (i thyink)
[00:53:32] <fenn> heh "Your cnc router will be as functional as a real milling machine"
[00:54:09] <fenn> well, i dont know how they figure that
[00:54:09] <BigJohnT> well looks like 420 IPM with 300 accl is the max I can get without loosing step with 24vdc
[00:54:12] <jessica> i'm not so stupid as to believe that, but the machine should do what I need it to.
[00:54:44] <jessica> I'm doing mostly art carving with my machine
[00:55:00] <fenn> BigJohnT: i noticed that when you push steppers to the edge like that, you'll lose steps while cutting at fast speeds
[00:55:01] <jessica> andsome RP/prototyping
[00:55:30] <archivist> BigJohnT, leave headroom for cutting forces etc
[00:56:00] <BigJohnT> plasma torch!
[00:56:25] <jessica> well folks... off to dinner. talk to you all soon.
[00:56:37] <jessica> thanks for the help!
[00:56:56] <fenn> it's a plasma torch?
[00:57:05] <BigJohnT> yep 48 x 48
[00:57:09] <BigJohnT> portable
[00:57:21] <archivist> BigJohnT, and pulling pipes cables weight of the head etc
[00:58:00] <BigJohnT> I'm just running 24vdc to test will have 40-50 vdc for real
[00:58:28] <BigJohnT> this is the first time it's alive so i'm exploring the unknown ...
[00:58:30] <fenn> voltage makes a large difference in max velocity
[00:59:24] <BigJohnT> looks like my base period will limit max speed but it it faster than I need so that is good
[01:00:30] <BigJohnT> I could use up to 550 IPM but I rarely cut that thin of material
[01:28:07] <BigJohnT> still getting RTAPI: ERROR: Unexpected realtime delay on task 1 at random even if I'm not running anything in EMC
[01:29:12] <toastyde1th> toastyde1th is now known as toastydeath
[01:49:41] <BigJohnT> Any one know anything about the RTAPI error?
[01:50:03] <fsdafsd> BURRRRRP
[01:50:20] <BigJohnT> must have been good
[01:52:10] <fsdafsd> Heineken
[01:53:13] <jepler> BigJohnT: emc has detected that a real-time deadline was missed. it means your computer is not suitable for realtime, or you've set BASE_PERIOD too low. The warning (specifically the one that says "on task #") is new in emc 2.2.
[01:54:42] <BigJohnT> jepler: I got the error even when my base period was set a lot higher than my latency test
[01:55:00] <BigJohnT> what makes a computer not suitable for realtime?
[01:55:19] <fsdafsd> its inability to think
[01:55:32] <BigJohnT> thanks, what causes that?
[01:55:34] <fsdafsd> and a non realtime kernel
[01:56:24] <BigJohnT> I've loaded the computer from the EMC2 CD with ubuntu
[01:56:28] <fsdafsd> ya needs RT kernel in order to rt unless you have a onboard gcode processor
[01:56:49] <BigJohnT> so I should have a rt kernel...
[01:56:52] <jepler> BigJohnT: some hardware can prevent emc from running when it needs to (e.g., very regularly every 20 microseconds if you have BASE_PERIOD 20000); when emc doesn't run on time, this can cause motors to stall or lose steps
[01:57:17] <jepler> fsdafsd: you are being deliberately unhelpful. stop or leave.
[01:57:36] <BigJohnT> like video cards or hard drive demands
[01:58:20] <jepler> BigJohnT: so emc continually measures whether it has met its BASE_PERIOD deadline. If it misses it, even once, it reports it.
[01:58:47] <BigJohnT> ok, how do I determine what the problem is?
[01:58:58] <SWPadnos> that isn't so wasy to do
[01:59:00] <SWPadnos> easy
[01:59:31] <BigJohnT> i've got the system monitor up and running and it shows the cpu running at 30 - 40%
[01:59:43] <SWPadnos> there are several things you can try, but if none of those causes the errors to stop, then there aren't many solutions other than changing motherboards
[01:59:44] <BigJohnT> no swap
[01:59:51] <jepler> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting#Unexpected_realtime_delay_check_dmesg_for_details
[01:59:58] <SWPadnos> do you have any USB memory sticks/cards plugged in?
[02:00:07] <BigJohnT> no
[02:00:24] <SWPadnos> is it a laptop or a small form factor PC (or one that's meant to run low power)?
[02:00:49] <SWPadnos> I think that wiki page will go over all the things I was about to try to remember ;)
[02:01:35] <BigJohnT> no it's a p4 tower
[02:01:53] <BigJohnT> with a 3.5 and a dvd and that's it
[02:02:00] <cradek> did you say what kind of video card it has? that's the most frequent cause
[02:02:29] <BigJohnT> no, I don't know what video card it has but I have another one or two that I can try
[02:03:01] <BigJohnT> just reading the link from jepler about the possible solutions
[02:04:06] <BigJohnT> It's kinda funny that it mostly happens when I start up emc but after it'
[02:04:15] <BigJohnT> s on for a while it don't happen
[02:04:23] <BigJohnT> might be the video card
[02:05:31] <BigJohnT> when I started axis the CPU went to 100% for a moment then I got the error...
[02:07:26] <BigJohnT> every time I start axis the CPU meter pegs out and shortly there after I get the error
[02:20:54] <BigJohnT> http://i47.photobucket.com/albums/f163/johnplctech/Plasma%20Cutter/P1010051.jpg
[02:21:01] <BigJohnT> http://i47.photobucket.com/albums/f163/johnplctech/Plasma%20Cutter/P1010049.jpg
[02:21:37] <cradek> BigJohnT: interesting setup
[02:21:58] <cradek> seems awfully geared down though
[02:22:33] <BigJohnT> I get 420 IPM with it...
[02:22:53] <BigJohnT> 0 backlash
[02:22:58] <cradek> very nice
[02:23:03] <BigJohnT> thanks
[02:23:13] <cradek> I guess leadscrews normally gear down the motion a bit too :-)
[02:23:16] <cradek> I wasn't thinking about it right
[02:23:21] <BigJohnT> been cyphering on it a while
[02:23:35] <BigJohnT> yea it's a plasma cutter, no load all speed
[02:23:40] <fenn> woah did you make that timing pulley?
[02:23:56] <BigJohnT> I lightened it, it was solid
[02:24:29] <BigJohnT> I made everything except the slide and 80/20 rails and the pulleys
[02:24:53] <fenn> why does the motor move? i mean why not do it like a printer where the motor is stationary and the belt is just clipped onto the part that moves
[02:25:52] <BigJohnT> It was simpler that way. If the motor was stationary then I needed an idler pulley and the X would move anyway
[02:26:13] <fenn> so the two axes are identical?
[02:26:15] <BigJohnT> I started out with the motor on the end
[02:26:24] <BigJohnT> pretty much the same
[02:26:30] <cradek> this looks easy to adjust but as it stretches the scale will change
[02:26:33] <BigJohnT> I've not built the x yest
[02:26:52] <fenn> i've thought about doing something like that with bicycle chain
[02:27:17] <BigJohnT> that would work too but the belt is easier for light loads
[02:28:32] <BigJohnT> the timing belt worked out better than I had hoped
[02:33:55] <BigJohnT> well gotta go
[02:46:13] <BigJohnT> the belts are Kevlar and don't stretch...
[02:48:30] <SWPadnos> what's your total gantry/carriage weight (roughly)?
[02:49:11] <BigJohnT> so far it's only a few pounds
[02:49:19] <BigJohnT> maybe 5
[02:49:34] <BigJohnT> but I'm not done LOL
[02:49:37] <SWPadnos> ok - what do you think it'll be when it's finished ;)
[02:49:55] <BigJohnT> SWAG 10 lbs
[02:50:02] <BigJohnT> at the most
[02:50:10] <SWPadnos> ok, that's pretty light
[02:50:15] <SWPadnos> including the torch?
[02:50:28] <BigJohnT> yea, the 80/20 is light
[02:50:43] <BigJohnT> torch is 1/2 pound maybe
[02:50:50] <Jymmm> Cool, my gf got me a 4pc grabit set... http://www.4grabit.com/Default.asp?bhcp=1
[02:50:50] <SWPadnos> ok, interesting
[02:51:51] <SWPadnos> I've got a friend who's making a light-duty router-type machine, and he's using a different approach to the fixed belt drive you have - I was wondering how similar your machine is to his
[02:51:55] <SWPadnos> not very, actually ;)
[02:52:18] <SWPadnos> (his will have a 10-20 pound gantry, and has a 4x12' work envelope, IIRC)
[02:52:59] <BigJohnT> The gantry will cover 50" of travel
[02:53:17] <SWPadnos> that's significantly less than 144" :)
[02:53:24] <BigJohnT> the Y will be 50" as well just so it is portable
[02:53:41] <SWPadnos> right - easier to get through doorways and such
[02:53:51] <Jymmm> BigJohnT: can you make it 54" instead, or too late?
[02:53:59] <BigJohnT> mine will be 4x4 but with only longer 80/20 could be 4x10
[02:54:12] <BigJohnT> the 80/20 is cheap
[02:54:25] <BigJohnT> why 54"
[02:54:49] <Jymmm> clamping
[02:55:05] <BigJohnT> of the material
[02:55:10] <Jymmm> yeah,
[02:55:23] <fenn> you dont need to cut the clamps :)
[02:55:52] <Jymmm> just seems on my machine when you need to clamp, there's not enough clearnace available
[02:55:56] <BigJohnT> the widest is 48" that I use and the clamps are under the material and only stick out a litt
[02:55:57] <BigJohnT> le
[02:56:35] <BigJohnT> The clamps I have designed slide in the T slot so you can place them where you want
[02:56:38] <SWPadnos> Jymmm, what town do you live in? (I want to look at hotel options for what dates are less expensive, if any)
[02:56:49] <Jymmm> or if you want to have an edge to place the material againest, there's not place to mount it.
[02:57:12] <Jymmm> SWPadnos: Tijana MehEko
[02:57:12] <BigJohnT> I'm not sure I understand
[02:57:18] <Jymmm> SWPadnos: SJ
[02:57:55] <SWPadnos> ok - north, south, near the airport ???
[02:58:04] <fenn> gps coordinates!
[02:58:12] <Jymmm> BigJohnT: Ok, say you have raw stock and you want to maintain that 90 angle, what are you going to use to align up the material to the machine?
[02:58:19] <fenn> so we'll know where to drop the bombs
[02:59:15] <BigJohnT> jymmm: I have some stop blocks that clamp to the 80/20 that I can remove after locking the material in place
[02:59:16] <Jymmm> fenn: drop em whereever YOUR GPS is currently pointing.
[02:59:45] <fenn> you could do the microscope thing.. if emc ever supports it
[03:00:12] <fenn> http://www.waterjets.org/waterjet_pictures_6.html
[03:04:26] <fenn> rotatekins.c could do it... if it took the rotation angle as a hal pin
[03:05:29] <BigJohnT> interesting fenn:
[03:05:49] <BigJohnT> I was born on Kodiak Island a long time agon
[03:05:53] <BigJohnT> ago
[03:07:07] <BigJohnT> well it
[03:07:12] <BigJohnT> 's getting late
[03:07:39] <fenn> * fenn jumps on the bedwagon
[03:29:49] <Gamma-X> is the digitizing algorithm up to par? can it be used for proffesional cnc work?
[03:31:05] <jepler> in emc 2.2 there is a single primitive probing operation, G38.2 "straight probe". It can write the probed coordinates to a file. It is pretty basic.
[03:31:43] <jepler> In the development version, there are various improvements which can be used to create smart probing programs. http://axis.unpy.net/01194541183
[03:31:59] <Gamma-X> thanks jepler
[03:32:09] <Gamma-X> jepler do u digitize urself?
[03:32:24] <jepler> nope, I don't have the equipment for it
[03:33:11] <Gamma-X> a probe with 2 wires?
[03:34:35] <jepler> yes it could be as simple as that .. doesn't change the fact that I don't have one (nor anything I'm interesting in mapping out with a probe..)
[03:35:01] <Gamma-X> gotcha
[03:37:08] <Gamma-X> jepler u a dev by any chance?
[03:37:19] <jepler> yes I am.
[03:38:13] <Gamma-X> jepler any chance u can recomend a i/o card that can be used for 4 axis cnc and digitizing aswell as tool touch off?
[03:39:10] <jepler> my own machine is a simple parallel-port system, so any advice I give will not be based on actual experience.
[03:39:36] <Gamma-X> ahhh i here paralel is way to slow for a big cnc.
[03:40:31] <jepler> yes, it is OK for step&direction motors with modest step rates, but not for high step rates or servo+encoder closed loop.
[03:41:15] <Gamma-X> how close is that program in copying a part?
[03:42:40] <jepler> I believe that the probe information was gathered from a .04 inch grid .. but you can choose an appropriate grid for your part
[03:43:14] <Gamma-X> can i put in a grid of .00001?
[03:44:36] <jepler> (that's a .04 inch grid on X and Y; the Z 'resolution' is roughly the distance moved in 1ms at the requested feed rate)
[03:44:56] <jepler> you could, but if you are doing a large area it would take a very long time.
[03:45:36] <Gamma-X> ahhh wow.... nice work... Thankyou for doing all this.
[03:45:40] <jepler> the 'smartprobe' program is just an emc gcode programming, using our looping and computation extensions to gcode, so there is the possiblity of programming it to use a coarse grid where the part being measured is relatively smooth, and a fine grid where it has more features. but you'd have to work out the details yourself.
[03:47:59] <Gamma-X> u mean like program it myself? lol
[03:55:47] <jepler> 'night all
[04:21:10] <jmk-st> time to shut this down and walk the dog - tomrrow I start wiring motors!
[04:23:58] <Gamma-X> lol good luck
[04:24:30] <jmkasunich> this week's progress: http://jmkasunich.com/cgi-bin/blosxom/shoptask/pc-cased-and-mounted-12-14-07.html
[04:31:13] <Gamma-X> anyone have a post proccessor for mastercam x2 and emc2?
[04:32:50] <toastydeath> lol
[04:33:22] <Gamma-X> why u laugh?
[04:43:14] <Gamma-X> ?
[04:43:18] <Gamma-X> toastydeath ?
[04:52:07] <toastydeath> i wasn't aware there were a large population of mastercam users who also used emc
[04:52:23] <toastydeath> you might have a hard time finding one
[04:52:32] <toastydeath> not that i'm an authority of any kind
[04:52:38] <cradek> I think lerneaen_hydra uses mastercam
[04:52:48] <CIA-8> 03cradek 07w_tool_length * 10emc2/src/Makefile: build maxkins
[04:53:05] <toastydeath> hopefully he'll read this and reply
[04:53:16] <cradek> and how hard can it be? it's just gcode
[04:58:36] <Gamma-X> hola\
[04:58:48] <Gamma-X> sorry was uhhh " buisy" lol
[04:59:55] <Gamma-X> mastercam is a great piece of software
[05:00:26] <Gamma-X> is anyone ever sellin a used mesa card?
[05:00:44] <cradek> not that I've ever seen
[05:01:38] <Gamma-X> darn. im lookin for a cheap setup here.
[05:04:27] <cradek> their software looks great. but if I can't even find a price, I can't afford it
[05:08:55] <toastydeath> it's true.
[05:09:36] <toastydeath> does anyone use modular vises
[06:08:44] <skunkworks> seems to me anonimasu maybe had worked on a post for mastercam... From what I hear it is pretty close to the fanuc post.
[09:20:01] <lerneaen_hydra> 'lo, I'm afraid I don't use mastercam, I use edgecam
[12:18:57] <BigJohnT> any one awake?
[14:07:58] <BigJohnT> anyone awake?
[14:18:51] <Guest539> hi to all
[14:19:54] <fenn> hello guest
[14:20:33] <Guest539> hi fenn i'm tried this chat
[14:21:20] <BigJohnT> morning fenn
[14:23:30] <Guest539> my name is alessandro and i live in florence italy
[14:24:06] <alex_joni> hello
[14:24:40] <Guest539> is this a emc2 chat or is a simple chat ?
[14:25:08] <BigJohnT> emc2
[14:25:57] <Guest539> thank's and sorry for my bed english
[14:26:25] <BigJohnT> I created a new config and only get the RTATI error when I do something like start the IRC
[14:27:00] <BigJohnT> the only diff between the old and new that I know of is the splash screen time.
[14:29:54] <christel> [Global Notice] Hi all! I am about to re-route a few of our European servers, expect there to be a few splits as I move things around. Thank you for using freenode and have a great day!
[14:34:33] <fenn> BigJohnT: if you start the latency test and do whatever you did to get the error, you should see the latency_max jump up
[14:34:52] <fenn> then you can figure out exactly what is causing the realtime delay
[14:35:48] <Gamma-X> i need cheap computer parts...
[14:35:54] <Gamma-X> anyone got any?
[14:36:09] <BigJohnT> ok I'll give that a try at least it's not at start up any more... dunno if the 5 sec splash gives it time to get going or not
[14:36:13] <fenn> i've got a pile of crt monitors..
[14:36:44] <Gamma-X> fenn any motherboards or parts?
[14:36:48] <Gamma-X> where do u live at?
[14:37:01] <fenn> indiana
[14:37:37] <fenn> i had three banana boxes full of motherboards but they were useless so i gave them to a craft lady
[14:37:39] <Gamma-X> thats a long drive lol
[14:37:51] <fenn> Gamma-X: see if there's a freecycle in your area
[14:38:05] <Gamma-X> what the hell is freecycle lol
[14:38:51] <fenn> its basically a mailing list where people offer junk they are getting rid of
[14:38:55] <BigJohnT> I got a lat max of 23086 one time when I fired up irc
[14:39:16] <BigJohnT> the rest of the time it's < 12000
[14:40:08] <Gamma-X> anyone got an extra i/o card theyd wanna trade for?
[14:40:35] <fenn> not me, i'm too cheap for that stuff, so i'm making my own :\
[14:40:48] <fenn> well, something like that
[14:40:58] <alex_joni> BigJohnT: try running some more complicated programs
[14:41:00] <Gamma-X> fenn..... ur making ur own... lol
[14:41:06] <alex_joni> like firefox, openoffice, soemthing with graphics, etc
[14:41:12] <BigJohnT> ok
[14:41:20] <alex_joni> and remember the max_latency (that is the worst case you need to work with..)
[14:41:43] <Gamma-X> fenn u have any information on that i may be interested
[14:41:45] <fenn> isnt there a common x driver that causes realtime delays? (nv or something)
[14:41:59] <fenn> Gamma-X: http://www.freecycle.org/
[14:42:07] <cradek> I've had problems with nv and nvidia both. with nvidia cards, only vesa works safely
[14:42:15] <fenn> -ESTACKERROR
[14:43:26] <fenn> Gamma-X: this is the basic idea: http://fennetic.net/machines/index.php?motherchip
[14:44:25] <BigJohnT> I think it's a nvidia card
[14:44:45] <BigJohnT> I put another card in but what a mess trying to get it to work
[14:45:38] <Gamma-X> fenn that looks like a lot lol
[14:45:46] <BigJohnT> cradek: how do you change the driver to vesa?
[14:45:53] <cradek> I think the magic to reconfigure a new video card is 'dpkg-reconfigure -phigh xserver-xorg'
[14:46:07] <fenn> Gamma-X: it's a distributed modular system, rather than a single big card with fancy wires going to the encoders
[14:46:39] <BigJohnT> cradek: from the terminal?
[14:47:04] <fenn> you can also just edit xorg.conf and change nv to vesa
[14:47:15] <BigJohnT> whoa I just got a 39248 lat max
[14:48:15] <cradek> run a couple copies of 'glxgears' to see it really go up
[14:48:18] <BigJohnT> fenn: where do I start looking for xorg.conf
[14:50:09] <BigJohnT> a search turns up nothing
[14:50:40] <Gamma-X> anyone got a spare breakout board?
[14:50:54] <jmkasunich> BigJohnT: /etc/X11/xorg.conf
[14:51:59] <BigJohnT> jmkasunich: got it thanks
[14:52:07] <jmkasunich> in Linux, most system configuration stuff is somewhere in the /etc tree
[14:52:19] <BigJohnT> ok good to know
[14:53:03] <BigJohnT> and wrote that down in my linux book LOL
[14:54:47] <BigJohnT> so just change the Driver "nv" to "vesa" in the Section "Device" Identifier "NVIDIA Corporation NV10DDR [GeForce 256 DDR]"
[14:55:12] <jmkasunich> that part I'm less sure about, but that sounds reasonable
[14:55:18] <cradek> yes
[14:55:23] <jmkasunich> save the original file with another name just in case
[14:55:33] <BigJohnT> ok thanks and backing up now
[14:55:36] <jmkasunich> and you'll need "sudo" to edit it....
[14:56:00] <BigJohnT> got that on my desktop someone here gave me the tip a while back
[14:57:13] <BigJohnT> here goes nothing!
[15:01:36] <Gamma-X> I wonder if there is a parts i/o card kit i can buy
[15:02:08] <BigJohnT> I still get 32000 when loading IRC
[15:02:16] <jmkasunich> Gamma-X: exactly what are you trying to do
[15:02:26] <BigJohnT> and now the screen is 1024
[15:03:24] <Gamma-X> jmkasunich im tryin to somehow build a i/o anything pci card... i want one for cheap!
[15:03:49] <jmkasunich> you mean like the mesa?
[15:04:04] <Gamma-X> pretty much
[15:04:16] <jmkasunich> $199 _IS_ cheap!
[15:04:16] <Gamma-X> or sumthing that can take its place. with same performance
[15:04:28] <jmkasunich> gawd, I thought I was a fscking cheapskate
[15:04:52] <cradek> no matter how many times you ask here you won't get a mesa for less than the purchase price. there are no used mesa cards to buy. it's too specialized.
[15:05:05] <fenn> well, there's also the pluto-servo, but...
[15:05:29] <Gamma-X> cradek lol im jewish... what can i say
[15:06:32] <Karl1> hey all
[15:06:47] <cradek> hi Karl1
[15:07:20] <Karl1> just got my mesa 5i20 and 7i33T - anyone have sucess with these?
[15:07:38] <cradek> sure, lots of people have
[15:08:09] <BigJohnT> I actually get better lat max with the nv driver...
[15:08:43] <cradek> BigJohnT: the problem must be elsewhere in your system then
[15:09:01] <cradek> did you check bios settings for power saving options that can be shut off?
[15:09:14] <BigJohnT> yea, got all that stuff off
[15:09:51] <BigJohnT> other than starting gaim the lat max is < 10000
[15:10:00] <BigJohnT> brb breakfast calls
[15:10:08] <cradek> that sounds nice
[15:10:09] <jmkasunich> even when you run things like glxgears?
[15:10:19] <cradek> maybe it's a network card problem
[15:11:59] <jmkasunich> I assume he's tried a browser, which should generate at least as much net traffic
[15:12:34] <jmkasunich> but you know what they say about assuming....
[15:15:04] <jmkasunich> breakfast sounds like a good idea
[15:17:55] <Karl1> being new to the linux OS, I am in pursuit for info on my setup for the Mesa Cards; I plan on running 3 of the 4 axis to Advance Motion Control BLDC PWM Servo Amps
[15:20:08] <Karl1> I also have a Pixie P100 card I was going to use to *test* my used amps and motors before connecting to the Mesa
[15:21:06] <fenn> Karl1: you should start by reading the user manual, specifically the HAL sections
[15:21:24] <fenn> then take a look at the m5i20 sample config files
[15:22:52] <Karl1> I guess that's what I was overwelmed with, configuring something with the limited expirience I have with the OS
[15:26:21] <Karl1> even the Mesa Documents seem squirley as far as loading the firmware controls - I was hoping for a plug and pray solution at first, but am not shy about this venture
[15:27:31] <jepler> "If loadFpga is 1 (the default) the driver will load the FPGA configuration on startup. " -- emc documentation, http://linuxcnc.org/docs/2.2/html/hal_drivers.html#r1_5
[15:30:25] <Karl1> Thx jepler, now I am getting some direction - I guess I come to the right place!
[15:32:13] <BigJohnT> ok i'm back for a bit
[15:32:18] <BigJohnT> what is glxgears
[15:32:56] <BigJohnT> yes, I did some net surfing etc and dloading files
[15:33:19] <jepler> glxgears is a program that uses opengl to display some rotating gears. you can run it by opening a terminal and typing "glxgear".
[15:33:22] <jepler> er, "glxgears"
[15:33:45] <jepler> if your system's graphics card is incompatible with realtime, this will often show the problem immediately
[15:33:49] <BigJohnT> ok can I open another terminal while the lat test is running?
[15:34:07] <jepler> yes
[15:34:36] <BigJohnT> ok it's running and lat max is < 12000
[15:34:38] <jepler> you can do just about anything except run emc while the latency test is running -- and you *should* do anything you will ever do while running emc while running the latency test.
[15:36:10] <BigJohnT> is there a way to tell the cpu not to give up so much time when emc is running?
[15:36:12] <Karl1> Another is I am concerned with: I will be buying an new PC, most likely Dual core - is that an issue with EMC? just don't load the 64bit version of UBUNTU?
[15:38:05] <BigJohnT> downloading files does not change the lat max at all :( I'm on dialup
[15:38:05] <jepler> BigJohnT: on a system that is good for RT, the CPU meter can show 100% without triggering the "unexpected realtime delay" message, because the "real time" part of emc gets its chance to run before anything else in the system.
[15:38:27] <BigJohnT> ok
[15:38:56] <jepler> Karl1: I believe that SWPadnos has successfully used emc on core 2 machines, but I am not aware of the details. If you want to use the precompiled version of emc, you have to install 32-bit ubuntu dapper, not any other version.
[15:39:02] <BigJohnT> I think I have another video card at the shop I might try
[15:39:24] <BigJohnT> otherwise I'll not join the chat while it's running
[15:40:50] <BigJohnT> one note about the error it should at least pause the file execution in emc when it comes up... if you were just going for the stop and you get it I might get confused and let my stepper run too far...
[15:43:58] <jepler> that's why you should have a physical big red button; you should never try to click a mouse button or press a key to stop emc in an emergency situation.
[15:46:01] <jepler> there's no guarantee that pressing a key will make emc stop in a timely fashion, even aside from the problem that this pop-up might appear just at the wrong moment and stop the 'esc' key from having its normal meaning in axis.
[15:47:13] <BigJohnT> I know that but the pop up should at least pause the program if you not watching the screen and you get the error...
[15:57:59] <BigJohnT> Is an AMD or an Intel motherboard better for use with emc
[15:59:30] <jepler> I personally have AMD but both AMD and Intel processors and motherboards are being used for emc.
[16:00:35] <BigJohnT> my win machine which I bought piece by piece is AMD but the hand me down from my shop is Intel...
[16:06:48] <BigJohnT> how can you tell what processor you have in Linux?
[16:07:33] <Gamma-X> anyone usin a vfd in here?
[16:07:37] <jepler> System > Preferences > Hardware Information (unless they renamed the option in the version of ubuntu I'm on now)
[16:08:25] <jepler> though in the terminal, "cat /proc/cpuinfo" may give you additional useful information
[16:08:30] <BigJohnT> that's not on the ubuntu menu
[16:08:37] <jepler> OK, then use "cat" in the terminal
[16:09:57] <BigJohnT> ok looks like it's a genuineintel 2409.126 MHz cpu
[16:10:13] <BigJohnT> with 512 KB cache
[16:15:32] <Gamma-X> BigJohnT are u using a vfd/
[16:17:02] <BigJohnT> on my plasma cutter?
[16:17:12] <jmkasunich> BigJohnT: you might want to try a different IRC client, if gaim always gives you error - try xchat instead, for example
[16:17:26] <BigJohnT> ok I'll try that
[16:20:05] <BigJohnT> Gamma-X: on my CNC plasma cutter?
[16:20:22] <jmkasunich> ;-)
[16:20:25] <Gamma-X> BigJohnT didnt know it was a plasma cutter lol
[16:21:05] <BigJohnT> I use vfd's at the shop for work projects
[16:25:21] <BigJohnT> well got xchat loaded so I'll see if I can connect with it :)
[16:31:30] <Gamma-X> Bigjohnt: how do u like teh vfd? can u use urs for a vfd aswell as 3 phase converter
[16:36:45] <fenn> rotatekins.c accepting a hal input for the amount of rotation: http://fennetic.net/pub/irc/rotatekins.png
[16:39:09] <fenn> BigJohnT: remember the microscope thing from last night?
[16:39:20] <fenn> BigJohnT: http://fennetic.net/pub/irc/rotatekins.png
[16:39:38] <BigJohnT> kinda
[16:40:11] <fenn> they aligned the machine to the workpiece in software, lined up to scratch marks with a microscope mounted to the tool
[16:40:28] <BigJohnT> ok
[16:40:49] <BigJohnT> like a setup scope for a mill?
[16:41:04] <fenn> yeah to help with setup
[16:41:11] <BigJohnT> ok
[16:41:30] <cradek> fenn: a better solution would put rotation in the interpreter, since you get a position jump when you rotate with your solution
[16:41:38] <anonimasu> yep
[16:42:10] <cradek> but you can get away with it if you remember to go to the origin before rotating
[16:42:20] <fenn> yeah but this is so easy
[16:42:26] <cradek> yeah I understand
[16:42:43] <anonimasu> heh
[16:43:14] <fenn> it would work with a jogwheel
[16:43:47] <fenn> then you can just pretend you have a big rotary table
[16:44:26] <cradek> you could just rotate by EmcPose->c then
[16:44:48] <cradek> bbl
[16:48:02] <BigJohnT> you mean something like this http://www.penntoolco.com/catalog/products/products.cfm?categoryID=5873
[16:48:32] <fenn> well, you could do it with an edge finder
[16:49:22] <BigJohnT> was that the guy making ice cream cone molds?
[16:52:15] <fenn> yeah
[16:52:59] <BigJohnT> ok I'm a little foggy on it I was celebrating the first running of my plasma table... :)
[16:57:36] <BigJohnT> anyone using antek power supplies?
[16:57:52] <SWPadnos> not me, but they look nice
[16:59:40] <BigJohnT> I like the regulated 5v and 12v that is on board
[17:04:07] <Gamma-X> does emc2 have an auto key to set the tool length if using a tool setter?
[17:04:20] <SWPadnos> no
[17:04:58] <SWPadnos> with EMC2, many things are possible, but not many of the features have a simple button to activate them
[17:05:35] <SWPadnos> so you can do tool length setting, but I think you need to have a subroutine in the G-code file that gets called at each tool change (or something like that)
[17:18:58] <Jymmm> SWPadnos: http://cgi.ebay.com/NEW-LED-PILOT-LIGHT-22MM-GREEN-LED-BULB-AC-DC-NIB_W0QQitemZ150191507835QQihZ005QQcategoryZ42898QQcmdZViewItem
[17:19:07] <Jymmm> SWPadnos: HAWT DAMN!!!!!!
[17:19:11] <Jymmm> finally found em
[17:21:30] <Jymmm> SWPadnos: Not sure if these are the same, but 130VAC/DC replacement LED bulbs http://cgi.ebay.com/NEW-YELLOW-HQ-LED-LAMP-BULB-130V-AC-DC_W0QQitemZ150131687152QQihZ005QQcategoryZ42898QQrdZ1QQssPageNameZWD1VQQcmdZViewItem?_trksid=p1638.m118
[17:23:19] <Jymmm> SWPadnos: http://cgi.ebay.com/NEW-10-PIECES-MINI-PILOT-LIGHT-9MM-RED-NIB_W0QQitemZ150188864329QQihZ005QQcategoryZ42898QQcmdZViewItem
[17:23:55] <Jymmm> SWPadnos: That's what I was talking about before finding the 22mm ones
[17:24:38] <Jymmm> they even have proximity switches for $10
[17:24:41] <Gamma-X> SWPadnos and how would I go about doing that? lol
[17:29:58] <BigJohnT> some variation of holeprobe.ngc I would guess
[17:32:39] <BigJohnT> http://www.isd.mel.nist.gov/personnel/kramer/pubs/RS274NGC_3.web/RS274NGC_33a.html#1013738
[17:32:41] <SWPadnos> no, there's a tool length setting G-code file that comes with EMC2.2
[17:32:49] <SWPadnos> Jymmm, cool
[17:32:56] <BigJohnT> cool
[17:33:10] <SWPadnos> digi-key has replacement lamps as well
[17:33:30] <Jymmm> SWPadnos: LED replacement bulbs?
[17:33:38] <SWPadnos> yes
[17:33:48] <SWPadnos> I think so anyway
[17:33:49] <Jymmm> SWPadnos: Cool, in a pinch.
[17:34:02] <BigJohnT> yep /usr/share/emc/ncfiles
[17:34:12] <Jymmm> SWPadnos: But now can you understnad why I was so adiment about these?
[17:35:04] <SWPadnos> err - sure. it's because you're a jerk ;)
[17:35:54] <Jymmm> SWPadnos: I'm going to bahamas tomorrow, till March!
[17:36:08] <SWPadnos> they have 10-packs of 16mm lights for $9.98
[17:36:27] <SWPadnos> Jymmm, cool, then I can stay at your place while I'm "fixing" your router :)
[17:36:29] <Jymmm> SWPadnos: I think I want the 22m ones
[17:36:42] <SWPadnos> yeah, they're probably better, and easieer to replace the bulbs
[17:36:44] <Jymmm> SWPadnos: Cool, you can bird sit!
[17:36:45] <SWPadnos> -e
[17:36:56] <SWPadnos> oh right. I'll be at the Hilton
[17:37:58] <Jymmm> SWPadnos: Nah, fuck you... you'll bird sit and you'll LIKE it!
[17:38:18] <SWPadnos> nah. I'll sit on the birds, *THEN* I'll like it!
[17:38:19] <Jymmm> SWPadnos: You'll like bebe.... he just wants his head scratched
[17:38:23] <SWPadnos> heh
[17:38:41] <SWPadnos> I'm sure machine work goes faster when there aren't any parakeets around (that think they're eagles)
[17:39:03] <Jymmm> SWPadnos: Cockatiels and parrots
[17:39:28] <Jymmm> Only Haley looks like an Eagle
[17:39:33] <SWPadnos> actually, I think it was a cockatiel that thought it was an eagle (a friend had a little birrd that thought it was a big bird)
[17:39:49] <Jymmm> SWPadnos: All these guys act like it =)
[17:40:21] <Jymmm> Bebe WILL let you he doens't approve of you not paying attnetion to him NOW NOW NOW
[17:40:49] <Jymmm> better photo http://cgi.ebay.com/NEW-COMPACT-PILOT-LIGHT-BLUE-22MM-LED-BULB-IS-AC-DC_W0QQitemZ150137819603QQihZ005QQcategoryZ42898QQcmdZViewItem#ebayphotohosting
[17:48:01] <Jymmm> I wonder if these are full sized covers? http://cgi.ebay.com/NEW-20-PCS-TOGGLE-SWITCH-WATERPROOF-COVERS_W0QQitemZ150192015148QQihZ005QQcategoryZ26211QQcmdZViewItem
[17:48:20] <Jymmm> they look a tad small
[17:50:57] <SWPadnos> here's your contactor, unless you need to use 12V for the controls: http://cgi.ebay.com/NEW-CN-PBC252-24V-CONTACTOR-24V-coil-25-35amp-2p_W0QQitemZ150098074205QQihZ005QQcategoryZ42911QQcmdZViewItem
[17:51:11] <SWPadnos> in which case it's this one: http://cgi.ebay.com/NEW-AC-HEATING-DP-CONTACTOR-12V-DC-coil-40-50amp-3p_W0QQitemZ150174492145QQihZ005QQcategoryZ42895QQcmdZViewItem
[17:52:00] <Jymmm> 3phase is a wee bit overkill =)
[17:52:28] <SWPadnos> no, you want to switch both legs of the 1-phase, and the third contact gives you an aux output that can be used for a pilot light or something
[17:53:00] <Jymmm> I'd REALLY liek to use SSE if I can.
[17:53:04] <Jymmm> SSR
[17:53:50] <SWPadnos> sure. though I find the click of a contactor reassuring - I know it's trying to do something :)
[17:54:26] <Jymmm> estop button http://cgi.ebay.com/NEW-NIB-EMERGENCY-PUSHBUTTON-TWIST-RELEASED-MAINTAINED_W0QQitemZ150188851403QQihZ005QQcategoryZ14883QQcmdZViewItem
[17:54:46] <Jymmm> That's what the idiot lights are for =)
[17:55:28] <SWPadnos> heh
[17:55:41] <SWPadnos> yep - that button looks like it has teh right action
[17:55:56] <SWPadnos> I wonder if it's any good in terms of quality
[17:57:01] <Jymmm> we'll find out
[17:57:13] <Jymmm> cheapest I've found so far.
[17:57:36] <Jymmm> pun intended
[17:59:50] <Jymmm> SWPadnos: heres the noise maker http://cgi.ebay.com/NEW-BUZZER-16MM-BLACK-24VAC-DC_W0QQitemZ150189223647QQihZ005QQcategoryZ100184QQcmdZViewItem
[18:00:10] <SWPadnos> I saw that :)
[18:03:57] <SWPadnos> I'd use this instead of a 3-position toggle: http://cgi.ebay.com/NEW-22-mm-Selector-Switch-3-Position-Maintained-2-NO_W0QQitemZ150187185551QQihZ005QQcategoryZ58166QQcmdZViewItem
[18:04:06] <SWPadnos> knobs are harder to bump into the wrong position
[18:07:29] <Jymmm> SWPadnos: It's NO, not NC though
[18:07:37] <SWPadnos> no problem
[18:07:59] <Jymmm> how do you figure that?
[18:08:03] <SWPadnos> left position (switch #1 closed) = auto - the switch goes in series with the EMC relay output
[18:08:09] <SWPadnos> middle position = off
[18:08:42] <SWPadnos> right position (switch #2 closed) = on, switch goes across the series connection of EMC relay and contacts from #1, which turns on the motor
[18:09:44] <Jymmm> assuming it's ON/OFF/ON
[18:09:47] <SWPadnos> it is
[18:10:56] <Jymmm> could be ON/ON/OFF it is only $5
[18:11:05] <SWPadnos> no, I don't think so
[18:11:20] <SWPadnos> the center position is unlikely to activate either switch
[18:11:43] <SWPadnos> if you look at the bottom photo, you can see that it's two sets of contacts that are identical, just one on either side
[18:11:59] <Jymmm> http://ecatalog.squared.com/fulldetail.cfm?partnumber=ZB2BE101
[18:12:23] <danielbr> hi all
[18:12:26] <danielbr> it seems that there is a limit to max pluto-step.stepgen.M.position-cmd: 2²° this is a sofware or hardware limit in the pluto FPGA?
[18:12:45] <Jymmm> SWPadnos: I sent em an email to confirm - it's ebay, NEVER ASSUME ANYTHING
[18:13:09] <danielbr> in the USC this limit is 2³¹
[18:13:34] <SWPadnos> position-cmd or encoder feedback counts?
[18:13:49] <danielbr> well the two is equal
[18:13:56] <SWPadnos> no they're not
[18:13:56] <Jymmm> Damnit SWPadnos, if I use 22m control for A/B, now I can't use toggle switches! fscker!
[18:14:15] <SWPadnos> position-cmd is supposed to be in user units, I think
[18:14:26] <danielbr> until you get a error the two is equal
[18:14:30] <SWPadnos> whereas feedback counts (not position) is in encoder ticks
[18:14:47] <SWPadnos> feedback position is not the same as feedback counts
[18:14:54] <SWPadnos> what are your user units?
[18:14:59] <SWPadnos> mm, inch, ...?
[18:15:01] <danielbr> in usc and pluto the feedback is emulated
[18:15:10] <danielbr> mm
[18:15:40] <SWPadnos> ok, and do you have a table that's 2^20 mm long? (ie, 1 km or so)
[18:15:49] <SWPadnos> (answer: no :) )
[18:16:03] <danielbr> no but if I need rotate an axis?
[18:18:03] <SWPadnos> there is a 2^20 or 2^21 limit in the pluto driver, but it's for step rate, not position
[18:18:17] <SWPadnos> as far as I can see by taking a quick look at the code
[18:19:36] <danielbr> when this cmd value reach this 2^20 the fb change the start crazy changes in the signal and i get errors
[18:20:05] <SWPadnos> well, I'd have to look at the code a bit more to see that, but unfortunately I'm about to leave for a bit
[18:20:17] <jmkasunich> I'm having a hard time understanding what you are describing
[18:20:49] <jmkasunich> the position command and feedback values are floating point - I would be astonished if anything weird happens when they hit the integer value 2^20
[18:21:21] <SWPadnos> no -there's a max delta value in pluto_step.comp that's at the 2^20 limit
[18:21:24] <danielbr> but this weird thing happens
[18:21:43] <jmkasunich> danielbr: "weird thing" has zero information content
[18:21:44] <SWPadnos> I haven't looked at exactly how all the masks and shifts are used though
[18:22:32] <danielbr> ok just informing a thing i see today
[18:22:36] <SWPadnos> Jymmm, you can swap those contact blocks around, so you could get some NO and NC-type switches, and make yourself a rotary with one NO and one NC, or two NC ...
[18:22:43] <SWPadnos> anyway, gotta run. bbl
[18:22:54] <Jymmm> laters
[18:24:14] <Jymmm> Hmmm, i wish there was a way to have the panel lamps BLINK when it AUTO mode
[18:24:15] <danielbr> I just asking if this max value can be setted to something like 2³¹ like the USC
[18:24:33] <jmkasunich> WHAT MAX VALUE?!?
[18:24:47] <jmkasunich> would it kill you to actually include a hal pin name or a variable name?
[18:25:26] <danielbr> it seems that there is a limit to max pluto-step.stepgen.M.position-cmd: 2²°
[18:26:44] <jmkasunich> * jmkasunich looks at the code
[18:26:57] <danielbr> thanks
[18:30:10] <jmkasunich> when pluto-step.stepgen.M.position-cmd exceeds 2^20, what happens?
[18:30:24] <Jymmm> If I plan on having a ON/OFF/AUTO switch, and an idiot light go go with it. Do you think the lmap being ON/OFF/BLINK would be a good thing?
[18:30:32] <Jymmm> LAMP
[18:30:39] <jmkasunich> (I'm not seeing any code that would imply a limit, but I don't have a pluto card, so I can't actually run it)
[18:30:40] <danielbr> the fb pin change the signal and start change the value like a crazy
[18:31:00] <jmkasunich> can you capture a halscope image and post it somewhere?
[18:31:24] <jepler> SYNOPSIS
[18:31:24] <jepler> Note: In this release of emc2, this driver is alpha-quality and not
[18:31:24] <jepler> suitable for use on production machines.
[18:31:27] <jmkasunich> probably want position-cmd, position-fb, counts
[18:31:28] <danielbr> not now until this is another city
[18:31:40] <danielbr> until next week
[18:32:05] <danielbr> ok i know this jeff
[18:32:13] <jmkasunich> in that case it is unlikely that anyone can help until next week
[18:32:36] <danielbr> i just asking if this can be easy to change
[18:32:40] <jmkasunich> I can say that there should not be a 2^20 limit - if there is it is probably a bug
[18:33:14] <jmkasunich> I'm curious what units you are using though
[18:33:33] <jmkasunich> 2^20 is 1 million of whatever unit - that is a lot of movement
[18:33:38] <danielbr> angular axis
[18:34:20] <jmkasunich> 2^20 degrees is about 2700 revolutions
[18:35:15] <danielbr> the value that cause problem is degrees * steps/degrees
[18:35:23] <danielbr> scale
[18:35:42] <jmkasunich> huh?
[18:36:06] <jmkasunich> you said "pluto-step.stepgen.M.position-cmd: 2²°"
[18:36:18] <danielbr> yes
[18:36:25] <jmkasunich> pluto-step.stepgen.M.position-cmd is in degrees isn't it?
[18:36:50] <danielbr> my error the problem is counts
[18:37:10] <jmkasunich> * jmkasunich looks at the code again
[18:38:47] <jmkasunich> damn jepler, you don't like comments so much do you? ;-)
[18:39:16] <jmkasunich> what is the function "extend()"?
[18:39:23] <jmkasunich> its invoked, but not defined
[18:39:28] <jepler> extend() is in pluto_common.h
[18:39:38] <jmkasunich> diuh
[18:40:07] <jepler> it performs extension from the short width value transmitted from the fpga to the 32 bit value used for feedback counts
[18:40:22] <jmkasunich> thats kind of what I expected
[18:40:37] <jmkasunich> and probably the most likely place for this issue to be coming from
[18:41:17] <jepler> ah: it works on 'long's which are 32 bits .. one bit is sign, 11 bits are the fractional part. that leaves 20 whole bits for position
[18:41:41] <jmkasunich> fractional parts?
[18:41:50] <jmkasunich> isn't count an integer, and extend an integer function?
[18:42:04] <jepler> it's fixed-point
[18:42:12] <jepler> F bits are fractional
[18:42:29] <jmkasunich> oh, this is the stepgen accumulator value?
[18:42:34] <jepler> right
[18:43:01] <jmkasunich> ah, in that case, it needs to be long-long
[18:43:08] <jepler> that might fix it
[18:43:16] <jmkasunich> so you have at least 32 good bits after you drop the fractional part
[18:45:15] <jepler> unfortunately I can't test pluto_step or pluto_servo (both of which would be affected by this change) at the present time.
[18:45:26] <jmkasunich> no pluto board handy?
[18:45:33] <jepler> nor anything useful to hook it up to
[18:45:49] <jepler> e.g., I'd need to test that pluto_servo still reads encoders properly
[18:45:54] <jmkasunich> oh
[18:46:16] <jmkasunich> with encoders, the counts are integers, no fractional part, right?
[18:46:19] <jepler> yes
[18:46:20] <Gamma-X> yay! i just got called in for overtime!.
[18:46:30] <Gamma-X> 500 bucks just for tomorow
[18:46:38] <cradek> jepler: I can let you know when I hook up the lathe next time
[18:46:40] <jepler> not a bad windfall
[18:46:46] <jmkasunich> I can think of a few ways to test it
[18:47:04] <jmkasunich> sim-encoder -> parport -> pluto (the latter link is actual hardware)
[18:47:05] <jepler> actually I guess I have the diagnostic mode where I loop back some digital outputs onto the first encoder
[18:47:22] <jepler> but that can only step at 1kHz, so it'll take awhile to get to any position which might be affected by the change
[18:47:28] <jepler> er, "step" isn't the right term
[18:47:50] <jmkasunich> well, I guess the first step would be to see if danielbr's report can be duplicated
[18:48:02] <jmkasunich> pluto can step pretty quick, right?
[18:48:15] <jepler> I suppose it would behave the same way with or without any actual motor
[18:48:32] <jmkasunich> yeah, step systems with no fb don't need hardware
[18:48:54] <jepler> well now all I need is a system with working realtime
[18:49:36] <jepler> (well I've got one, i'd just have to take it apart a bit to test pluto)
[18:50:23] <jmkasunich> if I had a pluto I'd take a stab at it
[18:50:45] <jepler> I have verified that this compiles, but not that it fixes anything: http://emergent.unpy.net/files/sandbox/long-long-for-pluto.patch
[18:50:56] <cradek> jepler: yell if you need any hardware :-)
[18:51:46] <danielbr> jepler: another question,i tried hook a software stepgen to some dout in pluto and I got very bad looking steps in osciloscope
[18:52:04] <danielbr> this is normal?
[18:52:15] <jmkasunich> danielbr: probably
[18:52:22] <jmkasunich> stepgen generates steps in a fast thread
[18:52:34] <jmkasunich> I bet pluto dig in and out are being updated in the slow thread
[18:52:37] <jepler> pluto's outputs are updated in the servo-thread. If you add stepgen to servo-thread you will get very low step rates but it should be OK. If you add stepgen to base-thread, you get unusable outputs on the step/direction pins
[18:52:46] <jmkasunich> the EPP interface isn't really fast enought to update that stuff in the fast thread
[18:53:18] <jepler> bbl
[18:53:19] <danielbr> ok so the max is 1khz
[18:53:41] <danielbr> using servo-thread
[18:53:43] <jmkasunich> if you want to do software stepping with the pluto digital outs, yes
[18:54:08] <jmkasunich> if you are using software stepping, use a plain parport instead - it can be updated from the fast thread
[18:55:19] <danielbr> i tried replace this rotatory channel with software stepgen but got this problem
[18:56:10] <danielbr> also the software stepgen have a nice option velocity input cmd
[18:58:39] <danielbr> i think all you know about my video in the wiki http://www.youtube.com/watch?v=0RH_H7jgQQY
[19:04:34] <jmkasunich> danielbr: I've seen the video, didn't know it was your machine
[19:04:38] <jmkasunich> that is using pluto?
[19:05:50] <danielbr> no using USC
[19:06:14] <danielbr> the pluto is just some test
[19:07:10] <danielbr> that machine is a retrofit we did
[19:07:37] <danielbr> the proprietary control was bad
[19:07:48] <danielbr> problem with huge steppers
[19:07:57] <danielbr> 14A
[19:08:12] <danielbr> replaced by ac servos
[19:20:00] <alex_joni> hi danielbr
[19:21:24] <danielbr> hi alex
[19:21:39] <danielbr> got my last email?
[19:21:47] <alex_joni> yup, I did :)
[19:22:05] <alex_joni> sounds reasonable to me.. no need to make all your work public
[19:22:31] <danielbr> ok so i'll try
[19:23:55] <alex_joni> danielbr: it has time..
[19:24:32] <danielbr> ok :)
[19:37:50] <Gamma-X> what kind of signal would an i/o card need to recieve when using a tool touch probe? electric wise?
[19:38:23] <BigJohnT> wouldn't that depend on the io card?
[19:38:57] <Gamma-X> a mesa.
[19:39:39] <BigJohnT> hmmm don't have one of those but it should be the same as any other input voltage
[19:39:40] <FinboySlick> Hello gang. On the 'Experimental' level, is there any USB support at all in the HAL? I've had a silly-that-just-might-work idea lately.
[19:40:17] <Gamma-X> BigJohnT im assuming 5v?
[19:40:18] <BigJohnT> yea do a google on joystick
[19:40:29] <cradek> not realtime, but there is some userspace stuff for all kinds of input devices
[19:40:48] <BigJohnT> if your other inputs are that I'd assume that
[19:41:23] <BigJohnT> finboy google handwheel too
[19:42:36] <FinboySlick> cradek: I know it's almost blasphemous in terms of machine control, but Logitech has mice that supposedly do 3200dpi sampling with 1khz axis reports... I'm wonderin if there'd be a way to use something like that as an 'encoder' of sorts.
[19:42:41] <lerneaen_hydra> cradek, how slow can userspace/usb be, worst case (assuming RT doesn't take up 100% cpu)
[19:44:09] <alex_joni> lerneaen_hydra: USB is 1msec packet rate
[19:44:15] <cradek> lerneaen_hydra: processes running in userspace can fail to run for an infinite amount of time, but in practice pauses are usually much shorter than that
[19:44:23] <alex_joni> so 1000 updates/second in the best case
[19:44:50] <lerneaen_hydra> so there's no real guarantee that anything will happen at all? not even soft-rt?
[19:45:48] <alex_joni> lerneaen_hydra: does USB work with soft-rt?
[19:45:53] <alex_joni> * alex_joni doesn't know
[19:45:57] <lerneaen_hydra> I don't have any idea
[19:46:06] <lerneaen_hydra> I just assumed it was soft-rt
[19:47:21] <Gamma-X> does anyone here have a tool touch off sensor?
[19:47:46] <FinboySlick> lerneaen_hydra: My reasoning behind is is that since a mouse does some of its own 'processing' it'll know where it is even if it skips a report and the next set of coordinates will still be accurate.
[19:49:38] <BigJohnT> gamma you hooking a tool touch off sensor up
[19:49:46] <lerneaen_hydra> FinboySlick, true, though you won't know for sure when the data reported by the mouse is available to realtime programs
[19:50:10] <lerneaen_hydra> though mostly it will be acceptable, seeing as how the joystick functionality seems to work
[19:50:41] <lerneaen_hydra> I'd be rather cautious when controlling iron with non-rt based controls though
[19:50:58] <FinboySlick> lerneaen_hydra: My goal would be some sort of a way to cheaply close the loop for, say, a stepper drive.
[19:51:36] <FinboySlick> EMC could notice that it skipped a step here and there, etc.
[19:51:53] <lerneaen_hydra> oh, I see
[19:51:58] <JymmmEMC> FinboySlick: jmkasunich's standard response to that.... "What do you expected emc to do if it does miss a step?
[19:52:41] <FinboySlick> JymmmEMC: Repeat it?
[19:52:48] <JymmmEMC> FinboySlick: how?
[19:52:53] <lerneaen_hydra> when it misses a single step it will* miss all consecutive steps after that too, untill the end of the move. *maybe not all moves, like moves on a machine without acceleration
[19:53:21] <lerneaen_hydra> but in general I've found that when a motor stalls it stalls completely
[19:53:28] <lerneaen_hydra> and it's very clear when that happens
[19:53:34] <lerneaen_hydra> a bit too clear sometimes
[19:53:43] <BigJohnT> that's been my experiance when it stalls it quits'
[19:54:05] <JymmmEMC> lerneaen_hydra: I've found that sometimes when it does stall, the motor keeps going.
[19:54:20] <JymmmEMC> and I can fix it if I catch it RIGHT THEN
[19:54:20] <lerneaen_hydra> what do you mean by keeps going?
[19:55:08] <JymmmEMC> lerneaen_hydra: It stalls (I hear it), then keeps chugging thru the rest of the gcode, just usually 1/2" off from where X (or y) thinks it should be.
[19:55:33] <lerneaen_hydra> JymmmEMC, hmm, interesting
[19:55:52] <BigJohnT> jymmm what kind of machine?
[19:56:03] <JymmmEMC> gantry router
[19:56:09] <JymmmEMC> 24" x 24" x 6"
[19:56:13] <JymmmEMC> travel
[19:56:39] <BigJohnT> does it stall because of overloading or too much speed?
[19:56:58] <JymmmEMC> good question
[19:57:44] <BigJohnT> my 48 x 48 plasma cutter will stall if I get the accel up too high during direction reversing
[19:58:08] <BigJohnT> but i'm going 400IPM...
[19:58:49] <BigJohnT> once I got the accel correct no more stalling
[20:00:13] <BigJohnT> jymmm what are you using to drive the gantry?>
[20:18:34] <danielbr> bbl
[20:58:44] <JymmmEMC> ballscrews
[21:07:24] <JymmmEMC> These are good, just made them.... http://www.foodnetwork.com/food/recipes/recipe/0,,FOOD_9936_31593,00.html
[21:37:26] <Gamma-X> if I made a touch probe would emc2 be able to auto sense the length of the probe device?
[21:48:55] <jepler> if you have a place on the table that you want to define as Z=0 (or Z=1 or whatever), then one way to do that is to probe down towards that location then use the resulting probe value to calculate a tool length for G43.1.
[21:50:30] <Gamma-X> jepler good idea
[21:58:09] <JymmmEMC> LawrenceG: ping
[21:58:42] <LawrenceG> hey
[21:58:48] <JymmmEMC> LawrenceG: Have you updated the wiring diagram for your system by chance? Got a newer PDF I could look ar?
[21:58:48] <JymmmEMC> at
[21:59:33] <JymmmEMC> LawrenceG: I'm still stuck at which BOB to get, but I still need to order the other controls/lamps and stuck for soem ideas on the design.
[22:00:11] <LawrenceG> no problem... I'll stick the pdf file somewhere
[22:00:36] <JymmmEMC> thanks
[22:01:12] <LawrenceG> I can email it to you....
[22:03:25] <JymmmEMC> LawrenceG: I found these neat LED lamps I will be using... http://cgi.ebay.com/NEW-COMPACT-PILOT-LIGHT-22MM-BLUE-LED-AC-DC_W0QQitemZ150191506961QQihZ005QQcategoryZ42898QQcmdZViewItem
[22:03:42] <JymmmEMC> LawrenceG: lens colors available are red grn, blu, yel, and wht
[22:04:11] <LawrenceG> those look very nice
[22:04:22] <JymmmEMC> LawrenceG: and they're LED's
[22:04:32] <JymmmEMC> various voltages up to 220VAC
[22:04:37] <LawrenceG> I see that guy has contactors, slectors etc
[22:05:29] <LawrenceG> you could set up your control power circuit using 12,24,120 volts...
[22:06:25] <LawrenceG> what voltage do you drive your steppers with 48volts?
[22:06:28] <JymmmEMC> LawrenceG: Yeah, well all the lights will be the same look/feel, but can have various voltages for the bulbs
[22:06:36] <JymmmEMC> Yeah, 48V
[22:07:30] <LawrenceG> you should get one email direct and one from usendit or something like that
[22:07:36] <JymmmEMC> LawrenceG: here's teh replacement bulbs... http://cgi.ebay.com/NEW-HQ-BLUE-LED-LAMP-BULB-24V-AC-DC_W0QQitemZ150189224738QQihZ005QQcategoryZ42898QQcmdZViewItem
[22:08:35] <JymmmEMC> LawrenceG: here's the blue pilot lamp disaseembled.... http://cgi.ebay.com/NEW-COMPACT-PILOT-LIGHT-BLUE-22MM-LED-BULB-IS-AC-DC_W0QQitemZ150137819603QQihZ005QQcategoryZ42898QQcmdZViewItem
[22:09:53] <LawrenceG> nice
[22:10:19] <JymmmEMC> I thought so. No bulbs to replace all the time.
[22:10:55] <JymmmEMC> maybe not oil tight, but much cheaper for hobby grade
[22:11:31] <LawrenceG> yes... whole assembly is only $1 more than bare bulb
[22:11:55] <JymmmEMC> spare parts?
[22:11:58] <JymmmEMC> shurg
[22:12:01] <JymmmEMC> shrug
[22:12:34] <JymmmEMC> 22mm is a nice size too
[22:12:50] <JymmmEMC> 35.4mm == 1"
[22:12:54] <JymmmEMC> 25.4
[22:17:03] <fenn> looks like a two-liter cap
[22:29:45] <alex_joni> good night all
[22:30:53] <JymmmEMC> nite alex
[22:31:04] <JymmmEMC> fenn: yeah, seems about right
[23:23:08] <BigJohnT> when I calculated the SCALE it came out to 4000 and the OUTPUT_SCALE = 1.000 but when testing the SCALE needed to be 3970 to get perfect travel
[23:23:23] <BigJohnT> is this common to have to tweek the scale
[23:25:12] <Ziegler> no
[23:25:24] <Ziegler> you might have some black lash problems
[23:25:28] <BigJohnT> this is a timing belt application
[23:25:37] <BigJohnT> 0.0005" backlash
[23:27:12] <BigJohnT> I'm thinking it was not as perfect of pitch diameter as stated of the timing belt pulleys
[23:27:32] <Ziegler> that might be
[23:27:51] <Ziegler> toothed belts?
[23:28:18] <Ziegler> if they are... then diameter shouldn't be it
[23:29:29] <BigJohnT> yes xl belts
[23:29:31] <BigJohnT> http://i47.photobucket.com/albums/f163/johnplctech/Plasma%20Cutter/P1010049.jpg
[23:31:08] <BigJohnT> I'm not loosing steps as I can go in and out with a dial indicator on the frame and return to zero after moving .800 and back
[23:31:17] <BigJohnT> 1" travel dial indicator
[23:31:33] <Ziegler> neat
[23:31:52] <BigJohnT> at 4000 it would move .101 for a commanded move of .100
[23:31:54] <BigJohnT> thanks
[23:32:16] <Ziegler> I dont have a clue then
[23:32:16] <BigJohnT> it really works slick, I can get 390 IPM from it
[23:33:07] <BigJohnT> at 3970 it moves .1000 for a command of .1000
[23:34:33] <fenn> that's less than one percent error, which is pretty good
[23:35:11] <fenn> anyway, what matters most is repeatability
[23:35:51] <BigJohnT> I think the combination of three pulleys and two belts just has some not quite perfect dimensions...
[23:35:59] <fenn> xl belts are designed for power transmission, not motion control
[23:36:02] <Skullworks-PGAB> B John: you put some time in building that
[23:36:08] <BigJohnT> I can run it out yea I know LOL
[23:36:27] <BigJohnT> yea uploading some more pics in a sec
[23:36:28] <fenn> use GT (round tooth) for zero backlash, and they're probably tighter control on the dimensions
[23:37:00] <BigJohnT> didn't see gt belts
[23:37:19] <fenn> er, maybe it's called "powergrip" or something
[23:37:26] <fenn> the point is the teeth are round
[23:37:30] <BigJohnT> i've seen them
[23:37:47] <BigJohnT> they are better for motion control...
[23:38:23] <fenn> would .0005 even matter for plasma cutting?
[23:38:50] <BigJohnT> no
[23:39:09] <BigJohnT> +- 0.010 is good
[23:40:11] <BigJohnT> i'm getting +- 0.001 now with the xl belts
[23:43:34] <BigJohnT> here is the belt connection and end block
[23:43:35] <BigJohnT> http://i47.photobucket.com/albums/f163/johnplctech/Plasma%20Cutter/P1010052.jpg
[23:43:48] <BigJohnT> the belt gripper
[23:43:49] <BigJohnT> http://i47.photobucket.com/albums/f163/johnplctech/Plasma%20Cutter/P1010053.jpg
[23:45:25] <BigJohnT> I know why I didn't pick the GT belts now they are MM
[23:49:49] <fenn> that doesn't matter for CNC
[23:51:03] <BigJohnT> McMaster Carr didn't have them in belting by the foot so that swayed me the most.
[23:51:40] <BigJohnT> my thinking at the time was to get an even division of steps to the inch
[23:59:55] <JymmmEMC> LawrenceG: got it thanks, will look at it tonight.