#emc-devel | Logs for 2007-12-13

[00:59:10] <skunkworks_> skunkworks_ is now known as skunkworks
[02:24:54] <cradek> woo!
[02:24:58] <cradek> C axis done
[02:36:41] <SWPadnos> hmmm. C?
[02:37:12] <SWPadnos> so it's parallel to Z, not parallel to the spindle (which would otherwise be the same thing)?
[02:42:20] <cradek> I don't understand the question
[02:42:37] <SWPadnos> well, C is usually parallel to Z, which on a 3-axis mill also means it's parallel to the spindle
[02:42:45] <cradek> right
[02:42:51] <cradek> Ifor some meaning of parallel)
[02:42:52] <cradek> (
[02:43:04] <SWPadnos> but that also means it's not as useful as other rotaries, except for something like knife orientation
[02:43:29] <cradek> right that would give you the degree of freedom you don't care about (around the axis of the tool)
[02:43:30] <SWPadnos> parallel == axis of rotation parallel to specified cartesian axis
[02:43:34] <SWPadnos> right
[02:43:36] <cradek> since tools are round it's wasted
[02:43:44] <SWPadnos> except for non-round tools ;)
[02:44:14] <SWPadnos> but this one is actually on the table, and its axis of rotation is parallel to the Z axis
[02:44:24] <cradek> yes
[02:54:45] <cradek> what excellent advice from gene
[02:56:07] <SWPadnos> indeed
[02:56:39] <cradek> I would not have thought to warn him about grounding the motor leads. It's just second nature to me but that's hard-won habit
[02:57:05] <SWPadnos> I thought of it briefly when I read the original pot, but I didn't look hard enough to actually respond
[02:57:07] <SWPadnos> post
[03:51:22] <steves_logging> steves_logging is now known as steve_stallings
[03:55:06] <cradek> jepler: I'm not so sure...
[03:55:42] <cradek> well it's probably not emc2.
[03:56:17] <jepler> cradek: yes, I wouldnt' be surprised if it was from nist/pd emc
[03:57:53] <jepler> here's an interesting gcode ajaxcnc supports: G29 Return from Reference Point
[03:57:54] <SWPadnos> odd. that may be a derivative of XEMC
[03:58:15] <cradek> huh
[03:58:32] <SWPadnos> they do say it's got a 1GB CF hard drive preloaded with Linux
[03:58:34] <SWPadnos> ...
[03:59:33] <cradek> http://ajaxcnc.com/images/kits_millkit.jpg
[03:59:39] <cradek> some isa card
[04:00:00] <jepler> ISA? wow.
[04:00:03] <SWPadnos> and fiber optic cables, don'tcha know
[04:00:21] <SWPadnos> for pure digital communication, of course
[04:00:28] <jepler> eh? they look like ribbon cables to me.
[04:00:42] <SWPadnos> the ebay auction says fiber / digital ...
[04:01:10] <cradek> is that an stg?
[04:01:29] <cradek> no they had a big connector on the outside I think
[04:03:05] <jepler> "Our latest software release for Linux also features a new high-resolution (1024x768) display for today's LCDs and monitors."
[04:03:50] <jepler> (yep they are still on isa. FAQ 13: Where can I buy a new ISA motherboard?)
[04:03:58] <SWPadnos> heh
[04:05:16] <jepler> ahahaha if you want a different function in the PLC you fill out a form and presumably pay for each revision you get wrong. http://ajaxcnc.com/downloads/plcbuildsheet.pdf
[04:05:51] <cradek> look at their supported network cards... ouch
[04:06:21] <SWPadnos> take a look at their accessories page if you want to see some really inflated prices
[04:06:29] <cradek> 3c509, Inter [sic] EtherExpress Pro
[04:06:34] <cradek> these are 10 year old cards
[04:06:45] <SWPadnos> as are ISA motherboards
[04:07:51] <cradek> neat, touch-off that includes your (configurable) edge finder offset
[04:09:05] <SWPadnos> only $850
[04:09:18] <cradek> ... or 3 lines of python code
[04:10:06] <SWPadnos> they do have a couple of interesting add-ons: coordinate system rotation, scaling and mirroring ...
[04:10:11] <cradek> I might have to do that (touch off from right/back, from left/front)
[04:10:20] <cradek> how to do it well in the gui is the hard part
[04:10:24] <SWPadnos> oh, and don't forget the extra $700 for "multiple work coordinate systems G54-G59)
[04:11:08] <SWPadnos> plus $1495 for "unlimited program size", instead of the 1M they allow in the standard package
[04:11:19] <SWPadnos> $1000 for rigid tapping
[04:11:32] <cradek> geeezzz
[04:11:32] <cradek> we could do (planar) rotation and (isotropic) scaling pretty easily
[04:11:32] <cradek> apparently nobody cares enough to do it
[04:11:33] <SWPadnos> man, we could be making some money if EMC2 were for sale! :)
[04:11:46] <SWPadnos> it's been requested several times
[04:11:59] <cradek> apparently nobody who can do it cares enough to do it
[04:12:01] <SWPadnos> but it's true, none of the develpers has cared enough to get it done
[04:12:04] <SWPadnos> right
[04:12:23] <SWPadnos> see, that's the problem with open source! :)
[04:12:49] <cradek> I'd sure as hell do it for $4000 or some other crazy amount
[04:12:55] <cradek> so, no more open source problem
[04:13:00] <SWPadnos> heh
[04:13:17] <SWPadnos> oh look, they charge $325 for the spindle "S" function
[04:13:40] <SWPadnos> "You will even have spindle speed control on the fly while running a program."
[04:14:26] <cradek> wowie
[04:42:53] <steve_stallings> I'm slow getting over to read developer tonight... but re: AJAX.... typical sell the perceived value
[04:43:23] <steve_stallings> they typically will bundle the expensive add-ons at greatly reduced prices in order to make a sale
[04:44:02] <steve_stallings> current systems do use optical only to connect computer to their control/interface/amplifier unit
[04:46:10] <steve_stallings> the Ajax stuff is mostly the same as the Centroid, just a matter of dollars paid and a magic setting in the softare configuration
[04:48:20] <SWPadnos> they list Centroid as one of their component suppliers
[04:51:13] <steve_stallings> interesting, they are just different arms of the same company
[04:51:47] <SWPadnos> I had run across their site(s) when I was looking for BP servo mounts
[04:52:04] <SWPadnos> they either had no prices, or exprbitant prices - I can't remember
[04:52:07] <SWPadnos> exorbitant
[04:52:13] <steve_stallings> Centroid has "dealers", AJAX is "do it yourself"
[04:52:36] <SWPadnos> so AJAX is the DIY Centroid dealer then?
[04:53:02] <steve_stallings> basically, there may be a few high end features reserved for Centroid only
[04:53:45] <SWPadnos> at the prices they're charging for "add-ons", I don't think they need to leave anything out
[04:53:46] <steve_stallings> setting the correct magic paramater on an AJAX control will get you a Centroid splash screen
[04:53:53] <SWPadnos> heh
[04:54:13] <SWPadnos> I think it's pretty funny (and nearly criminal) that they charge $325 to enable the S word
[04:55:03] <steve_stallings> if you are making your initial purchase, you can get a lot of freebies or discounts on the fancy stuff, but look out if you want it later
[04:55:23] <SWPadnos> well, I can see coordinate transforms as fancy, but not S
[04:56:26] <steve_stallings> the S word is not fancy itself, its the tech support that is required when the user tries to get it interfaced
[04:56:34] <SWPadnos> heh
[04:57:07] <steve_stallings> at least AJAX has some support on the phone
[04:57:28] <steve_stallings> my friend that installed one is no rocket scientist and they got him through it
[04:57:29] <SWPadnos> without taking a high-limit credit card number?
[04:58:33] <steve_stallings> as best I know, yes, and they have fixed price exchanges on modules for repairs, think I remember $450 for the interface/amp module
[04:59:28] <steve_stallings> friend paid for jog wheels on one machine, figured out how to do it himself on second
[05:00:01] <SWPadnos> at least a jog wheel is a hunk of hardware
[05:00:40] <steve_stallings> a small hunk, but yes the did have to ship something besides an email with a magic code
[05:01:26] <SWPadnos> that's a funny thing about the 6000 series of scopes from Agilent (the 1 GHz one I have)
[05:01:47] <SWPadnos> they actually have the 8M deep memory option installed, it's just a code to unlock it
[05:01:55] <steve_stallings> don't tell me they sell bandwidth by the MHz
[05:01:56] <SWPadnos> when you spend the $4000 or so on the option
[05:02:24] <SWPadnos> no, you can't get a 597Mz sxope if you're just short of $ for that 600 MHz one ;)
[05:03:04] <steve_stallings> old trick, Burroughs and IBM did it with mainframes back in the 60's and 70's
[05:04:03] <steve_stallings> pay $100K and tech comes out and changes crystal in clock ckt
[05:04:17] <SWPadnos> presto - faster mainframe!
[05:05:21] <steve_stallings> as a vendor I have mixed feelings, charge for hardware or charge for performance, which?
[05:05:40] <SWPadnos> I think I prefer the "fair markup" scheme
[05:05:55] <SWPadnos> but I'm sure I could make more money with the "what the market will bear" scheme
[05:06:30] <SWPadnos> I'm happier with myself (at least until I wish I had a retirement fund) charging what's fair vs. what I can gouge
[05:06:31] <steve_stallings> then the low end cost will rise to support the extra development to make the higher performance stuff work
[05:07:28] <SWPadnos> well, if you have a stratified market (with products based on similar platforms), you can certainly shift a little cost out of the low end and into the high end
[05:07:38] <SWPadnos> err - price, not cost ;)
[05:07:38] <steve_stallings> and you MUST have the higher performance stuff or Tim Allen will shop elsewhere
[05:07:45] <SWPadnos> heh
[05:11:46] <steve_stallings> steve_stallings is now known as steves_logging
[05:44:30] <CIA-42> EMC: 03seb 07TRUNK * 10emc2/src/hal/drivers/ (mesa7i43-gpio.h mesa7i43_gpio.comp): Improved timing of I/O operations during module load/board init.
[14:55:58] <alex_joni> good morning guys
[14:56:02] <alex_joni> jepler: buzz me when around
[14:56:16] <jepler> alex_joni: what's on your mind?
[15:00:53] <alex_joni> got a nice email from Daniel Sheeren
[15:01:06] <alex_joni> mind if I paste the relevant bits in here?
[15:01:19] <alex_joni> you probably know the stuff.. but it's my moral duty to report it :D
[15:01:52] <alex_joni> "Everythingl is working flawless now. I have had problems with very long loops (+1000) using o-word. I think Axis live plot overflow ram memory when using onboard video card and when using some 128mb ati pci card after some 15 min or so emc crash if if you not clear the live plot. When I disable emc live plot in Axis, I can work 8 hours running 8000 loops without problems. "
[15:02:36] <alex_joni> ""I running a compiled pre 2.2 in this machine. I don't know if this problem is present in the last 2.2 release."
[15:02:49] <jepler> I don't recall anything about the backplot changing
[15:04:02] <jepler> the backplot is supposed to be limited to consume no more than 1.6MB RAM, but before that limit is reached, redraw performance gets pretty low. Crashing after just 15 minutes is very surprising to me.
[15:04:46] <cradek> turning off backplot doesn't keep it from saving the points, it keeps it from displaying
[15:04:50] <jepler> that's true too
[15:06:45] <jepler> hm, hitting Escape to stop the preview plot loading seems to have broken for me :(
[15:19:10] <cradek> I did that yesterday when I wrote infinite repeat[] loops
[15:19:38] <jepler> huh
[15:20:02] <jepler> I just get EMC_TASK_ABORT in debug but it doesn't stop loading
[15:20:30] <cradek> I may have been in estop
[15:21:10] <jepler> I dunno why that would make a difference..
[15:43:00] <alex_joni> http://www.youtube.com/watch?v=0RH_H7jgQQY
[17:46:47] <steve_stallings> steve_stallings is now known as steves_logging
[19:26:20] <alex_joni> does override limits work for any of you?
[19:26:39] <cradek> sure
[19:26:46] <cradek> well I haven't tried it very recently
[19:28:11] <alex_joni> I got a user reporting it doesn't work with 2.2
[19:28:18] <alex_joni> and I just tried it with TRUNK
[19:28:32] <cradek> nice. on mcmaster you can say "I ordered here before" and enter only "your" email address, and it shows all your personal information
[19:28:34] <alex_joni> I see the message gets passed to motion
[19:28:59] <alex_joni> but the status somehow doesn't get updated?
[19:29:07] <alex_joni> (both with tkemc and axis)
[19:29:45] <cradek> huh, you're right
[19:29:56] <cradek> the checkbox doesn't come on
[19:30:29] <alex_joni> I hate beeing right :/
[19:30:46] <cradek> wtf
[19:30:54] <cradek> I know that worked semi-recently
[19:31:10] <cradek> jmk and I touched it at fest
[19:31:16] <cradek> but I don't remember what we did
[19:31:38] <cradek> I think we only override the limits that you're currently on. maybe it works if you're actually ON one
[19:32:33] <cradek> yes, it does work right
[19:32:38] <SWPadnos> I remember a discussion about that - only allowing motion off the limit that's activated
[19:32:41] <cradek> turn on a limit switch and you'll see
[19:32:42] <alex_joni> you're right
[19:33:00] <cradek> it also only lets you jog the right way
[19:33:06] <cradek> it's very good
[19:33:07] <alex_joni> ok, thanks for making it clear :)
[19:33:09] <alex_joni> yes..
[19:33:11] <SWPadnos> one issue was what to do when you have shared +/- limit switches - I don't remember what was done there
[19:33:20] <cradek> then it lets you jog either way
[19:33:25] <cradek> assuming we got it right
[19:33:27] <SWPadnos> well, that sounds easy :)
[19:33:30] <cradek> I know we discussed that
[19:34:09] <alex_joni> yes, it works as it should
[19:34:11] <alex_joni> cradek: thanks
[19:34:25] <cradek> welcome
[19:39:21] <cradek> I lied. I must have had a cookie or something.
[19:40:01] <alex_joni> ??
[19:40:29] <SWPadnos> mcmaster keeping personal info
[19:40:36] <alex_joni> ah
[19:40:48] <alex_joni> I actually thought about a homemade cookie
[19:41:02] <alex_joni> didn't see how that relates to limit overrides :D
[19:41:10] <SWPadnos> or donuts
[19:41:32] <cradek> wow, the 'precision' acme threaded rod is ".007 or better per foot"
[19:41:39] <cradek> that doesn't seem all that great
[19:41:48] <SWPadnos> then nook isn't the cheapo supplier
[19:41:56] <SWPadnos> their precision spec was 0.0003/ft
[19:42:07] <cradek> this isn't the cheapo, it's $25
[19:42:19] <SWPadnos> then nook isn't the non-cheapo supplier ;)
[19:45:19] <fenn> mmmm. cookies.
[19:46:00] <fenn> * fenn settles for a little debbie star-crunch
[19:46:21] <alex_joni> * alex_joni had some great homemade (grandma) cookies lately