#emc | Logs for 2009-08-06

[00:01:40] <toastatwork> geo01005_home I had 6 = Ax + A + Bx - B
[00:01:54] <toastatwork> i only used one value for x, giving me only one equation, obviously
[00:02:11] <toastatwork> needed to go back and do it again with another value, and now it works
[00:25:48] <gezar> howdy yall
[08:19:34] <pjm> good morning
[08:19:52] <archivist> its morning? damn
[08:23:47] <micges> good morning
[08:25:26] <alex_joni> 'lo
[09:41:57] <piasdom> g'mornin all
[12:49:09] <Valen1> Valen1 is now known as Valen
[13:27:17] <jymm> Could someone tell the width of the flex strip by this photo? http://dealextreme.com/customerphotos/5111-2332d228-982d-4b01-a3cd-26ab8180f434.jpg
[13:28:07] <archivist> you have a keyboard and a pair of calipers and open office
[13:28:38] <jymm> More so by the dimensions of the resistor
[13:29:00] <gezar> the flex strip is 1.1 keys wide
[13:29:27] <SWPadnos> the resistor could be an 0402, 0603, 0805, 1210, or 1812
[13:30:08] <cradek> 1cm
[13:30:34] <jymm> http://dealextreme.com/customerphotos/5111-e7ca05bc-540e-48b9-a7d7-d499ea81d2c4.jpg
[13:30:39] <jymm> does that help?
[13:30:43] <gezar> how have you guys all been?
[13:31:15] <SWPadnos> so fine
[13:31:19] <SWPadnos> err, just fine
[13:31:27] <SWPadnos> jymm, no, it doesn't help
[13:31:29] <jymm> cradek: (I'm hoping it's not that wide)
[13:31:39] <SWPadnos> 8mm, using the caliper method
[13:31:42] <gezar> hehehe, Im sorry, I didnt realize what time it was, ill be back
[13:32:16] <cradek> 10 +- 3 mm
[13:32:31] <SWPadnos> 11 +/- 5 mm :)
[13:32:41] <cradek> 10 +- 11 mm
[13:32:49] <SWPadnos> agreed!
[13:32:51] <archivist> .4" ish
[13:32:56] <jymm> http://dealextreme.com/details.dx/sku.5111
[13:33:32] <cradek> maybe you should ask them
[13:34:21] <SWPadnos> it.s 0.31 inches wide
[13:34:24] <SWPadnos> bonehead
[13:34:46] <SWPadnos> 43.31 long by 0.31 wide by 0.08 thick, weighs 0.49 oz
[13:35:16] <archivist> dont go reading the spec thats cheating
[13:35:31] <SWPadnos> too late. I guess I'm disqualified
[13:37:06] <archivist> cradek, next escape wheel I have to make 3.2mm od 11 teeth verge for a ladies pocket watch, worn out one arrived here today
[13:39:32] <jymm> Can that type of FlexPCB be bent to wrap around a 1/8" thick material? I haven't worked with FlexPCB like that
[13:40:07] <SWPadnos> you have to be really careful about two things
[13:40:18] <SWPadnos> 1, the traces like to break if the bend radius is too small
[13:40:44] <SWPadnos> (you didn't mention whether it's a 1/8" rounded corner, or just going around something like a flat bar)
[13:41:01] <SWPadnos> 2, the components tend to come unstuck when you flex the pads
[13:41:05] <jymm> I an deal with corners accordingly
[13:41:18] <jymm> s/an/can/
[13:41:22] <SWPadnos> that also happens with standard PCBs, but it's much worse on flex
[13:41:46] <SWPadnos> 1/8 is pretty small, so I'd be kind of surprised if it works well
[13:41:57] <jymm> I mean wrap the FlexPCB arond 1/8" sheet material like a Taco
[13:42:09] <jymm> the pcb begn the taco shell
[13:42:50] <SWPadnos> it's the bend radius that matters, so if you're talking about rounding a 1/8" thick piece of material, you have a radius of 1/16", which is probably not good
[13:43:47] <jymm> ok, thank you
[13:43:59] <jymm> * jymm throw the drawing board in the crash
[13:44:03] <jymm> trash
[13:44:09] <jymm> ;)
[15:30:00] <EbiDK|AWAY> http://thedailywtf.com/Articles/We-3-Concurrent-Engineering.aspx Duh!
[16:12:16] <mikeggg> is there a way to make the parallel port pins 1-9 inputs instead of outputs?
[16:12:49] <SWPadnos> set the direction to in
[16:13:02] <SWPadnos> it's in the hal_parport manual (man hal_parport)
[16:29:33] <geo01005> is the development on hagui dead?
[16:29:40] <geo01005> halgui....
[16:33:18] <cradek> do you mean halui?
[16:35:15] <geo01005> uhh...
[16:37:15] <geo01005> no, halgui... There is a note about it in the hal manual that says "GUI tool for configuration and tuning (Not implemented yet)"
[16:37:47] <geo01005> Looks the like code has not been worked on since 2006. /src/hal/utils/halgui
[16:39:05] <cradek> yeah, there is no such program that I know of
[16:39:55] <geo01005> pyvcp uses halui right?
[16:40:22] <cradek> they are not necessarily related - but sometimes used together
[16:41:12] <geo01005> ohh, halui does NML stuff...
[17:01:39] <alex_joni> geo01005: halui is supposed to be an emc2 interface driven by HAL pins
[17:01:55] <alex_joni> instead of clicking on things, you just hook up buttons & knobs & leds to it
[17:03:24] <geo01005> I see. So it doesn't really have anything to do with a gui, except for the fact that you can use it with virtual buttons and knobs with something like pyvcp.
[17:06:24] <alex_joni> right
[17:06:29] <alex_joni> it's actually halui not halgui :D
[17:10:30] <geo01005> but there is some code for something called halgui.
[17:12:06] <geo01005> Daniel Nilsson worked on it back in 2005 and 2006.
[17:12:26] <geo01005> looks like he didn't get too far.
[17:17:17] <alex_joni> nope, that should probably just get removed
[17:17:40] <archivist> or geo01005 finishes it :)
[17:18:29] <geo01005> some day :)
[17:19:40] <geo01005> for most application I'm not sure that having a gui hal editor or a schematic hal editor is all that helpful.
[17:21:20] <GammaX> I think a configuartion gui would be more astheticly pleasing.
[17:23:25] <alex_joni> geo01005: one of the reasons it never got done
[17:25:14] <geo01005> I suspected that.
[19:15:32] <geo01005> seb_kuzminsky, what happens when you have a joint running on a hm2 stepgen in velocity mode?
[19:15:52] <geo01005> Does the position f-b eventually overrun?
[19:20:45] <seb_kuzminsky> geo01005: i guess pos-fb *would* eventually overrun, but it's a ways out there, and it's no closer in than the same joint in position mode
[19:21:28] <geo01005> hmm, I don't want my machine to have windows syndrome.
[19:22:06] <geo01005> It shouldn't need to be restarted every couple of weeks because a floating point number is getting large.
[19:22:15] <seb_kuzminsky> well, there's only 32 bits to keep track of stepper counts, so it will overflow eventually
[19:22:22] <seb_kuzminsky> what are you trying to do?
[19:22:45] <geo01005> So I have a joint that drives a wire feed motor.
[19:23:32] <geo01005> every so often (30 mins or so) I rehome this joint.
[19:24:06] <geo01005> but I can't reset the stepgen at this time.
[19:24:18] <seb_kuzminsky> in vel mode, position-fb is not used by the stepgen driver at all, it's exported to HAL for whatever good it does up there
[19:24:24] <geo01005> So I was just calculating a offset value (Float)
[19:24:47] <geo01005> Yeah, but my homing procedure needs position feedback.
[19:25:42] <geo01005> I guess I should change my homing procedure to use the count value instead of the position-fb.
[19:26:14] <seb_kuzminsky> would a stepgen.reset pin, analogous to encoder.reset, be helpful?
[19:26:29] <seb_kuzminsky> wait
[19:26:36] <geo01005> it would...but I don't know if it is worth you doing it.
[19:26:38] <seb_kuzminsky> isnt homing done above hal?
[19:27:03] <geo01005> yeah, sorry... once again this system doesn't use emc.
[19:27:06] <geo01005> just the hal
[19:27:55] <seb_kuzminsky> ah, so "homing" is done in your code above hal
[19:28:30] <geo01005> Yeah, I created a hal component that does the homing and motion control for each of the joints in the machine.
[19:29:25] <geo01005> I suppose I should just change it to use counts... I suppose that is the more correct way anyhow.
[19:29:49] <seb_kuzminsky> pos-fb has sub-step precision, which is sometimes what you want and sometimes not what you want...
[19:30:42] <geo01005> it was just easy to use because it was in position units.
[19:31:06] <seb_kuzminsky> pos-fb is float, counts is s32, so counts is what may overflow and wrap around
[19:31:22] <geo01005> ahh, I see.
[19:31:26] <geo01005> That would be bad.
[19:31:40] <geo01005> well it would actually be ok.
[19:31:56] <geo01005> as long as it didn't happen in the middle of homing a joint.
[19:32:11] <seb_kuzminsky> what does homing mean in this context?
[19:32:47] <geo01005> Well so the wire is being melted at a fairly constant rate,
[19:33:15] <geo01005> so for the most part, for a short period of time I have a good estimate of where the end of the wire is.
[19:33:39] <geo01005> however after some time I need to back the wire up and "Home" it with a proximity sensor.
[19:34:07] <geo01005> make sense?
[19:34:54] <geo01005> ( the position of the wire determines how fast it is melted)
[19:34:55] <seb_kuzminsky> you move the wire-feeder thingy over to a touch-sensor thingy and feed it out until it trips the switch, then you call that home?
[19:35:03] <geo01005> yeah.
[19:35:26] <geo01005> or at least a know position.
[19:35:30] <geo01005> known
[19:35:48] <seb_kuzminsky> so if it zeroed the counts (and hence the position-fb) when the touch-sensor tripped, that'd be you taken care of?
[19:36:03] <geo01005> yeah.
[19:36:23] <geo01005> But I can do that will a little more effort in my component.
[19:36:36] <geo01005> I'm not sure that it is useful for anyone else.
[19:38:56] <geo01005> hmm, I just calculated how long it would take for the count s32 to overun....about 10 years...
[19:39:21] <geo01005> fairly slow feedrate.
[19:39:57] <geo01005> I think that it will be ok.
[19:40:12] <geo01005> I think that the lifetime of this machine will be about 5 years.
[19:40:29] <geo01005> And it will certainly be turned on and off more than once in that time.
[19:41:19] <geo01005> it is thinking like this that leads to y2k isn't it?
[19:41:54] <seb_kuzminsky> "by the time it overflows, we'll just switch to 64-bit machines"
[19:42:16] <archivist_attic> inc a flag to set it /me suffered that one
[19:42:59] <seb_kuzminsky> ok geo, i'll continue my time-honored tradition of not taking any action, then
[19:43:14] <geo01005> ok, thanks for the help :)
[19:43:25] <seb_kuzminsky> you mean, thanks for not helping ;-)
[19:43:28] <seb_kuzminsky> brb
[19:43:31] <geo01005> don't for get about SPI ;)
[20:46:36] <jymm> Hmmm, Did this get wefully expensive, or is it a different one than a year or so ago... http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=93885
[21:09:05] <skunkworks_> are you thinking of this? http://www.harborfreight.com/cpi/ctaf/Displayitem.taf?itemnumber=44991
[21:09:28] <skunkworks_> or this? http://www.harborfreight.com/cpi/ctaf/Displayitem.taf?itemnumber=47158
[21:11:32] <archivist> iirc the hoff thingy is based on 44991
[21:11:53] <archivist> flimsy looking column on it though
[21:18:36] <skunkworks_> it is what it is.
[21:18:44] <archivist> :)
[21:19:41] <skunkworks_> :)
[21:19:48] <archivist> I have stiffened the column on mine a lot, a serious difference
[21:19:54] <jymm> I dont see what the diff is other than chuck size
[21:20:12] <jymm> and a DC motor?!
[21:20:17] <archivist> you are not looking hard enough
[21:20:28] <jymm> nope
[21:20:47] <archivist> the column on the expensive on is far better
[21:21:20] <jymm> not $600 worth
[21:22:20] <archivist> when doing a real cut, one notices, 10/20 thou cut on the flimsy
[21:22:36] <archivist> man size probably on the other
[21:22:38] <jymm> Maybe $200, but not $600
[21:23:30] <archivist> there is a reason the expensive good stuff is bigger
[21:24:03] <jymm> Higher price != Better quality
[21:24:45] <jymm> Hmmm it's 200lbs MORE then the $400 one
[21:24:57] <skunkworks_> hmmm ;)
[21:25:13] <archivist> see what I mean by size
[21:25:32] <archivist> double the weight
[21:25:45] <archivist> 3 times the lightest
[21:26:17] <jymm> Well, for $700 more you get a WHOLE LOT MORE here, that I can see >http://www.harborfreight.com/cpi/ctaf/Displayitem.taf?itemnumber=42827
[21:26:44] <skunkworks_> round coulmns suck for cnc
[21:26:51] <archivist> round column == less good
[21:26:59] <jymm> 625lbs
[21:27:04] <archivist> but crap
[21:27:11] <archivist> I had one
[21:27:18] <jymm> It's HF, it's all crap =)
[21:27:31] <jymm> but we live with it =)
[21:27:45] <archivist> it makes a good drill
[21:28:15] <jymm> archivist: Honestly, I actually doubt that too. At least not without replacing the bearings
[21:29:14] <jymm> How difficult is it working with thin wall brass tubing? cutting drilling etc
[21:30:09] <geo01005> http://www.use-enco.com/CGI/INSRIT?PMAKA=325-5184&PMPXNO=5820408&PARTPG=INLMK32
[21:31:44] <Gamma-X> anyone own a benchtop multimeter?
[21:32:26] <archivist> jymm, depends on the brass, can be hard brass and easy to work with or soft and very sticky
[21:32:45] <archivist> most of us Gamma-X
[21:33:14] <jymm> * jymm handheld fits on the bench nicely
[21:33:17] <Gamma-X> archivist, im looking for one that is good for surface mount work. and specificly so I can hook up metering tweezers to it.
[21:33:43] <archivist> probe are irelevant to the meter itself
[21:35:00] <Gamma-X> hrmm
[21:35:05] <Gamma-X> i just want a nice set
[21:35:43] <jymm> Gamma-X: http://newboobs.com/
[21:36:18] <Gamma-X> nice
[22:16:21] <seb_kuzminsky> does this look sketchy to anyone else? Or am i just being a wuss? http://2.bp.blogspot.com/_sevIHb0O7UU/SnsZQ7lJooI/AAAAAAAAAVg/cmlMxCTVBzA/s1600-h/P8060447.JPG
[22:22:46] <bill2or3> heh, light cuts.
[23:04:46] <jymm> the heat marks?