#emc | Logs for 2007-06-06

[01:00:02] <JymmmmEMC> I'll be damn...
[01:00:33] <JymmmmEMC> I reflashed the BIOS on this mobo and now it retains this CMOS settings. Hawt Damn!
[01:01:00] <JymmmmEMC> Will have to check overnight, but has been powered off (switch, not button) for the last 3 hours.
[01:01:53] <JymmmmEMC> SWPadnos: are these FPGA cells reprogrammable?
[01:03:27] <JymmmmEMC> and whats the diff between a uC and a FPGA (seriosuly)?
[01:03:44] <JymmmmEMC> or a uC and a FPGA+Flash combo
[01:04:22] <jepler> JymmmmEMC: microcontrollers are sequential: they do one instruction after another
[01:04:46] <jepler> JymmmmEMC: FPGA is more like a circuit made of a bunch of gates (like AND, OR, NAND, etc): all the gates operate at the same time
[01:04:48] <JymmmmEMC> vs risk you mean?
[01:04:55] <JymmmmEMC> RISC
[01:05:50] <JymmmmEMC> jepler: fpga's being much harder to program?
[01:05:51] <jepler> inside the FPGA there are a fixed number of gates, and a fixed number of potential connections. At the lowest level, an FPGA binary program says which of the connections are used
[01:06:36] <JymmmmEMC> so unlike MAIN in C, there are little independant programs?
[01:06:59] <JymmmmEMC> (sorta kinda)
[01:07:02] <jepler> sorta kinda
[01:07:04] <jmkasunich> JymmmmEMC: they're only programs if you are pretending they are programs
[01:07:05] <jmkasunich> they are circuits
[01:07:20] <jmkasunich> FPGAs are for hardware designers ;-)
[01:07:39] <jepler> I wrote the firmware for the pluto-servo mostly by pretending it was a program
[01:07:45] <JymmmmEMC> can they interact with each other a la "master program" kinda thing?
[01:07:49] <jmkasunich> thats cause you're a programmer ;-)
[01:08:03] <jepler> it works but I'm certain it offends the sensibilities of a real FPGA professional, just like anything written by a programming novice offends me
[01:08:53] <jmkasunich> actually, I think the real FPGA pros are moving to (or have already moved to) the programming model, simply to deal with the complexity of the designs that you can put in todays monster FPGAs
[01:08:56] <JymmmmEMC> Hey now.... I'm DAMN PROUD of my BASIC program: 10 LPRINT CHR$(12) GOTO 10
[01:08:59] <jepler> JymmmmEMC: almost everything is done based on a clock signal -- for instance, one fragment of the source code might say "at every clock, make x be the old value of x plus one, unless x is already at its maximum"
[01:09:42] <jepler> in verilog, this would be something like: always @(posedge clk) if(x != max_x) x <= x+1;
[01:09:53] <jmkasunich> normal programming is inherently serial - execute this line of code, then the next, then the one after it.... you have to really work at it to make things appear to be happening in parallel.
[01:09:56] <JymmmmEMC> aw shit... time to goto work. Back in 90 to continue this conversation if you guys will be around.
[01:10:06] <jepler> hard to tell -- that's almost my bedtime :-P
[01:10:07] <jmkasunich> FPGAs are the opposite - they're inherently parallel - everything happens at once
[02:21:06] <Chris_sub_1> Anyone up for providing some advice on m5i20/7i33 servo PID tuning?
[02:24:27] <cradek> any specific question?
[02:28:38] <Chris_sub_1> I have the pidtest running fine, scales verified and such. What should I be looking for as I increase Pgains to indicate instability?
[02:29:13] <cradek> oscillation
[02:29:32] <cradek> is this a heavy/big or light/small machine?
[02:29:59] <Chris_sub_1> heavy/big. Bridgeport-sized Hurco factory CNC knee mill.
[02:30:23] <cradek> then your oscillation may be large and slow, and you may have to poke/jog it to get it to start
[02:30:29] <cradek> be careful :-)
[02:30:33] <Chris_sub_1> Scary fast, particularly in response to step control inputs. :)
[02:30:41] <cradek> you have estop tested and working right?
[02:31:03] <cradek> oh you are getting fast oscillation after a step?
[02:31:23] <Chris_sub_1> I was when I had an oops in the gain setting.
[02:31:43] <cradek> you're looking for barely sustained oscillation
[02:31:53] <cradek> if it dies out, increase it a tad more
[02:32:30] <Chris_sub_1> Looking how?
[02:32:51] <cradek> the screw, the axis moving, the position moving onscreen, etc
[02:33:18] <Chris_sub_1> Do you mean while it should be holding position, or while moving?
[02:34:38] <cradek> are you trying to do zeigler-nichols?
[02:35:22] <cradek> the oscillation you look for when doing ZN is instability around the setpoint while the setpoint is held constant
[02:38:25] <Ziegler> hehe
[02:39:00] <cradek> there he is now
[02:39:15] <Ziegler> my ears were ringing :-P
[02:47:28] <Chris_sub_1> cradek - Thanks. I have more reading to do. :)
[02:47:36] <cradek> welcome
[02:47:59] <cradek> if you're brave (and running trunk) you could try the pid autotuning
[02:48:10] <cradek> but that would take some reading too
[02:48:22] <Chris_sub_1> I'm not nearly that brave.
[02:48:25] <cradek> but it just does ZN, which you can do yourself
[03:31:04] <Jymmm> cradek: you use a breakout board?
[03:31:11] <Jymmm> if so, which one?
[03:35:21] <cradek> for the pluto I made one with the driver also on it
[05:13:46] <Jymmm> cradek: I dount I'll ever do anything with the pluto... I don't know ASM.
[09:11:07] <Jymmm> Hit the road jack... don't ya come back no more no more no more no more
[09:11:50] <Jymmm> great song =)
[12:16:40] <Guest843> http://www.cnczone.com/forums/showthread.php?p=305671#post305671
[12:16:50] <Guest843> Guest843 is now known as skunkworks_
[12:18:35] <skunkworks_> dan falck replied.. But my email notifactaion shows that he initally posted... http://pastebin.ca/544356
[12:18:48] <skunkworks_> :)
[12:36:05] <skunkworks_> it must be raining in jepler land.
[12:38:16] <maddash> why?
[12:38:55] <skunkworks_> cvs is down and jepler has disconnected.
[12:39:36] <maddash> what exactly is jepler, anyway?
[12:39:43] <maddash> what, not who...
[12:40:11] <skunkworks_> J. Epler.
[12:41:16] <skunkworks_> He is an artificial intellegence that comes off as being a great programmer on the IRC channel.
[12:42:47] <maddash> he wouldn't take my patches
[12:44:34] <maddash> hm, the homing system seems a bit lacking
[12:45:50] <alex_joni> that's a design feature
[12:46:09] <alex_joni> we did it like that to see how many people will notice
[12:46:31] <maddash> alex_joni: was that for me?
[12:46:45] <maddash> ^^huh?
[12:47:10] <alex_joni> maddash: yeah
[12:47:57] <maddash> alex_joni: to what "design feature" are you referring?
[12:48:11] <alex_joni> to the "bit lacking" you were referring
[12:48:44] <maddash> er, I didn't even mention what I thought was lacking..
[12:48:54] <alex_joni> yeah, I NOTICED that
[12:49:07] <maddash> oh. sarcasm.
[12:49:13] <alex_joni> only a tiny bit
[12:50:10] <maddash> well, there should be a (configurable) delay between each directional change in the homing (ie, do_homing())
[12:51:42] <skunkworks_> why?
[12:55:36] <maddash> if search_vel >> 0, for example, then an instantaneous directional change could damage some onboard equipment
[12:56:15] <maddash> ie, how your head would feel if I grabbed you by the shoulders and started shaking violently.
[12:57:04] <archivist> not if accel and decel is done properly
[12:57:08] <maddash> N.B.: ***could***
[12:57:42] <archivist> * archivist used to burn motors out with that trick
[12:57:45] <maddash> archivist: where are the appropriate config entries for the homing acceleration values?
[12:58:00] <archivist> dunno
[12:58:20] <skunkworks_> They use each axis configuration.
[12:58:45] <maddash> I don't think there are any, because, from what SWPadnos told me, homing occurs in a "free" mode of motion (as opposed to coordinated or jog)
[13:00:11] <skunkworks_> maddash: It uses the individual axiss constraints in "free" mode.
[13:00:40] <maddash> skunkworks_: where are the config entries for those?
[13:01:39] <skunkworks_> in the ini file - [AXIS_0]....
[13:02:36] <maddash> skunkworks_: that's nice. which entries in the AXIS_* section?
[13:04:22] <skunkworks_> maybe I am misunderstanding what you are looking for then. Each axis has its own max velocity,acceleration and such.
[13:05:19] <maddash> the documentation clearly states that the entries under AXIS* apply only in "coordinated" mode.
[13:06:28] <maddash> at any rate, once i've finished understanding control.c, i'll write a patch for it.
[13:08:51] <skunkworks_> The way I understand it (which could be wrong) is that the [traj] section is for coordinated mode. This gets trumped by the individual axiss constraints in each axis_* section. Each axis_* section is then used for 'free' mode.
[13:09:46] <alex_joni> at any time (free mode, coordinated mode, teleop mode) the axis_* limits should and will be obeyed
[13:09:49] <archivist> often people home at the minimum rate as they cannot decelerate but /me does not know emc fully yet
[13:09:51] <maddash> that's not what the pdf docs say
[13:09:54] <alex_joni> even during jogging
[13:09:58] <alex_joni> and during homing
[13:10:10] <alex_joni> maddash: can you be a bit more precise than "that's not what the pdf says" ?
[13:10:19] <alex_joni> URL and page?
[13:10:42] <maddash> any of the pdfs that describe the INI configs
[13:10:55] <maddash> eg, emc2 user manual.pdf
[13:11:02] <maddash> brb, battery is low.
[13:12:18] <maddash> .
[13:13:08] <alex_joni> maddash: I think you spotted a docs bug
[13:14:35] <cradek> [AXIS_*] limit the axis motion
[13:14:41] <cradek> [TRAJ] limits the tooltip motion
[13:15:25] <alex_joni> MAX_VELOCITY = 1.2 Per axis maximum velocity while coordinated motion is in effect.
[13:15:30] <maddash> then what's this "free mode" business that SWPadnos sold me?
[13:15:43] <alex_joni> cradek: that's obciously wrong
[13:16:03] <maddash> what's obviously wrong?
[13:17:14] <cradek> free mode is jogging, homing - anything that's not executing gcode
[13:17:42] <alex_joni> maddash: that statement from the docs
[13:17:47] <maddash> SWPadnos made a clear distinction between coordinated, jogging, and free...
[13:17:56] <alex_joni> maddash: there is a clear distinction
[13:18:03] <alex_joni> but they all obey the same limits
[13:18:13] <alex_joni> defined by the axis_* max_vel and max_accel
[13:18:14] <maddash> well, either cradek is wrong, or SWPadnos is wrong.
[13:18:18] <cradek> which statement from the docs is wrong?
[13:18:24] <alex_joni> MAX_VELOCITY = 1.2 Per axis maximum velocity while coordinated motion is in effect.
[13:18:36] <alex_joni> cradek: not necessarely wrong, but incomplete
[13:18:49] <cradek> which section's MAX_VEL is that?
[13:18:56] <alex_joni> that maximum velocity is also active while free mode is in effect
[13:19:02] <alex_joni> cradek: AXIS_*
[13:19:13] <cradek> yes it's always respected
[13:19:18] <cradek> the axis can never move faster than that
[13:19:27] <maddash> are you guys just pulling this info out of your asses, or do you have the code to back you up?
[13:19:50] <maddash> ^^ respectfully.
[13:19:59] <cradek> brb
[13:20:41] <alex_joni> maddash: respectfully: _we_ did code this, so I'd say we know what we're talking about
[13:21:33] <alex_joni> maddash: but I appreciate if you report things you feel are not clear enough explained in the docs
[13:21:47] <maddash> then why are you two telling me different things than cradek?
[13:21:59] <alex_joni> we're not
[13:22:06] <maddash> "you two" being SWPadnos and alex_joni
[13:22:31] <alex_joni> I'm not sure what SWPadnos told you, but cradek and I said the same thing above
[13:22:40] <maddash> logger_emc?
[13:22:47] <alex_joni> logger_emc: bookmark
[13:22:47] <alex_joni> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-06-06.txt
[13:23:11] <alex_joni> maddash: another try at clarifying it:
[13:23:31] <maddash> 09:17:44 <alex_joni> maddash: there is a clear distinction
[13:23:33] <maddash> vs.
[13:23:34] <alex_joni> the axis_* settings (max_vel and max_accel) define how fast each of the joints can move
[13:23:38] <maddash> 09:17:01 <cradek> free mode is jogging, homing - anything that's not executing gcode
[13:24:03] <alex_joni> maddash: I only said there is a clear distinction between free mdoe and coordinated mose
[13:24:06] <alex_joni> mode
[13:24:12] <alex_joni> maddash: but only that, nothing more
[13:24:49] <alex_joni> the AXIS_* settings (max_vel and max_accel) define the maximum speed and accel each joint can move
[13:24:57] <alex_joni> at no time those can be exceeded
[13:25:23] <alex_joni> not during homing (free), not during jogging (free/teleop), not during g-code (coordinated)
[13:25:34] <alex_joni> following me so far?
[13:26:32] <maddash> so, what you're (cradek and alex_joni) telling me is that coordinated mode == {gcode} and free mode == {jogging/homing}?
[13:26:44] <alex_joni> yes
[13:26:51] <maddash> SWPadnos: dude, wtf?
[13:27:05] <alex_joni> what did SWPadnos say?
[13:27:06] <maddash> alex_joni: brb, let me try out your theory.
[13:27:13] <archivist> how many steps does it stop in when it sees the home switch
[13:27:45] <maddash> alex_joni: like I said, he gave a clear distinction between all three of coordinated, jogging, and free (aka homing)
[13:28:10] <alex_joni> both homing and jogging are performed by the free mode trajectory planner
[13:28:21] <alex_joni> there is a distinction between jogging and homing though
[13:28:32] <alex_joni> different parts of the code that sets up the move
[13:28:47] <alex_joni> but there is a common part (the free mdoe TP) that performs the move
[13:29:01] <alex_joni> archivist: depends on the accel setting and on the scale
[13:29:44] <archivist> so one needs to allow for an overrun on the switch
[13:32:36] <skunkworks_> archivist - there are a bunch of ways you can setup homing. The one I use - is it hits the homing switch at a 'faster' velocity - then moves off at a 'slower' velocity.
[13:33:08] <skunkworks_> which then sets home
[13:38:29] <archivist> does emc have facility for two spaced home switches
[13:40:59] <skunkworks_> jepler: welcome back
[13:41:29] <skunkworks_> where you hit one - then slow down to hit the next one?
[13:42:07] <archivist> or hit one and emergency stop on second
[13:42:25] <archivist> yes either
[13:48:21] <maddash> alex_joni: if both homing and jogging are performed by the TP, then only the [TRAJ] section is relevant for homing, no?
[13:50:01] <cradek> that's incorrect
[13:50:13] <cradek> like I said above, an axis NEVER moves faster than its AXIS velocity limit
[13:50:44] <cradek> and, neither homing nor jogging use the TP; these are free mode operations
[13:51:24] <cradek> oh alex said "free mode", he's right
[13:51:40] <cradek> there's a simple "free mode planner" which moves the joints (axes) individually
[13:52:11] <cradek> it is not really a trajectory planner, but that's just semantics
[13:54:00] <maddash> cradek: let me restate: normally, if the TRAJ limits < AXIS limits, then only TRAJ applies during homing. no?
[13:54:22] <maddash> ^^ otherwise, AXIS limits would kick in.
[13:56:59] <jepler_> skunkworks: my ISP seems to have a little problem with incompetence this morning
[13:59:45] <cradek> maddash: I think, but am not sure, that TRAJ is ignored in free mode (homing, jogging)
[14:00:11] <cradek> homing is not really a trajectory
[14:00:45] <cradek> having TRAJ limits lower than any AXIS limit would be a strange configuration; I don't know why someone would want that.
[14:00:52] <cradek> brb again
[14:01:23] <maddash> 09:20:29 <alex_joni> maddash: respectfully: _we_ did code this, so I'd say we know what we're talking about
[14:05:39] <jepler_> jepler_ is now known as jepler
[14:08:43] <maddash> heh cool, the homing also accounts for the decceleration of the latch movement
[14:37:02] <lerneaen_hydra> cradek: can backlash exceed axis limits?
[14:37:55] <cradek> It's not clear to me how backlash and limits should interact. I also don't know how they currently DO interact
[14:39:11] <lerneaen_hydra> I seem to recall something about backlash being able to generate 2*maxvel and 2*maxaccel, though I don't recall if that was traj speed of axis speed
[14:39:29] <lerneaen_hydra> well, surely it shouldn't be able to exceed max axis speed?
[14:40:02] <cradek> well the hope is that when backlash compensation fires, it doesn't move the axis at all
[14:40:35] <cradek> we know that's not always how it works out, since the axis can get "pushed around" by other things
[14:41:22] <cradek> (cutting forces, inertia, etc)
[14:41:23] <lerneaen_hydra> well, the motor still has to spin at 2x it's normal max speed
[14:41:36] <lerneaen_hydra> and a stepper for instance may loose steps
[14:41:39] <cradek> yes, if using backlash comp, you have to consider that
[14:42:08] <lerneaen_hydra> hmm, I see
[14:42:28] <cradek> since backlash comp is always applied when reversing direction, the motor is nearly stopped anyway
[14:42:42] <cradek> BUT the axis may be accelerating heavily even though the motor is nearly stopped
[14:43:05] <cradek> so maybe it's not really the motor's max speed that's the concern
[14:43:26] <lerneaen_hydra> it seems to me that right now there seems to be a certain "which way is the best (single) way to do backlash comp", rather than doing a few different operating modes
[14:43:40] <lerneaen_hydra> yeah thats very true
[14:43:43] <cradek> there is no good way to do backlash comp :-)
[14:43:52] <lerneaen_hydra> a servo+backlash system seems to be easier to do
[14:43:58] <lerneaen_hydra> or at least closed loop
[14:44:01] <cradek> there are only ways that are slightly better than nothing
[14:44:08] <lerneaen_hydra> yeah
[14:44:48] <lerneaen_hydra> there's also the "let other axes wait untill backlash is removed" method, that would work for some things
[14:46:21] <SWPadnos> hmmm. it appears maddash had some confusion about what I had said about max_vel
[14:46:23] <SWPadnos> oh well
[14:47:00] <cradek> I couldn't tell if maddash had questions he wanted answered, or what as going on
[14:47:12] <cradek> so whatever
[14:49:50] <SWPadnos> yep. oh well. whatever.
[14:50:42] <SWPadnos> I was pretty clear (I thought) when I said that neither planner will exceed an AXIS* limit, and additionally in coordinated motion the TRAJ limit won't be exceeded
[14:51:07] <SWPadnos> though I may not have explicitly mentione that the TRAJ limit is on the vector sum (I did have an example though, so ...)
[14:52:19] <cradek> my understanding is the same as yours
[14:52:34] <SWPadnos> oh good. then we're both right or we're both wrong
[14:52:47] <SWPadnos> and since you wrote the code, I'm betting on the former ;)
[14:52:55] <cradek> if so, let's blame any mistake on imprecise language
[14:53:29] <SWPadnos> ok. done
[15:04:52] <skunkworks_> he seems to be a bit on the literal side.. I remember reading SWPadnos comments and they made sense. (not that means anything coming from me)
[15:05:21] <SWPadnos> what are you saying? that makes no sense ;)
[15:05:29] <cradek> huh? I don't follow
[15:05:52] <alex-_> alex-_ is now known as The_Ball_
[15:05:55] <skunkworks_> ;)
[15:06:51] <The_Ball_> got my ac servos today and got one mounted on the x axis, now i have to get it to move
[15:07:10] <SWPadnos> push REALLY hard!
[15:07:33] <lerneaen_hydra> O_o
[15:07:34] <lerneaen_hydra> wtf
[15:07:39] <The_Ball_> haha, i needed a good laugh, got the flu and it's 0100 with no move yet
[15:07:46] <lerneaen_hydra> me pokes The_Ball_
[15:07:46] <skunkworks_> so where they resolvers or encoders?
[15:07:50] <lerneaen_hydra> oops
[15:07:55] <lerneaen_hydra> * lerneaen_hydra kicks The_Ball_
[15:08:09] <SWPadnos> lerneaen_hydra, /you pokes ...
[15:08:17] <lerneaen_hydra> * lerneaen_hydra waits for +o :D
[15:08:28] <The_Ball_> they are resolvers, but the ge/fanuc controller can output a a/b/key emulated encoder as well
[15:08:38] <skunkworks_> Nice
[15:08:44] <lerneaen_hydra> oh, wait
[15:08:48] <lerneaen_hydra> -.-
[15:08:49] <skunkworks_> sound like nice little controllers
[15:08:55] <lerneaen_hydra> sorry The_Ball_, I thought you were someone else
[15:09:01] <skunkworks_> they also do step and direction - correct?
[15:09:35] <The_Ball_> yes, but i have to program them with a serial cable, no luck yet, got 24 discrete io's!
[15:10:24] <anonimasu> the drives?
[15:10:52] <The_Ball_> yes, they are "motion controllers" so they are smal plc's kinda
[15:11:42] <skunkworks_> Jepler: how where you going to get the pluto run any faster than the 1ms or so communication with the ecc port? or would you not use epp
[15:11:45] <skunkworks_> epp
[15:11:46] <skunkworks_> eep
[15:11:48] <skunkworks_> whatever
[15:12:07] <The_Ball_> oh, got the pluto today as well, but that's another project
[15:12:15] <skunkworks_> ;)
[15:12:56] <skunkworks_> well - crap. He isn't even in here.
[15:13:04] <The_Ball_> ecp/epp is it
[15:13:24] <skunkworks_> wondering why the auto-complete wasn't working
[15:14:46] <skunkworks_> the_ball: so there must be a way to configure the pid loop serialy?
[15:14:50] <skunkworks_> in the drive
[15:15:10] <skunkworks_> using step and direction
[15:15:45] <The_Ball_> i went to this metal place and a got a 1mx10cmx10cmx3mm box steel "tube" to mount the servos, then thought no way i'm attacking that thing with a hacksaw, so i got one of thouse 355mm disk grinders/cutters. never used one before, veeery effective ;)
[15:16:20] <cradek> skunkworks_: I'm using .5ms servo cycle on my lathe - and I'm pretty sure it could go faster
[15:16:22] <The_Ball_> yes, you can put it in command mode as well, so you can send step commands as well, wonder if emc has support for such a mode
[15:17:52] <skunkworks_> Cradek: I though jepler had mentioned running it in a 'fast' tread which makes me think 50us or so.
[15:17:58] <skunkworks_> thread
[15:18:05] <cradek> yeah that would be totally different
[15:18:34] <cradek> I'm just interested in it for servo control, but he wants to try other things too
[15:18:54] <skunkworks_> although now I remember desciding to work on the charge pump - so maybe that was nixed ;)
[15:19:53] <skunkworks_> charge pump/watch dog
[15:20:16] <cradek> yeah watchdog would be nice, if there's room to add it
[15:21:08] <skunkworks_> cradek: so do you have a good program to demo now? Threading was 'so' last year..
[15:21:25] <skunkworks_> ;)
[15:21:46] <cradek> any suggestions? I'm not sure what would be cool
[15:21:53] <cradek> I can change tools now
[15:22:05] <cradek> so I have left and right cutters, threading, parting
[15:22:26] <cradek> might make a boring bar there if I can find a grinder
[15:22:30] <skunkworks_> I don't know - something with a ball on the end - threads in the middle...
[15:23:18] <cradek> brb
[15:26:57] <archivist> A fusee
[15:28:43] <skunkworks_> http://en.wikipedia.org/wiki/Fusee
[15:28:50] <skunkworks_> I had to look it up
[15:38:03] <archivist> nice evil thing to make
[15:38:30] <cradek> I think I could actually do that with G33
[15:38:59] <cradek> that's a great idea (but nobody there will know what it is)
[15:40:56] <cradek> archivist: I know the curve is often determined emperically, but do you know what mathematical curve the shape is close to?
[15:42:06] <cradek> With a mainspring, the coil inside contains a pressurized force on the escapements and the balance wheel, pushing the watch's movement into submission. A fusee, instead of containing a coil inside the casing, contains a chain, that wraps itself around the cone-shaped pulley, therefore creating a "pull" force that drives the watch.
[15:42:21] <cradek> ^^ totally bogus text in the wikipedia article (IMO)
[15:42:43] <skunkworks_> I went by the picture :)
[15:43:18] <cradek> as the spring unwinds, the radius of action goes up to match the declining force
[15:43:29] <cradek> it's a very neat setup
[15:43:32] <skunkworks_> http://www.abbeyclock.com/gbk/gbkt1.html
[15:44:08] <archivist> Rawlings P296 gives r^2theta=RK/2
[15:44:31] <archivist> not hyperbola or parabola
[15:45:32] <cradek> can you explain the constants R,K?
[15:45:34] <archivist> but thats an aproximation as well see Preisendorfer ...
[15:45:55] <archivist> he refs another
[15:46:02] <cradek> the only 'scholarly' text I have is the Daniels book, I bet it has a formula too
[15:46:32] <cradek> I could cheat and measure one, and match it
[15:47:10] <archivist> this is the Science of clocks and watches by A.L. Rawlings (the only one I trust for the maths)
[15:47:24] <archivist> Danuiels is ok as well
[15:48:20] <cradek> I bet the curve is different for different springs - the old ones are for regular blued steel I suppose
[15:48:45] <cradek> even then I think spring and fusee were matched by hand
[15:48:56] <archivist> ah Preisendorfer R W Fusee Theory Horological Journal Vol 120 No 5 nov 1977
[15:49:45] <archivist> Rawlings goes on to say why that eq above was approx and go read the other ref
[15:50:01] <cradek> heh ok
[15:50:13] <cradek> have you made replacement fusees?
[15:50:19] <archivist> not yet
[15:50:37] <archivist> I need a cnc lathe first
[15:50:43] <cradek> do clocks have the internal spring for maintaining power like english watches do?
[15:50:47] <archivist> or a Fusee engine
[15:50:49] <cradek> or a rotary axis for your little mill
[15:51:16] <cradek> (assuming you could find rotary tooling small enough)
[15:51:29] <cradek> I think for clocks it wouldn't be too hard
[15:52:44] <archivist> we were dscussing that last night, the ownser of said miller says I can do what I want to it and he has the parts for a rotary axis (a bit crap iirc)
[15:53:16] <cradek> that sounds like a good start
[15:53:24] <cradek> a control retrofit would be the first step I assume?
[15:53:43] <archivist> yup
[15:54:10] <cradek> sounds fun
[15:54:11] <archivist> mainly an interfacing job
[15:54:32] <archivist> as stepper drivers are in it
[15:54:57] <cradek> cool, it's pretty simple then
[15:55:05] <skunkworks_> what is the software like that controls it now?
[15:55:15] <skunkworks_> pretty bad?
[15:55:16] <archivist> * archivist does a bit more cutter sharpening
[15:55:20] <archivist> very bad
[15:55:48] <archivist> for real use its bad
[15:56:41] <skunkworks_> You need to play with emc2/axis.
[16:00:19] <archivist> worst is the stop between quadrants on a circle
[16:02:38] <cradek> no motion blending at all (stops at the end of every move)?
[16:08:17] <archivist> yup
[16:08:25] <cradek> yuck
[16:09:04] <archivist> and no backlash and crashing comms
[16:09:31] <cradek> oh I figured it would have ballscrews
[16:09:48] <archivist> it feeds the g code one line at a time from the pc
[16:10:12] <archivist> no its got cheapo screws
[16:10:15] <cradek> whee
[16:10:37] <archivist> educational toy really
[16:10:59] <cradek> too bad about the screws - the control can be easily updated
[16:11:00] <archivist> but as / more solid than a shereline
[16:11:22] <cradek> I put acme threads and zero backlash nuts on my mill - it really helped
[16:11:38] <cradek> I often had .010 of backlash before that
[16:11:48] <archivist> ew
[16:12:08] <cradek> for any work with small tooling you really have to get rid of the backlash mechanically - otherwise the tools just break
[16:15:01] <cradek> archivist: http://www.dumpstercnc.com/leadnuts_acme.html
[16:15:03] <archivist> could do with automated cutter centering for gear making
[16:15:47] <cradek> archivist: we now have support for poking a switch with the tool and offsetting the tool based on the results - I bet you could think something up
[16:15:53] <anonimasu> :)
[16:15:56] <anonimasu> cradek: nice!
[16:16:08] <archivist> thems cheap!
[16:16:15] <cradek> or do you just need basic homing? that's easy
[16:16:16] <anonimasu> I just got some micro relays for setting up home switches..
[16:16:29] <cradek> archivist: I used the 1/4-16 acme
[16:16:30] <anonimasu> but I need to wire up the index pulses of my servos..
[16:16:55] <archivist> no got to find the center of the gear milling cutter
[16:17:23] <archivist> to a thou
[16:17:26] <cradek> archivist: I'm really happy with the results - I cut .006" features all the time (pcb traces) with no visible variation under the microscope
[16:17:53] <cradek> even with circular pads etc
[16:18:01] <archivist> a 6 leaf pinion with a badly set cutter is horible
[16:18:50] <cradek> you could poke a switch with just one side, since you know the thickness of the cutter right?
[16:19:13] <anonimasu> hm I found a good drawing on anti backslash nuts..
[16:19:14] <anonimasu> :
[16:19:16] <anonimasu> :)
[16:19:21] <anonimasu> ball screw nuts..
[16:19:22] <cradek> sounds like it's still just a variation of homing
[16:19:56] <archivist> no I dont know the thickness or dia as they are all diff and small
[16:20:21] <anonimasu> eh?
[16:20:28] <anonimasu> what kind of shitty cutters do you use
[16:20:32] <cradek> hm ok, I'd have to see the setup to suggest anything "smart" but you have a lot of options/flexibility with emc
[16:20:46] <cradek> (emc2)
[16:21:05] <anonimasu> there are laser measurement stuff to do this from reinshaw.
[16:21:36] <archivist> well as there are 12/14 teeth per cutter it would need to be rotated to find the best pos
[16:22:00] <archivist> I cant afford renishaw
[16:22:13] <anonimasu> archivist: I dont get why all your tools are different..
[16:22:32] <anonimasu> if you order something 6mm the endmill/tool are 6.000
[16:23:09] <archivist> each size of gear needs its own cutters
[16:23:19] <archivist> gear cutters not end mills
[16:23:36] <anonimasu> yeah, but if you order a cutter they are at the size you order them..
[16:23:40] <archivist> no
[16:23:57] <archivist> and the size changes with sharpening
[16:24:12] <anonimasu> yes ofcourse..
[16:24:34] <archivist> the dia is not fixed at all with epicyclice gear cutters
[16:25:07] <archivist> only the form is specified
[16:25:43] <anonimasu> ah yeah, that's true..
[16:26:34] <archivist> www.archivist.info/cnc top right pic for a 1950's gear miller
[16:26:56] <archivist> one to the left is the cutter grinder
[16:28:18] <anonimasu> nice
[16:28:21] <anonimasu> :)
[16:28:46] <archivist> i right royal pain to setup for one offs though
[16:28:54] <archivist> i/a
[16:29:09] <anonimasu> brb
[17:15:35] <skunkworks_> alex_joni: we don't really know where we are going next.
[17:15:41] <skunkworks_> next big trip
[17:17:24] <skunkworks_> Looks like my telescope tracking mount is out for delivery. Nice
[17:35:33] <SWPadnos> hmmm. it should be relatively easy to derive the radius equation for a fusee
[17:39:34] <archivist> I have not managed to draw one in solidworks yet
[17:39:45] <SWPadnos> that wouldn't be so easy, I imagine :)
[17:40:44] <archivist> dunno if the later versions can follow a profile yet
[17:40:59] <SWPadnos> it should basically be a spiral ramp with a specific rate of decrease in radius, and I'm not sure you can do the right math functions in SW
[17:41:25] <SWPadnos> they should be able to - I know the "piping" features are pretty advanced
[17:41:43] <SWPadnos> but that doesn't really help with spirals ;)
[17:45:19] <archivist> the real radius eq will also take into account the spring in the barrel and its state of wind etc so not that easy
[17:46:05] <archivist> and depending on spring type that curve will vary
[17:46:50] <SWPadnos> well, since you want equal force all the time, you need to keep the force from torque (T/r) equal to the spring force (K*theta^2)
[17:46:56] <SWPadnos> +/- some constants ;)
[17:47:21] <archivist> some clock springs now come with a reverse "set" in part of it
[17:48:00] <SWPadnos> the equation you posted looks "right enough" to me, knowing that K is the spring constant
[17:48:10] <SWPadnos> I don't know what R was supposed to be though
[17:48:24] <archivist> yes but the fella writing that also debunks it
[17:48:37] <SWPadnos> I also don't know why ir's r^2 - it should be theta^2
[17:48:43] <archivist> and he is good
[17:48:55] <SWPadnos> (assuming you're wrapping/unwrapping a spring with the rotration of the "cone")
[17:49:22] <archivist> yes but the spring is externally constrained
[17:49:36] <SWPadnos> ?
[17:49:46] <archivist> its in a barrel
[17:50:13] <SWPadnos> ok, that's not a problem. it's the rotation that exerts spring force, not extension or compression (right?)
[17:50:26] <archivist> yes
[17:50:47] <SWPadnos> ok, so the constraint is that you cant "over-unwind" the spring, because that would require too large a radius
[17:51:01] <archivist> and a wire/cord around the barrel to the fusee groove
[17:51:13] <SWPadnos> (or over-wind, depending on whether it's a rotaional compression or expanion spring)
[17:51:30] <SWPadnos> ok, so the fusee isn't the thing with the spring inside it
[17:51:56] <archivist> no, the spring is in the barrel
[17:51:54] <SWPadnos> you have a wind-up thing, and the fusee "consumes" a constant spring force, rather than providing a constant force
[17:52:05] <archivist> yes
[17:52:40] <archivist> it converts variable force from the spring to a constant
[17:52:46] <SWPadnos> and the chain unwinds from a constant radius?
[17:52:53] <SWPadnos> likea wire spooll
[17:52:57] <SWPadnos> -l
[17:53:33] <archivist> yes arount the barrel
[17:53:37] <SWPadnos> ok
[17:54:42] <SWPadnos> so that's where the R comes from, it's the barrel radius
[17:55:26] <archivist> www.archivist.info/clock theres three on this clock
[17:56:39] <SWPadnos> spring force is K*x, where x is the displacement
[17:56:51] <SWPadnos> I think for circular action, that becomes k*theta
[17:57:12] <SWPadnos> but then it's torque, so you divice by the "action radius" (my term)
[17:57:15] <SWPadnos> divide
[17:57:25] <SWPadnos> so you have K/R * theta
[17:57:45] <cradek> SWPadnos: now write the gcode and I'll cut one for us
[17:57:53] <SWPadnos> well, I'm still deriving ;)
[17:58:42] <SWPadnos> you want a constant force F, using a changing radius r
[17:59:05] <ds2> anyone been following the mwmill2 list on the gear hobber? it seems like something that can be emulated on CNC with a sync'ed spindle
[17:59:05] <archivist> hmm hang on a bit and Ill put up Rawlings few pages
[17:59:05] <SWPadnos> F=K/R*theta (like the equation already posted :) )
[17:59:17] <cradek> what's F? you want constant torque
[17:59:28] <cradek> F (force on the chain) will change
[17:59:49] <SWPadnos> right - constant torque :)
[18:00:07] <SWPadnos> so F*r = T
[18:00:17] <SWPadnos> F= force provided by barrel, T=constant desired torque
[18:00:59] <SWPadnos> theta is the integral of R/r, I think
[18:01:27] <SWPadnos> (which is where I start getting into all that stuff I've forgotten)
[18:02:55] <SWPadnos> substituting integ(R/r) for theta, you get F=(K/R)*(R*integ(r^-1))
[18:05:24] <SWPadnos> so the R's cancel, and you have F=K*integral(1/r)
[18:05:40] <skunkworks_> * skunkworks_ cheers SWPadnos on
[18:05:45] <SWPadnos> the integral of 1/r is ln(r)+C
[18:05:53] <SWPadnos> yay!
[18:06:47] <SWPadnos> (I had to look that up - I couldn't get past the x^(n+1)/n+1), where n was -1. I remembered that division by 0 is no good :) )
[18:07:50] <archivist> ew maffsss
[18:08:05] <SWPadnos> so F=K*ln(r), and F*r=T, so K*ln(r)*r = T
[18:08:32] <SWPadnos> isolating r is left as an exercise for the reader ;)
[18:08:53] <SWPadnos> actually, I may be able to do that. hmmm
[18:08:59] <archivist> Im waiting for the camera to download and Rawlings 3 pages will be up
[18:09:05] <SWPadnos> heh
[18:11:33] <archivist> hmm need to redo page 2
[18:12:35] <SWPadnos> oh crap. what's actually needed is dr/dtheta
[18:12:44] <SWPadnos> I haven't had enough coffee for that ;)
[18:13:04] <SWPadnos> I'll take the easy route and ask my mother when she calls
[18:15:56] <skunkworks_> that is still funny - makes me think of alien
[18:16:48] <alex_joni> heh
[18:18:33] <skunkworks_> Jepler: agian - welcome back :)
[18:20:35] <archivist> www.archivist.info/rawlings
[18:24:27] <cradek> "This is a favourite example in books on elementary mechanics and, so far as I have noticed, these books never give the right answer for the shape of the cone."
[18:24:31] <cradek> haha
[18:24:41] <SWPadnos> yeah, I liked that too
[18:24:48] <cradek> archivist: thanks for scanning this!
[18:25:31] <archivist> cradeck just reading about him, he worked for Bulove
[18:25:59] <archivist> in the military contracts division
[18:26:29] <cradek> I bet bulova never made a fusee!
[18:26:33] <cradek> (hamilton sure did)
[18:26:46] <archivist> ya never know
[18:27:09] <archivist> he was a smart cookie
[18:32:42] <archivist> ah the smaller type section is an addition by PW to the original work
[18:33:28] <archivist> Philip M Woodward
[18:35:16] <SWPadnos> ok, so a few incorrect assumptions plus some incorrect derivation gives me a completely wrong answer ;)
[18:35:33] <archivist> hehe
[18:35:40] <cradek> SWPadnos: the next step is using your wrong answer to write a wrong program to generate wrong gcode
[18:35:48] <SWPadnos> that would be a good demo!
[18:36:08] <SWPadnos> "blindly following wrong profiles, with great accuracy!"
[18:36:46] <archivist> if it gets close to the drawing its good enough
[18:36:53] <cradek> yeah
[18:37:27] <cradek> we should try to do this. It would be possible to make a pretty coarse one with the left cutting tool I have
[18:37:55] <cradek> it would be a great demo/test of the flexibility of G33
[18:38:17] <cradek> (CSS would help with this of course)
[18:38:59] <SWPadnos> it would be an excellent CSS demo
[18:39:24] <archivist> about 16 turnsfusee == about 8 turns on the spring barrel
[18:39:25] <cradek> if we cared, we could rig up css in hal
[18:39:59] <cradek> heck, or, we could do css
[18:42:41] <skunkworks_> cradek: you would sync a bunch of short line segments to the spindle?
[18:42:46] <cradek> skunkworks_: yes
[18:42:48] <skunkworks_> cool
[18:43:11] <cradek> adjacent synced moves keep sync through the blend
[18:43:39] <skunkworks_> I suppose it is similar to your exit/enter angles.
[18:43:43] <skunkworks_> in threading
[18:43:50] <cradek> yes that does the same thing
[18:44:05] <skunkworks_> I guess I never thought about doing something like that.
[18:44:38] <LawrenceG> hi chris... cvs.linuxcnc.org port 22: Connection timed out... I think this is at your place??
[18:44:52] <alex_joni> LawrenceG: jepler's place
[18:45:01] <LawrenceG> ah... np
[18:45:02] <cradek> LawrenceG: it blew away - it'll be back soon
[18:45:08] <alex_joni> LawrenceG: should be back soon
[18:45:20] <skunkworks_> cradek: is it raining there - or is it other issues?
[18:45:29] <cradek> must be something else today
[18:45:35] <alex_joni> skunkworks_: jepler said ISP foolishness
[18:45:38] <archivist> he was cursing the isp earlier
[18:45:41] <cradek> very windy, but that usually doesn't hurt anything
[18:46:29] <archivist> on a microwave link?
[18:46:36] <cradek> we were laughing about a ridiculous SUV in the parking lot here - I (jokingly) said it probably cost $40k, and he (even more jokingly) guessed $65k
[18:46:50] <cradek> guess what, it's either $85k or $108k according to the website
[18:46:58] <alex_joni> :/
[18:46:59] <skunkworks_> wow.
[18:47:09] <alex_joni> mmtb
[18:47:20] <archivist> some ppl are easily parted from their cash
[18:47:44] <cradek> car price $108000; down payment $20000; loan period 60 months; interest rate 5%: monthly payment $1660.67
[18:47:50] <cradek> ^^ their finance info
[18:48:25] <skunkworks_> alex_joni: mmtb?
[18:50:18] <skunkworks_> ah - more money than brains.
[18:50:38] <skunkworks_> boy - didn't know that one.
[18:50:57] <archivist> I unfortunately have less money than brains
[19:08:03] <alex_joni> archivist: that's how it goes for regular folks
[19:08:14] <alex_joni> and for smart folks too :P
[19:13:41] <archivist> trouble is one ends up doing things on the cheap and quality suffers
[19:14:07] <archivist> or you keep quality up and earn a pittance
[19:17:28] <lojz> hello. I installed emc2 and tried to run a stepper-xyza, but i get an error ( http://pastebin.ca/545266 ) Should this configuration work fine, or do i have some other problem?
[19:18:13] <jepler> lojz: hidden among all those lines, here is the real error:
[19:18:14] <jepler> _tkinter.TclError: Togl: X server has no OpenGL GLX extension
[19:18:26] <jepler> do other opengl applications (such as "glxgears") run on your system?
[19:18:56] <jepler> stepper-xyza uses the AXIS user interface which requires opengl. You can edit the .ini file to select one of the other GUIs, such as tkemc. Or you can figure out why your system doesn't run opengl apps.
[19:19:25] <lojz> ok, thank you
[19:38:03] <JymmmmEMC> jmkasunich_: Do you see my msg to you the other day about this? http://www.pilz.com.au/pulse.htm
[19:54:12] <jepler> http://en.wikipedia.org/wiki/Current_loop#Process-control_use
[19:59:59] <anonimasu> um yes?
[20:00:13] <anonimasu> :)
[20:01:46] <skunkworks_> we had a old teletype that had a current loop interface.
[20:02:19] <JymmmmEMC> jepler: In respect to using a current-loop with a relay, how would you detect/measure the current to trigger the relay as needed? Some kind of shunt?
[20:02:48] <anonimasu> we use it for sensors..
[20:03:37] <jepler> JymmmmEMC: just like with those pilz modules, there would be some amount of digital electronics inside .. the details are left as an exercise to the designer of the current-loop-estop-system :-P
[20:04:00] <anonimasu> hehe
[20:04:04] <JymmmmEMC> jepler: File System Check
[20:04:21] <anonimasu> the real 4-20 ma stuff sense how much current is needed to keep the output voltage constant at the receive side..
[20:05:34] <JymmmmEMC> I've been thinking about Mariss estop circuit, and I remembered that if you held the START button on, even if yo hit the STOP button, the circuit will still be engaged.
[20:06:28] <skunkworks_> JymmmmEMC: your still thinking too much about multible failures. Is this machine going to be run by others that would sue you if something went wrong?
[20:08:08] <JymmmmEMC> skunkworks: It's my nature. There may be a time/place where that just might be the case, and it's better to do it right and not half-ass.
[20:08:52] <skunkworks_> JymmmmEMC: you missed this one :( http://cgi.ebay.com/6-PC-LOT-OF-PARKER-COMPUMOTORS-MODEL-NO-OEM650_W0QQitemZ230133106438QQihZ013QQcategoryZ111543QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
[20:09:25] <JymmmmEMC> skunkworks the 650's do NOT have mid0band compensation, the 750's do.
[20:11:38] <JymmmmEMC> skunkworks but thanks =)
[20:13:43] <anonimasu> :)
[20:13:49] <skunkworks_> I think it does. You can tune the OEM650X to minimize resonance and optimize
[20:14:02] <skunkworks_> smoothness by adjusting the small potentiometers (pots) on the top
[20:14:12] <skunkworks_> of the unit. Figure 3-1 shows the location of the potentiometers and
[20:14:22] <skunkworks_> of the unit. Figure 3-1 shows the location of the potentiometers and
[20:14:28] <skunkworks_> their functions.
[20:14:27] <skunkworks_> sorry
[20:14:58] <JymmmmEMC> skunkworks Well, I can call Parker again I guess.
[20:15:10] <skunkworks_> page 33 of the oem650 manual
[20:15:31] <JymmmmEMC> The 750 specifically state they have mid-band compensation.
[20:15:53] <skunkworks_> hmm. I guess we never adjusted the pots on ours.
[20:16:23] <skunkworks_> At 1000 steps per rev we never seemed to have a problem
[20:27:21] <skunkworks_> JymmmmEMC: http://cgi.ebay.com/3-Parker-Compumotor-Stepper-Drives-and-Power-Supply_W0QQitemZ230137147843QQihZ013QQcategoryZ78193QQrdZ1QQssPageNameZWD7VQQcmdZViewItem
[20:28:37] <skunkworks_> power supply is worthless - 2.7a cont - 4a peak.
[20:40:22] <JymmmmEMC> Well, I have a 48V@5.6A PS
[20:40:45] <skunkworks_> but 2 750s and 1 650 :)
[20:40:52] <JymmmmEMC> I also have a call into Parker too, should callback within the hour
[20:41:01] <JymmmmEMC> well 650 on Z
[20:44:30] <JymmmmEMC> The current-loop thingy sounds reasonable though, could add a resistor at each mushroom.
[20:45:09] <skunkworks_> The only thing the auction has going for you is it doesn't have 'CNC' in the title.
[20:45:44] <skunkworks_> it does have stepper though
[20:45:56] <JymmmmEMC> Heh, no doubt. But I have a trick up my sleeve for that.
[20:46:28] <JymmmmEMC> Oh good, it ends at 9pm
[20:47:02] <JymmmmEMC> I'll be at work on the 10GigE pipes =)
[20:48:39] <skunkworks_> that helps - dad uses an atomic clock - then uses a few test bids to estimate the delay. So he know about how much time before the end he has to bid
[20:49:12] <ds2> that's a lot of BW to use for DoS
[20:50:03] <JymmmmEMC> ds2: Yes, Yes it is =) And for those that they to DoS us, it's impossible and we just laugh and null route =)
[20:50:03] <alex_joni> * alex_joni heads to bed
[20:50:06] <alex_joni> good night all
[20:50:11] <JymmmmEMC> nite alex
[20:50:13] <skunkworks_> Night alex
[20:50:24] <JymmmmEMC> alex_joni: say hi to your bride from us!
[20:54:18] <alex_joni> JymmmmEMC: when I'll get home ;)
[20:54:25] <alex_joni> * alex_joni is still on a business trip
[20:54:47] <alex_joni> g'night
[20:54:54] <skunkworks_> :)
[20:56:31] <JymmmmEMC> alex_joni: I'll be sure to get you a divorce attorney... she's gonna kick ya to the curb ! LOL
[20:56:50] <JymmmmEMC> BBQ lit!
[20:57:36] <skunkworks_> what you grilling?
[20:59:06] <JymmmmEMC> marinated Steak schikabobs
[21:00:51] <skunkworks_> mmmmmm. We do mostly marinated chicken.
[21:01:53] <JymmmmEMC> skunkworks For chicken, one marinated we use is lime juice, crushed garlic, a bit of olive oil, and a touch of worchester sauce.
[21:02:15] <JymmmmEMC> juice of two limes is usually enough. Just toss everything in a ziplock bag
[21:03:01] <skunkworks_> yep. Garlic and herb is what the wife likes.. I prefer teryiocky(sp)
[21:03:24] <JymmmmEMC> no, no herbs, too overpowering.
[21:04:28] <JymmmmEMC> We just found (on the third try) a great Satay marnade.
[21:04:46] <JymmmmEMC> REALLY REALLY GOOD on chicken
[21:05:44] <skunkworks_> never come across satay - with us being midwesterners and all.
[21:05:54] <JymmmmEMC> skunkworks you like Thai food?
[21:06:10] <skunkworks_> I don't know if I have had much
[21:06:21] <JymmmmEMC> you like spicey but flacorful?
[21:06:25] <JymmmmEMC> flavorful
[21:06:58] <skunkworks_> yes - I guesss I have been to a local thai place.. it was good.
[21:07:11] <JymmmmEMC> Ever had Thai Peanut suace?
[21:07:29] <skunkworks_> no - sure I would have remembered that
[21:08:38] <JymmmmEMC> Probably, We buy it from the restraunt supply in 4.5 cup containers. then just buy cans of coconut milk and make it as we need it. Just emptied the container after 1.5 years - GOOD STUFF!
[21:08:59] <JymmmmEMC> Peanut Sauce Mix (not premade) http://www.atasteofthai.com/peanut.cfm
[21:10:09] <JymmmmEMC> Those are the retail pkging, but if you hit a restraunt supply it comes in a plastic tub with a goldish label on it.
[21:10:55] <skunkworks_> Cool - thanks
[21:11:21] <JymmmmEMC> Cash and Carrry (a subsiderary of Smart and Final) carry it.
[21:12:11] <JymmmmEMC> Ah, here we go.... 3rd item down http://www.andreprost.com/thai_peanut_items.htm
[21:13:22] <JymmmmEMC> If you're not sure, you can go for the 3.5oz pkg, but it's good stuff.
[21:15:08] <JymmmmEMC> Everywhere we've bought coconut milk, it comes in 13.5oz cans and you just add 3.5Oz on peanut suace mix. A lil more or less to your taste.
[21:16:11] <JymmmmEMC> But you REALLY have to let it come up to heat slowly. and don't let it sit on the burner too long after that or the oils will start to seperate. AND STIRE CONSTANTLY.
[21:16:12] <skunkworks_> Thank you.. Something more to try. We have been making Zuppa Toscana and Pasta E Fagioli
[21:16:17] <skunkworks_> soup
[21:16:42] <JymmmmEMC> Yeah, we've been avoiding all startches as much as possible.
[21:16:56] <skunkworks_> wife is a pasta junky
[21:17:27] <JymmmmEMC> I have't had spaghetti in a year at least. I just miss the garlic bread =)
[21:17:39] <skunkworks_> yah - that sounds like me
[21:18:10] <JymmmmEMC> The peanut suace is good over white rice too (my one minimal except to the startches)
[21:18:22] <JymmmmEMC> or lightly steamed veggies.
[21:18:41] <JymmmmEMC> it's REALLY rich too, don't need a lot of it.
[21:21:42] <JymmmmEMC> skunkworks I picked up a Directway dish with 2 (3?) LNB's on it the other day.
[21:22:22] <skunkworks_> Nice. parents are still working.
[21:22:37] <JymmmmEMC> huh?
[21:22:48] <JymmmmEMC> In the lbrary with the candlestick
[21:22:58] <skunkworks_> it is still working ;) wink wink nudge nudge
[21:23:33] <JymmmmEMC> skunkworks DOH! Gotcha. I Need to test this thing and see if it even works.
[21:24:02] <JymmmmEMC> It came with a modem that has lan port, might try to see if it has a db meter of some sort.
[21:24:48] <JymmmmEMC> I have to finish clearing out the back yard of the 4 years of pack weeds before I even touch it.
[21:25:53] <skunkworks_> Hmm - a DW4020?
[21:26:01] <skunkworks_> modem
[21:26:57] <skunkworks_> I found that a cheap directway reciever works great as a signal level device. It also gives the name of the satellite your pointing to.
[21:27:11] <JymmmmEMC> When we moved into this place it had been vacant for 4 years from what the utility companys have told us. Right now, it was just a fire hazard.
[21:27:12] <skunkworks_> cheap as in free
[21:27:27] <JymmmmEMC> Let me look, parker just called back
[21:28:03] <JymmmmEMC> dw6000 or dw8000
[21:28:49] <skunkworks_> I got out of directway just as they where getting the dw4020 out. I had the modems that needed to be hooked into a computer.
[21:28:50] <skunkworks_> usb
[21:29:08] <JymmmmEMC> ethernet
[21:29:18] <skunkworks_> I know.. mine system was usb
[21:29:22] <JymmmmEMC> ah
[21:29:28] <skunkworks_> so a computer was needed to share the internet
[21:29:39] <JymmmmEMC> yeah, I remember that
[21:30:21] <skunkworks_> better than dial up - but that was about it ;)
[21:30:40] <JymmmmEMC> did that have a telco jack on it?
[21:30:57] <skunkworks_> no - it was 2way
[21:32:19] <skunkworks_> so ping times where a good second
[21:32:21] <skunkworks_> :)
[21:33:15] <skunkworks_> * skunkworks_ goes home. bbl
[21:33:30] <JymmmmEMC> Yeah all wireless links have huge latency
[21:34:38] <maddash> the homing manual page mentions "HOME_ALL" and "HOME_SEQUENCE" -- how do I achieve this?
[21:35:11] <maddash> I grepped the entire emc2/src directory for both terms and came up with nothing.
[21:37:25] <JymmmmEMC> skunkworks: Ok, confirmed by Parker T/S... the 650's do NOT have mid-band compensation, that feature was added to the 750's.