#emc | Logs for 2009-04-04

[00:00:05] <archivist> I cheat and rotary mill :)
[00:07:21] <John_s> Same here, small stuff i use an old tap with only one flute and larger stuff I use one doe out of a Coventry Diehead, http://www.stevenson-engineers.co.uk/files/threadmill.jpg
[00:09:22] <archivist> I made a holder for single point lathe insert to fit the spindle so I can do any odd pitch
[00:09:48] <archivist> Im mainly matching antique stuff
[00:21:32] <John_s> Signing out now, off for a shower and the ig black, gotta be in stoke at 8:00 am today.
[00:21:49] <archivist> have fun
[00:22:47] <archivist> time I went home
[00:26:44] <UncleG> hey archivist, you around?
[00:26:49] <archivist> yes
[00:26:55] <UncleG> awesome
[00:27:10] <archivist> but was about to burger orf
[00:27:18] <UncleG> Are you aware of what hardware I am using?
[00:27:41] <UncleG> oh
[00:27:41] <UncleG> ok
[00:27:48] <archivist> apart from the servo no
[00:28:32] <UncleG> Well, I just mean the amp for the servo
[00:30:32] <UncleG> so do you think it is not worth my efforts to use an lpt with it?
[00:30:58] <archivist> go for proper control
[00:31:08] <UncleG> explain pls
[00:32:40] <archivist> Im not sure what you need to know,
[00:33:35] <UncleG> well, basically I have all the amps and servos, encoders and mechanical side to the system, what should I use between emc linux and my servo amp. Currently I am using an lpt port.
[00:33:37] <archivist> Im not the best person to explain servo connections
[00:34:15] <archivist> are you using PWM on the lpt port
[00:35:34] <UncleG> yes
[00:36:01] <UncleG> pwmgen.up/pwmgen.down
[00:36:09] <UncleG> for cw and ccw
[00:36:52] <archivist> ok then you just need to amplify/filter the resultant signal to feed the amps
[00:38:09] <UncleG> this is what I expected. however I do not know what is required to do so
[00:39:24] <archivist> I seem to remember someone working on a circuit Im just looking on the wiki
[00:39:44] <UncleG> great
[00:40:50] <archivist> cant find at the moment
[00:40:50] <UncleG> I do have a some ic's that do voltage doubling however they use capacitors to function and it appears to work up untill I try to stop the pwm signal and then it appears the capictors finishines discharging forcing the motor past fault
[00:52:00] <archivist> either filter (1 R and 1 C) and amplify or drive a transistor from the parport to give you 0-10 v pwm
[00:52:03] <skunkworks> no money at all to spend? A mesa printer port 7I43 card + 7I33 for $89 + $69 gives you 4 analog axis interface with a bunch of i/o (and encoder counter
[00:52:14] <skunkworks> s
[00:52:37] <archivist> for a real machine thats cheap to do it properly
[00:54:59] <UncleG> yea, that is direction I plan to go, however right now I was shooting for the cheap method until I understand what I have already more.
[00:55:39] <UncleG> I tried using a transistor to boost the signal input but to no avail :( I should I assume I am doing something wrong.
[00:56:38] <UncleG> I just need some more torque out of the motor so I though I could wire a solid 5v to the cw and ccw terminals through a transistor that was activated by the par port
[00:56:39] <archivist> cheap method only wastes time sometimes
[00:57:51] <UncleG> parport + the constant 5v I assumed would do it. but no luck I think I am missing resistors in places they need to be. the datasheets I found for the transistors tell me their specs and not how to use them.
[01:03:00] <archivist> http://controls.ame.nd.edu/microcontroller/main/node40.html
[01:03:11] <archivist> see above adjust to suit
[01:04:39] <UncleG> that looks perfect... now all i gotta do is learn what all those lil symbols n such mean :P
[01:04:48] <UncleG> thank you!
[01:05:35] <dareposte> hello all
[01:07:28] <dareposte> what happens when you mix sodium hydroxide, ammonium nitrate, water, and boil for a few hours?
[01:08:10] <Valen> sounds like you might get some crystals building up that would be some kind of explosive
[01:08:30] <dareposte> hm i hope not
[01:08:45] <dareposte> i've heard it's a way to form black oxide coatings on metal parts
[01:09:02] <dareposte> don't really want anything to explode
[01:09:44] <Valen> could well be, just be a little carefull playing with nitrates
[01:09:58] <Valen> they are the thing that puts the boom in stuff
[01:09:58] <dareposte> i have a bit of fertilizer left over
[01:10:07] <dareposte> just reseeded my lawn
[01:10:07] <archivist> or go out and get ready made fluid
[01:10:10] <UncleG> archivist: will this do what I need at all? http://www.intersil.com/data/fn/fn3072.pdf pg 10
[01:10:15] <UncleG> page 10
[01:10:53] <dareposte> archivist: I have some zinc chromate conversion solution
[01:11:36] <dareposte> never tried the black iron oxide conversion though
[01:12:07] <archivist> we buy in the stuff
[01:12:16] <dareposte> really, where do you get it from
[01:12:21] <dareposte> that sounds a bit safer...
[01:12:48] <archivist> engineering suppliers
[01:12:53] <UncleG> I need to build this http://controls.ame.nd.edu/microcontroller/main/node40.html
[01:13:21] <UncleG> Anyone want to walk me through what I need because I am rather lame when it comes to understanding these symbol
[01:13:22] <UncleG> s
[01:13:43] <archivist> read the rest of the pages its teaching you
[01:14:05] <archivist> we cant do that in irc as well
[01:15:45] <Valen> mesa has a board for doing that dont they?
[01:15:51] <archivist> yup
[01:15:56] <Valen> PWM + direction in > +-10v out
[01:16:08] <archivist> far easier
[01:16:27] <UncleG> I was meaning for that to go in the electronics room :P
[01:16:53] <dareposte> archivist: this stuff seems pretty nasty
[01:17:18] <Valen> if its oxidising things then generally they are pretty agressive chemicals
[01:17:21] <archivist> luvly stuph, dont drink it all at once
[01:17:59] <dareposte> says here it will permanently blind you, or convert you to soap??
[01:18:09] <dareposte> (or both, I guess)
[01:18:21] <Valen> yuh agressive
[01:18:43] <Valen> rember acid to water not water to acid ;->
[01:18:56] <Valen> or in this case , nasty oxidiser to water
[01:19:01] <dareposte> yeah
[01:19:12] <dareposte> it seems the salts raise the boiling temp of the solution quite a bit
[01:19:26] <dareposte> i guess no topping-off of the liquid :)
[01:19:28] <Valen> cool
[01:19:40] <dareposte> can you spell steam explosion?
[01:19:41] <Valen> once its diluted it probably wont be so bad
[01:19:59] <Valen> mix well and add slowly
[01:20:13] <archivist> http://www.black-it.co.uk/?gclid=CJ3ry7eH1pkCFYU-3godz015aw
[01:20:31] <dareposte> hm
[01:20:56] <archivist> google term was chemical blacking solution steel
[01:21:04] <archivist> there are others
[01:21:09] <dareposte> archivist: somehow forming an iron oxide, with "safe, re-usable, non-toxic chemicals" seems like a bit of a .. hoax?
[01:21:18] <dareposte> ever used it?
[01:21:41] <archivist> I used one here and it rubbed of
[01:21:58] <archivist> but was old solution
[01:22:14] <UncleG> archivist: I see R for resistor and C for capacitor, but I don't have the slightlest clue as to which ones I use. where might I find a part list for such a device?
[01:22:26] <archivist> we blue steel here the old fashioned way
[01:22:42] <Valen> UncleG it depends on the PWM frequency
[01:22:59] <archivist> UncleG, rou read the next page and you work it out
[01:23:09] <UncleG> the freq is a variable
[01:24:09] <Valen> you sure the frequency is variable not the duty cycle?
[01:24:12] <Valen> (IE on time)
[01:25:22] <UncleG> I thought I could change the pwm freq within hal
[01:25:28] <UncleG> atleast Im pretty sure I can
[01:25:45] <Valen> yeah there you could but it won't change in operation
[01:25:53] <Valen> IE while your running it
[01:26:02] <Valen> (well unless you are doing something funky(
[01:26:30] <Valen> You probably want to use something like a multi-turn trimpot for the R in the RC network on that page
[01:26:54] <Valen> it lets you vary the resistance over a range to get a value your happy with
[01:27:00] <Valen> do you have an oscilloscope?
[01:27:12] <UncleG> it's starting to sound like im going to buy a mesa board
[01:27:24] <UncleG> no, just a digital scope with emc
[01:27:42] <archivist> dareposte, there is also a heat treatment method
[01:27:45] <Valen> i dont know much about that, i'm just getting into doing CnC
[01:28:01] <Valen> but i repair speed controller boards and the like
[01:28:11] <dareposte> archivist: I think i'm going to try this recipe I found, on some test pieces
[01:29:07] <UncleG> I just don't understand all these equations
[01:29:26] <dareposte> basically NaOH:NH4NO3 in a 2:1 ratio, boil, submerge parts for 10 minutes, cool and remove parts
[01:29:38] <dareposte> outdoors, with a rubberized apron and gas mask
[01:29:51] <Valen> uncleG does the scope thing let you see analog voltages?
[01:30:15] <UncleG> no
[01:30:20] <Valen> ahh
[01:30:27] <UncleG> it uses lpt port
[01:30:34] <archivist> dareposte, description of oil black http://books.google.co.uk/books?id=vb3dDN9FiekC&pg=PA543&lpg=PA543&dq=oil+black+heat+treatment&source=bl&ots=M9YlPuC0w_&sig=W0KZSvueJdzUtwOmbiMjZ5ZcK9U&hl=en&ei=yLfWSeC6J8Ow-AbinbTKBQ&sa=X&oi=book_result&ct=result&resnum=2
[01:31:10] <Valen> In that case I'd just use that multi turn pot, wind it up to maximum resistance and then slowly start winding it back down until you get a good responsivnes without "vibration"
[01:31:29] <dareposte> archivist: impressive find
[01:31:43] <dareposte> don't have nitric acid though
[01:31:48] <archivist> :)
[01:32:23] <dareposte> that's probably a good thing, from the soud of it i probably don't want any :)
[01:32:45] <UncleG> I will need both cards? " A mesa printer port 7I43 card + 7I33 for $89 + $69 gives you 4 analog axis interface with a bunch of i/o (and encoder counter"
[01:32:46] <BMG> Nitric acid from lab supply - we use it in cement chemistry labs alot
[01:32:57] <Valen> The mesa board is only $70 though
[01:33:13] <UncleG> regardless, will I need both? or just 1?
[01:33:28] <Valen> what are you using for servo drives?
[01:34:02] <UncleG> a servo dynamics brush 815 differential input amp
[01:34:21] <UncleG> +10 cw +10 ccw
[01:35:02] <UncleG> and also a copley controls, same thing.
[01:35:19] <Valen> i'm no expert but thats probably the easiest way to get it going
[01:35:34] <Valen> what did they set you back in terms of $?
[01:35:44] <UncleG> $150 roughly
[01:36:01] <UncleG> I still don't know if I need one or both of them
[01:36:26] <Valen> both cards togther gives you a nice analouge output voltage
[01:36:38] <UncleG> skunkworks: do I need both of those cards or just one?
[01:51:33] <jmkasunich> if you are interested in black-oxide, browning, or blueing, you might want to take a look at brownells
[04:22:58] <SWPadnos> http://www.youtube.com/watch?v=D2FX9rviEhw
[05:19:17] <JymmmEMC> SWPadnos: On time?
[08:55:15] <pjm_> good morning
[09:07:13] <OoBIGeye_> good morning
[11:32:43] <Martinp23> [Server Notice] Hi folks! The server you are on (kubrick) will be going down for around one hour for maintenance on Monday 13th April at 1400 UTC. We suggest that you connect to another server in the irc.freenode.net rotation to avoid disruption, asap. Thanks for using freenode and have a nice day!
[12:21:41] <BigJohnT> http://arstechnica.com/open-source/news/2009/03/google-launches-project-to-boost-python-performance-by-5x.ars
[12:26:49] <fenn> i still dont see how that's any different from psyco
[12:34:07] <BigJohnT> I don't know anything about either one but thought it might be interesting to someone
[12:54:27] <BigJohnT> hmmm, how can I have an input number freeze when I do a M5 and motion.spindle-on goes false
[13:00:13] <jepler> you mean you have some input X, and while the spindle is running you want Y to change as X changes, but while it's not running you want Y to stop changing and just retain its last value?
[13:01:12] <BigJohnT> well kinda sorta... I have an offset to Z and when the spindle stops I want to retain that offset until I move Z up a distance
[13:01:30] <BigJohnT> and gradually remove the offset as I raise Z
[13:01:58] <jepler> for the simple thing I said, 'tristate_float' would work
[13:02:07] <jepler> for the more complicated thing you just said, it's not a solution
[13:03:44] <jepler> personally I'd figure out what's required and write a comp, since it sounds like you need to deal with non-whole numbers and as far as I know classicladder isn't good for that.
[13:04:59] <BigJohnT> the offset changes as I move X and Y (torch height control) but at the end of a cut I need to safely raise the torch
[13:06:11] <BigJohnT> I have the offset working kinda using a mux2 and sum2 to hijack the axis.2.motor-pos-cmd
[13:06:44] <BigJohnT> but when the torch goes off with M5 I will loose the offset signal...
[13:07:06] <BigJohnT> and I don't want the torch to slam into the material if the offset was up
[13:12:26] <BigJohnT> maybe sample-hold will do it
[13:13:36] <jepler> yes -- sample_hold will do like tristate_float and allow you to preserve the last height value when the spindle turns off
[13:13:36] <jtr> Could you just add another offset when torch is off?
[13:13:54] <jepler> but you wanted something more complicated than that :-P
[13:15:10] <BigJohnT> once I hold it then I can take the action to reduce it ... one step at a time :)
[13:16:21] <jepler> sure
[13:36:30] <BigJohnT> actually the tristate_float looks like more of what I need for this step
[13:40:53] <BigJohnT> and writing a comp will be the final solution, but I wanted to explore as far as I can with existing components it might help me understand better what I need in the end
[14:09:42] <skunkworks> mach is a virus! ;) http://cnczone.com/forums/showthread.php?t=77427
[14:10:14] <JymmmEMC> skunkworks: You REALLY need to get a life and stop spending so much time on cnczone.
[14:10:16] <skunkworks> I also like the ger21's explaination of machs max steps/second
[14:10:45] <skunkworks> JymmmEMC: maybe I need a 12 step program...
[14:10:49] <BigJohnT> * BigJohnT is off for a ride on the bike
[14:10:57] <JymmmEMC> BigJohnT: bastard!
[14:11:01] <JymmmEMC> enjoy!
[14:11:19] <BigJohnT> how do you make the smiley with the tounge sticking out :)
[14:11:44] <BigJohnT> just a short ride to mammoth springs ar
[14:11:51] <JymmmEMC> BigJohnT: That depends, are you sitting bitch on your ride?
[14:12:02] <cradek> I like how cnczone has all their ads on ads.cnczone.com to make it easy to block all the blinking crap
[14:12:04] <BigJohnT> never
[14:12:19] <skunkworks> heh
[14:12:32] <JymmmEMC> BigJohnT: Such a shame, it leaves your hands "available"
[14:12:44] <BigJohnT> LOL
[14:13:50] <tomp> ooh bmovedirive in just started hunchback of noter dame with lon chaney ( remember the dr pepper ad 'read my lips? )
[14:14:08] <JymmmEMC> cradek: AdBlock Plus FTW!
[14:14:32] <tomp> mplayer -autosync 30
[14:15:31] <JymmmEMC> So, I goto cancel my paid yahoo domains account. Click on CANCEL, damn thing takes me to a CALL US page.
[14:15:58] <skunkworks> sounds like aol ;)
[14:17:54] <JymmmEMC> I personally wouldn't know. Many moons ago Yahoo Domains offered DNS and IMAP, They discontinued IMAP so screw em.
[14:18:15] <JymmmEMC> I've had it for maybe 10 years
[14:19:45] <JymmmEMC> The vhosting I have now offers everything except full control of DNS, but registar now offers that, so it's all good. I maintain control of DNS
[14:34:52] <tomp> hard real time networking based on firewire http://www.rtfirewire.org/paper.pdf
[14:35:13] <tomp> uses xenomai
[14:37:22] <tomp> its 'deterministic' :)
[14:39:08] <tomp> i see where mel broks got the stuff for 'start the revolution without me'... all in hunchback ( the B&w silent version)
[14:41:07] <tomp> the paper on fire wire specificly address pwm and encoder feedback
[14:43:04] <JymmmEMC> tomp: Issue being... firewire is dead
[14:43:39] <jepler> tomp: the timings in figure 11 look promising, if I understand correctly how they relate back to the left hand part of figure 3. If you can request to read inputs and then get the response within 160us 100% of the time, you're in the realm that emc2 requires without changing the fundamental way its servo loop works.
[14:45:15] <JymmmEMC> See figure 14
[14:45:24] <JymmmEMC> pg 7
[14:45:25] <jepler> payloads of 56 bytes are pretty small; I don't even think that's enough for 2 hostmot2 encoder reports
[14:46:20] <tomp> yes, the graphs show 100% and less, they admot they miss stuff at > 5khz ( if i read correctly)
[14:46:35] <tomp> firewire dead? huh? i see it in shops all the time
[14:47:06] <tomp> paport is more deader than firewire, firewire was just never as alive ( if thatr makes snese )
[14:47:27] <JymmmEMC> tomp: Even the new apple macbooks dont come with firewire anymore, I think they are holding out for USB3
[14:48:02] <tomp> is usb3 suited to realtime? firewire is
[14:48:40] <tomp> yes 56 bytes is small, hard to shove much in that space ven binary
[14:48:50] <JymmmEMC> FW just isn't commonplace, as is ethernet
[14:49:05] <tomp> true ( not as alive as...)
[14:49:16] <JymmmEMC> I use/like FW for ext devices, but the industry is just moving away from it
[14:49:19] <jepler> (oh, I was confusing bytes and bits; it's not quite as bad as I said, but it's still not great)
[14:49:31] <tomp> the paper compare rt ethernet to rt firewire, and ether had an edge
[14:51:57] <JymmmEMC> woohoo! all domains moved over and are active!
[14:53:23] <tomp> cool, whose the new provider?
[14:53:52] <JymmmEMC> actually old vhost, just migrated all the domains into one
[14:54:22] <JymmmEMC> well, I mean revived dead ones, put DNS for all domains to one registar, etc
[14:58:36] <tomp> oh i thought old was aol
[14:58:59] <JymmmEMC> I've NEVER used aol
[14:59:10] <tomp> oops
[14:59:55] <tomp> oh yahoo maybe
[15:06:16] <tomp> haha, no audio on silent movies (hinchback of nore dame) , so this guy keeps stoppingthe other guy from eating, all silent, as good as any dialog would have done
[15:18:02] <tomp> he rtethernet is descibed here http://rtfirewire.org/RTnet-ETFA05.pdf
[16:45:19] <als> on the Beta release halvcp has been removed, use pyvcp instead does this mean the old vcp will not work in 2.3?
[16:45:46] <jepler> yes, that's right.
[16:47:03] <als> there was talk of a translator being wrote to port them over?
[16:47:28] <jepler> as far as I know, nobody ever wrote one
[16:48:15] <als> when will 2.3 be released?
[16:49:00] <jepler> I hope it'll be sometime this month
[16:51:52] <als> did someone write a new halui test panel using pyvcp?
[16:52:38] <jepler> yes, I believe that's what the sample configuration halui_pyvcp demonstrates.
[16:53:24] <als> ok that's on the beta release
[17:19:15] <skunkworks> * skunkworks wishes he had 220v in his manroom.
[17:20:16] <skunkworks> or a 1:1 transformer
[17:23:32] <skunkworks> * skunkworks is not going to rectify the mains!
[17:30:01] <cradek> yeah, that's a bad idea.
[17:30:15] <skunkworks> :)
[17:31:30] <skunkworks> wow - we actually have 130v here
[17:32:35] <LawrenceG> skunkworks, whats the problem... the transformer is on the pole!
[17:32:52] <skunkworks> that would be a bit too much. (183v rectified and filtered)
[17:33:26] <skunkworks> looking for around 160
[17:34:06] <skunkworks> at 95%dc that would give around 150v at full
[17:34:55] <skunkworks> Well - should I unhook the ground and neutral at the box? ;)
[17:35:06] <LawrenceG> http://imagebin.ca/view/1b7EG6.html
[17:36:22] <skunkworks> LawrenceG: nice - what are you looking for performance?
[17:37:12] <skunkworks> R1 is current sense?
[17:37:37] <LawrenceG> latest proto ... now I have to try and mill it... 15amp cont, up to about 80v... also looking at adding a 3rd half brdge so I can try and spin my brushless motors
[17:38:20] <LawrenceG> using homemade shunt on back of board
[17:38:48] <jepler> oh now I spotted r1
[17:39:26] <LawrenceG> probably shim stock soldered between the 2 traces in a loop
[17:40:06] <skunkworks> logic goes on a different board?
[17:40:11] <LawrenceG> board is pretty small.. somewhere around 1.75x3.5"
[17:40:37] <jepler> it seems like there must be some tricks to getting good behavior of the current comparator. ISTR that the 20A current limit on skunkworks's board is 300mV (.015ohm sense resistor) but the comparators I was looking at had statistics like 200mV input difference for guaranteed output saturation
[17:40:44] <skunkworks> If I could replace all the logic with an avr or clpd - the board would be tiny
[17:40:49] <LawrenceG> yes... dspic drives 1/2 brdges and shunt voltage processed off board
[17:41:05] <jepler> skunkworks: I agree, you should do that
[17:41:54] <skunkworks> jepler: I am really seriously thinking about it. (for a good learning proccess)
[17:41:57] <LawrenceG> the plan is to bolt 3 of these to a breakout board that has parallel port buffering, 3 dspics
[17:42:19] <LawrenceG> leds and other fluff
[17:43:07] <skunkworks> LawrenceG: the 393 seems to be working great. (includes ground)
[17:43:20] <skunkworks> with a 3.3k pull up.
[17:43:37] <skunkworks> others use a opamp with +/- supply
[17:44:15] <LawrenceG> excellent... I want to avoid another supply.... already need +5, +12 and motor volts
[17:44:24] <skunkworks> I can adjust the current limit down to 0.
[17:44:26] <skunkworks> which is cool
[17:45:47] <LawrenceG> this was first board after upgrading to eagle 5.4.... they changed a few things!
[17:45:56] <skunkworks> Jepler: could I measure the voltage across the sense resistor directly with a avr?
[17:46:04] <skunkworks> * skunkworks should really do some research.
[17:46:37] <skunkworks> I see they have a a/d
[17:46:50] <LawrenceG> skunkworks, what was the design voltage drop for your shunt?
[17:48:10] <skunkworks> max of .3v - the sense resistor has a max current rating of 22a (Is that what you mean?)
[17:49:51] <jepler> it's a .015ohm, 7W resistor, 300mV at 20A, correct?
[17:52:00] <skunkworks> yes
[17:53:07] <skunkworks> that is 6w
[17:53:55] <skunkworks> although I really don't think normal usage will aproach that. And acc should be saftey under that
[17:54:28] <skunkworks> they rate the resistor at something like 5 times the watage for 5 seconds
[17:56:22] <skunkworks> wattage
[17:57:41] <skunkworks> and if this works well - I could theoretically put another resistor on the back of the board in parallel. That alow for a few more amps.
[17:59:40] <skunkworks> I do have some .005ohm resistors also - but the they still have the 22a max rating
[18:00:12] <skunkworks> The last redesign was to really try to make everything as symetrical as I could.
[18:00:36] <skunkworks> I think it really helped with the symitry of the current limit.
[18:01:50] <skunkworks> symmetry
[18:02:54] <skunkworks> (and blanking)
[18:14:24] <skunkworks> as of right now - I am running .95 dc into the drive at max. The current limit voltage is set to 245mv and the current is tripping at aprox 15a. That is really close. The calculated trip is 16.3. that is using a clamp dc amp meter at the motor. (And I really don't know how accurate that is)
[18:14:51] <skunkworks> * skunkworks is really happy.
[18:15:45] <skunkworks> I set the divider so I could run it up to .363v
[18:20:57] <LawrenceG> skunkworks, my board design is for very low cost servo... the most expensive part is the 1000uf/100V cap.. could be cheaper for lower voltage motors
[18:21:38] <skunkworks> yes. - You can use mosfets that have lower rds (Rds goes up with operating voltage) For the most part.
[18:21:49] <LawrenceG> minimal use of terminal/connectors
[18:22:49] <skunkworks> Yah - I am using the cheap screw terminals for the power/pwm/enable. the motor connectors are soldered directly to the board. Power in is sandwiched into the cap scres
[18:22:51] <skunkworks> screws
[18:23:01] <LawrenceG> the IRF540's have a very nice range and are not that expensive.... I have a bunch of sample fets from on semi that are rated at 60V... they would work for 24v servo motors just fine
[18:24:24] <LawrenceG> but a design is not finished until the smoke has leaked out a few times
[18:25:29] <LawrenceG> I have a nice 500w/24v motor here I am going to use for worst case testing.... it used to run the blade on a battery powered lawn mower!
[18:26:04] <skunkworks> yah - the igbts I am using are like 8.45 new. (I got them for 4 for 3.49)
[18:27:29] <LawrenceG> I saw the ebay link.... a good price... tempted to buy a tube from the guy, but I dont really have a target system in that power range.. I am thinking bridgeport or smaller size
[18:28:21] <LawrenceG> mostly hobby router table drives
[18:29:48] <LawrenceG> I want to replace the output stage on my dspic servo design... this board should have a parts cost < half the price of the opa549 it currently uses
[18:30:01] <skunkworks> Neat
[18:30:27] <skunkworks> that would work good for the puma arm. I was planning on making a small version with something like irf540s.
[18:31:09] <Optic> moop
[18:31:22] <skunkworks> I think I have a bunch from my try at stepper drives. It was a lot less successful. ;)
[18:32:45] <LawrenceG> now you have the current limit stuff working, it may be usable for steppers.... it would be interesting to have a computer generated reference for the current limit
[18:34:31] <skunkworks> It was a long time ago. Like 15 years
[18:35:31] <skunkworks> I really have no reason to make a stepper drive anymore. We have a bunch of oem650s and assorted other comercial drives we can use. They work a ton better
[18:36:54] <LawrenceG> yea.. I really like the oem750's I got of epay... pretty painless way to get steppers going.... I would hate to clone one of them.... there are thousands of parts on their PCB's
[18:37:34] <skunkworks> yes - and they microstep and everything
[18:38:44] <LawrenceG> I was watching for some of the Parker servo drives, but never really found enough of one model to do a system with
[18:38:57] <skunkworks> heh
[18:39:52] <LawrenceG> in another couple of years, we will be able to salvage drives and motors out of washing machines
[18:40:41] <LawrenceG> our new washer has a really cute direct drive system and controller
[18:41:10] <LawrenceG> but I need clean shorts more than a 1 axis servo drive!
[18:41:26] <skunkworks> heh
[18:41:43] <skunkworks> someone is using one - sort of a large stepping motor.
[18:41:50] <skunkworks> I cannot remember where I saw that.
[18:42:45] <LawrenceG> If I get the 3phase bridge done, I want to try using a car alternator.... it might make a nice high speed spindle drive
[18:43:30] <LawrenceG> the alternators must spin >10,000rpm
[18:45:35] <skunkworks> http://www.instructables.com/id/Make-your-own-cnc-controller/?comments=all
[18:46:15] <LawrenceG> and > 1hp output 13v/100amp is close to 2hp
[18:50:59] <LawrenceG> that link has a pretty wild looking constant current driver!
[18:51:27] <LawrenceG> that would probably run an alternator
[18:52:50] <LawrenceG> Q1 would be better replaced with a lm393 and a reference that is adjusted by the CPU for pseudo sine wave current drive
[19:38:36] <skukworks> skukworks is now known as skunkworks
[19:56:39] <jepler> hm, this datasheet calls for an impossible temperature profile. http://emergent.unpy.net/index.cgi-files/sandbox/solder-profile.png -- to get from 220 to 255 degrees at 4deg/second takes (255-220)*4 = 35*4 = 140 seconds, so how can you spend only 60 seconds above 220?
[20:00:48] <fenn> that way it's not their fault when it doesn't work :)
[20:02:41] <fenn> btw it's 4 deg/s not 4s/deg so that would be 2*(255-220)/4 <= 17.5 sec above 220
[20:02:49] <fenn> oops >= i mean
[20:03:06] <jepler> oh duh
[21:13:40] <skukworks> logger_emc: bookmark
[21:13:40] <skukworks> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-04-04.txt
[22:32:15] <cradek_> cradek_ is now known as cradek
[23:50:05] <shayter> hey -- is there anyone out there that can answer a torque-related question?
[23:54:48] <shayter> Ok. I have another question re: latency and video drivers.