#emc | Logs for 2008-11-19

[00:00:31] <dmess> but then again he had over 60 screw machines in his shop at 1 time... only ever 1-3 running..
[00:00:54] <Gamma-x> sometimes i think of just getting a centroid retro kit just for the reliability and also the features. anyoen agree. not putting down emc at all....
[00:01:38] <dmess> completely DIS agree... that my vote
[00:01:54] <Gamma-x> but for what reason?
[00:03:01] <dmess> features are thing you dream and build... i dont want to be handed "features" i have no use for.. gimme the engine to make what i need.. and support it
[00:03:31] <Gamma-x> ok
[00:03:43] <dmess> im a DIY kinda cnc programmer
[00:04:27] <dmess> if i need a new CUSTOM G code i wanna be able to do it...
[00:05:04] <Gamma-x> true
[00:08:58] <archivist> and updates to software really happens with EMC with devs listening
[00:15:07] <fenn> Gamma-x: if you dont understand how your machine is built, you wont be able to fix it _when_ it breaks
[00:15:16] <Gamma-x> true
[00:16:59] <Gamma-x> there is gunna be a lot of extra parts left over when im done.... gunna be weird! lol
[00:21:47] <Gamma-x> any tool touch sensor in emc yet?
[00:23:38] <archivist> sensor is a hardware device
[00:24:17] <jepler> g38.2 has been usable for that purpose for ages. in the next version it will get more useful because there will be a new "set tool table entry" code to actually put a value computed from the probed value in the tool table. http://linuxcnc.org/docs/devel/html/gcode_main.html#G10 L1 Set Tool Table
[00:24:42] <jepler> (hm that link may not go to the right spot; trust me, it's in there)
[00:25:02] <fenn> jepler: is there a way to use probed values in arbitrary g-codes?
[00:25:42] <jepler> fenn: sure. the probed value goes in #5061..#5069 so you can do anything with it
[00:25:51] <fenn> aha
[00:38:32] <dmess> touch probe is all about the high speed skip signal... is that implemented yet??
[00:40:58] <dmess> as long as i know where it gets put our macros will spin and relocate the to hells back acre if you want it there...
[00:46:01] <dmess> we also have a center of rotation macro... so fixures dont have to be @ COR and still line bore from bothe sides and meet .001"-.002" true position in line...
[00:47:11] <dmess> that one a Very good friend of mine's macro.. and is bulletproof across many fanuc controlers
[00:48:00] <dmess> it would only take minor tweaks to run in emc
[01:02:21] <skunkworks> jmkasunich: http://imagebin.ca/view/uHmAw1qe.html
[01:02:56] <jmkasunich> skunkworks: yes?
[01:03:10] <jmkasunich> new version?
[01:03:26] <skunkworks> I am trying to tighten the ciruit up... I don't like how I am getting the driver ic ground.
[01:03:59] <skunkworks> for the lower current version. (not using the extra free-wheeling diodes)
[01:04:02] <jmkasunich> this is a different circuit? diodes integrated with the fets?
[01:04:07] <skunkworks> yes
[01:04:50] <skunkworks> .050 isolation
[01:05:40] <skunkworks> the lower right pad of the resistor is logic ground. Now that has to get back up to the driver ic. I don't see any way to make it any better than running it around the outside.
[01:05:51] <skunkworks> (not having more than 2 layers.)
[01:06:00] <skunkworks> and the other question is - does it matter :)
[01:06:10] <jmkasunich> layout always matters
[01:06:24] <jmkasunich> be carefull with the term "logic ground" tho
[01:06:28] <jmkasunich> there are several grounds
[01:07:02] <jmkasunich> the driver ground needs to provide a return path for gate drive current
[01:07:05] <skunkworks> reading the app notes - make things as short as possible. (why I put only the driver ic's in the h-bridge)
[01:07:12] <jmkasunich> that has nothing to do with the current sense ground
[01:07:41] <jmkasunich> you have less than a volt across the sense resistor, right?
[01:07:49] <skunkworks> yes
[01:08:01] <skunkworks> could I hook ground to the bottom of the mosfets?
[01:08:12] <jmkasunich> that's what I'm pondering
[01:08:13] <skunkworks> (seems not right) ;)
[01:08:24] <jmkasunich> there are potential risks and benefits
[01:08:27] <skunkworks> it is .015 ohms
[01:08:40] <jmkasunich> plus probably a hundred or two nano-henries
[01:08:45] <skunkworks> yes
[01:09:01] <skunkworks> well - it says at that resistance it has 'no' inductance
[01:09:14] <jmkasunich> the most important thing that connects to the bottom right resistor pin is the current limit comparator
[01:09:18] <jmkasunich> nothing has no inductance
[01:09:24] <skunkworks> :)
[01:09:36] <skunkworks> yes.
[01:10:02] <jmkasunich> I've measured 50V across a 6" length of 1/4" thick x 2" wide busbar, due to inductance * dI/dT
[01:10:18] <skunkworks> yecky
[01:10:36] <jmkasunich> can be dealt with, as long as you are aware of it
[01:10:44] <jmkasunich> anyway, you fortunately don't have that problem
[01:11:51] <skunkworks> .3v at 20a across the sense resistor
[01:12:08] <jmkasunich> I'd put the current limit comparator below (preferred) or to the right of the current sense resistor
[01:13:11] <jmkasunich> I'd also extend the blue flood that is attached to the top left resistor terminal so it is under the entire resistor, and goes around the bottom of pad4 and connects to the source of the bottom right fet
[01:13:12] <skunkworks> I was wondering if I had the drivers ground hooked to the bottom of the mosfets and the rest hooked to the sense resistor. (puting the comparator right at the resistor.)
[01:13:26] <jmkasunich> thats about what I was thinking
[01:13:35] <jmkasunich> gotta look closer tho - where is the schematic
[01:13:49] <skunkworks> give me a second.
[01:17:02] <skunkworks> wow - my site is slow.
[01:22:32] <skunkworks> jmkasunich: sorry http://www.electronicsam.com/images/KandT/servostart/schemmore.png
[01:26:20] <skunkworks> hmm - I am trying to understand - 'I'd also extend the blue flood that is attached to the top left resistor terminal so it is under the entire resistor, and goes around the bottom of pad4 and connects to the source of the bottom right fet'
[01:34:24] <JymmmEMC> skunkworks: I think that means extend the copper area under the resistor to act like a heatsink
[01:34:47] <JymmmEMC> or clear all of it under the resistor
[01:34:49] <skunkworks> now that I know the quirks of eagle.. I really like it.
[01:34:51] <JymmmEMC> one of the twp
[01:36:13] <JymmmEMC> I just got back from the range... I'm gonna have to find some other stress reliever now; guess I'm a lil too complacit (sp?) now.
[01:37:20] <JymmmEMC> which really really sucks btw!
[01:37:51] <skunkworks> ?
[01:38:22] <JymmmEMC> skunkworks: When I'm really stressed, I hit the range. I hit the range today, it didn't help this time =(
[01:38:48] <skunkworks> ah
[01:39:39] <jmkasunich> skunkworks: I mean, put copper under that resistor
[01:39:40] <JymmmEMC> When I hit the range, I focus on the task at hand (lethal weapon and all), so usually everything else kinda loses focus, and the mind has a chance to relax.
[01:40:08] <jmkasunich> and wrap the copper area around below Pad4, connect to the blue area on the lower right
[01:40:13] <jmkasunich> inductance thing ;-)
[01:41:32] <JymmmEMC> Maybe I just need to fire bigger bullets, gawd knows I've tried quantity =)
[01:42:28] <skunkworks> ah - ok - that makes sense. I would just have to use the red layer to connect logic then.
[01:43:19] <skunkworks> (to the resistor)
[01:50:33] <jmkasunich> right
[01:50:56] <jmkasunich> JymmmEMC: you need reactive targets
[01:51:07] <JymmmEMC> jmkasunich What's that?
[01:51:20] <JymmmEMC> the kind that explode?
[01:51:28] <jmkasunich> something that shows when you make a hit
[01:51:36] <jmkasunich> yes, or steel plates that fall down, etc
[01:51:37] <jmkasunich> http://www.boomershoot.org/
[01:52:11] <K`zan> Anyone here know anything about these drives: http://imshome.com/im483.html
[01:52:27] <JymmmEMC> jmkasunich ah, They only allow paper targets at this range I went to. I'll try a different range next time.
[01:52:48] <jmkasunich> I can only shoot paper too, at all the local ranges I'm aware of
[01:52:49] <K`zan> Seem to be pretty good from what little I know and are bipolar which would mean I don't have to derate the steppers like I do for unipolar?!
[01:52:58] <jmkasunich> haven't gone shooting in over a year
[01:53:11] <jmkasunich> K`zan: what do they cost?
[01:53:30] <JymmmEMC> jmkasunich At another range, you can bring anythign except food or glass. Got an old washing machien you want to turn into swiss cheese? no problem =)
[01:53:34] <K`zan> Any responses, please prefix with my nick, I have been running around like a madwoman all damn day.
[01:53:51] <K`zan> jmkasunich: About $45 each used....
[01:54:07] <jmkasunich> used?
[01:54:26] <jmkasunich> for that power level, you ought to be able to get new microstepping drives for <$50 per axis
[01:55:13] <JymmmEMC> K`zan: Hey gal, why not one of these? http://www.geckodrive.com/product.aspx?c=3&i=14469
[01:55:14] <jmkasunich> and I wouldn't buy used electronics unless I can either test it first, or have a money-back guarantee, or it is very cheap (like 10% of new price)
[01:55:46] <jmkasunich> the G540 is about $75 per axis, or $100 per axis if you only need 3
[01:55:53] <jmkasunich> not quite in the same budget catagory
[01:56:10] <JymmmEMC> The G540 is 4 axis for $300
[01:56:25] <jmkasunich> 300 divided by 4 = $75
[01:56:31] <JymmmEMC> wait, you just said that =)
[01:56:38] <JymmmEMC> sorry, long day
[01:56:38] <jmkasunich> if you only need three, the fourth is a waste, so that means $100 per axis
[01:56:50] <JymmmEMC> spare?
[01:57:08] <jmkasunich> I believe K`zan is on a budget
[01:57:36] <jmkasunich> the G250 is a single axis for $46 - basically the same for a new drive as the used ones she originaly posted
[01:58:07] <jmkasunich> the G250 has basically the same power ratings, and comes from a company with a good reputation for product quality and support
[01:58:09] <K`zan> Uh, yah we are :-/.
[01:58:15] <JymmmEMC> Yeah, I understnad, but she's been on a budget for years try to go the cheapest route. If these were available when I bought the xylotex... IN A HEARTBEAT
[01:58:22] <K`zan> $135 vs $300 make BIG difference to me.
[01:59:06] <K`zan> Yes, I have, life has this nasty way of interfering with what I WANT to do :-) :-/ :-(.
[01:59:14] <SWPadnos> I think there's a G251, which nas screw terminals instead of pin headers (or it's the other way around)
[01:59:17] <JymmmEMC> k` You've had the same $150 budget for a couple years now. Couldn't ya collect cans and bottles or sumthin ? =)
[01:59:18] <SWPadnos> hes
[01:59:20] <SWPadnos> has
[01:59:25] <jmkasunich> the G450 is just a fancy way of interconnecting 4 of the G250's
[01:59:28] <K`zan> Do like the idea for that thing though, a lot.
[01:59:36] <jmkasunich> do the interconnect yourself and save $100
[01:59:42] <JymmmEMC> jmkasunich saves wiring headaches
[02:00:01] <K`zan> What happens if you blow one? Replace a 250 or toss it all and get another 450?
[02:00:02] <jmkasunich> JymmmEMC: yeah, and if you value time more than money, its a great idea
[02:00:32] <K`zan> Looking over the datasheet on that thing, it looks a lot like a gecko...
[02:00:36] <jmkasunich> K`zan: I'm sure if you toast one of the 250's inside the 540, you can replace just that drive
[02:00:39] <JymmmEMC> K`zan: I think you can just replace the blown one, plus you would already ahve a spare
[02:00:55] <K`zan> That is good then.
[02:01:02] <jmkasunich> but the only thing I know about that product is what I read on the webpage, you should read the fine print yourself
[02:01:16] <K`zan> Hate it when one of something combined blows and it becomes (expensive) trash...
[02:01:24] <jmkasunich> I do know that gecko is good folks - I use their other drives, and I've talked with the owner
[02:01:30] <JymmmEMC> k`And if you fess up that you made a mistake and blew th drive, Mariss will replace it (one time) for free
[02:01:38] <SWPadnos> if you blow one axis drive, just toss it and plug in another
[02:01:52] <K`zan> Did before I asked here. Looks damn good to me, esp at the price. Should beat hell out of the HCNC one I have.
[02:01:53] <SWPadnos> if you blow something on the "mainboard", I don't know that there's a replacement part
[02:02:06] <SWPadnos> but theoretically you don't have to replace the drives or the enclosure
[02:02:08] <K`zan> jmkasunich: Can you get me 4 samples :-)?
[02:02:35] <JymmmEMC> K`zan: I am serious about the "fessing up" part.
[02:02:51] <K`zan> No, the drive works fine, but one has to derate the steppers quite a bit with the unipolar drivers.
[02:03:38] <JymmmEMC> SWPadnos: Compare the G540 with JUST breakout boards alone. Seems worth it to me
[02:03:45] <K`zan> I go from a 284 to a 485 (IIRC, but close to that) because it is unipolar to get the equivalent oz/in rating for the stepper.
[02:04:19] <jmkasunich> neat, the 540 even comes with the mating connectors for the motors
[02:04:20] <SWPadnos> I absolutely hate the new gecko site
[02:04:28] <K`zan> I may just byte the bullet and get huskier steppers for the time being. Dunno, seems the more I learn the less I know :-/.
[02:04:51] <JymmmEMC> jmkasunich Yeah, just db9's
[02:04:56] <SWPadnos> finally noticed that about the world, eh? :)
[02:05:34] <SWPadnos> ok yes, the G251 has sscrew terminals, and is $49
[02:05:47] <K`zan> JymmmEMC: I have no problems with the HCNC drive other than the stepper derating issue. More expensive steppers than if it was bipolar.
[02:06:09] <SWPadnos> that's due to the annual sale, where you get the 100-piece prive
[02:06:13] <SWPadnos> normally $69
[02:06:17] <K`zan> SWPadnos: :-), No not just, but does it have to infect EVERYTHING :-) LOL.
[02:06:24] <SWPadnos> heh
[02:06:43] <K`zan> :-) *NEVER* a dull moment.
[02:06:55] <SWPadnos> note that the G250 doesn't have a heatsink either, though the G251 does
[02:07:07] <K`zan> Sigh, my brain hurts, perhaps I'll just contemplate a new video card for the moment...
[03:11:49] <JymmmEMC> http://www.via.com.tw/en/products/embedded/vp7710/
[03:58:40] <Gamma-x> i need help findin an encoder for me servo
[03:58:45] <Gamma-x> sem mt39 i believe
[03:58:52] <Gamma-x> i know its a 27pnd servo
[04:10:28] <SkullWorks_PGAB> what is the rear shaft size?
[04:12:41] <Gamma-x> 5mmi believe
[05:07:00] <maddash> what's the best (simplest) way to compensate for noise in an ADC conversion?
[05:07:30] <JymmmEMC> twisted pair
[05:07:44] <JymmmEMC> faraday cage =)
[05:07:59] <SWPadnos> averaging, but it doesn't work near the top or bottom of the range
[05:08:16] <maddash> SWPadnos: i tried that, but even the average moves up and down
[05:08:28] <maddash> SWPadnos: a rolling average
[05:08:31] <SWPadnos> then it's likely that the signal is moving up and down
[05:08:47] <SWPadnos> you did a rolling average?
[05:08:52] <maddash> yep
[05:09:14] <SWPadnos> ok - that should filter high frequency noise, but not low frequency
[05:09:22] <SWPadnos> (any averaging does that)
[05:09:52] <SWPadnos> but if you have one really high sample, it will "pollute" N-1 results (where N is the number of samples averaged)
[05:09:57] <SWPadnos> or maybe N results
[05:10:55] <SWPadnos> is this an internal A/D in a microcontrller?
[05:11:06] <maddash> yep
[05:11:38] <SWPadnos> well, first off, do you know for sure that the noise isn't actually there?
[05:12:04] <maddash> see, there isn't really that much noise (+/- 2 LSB), but I'm trying to display the results on an LCD...and because the values change so quickly, the LCD results are almost unreadable
[05:12:05] <SWPadnos> ie, if you look at the input with a scope, do you see noise?
[05:12:19] <maddash> no noise in the scope
[05:12:27] <SWPadnos> oh, you should do ~1/8-1/2 second averages for display
[05:12:39] <maddash> whoa
[05:12:48] <SWPadnos> there's no reason to update an LCD faster than people can read
[05:12:53] <SWPadnos> or not much faster anyway
[05:13:55] <maddash> would a low pass filter ( the kind that is used to translate PWM to analog output) "even out" the noise?
[05:14:03] <JymmmEMC> 2x16 ?
[05:14:13] <maddash> JymmmEMC: 20x4
[05:14:19] <JymmmEMC> ah
[05:14:36] <JymmmEMC> no serial backpack?
[05:14:45] <maddash> nope
[05:14:53] <JymmmEMC> you suck =)
[05:14:57] <SWPadnos> if you don't see noise on a scope (that's set properly), then conditioning the signal won't help
[05:15:15] <SWPadnos> a couple of LSBs is about as good as an internal ADC will get
[05:15:43] <SWPadnos> the things to look at are proper power supply bypass caps and proper decoupling between the analog and digital supply pins
[05:16:51] <SWPadnos> also, if there's a separate Vref pin, make sure it's bypassed nicely, following the datasheet instructions
[05:17:36] <maddash> so no amount of averaging, whether in hardware or software, will help?
[05:17:59] <SWPadnos> it could
[05:18:21] <SWPadnos> how many samples are you averaging, and is the 1-2 lsb variation before or after averaging?
[05:18:30] <SWPadnos> err, +/-2 LSB
[05:18:47] <maddash> after
[05:18:55] <SWPadnos> #samples?
[05:19:29] <maddash> six, at a 1.44 kHz sample rate
[05:20:07] <maddash> should I average over more samples? or even oversample?
[05:20:10] <SWPadnos> six - odd number :)
[05:20:56] <SWPadnos> so presumably the value you're sampling varies at <500 Hz or so?
[05:21:05] <SWPadnos> or is intended to ata ny rate
[05:21:08] <SWPadnos> at any
[05:21:43] <maddash> 500 Hz is a good number
[05:22:01] <SWPadnos> ok, it should be <100 Hz actually
[05:22:16] <SWPadnos> if you average 6 samples, those should all be considered one sampling period
[05:22:38] <SWPadnos> so you're actually at 1.44k/6 = 180 Hz
[05:22:50] <SWPadnos> uh, no - hold on
[05:23:03] <SWPadnos> (I should have been in bed an hour ago - not thinking well)
[05:23:11] <SWPadnos> 240 Hz
[05:23:23] <maddash> um, I don't think I was clear: I take 1 sample every 1/1440 seconds, then I add it into a rolling avg over the past 6 samples...
[05:23:31] <SWPadnos> ah, ok
[05:23:42] <maddash> SWPadnos: don't let me keep you away from bed
[05:23:43] <SWPadnos> no, that was what I understood
[05:23:48] <SWPadnos> 1.44 kHz
[05:24:03] <SWPadnos> but the sample period is actually 1/(1440/6)
[05:24:15] <SWPadnos> it just slides along at a higher resolution
[05:24:29] <SWPadnos> so you're effectively sampling at 240 Hz
[05:24:44] <SWPadnos> eh, no biggie. you'll get the smoke if I screw up :)
[05:24:59] <maddash> D:
[05:25:02] <maddash> heh
[05:26:19] <maddash> so you're saying that by taking an average, I'm lowering my sampling frequency?
[05:26:40] <SWPadnos> you're lowering the bandwidth
[05:26:51] <SWPadnos> that's what noise filtering does, since most noise is high frequency
[05:27:01] <SWPadnos> stick a cap on the line, and high frequency noise can't get through
[05:27:15] <maddash> a capacitor?
[05:27:18] <SWPadnos> yes
[05:27:33] <SWPadnos> averaging is the equivalent of a parallel capacitor on the input
[05:27:48] <maddash> so sticking a cap will also lower bandwidth?
[05:27:51] <SWPadnos> yes
[05:28:04] <SWPadnos> if you go from the input to ground (or Vcc)
[05:28:19] <SWPadnos> if you put it in series, it will block DC and pass high frequencies
[05:29:12] <maddash> meh, so why waste computation power by avging? I could just take the easier hardware road
[05:29:40] <SWPadnos> heh. for cheap stuff, the question is "why add component and assembly cost when I can put it in the software for free?"
[05:29:51] <SWPadnos> but for one-offs, either way will work
[05:30:38] <SWPadnos> incidentally, if I had the sample rate available, I'd do 4 or 8 samples, so I could divide by shifting instead of actually dividing
[05:31:10] <maddash> oh, heh. I don't actually divide; I calculate a trapezoidal sum
[05:31:35] <SWPadnos> or even better, don't bother dividing at all. if you have a 10-bit A/D and you do 8 samples, just consider the number to be a 13-bit quantity
[05:31:46] <maddash> yep
[05:31:57] <SWPadnos> trapezoidal? so you're weighting the edge values lower than the center?
[05:32:02] <maddash> but isn't that over sampling?
[05:32:18] <SWPadnos> well, no, not really
[05:32:28] <maddash> no, weighting the center 4 values twice more than the first and last values
[05:32:43] <maddash> weighing*
[05:32:43] <SWPadnos> heh, same thing :)
[05:32:58] <SWPadnos> but shifted by 1 bit
[05:33:27] <maddash> if I stuck the cap, but jacked up my sampling frequency to, say, 2.8kHz, I would increase my bandwidth, right? would the noise come back?
[05:33:30] <SWPadnos> eh - look up oversampling on wikipedia. I probably don't know the actual definition :)
[05:33:42] <SWPadnos> you would not increase your bandwidth
[05:33:49] <SWPadnos> due to the cap
[05:34:32] <SWPadnos> if you have a series of filters (the averaging, the cap, another filter ...), then the bandwidtth will be determined by the lowest bandwidth stage (for the most part)
[05:35:25] <maddash> so what would raising the sampling freq do?
[05:36:15] <SWPadnos> it would raise the effective sampling bandwidth, assuming that everything remains the same
[05:37:00] <SWPadnos> if you need to measure 500Hz signals, and you want to use 6 samples per average, then you need 500 * 2 * 6 = 6 kHz or faster sampling
[05:37:00] <maddash> "SWPadnosyou would not increase your bandwidth"
[05:37:24] <SWPadnos> you mentioned sticking a cap (in the circuit, I assumed)
[05:37:48] <SWPadnos> if you add a cap, that might become the dominant factor in system bandwidth
[05:38:11] <SWPadnos> but *only* increasing the sample frequency would increase the effective bandwidth
[05:39:11] <maddash> ah
[05:43:45] <maddash> SWPadnos: thanks
[05:43:51] <maddash> SWPadnos: and good night
[05:43:54] <SWPadnos> oh, sure. good luck with it
[05:43:56] <SWPadnos> heh. night
[09:59:37] <Fisia> check this out: http://www.lockergnome.com/knitter/2007/10/08/boot-linux-faster-with-an-open-bios/
[09:59:44] <Fisia> :)
[10:00:09] <Fisia> http://www.openfirmware.info/OpenBIOS
[10:00:19] <Fisia> :)
[10:01:30] <Fisia> it say:
[10:01:36] <Fisia> "LinuxBIOS takes the approach of loading a small Linux kernel directly into the boot ROM. This approach is increasingly practical now that boards have 1-2MB of flash ROM onboard, according to Seebach."
[10:02:10] <Fisia> ..
[10:07:17] <Fisia> http://www-128.ibm.com/developerworks/linux/library/l-bios.html?ca=dgr-lnxw09OpenBIOS
[10:07:29] <Fisia> "What has the BIOS ever done for us?"
[10:10:04] <Fisia> ...
[10:10:27] <Fisia> in my opinion, if we had only little bit spaces on our ROM-Bios, just direct initialize on our Flashdisk-storage
[10:11:07] <Fisia> then flashdisk mount out the load TO RAM and initialize all the REST ...!!
[10:12:38] <Fisia> ..eliminate unwanted periperal PROBE...that would speed all up
[10:42:27] <fragalot> isn't that basically what EFI is?
[10:42:31] <fragalot> (the kernel thing)
[11:47:15] <archivist> * archivist is bored, drilling a dividing plate
[11:47:47] <archivist> cnc , set number, sit back ...wait :)
[14:44:30] <cradek_> cradek_ is now known as cradek
[14:58:08] <stustev> cradek: the cinci acts just like yours - I don't know what I was doing yesterday - I was on the phone with Aram and working on the machine - I must not have been watching close enough - thanks
[15:00:41] <stustev> the only problem is Aram probably will not understand yet - maybe I will call him today - It will have to be late today - after he gets home
[15:02:22] <cradek> you are our hero for calling him and helping
[15:02:33] <archivist> :)
[15:03:52] <SWPadnos> heh
[15:06:57] <jymm> Mornin
[15:08:27] <SWPadnos> hiya
[15:08:37] <jymm> * jymm just had an evil thought... Disgruntaled employee programs PBX to randomly move phone numbers to various phones every 30 seconds
[15:08:58] <SWPadnos> you don't know any disgruntled employees now, do you?
[15:09:05] <jymm> just cradek
[15:09:10] <archivist> evil boy
[15:09:16] <jymm> but I dont think he has a pbx
[15:11:42] <jymm> Imagine if the gf called and the random guy that answers sounds like her bf....
[15:18:16] <jymm> and the girl that calls sounds like random guy's gf, and she starts bitching him out... oh man LOL
[15:29:27] <skunkworks_> hmm - so - a max of .5V +/- between logic ground and driver ground.. Discuss.
[15:29:44] <skunkworks_> good? - bad? - not a problem?
[15:30:05] <skunkworks_> (actually probably 350mv at the most.
[15:30:41] <archivist> =amps and resistance
[15:30:54] <archivist> and a bit of inductance
[15:31:32] <SWPadnos> um. if the driver has logic inputs, then that's bad
[15:31:45] <skunkworks_> I am trying to decide if I can hook the gate driver's ground to the lower mosfets - and the rest of the logic to the actuall ground of the current sense resistor.
[15:31:46] <SWPadnos> if it's meant to be an isolated ground, then it doesn't matter
[15:32:10] <SWPadnos> oh - 0.5V - I saw that as 5V :)
[15:32:14] <skunkworks_> heh
[15:32:36] <SWPadnos> you do need to know the impedance of the source of the difference
[15:32:58] <jepler> SWPadnos: back at fest you mentioned that you put a switching power supply on many of the boards you design. what chip do you prefer? is it a buck or a boost topology?
[15:32:58] <SWPadnos> if it's a low impedance source (say a motor returning energy), then it's bad
[15:33:06] <SWPadnos> LM2675
[15:33:31] <SWPadnos> it's a - err - it needs a higher supply voltage, it won't boost
[15:33:37] <jepler> buck
[15:33:43] <SWPadnos> :)
[15:34:12] <SWPadnos> there are similar ones that have buck/boost designs, so they're more universal
[15:34:16] <skunkworks_> 01:13:12 <skunkworks> I was wondering if I had the drivers ground hooked to the bottom of the mosfets and the rest hooked to the sense resistor. (puting the comparator right at the resistor.)
[15:34:27] <skunkworks_> 01:13:26 <jmkasunich> thats about what I was thinking
[15:34:40] <SWPadnos> same shit, different day :)
[15:35:29] <skunkworks_> that is about where I am at. the drivers show a logic min/max of min-2.7 for logic 1 and max of .8v for logic 0
[15:38:33] <jepler> hm, this drawing of a TSSOP-14 package doesn't quite make sense. It has the pin spacing labeled as "12x 0.65", but there aren't 12 places where there's an 0.65 space between two adjacent pins
[15:38:44] <jepler> er, wait, I guess there are
[15:38:49] <jepler> * jepler can't count
[15:38:50] <SWPadnos> yep, 12 gaps
[15:39:05] <jepler> nevermind
[15:39:07] <jepler> * jepler crawls back in his hole
[15:40:36] <SWPadnos> the thing with those switchers is to be sure you have the right kind of inductor
[15:40:51] <SWPadnos> you need low resistance high current power inductors
[15:40:55] <SWPadnos> and the diode too
[16:41:53] <cradek> huh, firefox on dapper healed - now it opens a link on the first try
[16:42:12] <cradek> must have been that update on the 10th
[16:42:28] <BigJohnT> good, I was getting tired of that...
[16:42:41] <fragalot> I believe the term is "exellent"
[16:42:53] <maddash> did you guys get the mail from sebastian about coverity?
[16:43:00] <SWPadnos> yes
[16:43:00] <archivist> yes
[16:43:00] <fragalot> nop
[16:43:16] <archivist> fragalot, mailing list
[16:43:20] <maddash> coverity, as in static code analysis coverity.com?
[16:43:25] <archivist> yes
[16:43:28] <fragalot> * fragalot checks his emails
[16:43:41] <maddash> why??
[16:43:44] <SWPadnos> it was on the developer list
[16:44:01] <SWPadnos> to check the code for potential problems ...
[16:44:18] <SWPadnos> he's fixed a half dozen issues it turned up
[16:46:24] <fragalot> is cnczone.com down?
[16:46:43] <fragalot> nvm,.. www.cnczone.com doesn't work, cnczone.com does
[16:48:06] <Lerman_____> Lerman_____ is now known as Lerman
[16:48:19] <maddash> their site screams "eliteness" to me
[16:48:19] <maddash> is there a list of these issues somewhere?
[16:48:34] <alex_joni> only if you get a devel access
[16:48:48] <SWPadnos> unfortunately, they require logins to get the scan reports
[16:49:04] <alex_joni> not really unfortunately
[16:49:08] <SWPadnos> I think it's a security thing - reporting overflows is best done to the developers, not crackers
[16:49:15] <alex_joni> * alex_joni understands the decision to keep things private
[16:49:29] <SWPadnos> (and other exploitable things)
[16:49:52] <SWPadnos> the scan tool used to be the "Standford checker", which was used to find hundreds of bugs in the Linux kernel
[16:50:04] <maddash> can't sebastian post his copy onto the wiki or something?
[16:50:15] <alex_joni> why should he?
[16:51:01] <alex_joni> it's not really something for the wiki.. at least I don't see why it should be on the wiki
[16:51:09] <maddash> SWPadnos: security through obscurity?
[16:52:07] <archivist> if bugs fixed and further scan done and ok, only then....
[16:53:11] <archivist> ie not a real need as long as bugs get fixed (and they do in these parts)
[16:53:29] <cradek> secrecy about bugs is not the free software way - I'm afraid I'm on maddash's side of the fence here
[16:54:30] <archivist> I also can see in this type of code that openness helps others see problems and cures
[16:57:22] <cradek> but coverity gets to give away whatever they want to give away - they have a product to sell, after all.
[16:58:21] <alex_joni> there are other software packages that scan for bugs/problems
[16:59:13] <SWPadnos> I don't believe in security through obscurity either, but as a developer, I'd like to have a head start on fixing things
[16:59:51] <cradek> yes there's an honored tradition of reporting a security bug to the developers first.
[16:59:52] <SWPadnos> it would be nice if they'd give the option for individual projects to have their reports published
[17:00:32] <fragalot> security trough obscurity might work, but the real question is... is it worth it?
[17:00:51] <fragalot> (ofcourse, that obscurity thing only helps if they have the source code)
[17:01:05] <fragalot> but if they do, and you're closed source, you've got bigger problems :p
[17:01:23] <SWPadnos> uh
[17:01:25] <SWPadnos> right
[17:02:12] <fragalot> Downloaded 13.2 megabytes in 6 seconds. (1989.21 KB/s)
[17:02:13] <fragalot> whee
[17:17:07] <archivist> cradek, know of any characters with isolated separate glyphs inside a glyph?
[17:18:07] <cradek> do you mean separate paths? lots like 'o' have that
[17:18:30] <archivist> no I mean a dot inside an o
[17:18:35] <cradek> anything with a hole in it
[17:18:41] <cradek> oh I see
[17:19:11] <archivist> Im thinking chinese /whatever
[17:19:16] <cradek> some fonts surely have that - I think vera sans mono has a dotted zero
[17:20:00] <archivist> something I need to detect and deal with
[17:22:10] <archivist> I note kerning in your code seems to be from last point
[17:22:19] <jepler> U+00ae 174 REGISTERED SIGN “®” Symbol, other
[17:22:33] <jepler> this might be an example of the kind of character you're concerned about: ^^
[17:22:44] <archivist> ah good example jepler
[17:23:00] <jymm> %
[17:23:01] <cradek> http://timeguy.com/cradek-files/emc/vera-sans-mono-zero.png
[17:23:36] <cradek> I remembered that zero because of how it sucks (looks like an 8)
[17:23:43] <archivist> try a following M after that cradek
[17:23:55] <cradek> but © and ® are good examples too
[17:24:20] <archivist> or a T
[17:24:49] <cradek> 0M and 0T look fine at first glance - what am I supposed to see?
[17:26:03] <archivist> incorrect inter character spacing if kerning is wrong
[17:26:04] <cradek> load again: http://timeguy.com/cradek-files/emc/vera-sans-mono-zero.png
[17:26:46] <cradek> oh I didn't think the tracer does kerning at all - it just moves over by the advance for each character
[17:27:34] <archivist> mono spacing ?
[17:28:00] <cradek> no - each character has its own advance
[17:28:22] <archivist> ok
[17:28:27] <cradek> /* xoffset will get bumped up by the x size of the char just plotted */
[17:28:27] <cradek> return face->glyph->advance.x;
[17:29:06] <archivist> good /me can ignore kerning as well :)
[17:30:19] <archivist> although to correctly position engraving we need total width of a "sentence"
[17:31:11] <archivist> add a variable/constant to gcode
[17:31:32] <cradek> jmkasunich gave me a patch that adds some comments with the total sizes. I hope I didn't lose track of it.
[17:32:05] <jepler> U+263a 9786 WHITE SMILING FACE “☺” Symbol, other
[17:32:12] <cradek> date: 2008/04/27 02:55:53; author: chris; state: Exp; lines: +87 -11
[17:32:12] <cradek> changes by jmk
[17:32:26] <cradek> at least I have them - no idea if I made a release since then
[17:33:07] <jepler> so many fun characters in unicode
[17:33:28] <jepler> U+270c 9996 VICTORY HAND “✌” Symbol, other
[17:34:07] <jymm> ASCII RULES!!!
[17:34:51] <jymm> jepler: you looking at UTF 8 or 16 ?
[17:34:56] <archivist> as I preload chr array etc outputting stuff for sizes etc wont be a problem
[17:35:28] <jepler> jymm: I believe that what I pasted above is encoded in UTF-8 by my IRC client, if that's what you're asking
[17:36:16] <cradek> jepler> U+270c 9996 VICTORY HAND "270c" Symbol, other
[17:36:21] <cradek> I see ascii
[17:36:30] <jymm> jepler: It was =) I can see the correct symbol in WeeChat (kinda surprised too)
[17:36:31] <cradek> err, jepler> U+270c 9996 VICTORY HAND "\u270c" Symbol, other
[17:36:52] <cradek> but I can type and then see «®»
[17:37:00] <cradek> oh well
[17:37:31] <jymm> cradek: double-left arrow, (c), double right arrow ???
[17:37:52] <jymm> (small type, could be (R) )
[17:39:37] <jepler> http://emergent.unpy.net/files/sandbox/unicode-argument.png
[17:40:02] <jepler> here's how it looks in my terminal, font size increased to make the symbols more legible
[17:41:20] <cradek> but your WHITE SMILING FACE looks black to me
[17:41:33] <jepler> yeah that's a real puzzle innit
[17:42:45] <jymm> ah (R)
[17:44:12] <jepler> dangit, why's my package still in davenport, ia
[17:48:05] <archivist> sɹǝʇɔɐɹɐɥɔ ʎuunɟ ɹǝɥʇo puɐ
[17:48:35] <jepler> heh, yeah
[17:49:17] <jymm> http://imagebin.org/31603
[17:50:04] <jymm> archivist: how you do that =?
[17:50:44] <archivist> ɯɯʎɾ ɔıƃɐɯ sʇı
[17:50:51] <jymm> that one I can have fun with
[17:50:58] <jymm> archivist: magic my ass
[17:51:11] <archivist> http://www.sevenwires.com/play/UpsideDownLetters.html
[17:51:43] <jymm> ǝɯosǝʍɐ sı sıɥʇ
[17:51:48] <jymm> archivist: TY!!!
[17:52:20] <archivist> I must put some in the #mysql bot
[17:53:01] <alex_joni> jepler: did you do anything to irssi to make it work with utf?
[17:54:13] <LawrenceG> alright you guys... Its a real pain to turn this 19" CRT upside down to read whats going on
[17:54:41] <archivist> ǝɔuɐɹnɐן ɹood
[17:54:42] <fragalot> alex_joni: it Should JustWork in theory
[17:54:44] <alex_joni> pǝsɹǝʌǝɹ
[17:54:57] <fragalot> given that the terminal you use supports UTF8
[17:55:03] <alex_joni> doesn't here .. but I have a putty inbetween
[17:55:12] <alex_joni> and screen
[17:55:25] <alex_joni> I just set putty to use utf-8 encoding
[17:55:42] <alex_joni> «®»
[17:55:44] <fragalot> :) putty is the terminal.
[17:56:04] <jymm> alex_joni: I'm ssh'ed and using weechat
[17:56:38] <jymm> alex_joni: your 'reversed' came in fine
[17:57:16] <alex_joni> doesn't look fine here though
[17:59:46] <jepler> alex_joni: I may have changed some settings, I dunno
[17:59:53] <fragalot> doesn't here either because i was too lazy to enable utf support
[17:59:54] <fragalot> :D
[18:00:17] <jepler> recode_out_default_charset = "UTF-8";
[18:00:17] <jepler> recode = "yes";
[18:00:18] <alex_joni> eh, n/m.. utf isn't that important around #emc and irc
[18:00:43] <jymm> alex_joni: you have to reset putty I think - try opening a new session after making the change
[18:01:12] <archivist> I need utf in #mysql as character set problems are common
[18:01:22] <alex_joni> jepler: that is in the irssi config?
[18:01:38] <fragalot> archivist: I can't even use backticks lol
[18:02:04] <archivist> ' ` '
[18:02:09] <jepler> alex_joni: that's a bit of my ~/.irssi/config fil
[18:02:10] <jepler> e
[18:02:14] <archivist> ` ' `
[18:02:20] <alex_joni> ok, thanks ..
[18:02:21] <jepler> inside settings = { core = {
[18:02:23] <alex_joni> * alex_joni googles
[18:02:53] <fragalot> alex_joni: /set <setting>
[18:02:54] <jepler> I think you can /set recode yes
[18:02:57] <jepler> and so on
[18:03:00] <jepler> bbl, lunchtime here
[18:03:07] <fragalot> jepler: it's recode ON
[18:03:09] <fragalot> not yes
[18:03:21] <alex_joni> Irssi: Unknown setting recode
[18:03:23] <alex_joni> ;)
[18:03:31] <fragalot> eh
[18:03:40] <fragalot> 19:03 recode = ON
[18:04:11] <alex_joni> might be a slightly older irssi :D
[18:04:18] <alex_joni> Client: irssi 0.8.9 (20031210 2316)
[18:04:54] <fragalot> 19:05 CTCP VERSION reply from fragalot: irssi v0.8.12 - running on Linux x86_64
[18:05:19] <fragalot> Installed versions: 0.8.12(13:35:53 10/19/08)(ipv6 perl ssl -socks5)
[18:05:24] <alex_joni> yeah, it got introduced in 0.8.10 :)
[18:05:26] <fragalot> yeah, you're pretty outdated :p
[18:12:24] <alex_joni> they even removed sarge from the debian repo's
[18:18:37] <alex_joni> * alex_joni gives up on a newer irssi :)
[18:19:37] <jymm> alex_joni: sudo apt-get install weechat
[18:22:24] <fragalot> alex_joni: download irssi, ./configure && make && make install done.
[18:22:30] <fragalot> actually
[18:22:42] <fragalot> compile irssi, and point the current session to the binary with /update
[18:22:45] <fragalot> irssi can update in-place
[18:27:17] <archivist> cradek /me spots a buglet ./truetype-tracer -f /usr/share/fonts/truetype/ttf-bitstream-vera/VeraMono.ttf '®'
[19:42:33] <LawrenceG> archivist, was having a look at the problem with '®' in my version of ttt.... failure to handle multibyte characters properly... I have ignored the problem up until now... any suggestions on how they are encoded?
[19:43:52] <archivist> I thought (didnt look) that it may have been a missing character in that font
[19:44:42] <LawrenceG> I actually got it to make the '®' but it had a bogus other char in front of it
[19:45:22] <archivist> and this sort of job needs a hard error perhaps for missing characters in a set
[19:45:28] <LawrenceG> '®' seems to encode as the 2 byte sequence 0xc2 0xae
[19:45:44] <archivist> ah ok
[19:46:19] <cradek> LawrenceG: ttt -u should deal with those fine
[19:46:27] <archivist> I can see other languages loving it id we get multibyte working
[19:46:36] <archivist> if if
[19:46:51] <archivist> tyops in qty :)
[19:47:05] <cradek> "TTT 3.0 can use Unicode fonts such as Kanji and accepts Unicode strings as input."
[19:47:17] <LawrenceG> cradek, ah... I am running an older version of ttt that I put rotation into
[19:47:22] <cradek> ah
[19:47:50] <LawrenceG> archivist, what version of ttt are you running?
[19:49:13] <archivist> recent/latest needs a -v option :)
[19:51:13] <archivist> * archivist adds -v here
[19:59:34] <fragalot> who here feels like making a pulsejet?
[20:00:27] <archivist> * archivist used to make rockets when a naughty boy
[20:01:21] <fragalot> i've only made a jar "ramjet"
[20:01:44] <fragalot> which had about *drums* no thrust, but looked awesome.
[20:02:28] <fragalot> archivist: anyhow, a pulse jet should be fun to make
[20:02:53] <fragalot> no moving parts, reasonable amount of thrust,.. pretty intimidating noise.. :D
[20:03:08] <fragalot> and the glow it gets when its running is also pretty sweet
[20:11:40] <archivist> my ttt has a version :) and a -v option
[20:13:27] <anonimasu> :)
[20:19:19] <LawrenceG> cradek, I see you added "setlocale(LC_CTYPE, "");" to support unicode option.... what was the effect of this?
[20:20:48] <LawrenceG> nm, LC_CTYPE
[20:20:48] <LawrenceG> Affects behavior of regular expressions, character classification, and conversion functions (see ctype(3C), wctype(3C), wconv(3C), conv(3C), and regexp(5)). The LC_CTYPE category also affects the behavior of all routines that process multibyte characters (see multibyte(3C)).
[20:23:43] <archivist> I wonder if ALL fonts start on the outer polly of a glyph
[20:24:08] <jymm> why you playing with TT instead of ClearType?
[20:24:17] <archivist> Im assuming that at the moment
[20:24:31] <SWPadnos> because the program is "truetype tracer", not "cleartype tracer"
[20:24:45] <jymm> SWPadnos: and WHY is that?
[20:24:54] <archivist> it just is
[20:24:57] <SWPadnos> and cleartype has no bearing on machining fonts
[20:25:11] <SWPadnos> cleartype is a rendering scheme, not a glyph encoding scheme
[20:26:19] <jymm> SWPadnos: sew?
[20:26:31] <SWPadnos> not recently
[20:27:00] <SWPadnos> though I did remove some embroidery from a shirt a week or two ago
[20:31:36] <jymm> SWPadnos: sew?
[20:31:47] <SWPadnos> no, I just told you :)
[20:32:00] <jymm> SWPadnos: sow?
[20:32:11] <SWPadnos> no, but I did have some ham with my eggs today
[20:32:29] <jymm> green?
[20:32:39] <SWPadnos> no, but it was getting there
[20:32:43] <jymm> lol
[20:32:46] <SWPadnos> good thing I ate it today
[20:33:16] <jymm> SWPadnos: Hey, that 32port serial concentrator.... what can I use to control things cheaply enough with it?
[20:33:35] <SWPadnos> uh - what?
[20:33:52] <SWPadnos> what "things" do you want to control?
[20:34:33] <jymm> SWPadnos: well, No/NC, but also simple stepper control would be nice.... RC servo, not NEMA type servo
[20:34:52] <jymm> or both =)
[20:35:05] <SWPadnos> ok, so relays and maybe some motors (like turning lights on and off or moving window shades)?
[20:35:11] <jymm> yeah
[20:35:46] <SWPadnos> no idea :)
[20:35:55] <jymm> bastard!
[20:36:03] <SWPadnos> I bet there are some X10-like things available though
[20:36:17] <SWPadnos> it also depends on the amount of power you might want to switch with the relays
[20:36:28] <jymm> Yeah, but they are not serial, only the control is
[20:36:43] <SWPadnos> http://www.pencomdesign.com/8ch_relay.htm
[20:36:47] <jymm> (I have one of the X10 serial controllers)
[20:36:49] <SWPadnos> http://www.electronickits.com/kit/complete/elec/ck1610.htm
[20:37:13] <alex_joni> there is a thing called fs20
[20:37:20] <alex_joni> for wireless control of such things
[20:37:26] <alex_joni> relays, blinds, etc
[20:37:33] <SWPadnos> it's pretty trivial to make a microcontrolelr-based board to control some relays too
[20:37:37] <SWPadnos> microcontroller
[20:38:05] <jymm> I said cheaply
[20:38:28] <jymm> I have a $1400 serial concentrator, why would I want to buy a $100+ serial relay baord =)
[20:38:49] <SWPadnos> that would be cheap. you can probably make a serial uC board with relays for $50 or less
[20:38:58] <SWPadnos> especially if you source from places like Halted
[20:39:17] <SWPadnos> or, since you have 32 ports: http://www.windmeadow.com/node/4
[20:39:57] <SWPadnos> and, linked from that page: http://stats.the-abbeys.com/blog/index.php?entry=entry060116-095620
[20:41:33] <jymm> much better =)
[20:42:29] <jymm> each port connect via RJ45, easy enough to wire up at least =)
[20:43:01] <SWPadnos> RJ45 won't necessarily work, the voltages may not be as high as RS232
[20:43:29] <jymm> No no Ithe serial concentrator uses RJ45 instead of DB9 connectors
[20:43:42] <jymm> SWPadnos: Think cisco console
[20:43:50] <SWPadnos> oh duh - I read RS485 there, my bad
[20:44:00] <jymm> it's all good =)
[20:44:01] <SWPadnos> and I even typed it :)
[20:44:07] <jymm> lol
[20:45:12] <jymm> SWPadnos: ןןɐ sı ʎɐp ɹıɐɥ pɐq ɐ ƃuıʌɐɥ ʇsnɾ ǝɹɐ noʎ
[20:45:25] <SWPadnos> sorry, this client doesn't do unicode
[20:45:37] <jymm> it's all good
[20:45:52] <SWPadnos> I assume that says soemthing witty upside down :)
[20:45:57] <archivist> hehe
[20:46:07] <SWPadnos> so lol
[20:46:24] <jymm> SWPandos: You are just having a bad hair day is all
[20:46:29] <SWPadnos> yay, time to put the new motherboard in and maybe actually get my workstation up and running
[20:46:31] <jymm> is what it said
[20:46:36] <SWPadnos> at least I have hair
[20:46:44] <SWPadnos> oh
[20:46:46] <SWPadnos> ;)
[20:46:55] <jymm> you just havne't played with enough explosives is all
[20:47:11] <SWPadnos> no, I guess not
[20:48:32] <jymm> whats the distance for RS232.... 300ft?
[20:49:13] <SWPadnos> 25 is more like it
[20:49:24] <jymm> is 35ft pushing it?
[20:49:27] <jymm> err 25ft
[20:49:36] <alex_joni> depends on the baud rate
[20:49:45] <jymm> 9600
[20:49:49] <cradek> http://www.google.com/search?q=rs232+max+distance
[20:49:50] <SWPadnos> it depends on what you're doing, and the type of wire you use
[20:49:53] <alex_joni> you can go pretty far using twisted pair and low baudrate
[20:50:13] <jymm> I wonder how much cross-talk with 32 ports
[20:50:31] <fenn> * fenn just hopes there are no explosives involved
[20:50:33] <SWPadnos> isn't that box ethernet on one side and 32 serial ports on the other?
[20:50:41] <jymm> SWPadnos: yep
[20:51:10] <jymm> fenn: you're no fun
[20:51:16] <alex_joni> 250ft for 9600
[20:51:25] <archivist> * archivist adds explosives if they are missing
[20:51:32] <jymm> oh cool, that's far more than enough distance
[20:51:33] <alex_joni> * alex_joni adds more explosives
[20:51:41] <SWPadnos> ok, so if you're using the ports as static switches then you need to look at wire gauge/resistance and drive levels, not baud rates
[20:51:53] <alex_joni> jymm: that's max distance.. not surely guaranteed.. gotta try it
[20:52:00] <alex_joni> use some cat-5 or so cable
[20:52:14] <jymm> alex_joni: Yeah, np. Just didn't want some odd limitation of 12ft or something
[20:52:30] <jymm> at east not with 32 ports
[20:52:44] <alex_joni> it depends on the chips inside
[20:52:55] <jymm> alex_joni: just playing here is all
[20:53:03] <alex_joni> some are crappy like the max232 which generate the +/- voltages from 5V using charging caps
[20:53:06] <jymm> It's a Digi box
[20:53:15] <alex_joni> others use real +15/-15V PSU's
[20:53:22] <jymm> they usually make quality stuff
[20:53:54] <jymm> I can check tonight.
[20:54:26] <fenn> why dont you just use ethernet?
[20:54:41] <jymm> fenn: for controlling relays?
[20:54:54] <fenn> well, rs232 isn't gonna drive relays either
[20:55:08] <jymm> fenn: see the link SWPadnos posted
[20:55:12] <alex_joni> http://www.hw-group.com/products/ip_relay/index_en.html
[20:56:00] <jymm> Guys, I already have the $1400 32 serial port concentrator... I'm just playing around with it.
[20:56:14] <fenn> what is a "serial port concentrator" for anyway?
[20:56:51] <archivist> concrtration, not a jymm speciality
[20:57:13] <jymm> fenn: http://www.digi.com/products/serialservers/etherlitespecs.jsp
[20:57:22] <fenn> i gather it's something like an ethernet switch, but doesnt $1400 seem a bit steep for that?
[20:57:29] <archivist> not used since the dark ages
[20:57:53] <jymm> fenn: When I stall the drivers, each of the ports show up under /dev/tty
[20:58:16] <jymm> all across ethernet =)
[20:58:19] <skunkworks_> 'stall
[20:58:31] <skunkworks_> redneck computer talk?
[20:58:53] <alex_joni> wheee.. 80C186 embedded microcontroller
[20:58:56] <fenn> its what happens when your etherspeed is too low and you start generating ether vortices
[20:59:19] <jymm> alex_joni: You like that huh =)
[20:59:59] <jymm> eh, it was a free toy... just playing
[21:01:40] <iso> to rescan HBA's without rebooting on solaris 9 I ran cfgadm -c configure c2 c3 c4 c5
[21:01:58] <fenn> iso: wrong emc
[21:02:08] <iso> sigh...
[21:02:49] <archivist> hehe
[21:02:54] <fenn> at least he didnt ask to ask
[21:03:24] <archivist> oo I think I can invalidate a US patent http://www.freepatentsonline.com/6901672.html
[21:03:33] <jepler> archivist: too bad they don't give you a cookie when you do that
[21:03:37] <alex_joni> I suspect wrong network
[21:03:48] <alex_joni> cookie? someone said cookie?
[21:04:26] <jymm> alex_joni: you must be hungry =)
[21:04:32] <archivist> I have an english caliper setting guage made by Chesterman
[21:05:04] <alex_joni> jymm: only 1/4
[21:05:22] <alex_joni> one of my grandma's was born in budapest
[21:08:53] <archivist> which before they merged with Rabone circa 1963
[21:09:41] <jymm> alex_joni: you forget the <rim_shot>
[21:20:17] <alex_joni> see you all
[21:21:56] <jymm> G'Night Alex
[21:39:01] <SWPadnos> holy shit
[21:39:36] <SWPadnos> the thermal paste on my CPU cooler is so sticky, the CPU pulled out of the socket rather than the heatsink coming off the CPU
[21:39:58] <SWPadnos> this is without me releasing the ZIF latch
[21:40:13] <SWPadnos> luckily it doesn't look like any pins got bent or torn off
[21:42:35] <jymm> ouch
[21:43:01] <jymm> SWPadnos: If you're heat sink grease comes in two tubes, THAT AINT IT! =)
[21:45:50] <SWPadnos> no, this is the HS/fan that came with the CPU, a nice AMD one with heat pipes and stuff
[21:46:10] <SWPadnos> the thermal stuff is pre-applied for you, and it's darned sticky
[21:46:26] <SWPadnos> I got the CPU off by heating the sink with a hair dryer :)
[21:46:30] <jymm> ah, gotcha
[21:47:46] <SWPadnos> the weird thing is that it's cleaning up nicely with Kleenex, so it's not actually all that sticky
[21:48:43] <SWPadnos> I just hope the CPU survived being ripped out of its socket, heated through its heatsink, then dropped on the floor when it finally popped off :)
[21:49:15] <SWPadnos> oh, and the cleaning with non-static-control kleenex
[21:49:44] <jymm> SWPadnos: Jsut rub it againest the shug carpet to clean it
[21:49:48] <jymm> shag
[21:49:56] <SWPadnos> yep, that's the next polishing step
[21:50:10] <SWPadnos> plus there's some nice sand in there to make sure all the numbers wear off
[21:50:28] <SWPadnos> hmmm. now to find that tube of Arctic Silver
[21:50:47] <jymm> no, no, silicon RTV
[21:51:36] <SWPadnos> hmmm, maybe this spray can of expanding foam would work ;)
[21:55:51] <jymm> Nice choice... High density I assume?
[22:01:02] <fenn> archivist: maybe these guys will give you a cookie (note: i havent read the page) http://yro.slashdot.org/article.pl?sid=08/11/18/1725231
[22:02:31] <archivist> fenn hehe, but I think a caliper gauge does not have the same value
[22:03:21] <archivist> must be nuts patenting a caliper setting gauge as late as they did
[22:06:14] <archivist> I see the company offering the money wants 95% with a cap at 50k
[22:06:39] <archivist> worth working against them!
[22:06:51] <archivist> as well
[22:52:10] <chr0n1c> hola banditos!
[23:01:28] <Gamma-x> ok so is it bad for the tachs on a servo to connect directly to the amps?
[23:01:43] <Gamma-x> oh and the amps are older westamps... ive read a lot that they are not that good.....
[23:06:06] <SWPadnos> if you have analog velocity-mode amps, the best way to make them work is to connect the tach directly to the amp
[23:06:30] <SWPadnos> in fact, barring massicely complex DSP-type systems, that's the only way to make them work :)
[23:06:33] <SWPadnos> massivelty
[23:06:35] <SWPadnos> -t
[23:08:43] <Gamma-x> SWPadnos, i think i know why my table is running away
[23:08:58] <Gamma-x> i think i need to swap the tach signal,
[23:09:49] <Gamma-x> SWPadnos, oh and ive been trying to look for encoders for my sem mt series servos any ideas? i think they have a 5mm shaft,
[23:11:05] <SWPadnos> you need to know the mounting type also. there are a few standard sized for servos
[23:11:22] <SWPadnos> some people like USDigital and Renco
[23:11:31] <archivist> http://usdigital.com/products/encoders/incremental/rotary/shaft/s1/
[23:11:35] <SWPadnos> if you have loads of cash, there are also companies like BEI
[23:11:55] <Gamma-x> no i want somewhere in the middle, great precision but low price..
[23:11:59] <Gamma-x> or as low as possible i should say
[23:14:05] <Gamma-x> i dont get how thats gunna connect to my servos... i took off the plate on there is only a little tiny nug to stick sumtin onto
[23:14:30] <archivist> make buy an adapter
[23:15:57] <Gamma-x> hmmm
[23:15:58] <Gamma-x> wierd
[23:16:04] <Gamma-x> thats what im affraid of haha
[23:16:58] <SWPadnos> if there's a depressed ring and a couple of little clip-like things around the circumference, that's a standatd servo encoder mounting scheme
[23:17:04] <SWPadnos> standard
[23:17:16] <SWPadnos> you just need to measure the diameter, since there are several standard sizes
[23:19:35] <SWPadnos> wait, does the encoder have a 5mm shaft? (ie, is there a coupler connecting the encoder shaft to a back motor shaft?)
[23:31:47] <Gamma-x> uhhh
[23:31:50] <Gamma-x> que?
[23:32:46] <Gamma-x> i see a lot of people make back plates for the thing.
[23:33:32] <Gamma-x> anyone got some scrap 1/4 inch or 1/2 inch alluminum sheets? wanna make me an adapter? lol realy easy i promise!
[23:34:02] <archivist> you want to play machining, get some practice in!
[23:34:33] <Gamma-x> haha i cant cause the thing is in ny and i dont have the spindlle running cause i dont wanna configure my vfd yet lol
[23:34:37] <BigJohnT> Gamma-x: I have some scrap 6061 come on down and get a piece...
[23:34:56] <Gamma-x> hey if someone make them ill make a trade?
[23:34:59] <Gamma-x> lol
[23:35:12] <Gamma-x> i would make them but i cant haha
[23:35:23] <BigJohnT> you got a file?
[23:36:31] <Gamma-x> hmmm
[23:36:34] <archivist> perhaps Gamma-x time you went to classes, and do your parts there
[23:36:47] <Gamma-x> ehhh
[23:36:55] <Gamma-x> i guess ill jsut wait it out cause no one wants to helP! lol
[23:37:09] <dmess> i wanna configure mine but just cant seem to get the vdf to turn the spindle on
[23:37:17] <archivist> * archivist is tooooo far away
[23:37:27] <Gamma-x> dmess, what kinda vfd lol
[23:37:30] <BigJohnT> we are trying to help but ...
[23:38:08] <archivist> we cant work files, drills and lathes via irc
[23:38:17] <Gamma-x> lol
[23:38:31] <archivist> your job
[23:38:54] <chr0n1c> you could work cncs with remote-x?
[23:39:22] <Gamma-x> true
[23:39:26] <archivist> we would have to trust Gamma-x to tighten the chuck etc
[23:39:30] <Gamma-x> i could dnc over the itnernet! lol
[23:39:45] <Gamma-x> 9600 baud?
[23:40:03] <chr0n1c> wow, my first modem was 9600
[23:40:15] <BigJohnT> mine was 1200 hayes
[23:40:24] <archivist> fast stuff!
[23:40:55] <BigJohnT> Gamma-x: you need some of these http://en.wikipedia.org/wiki/Caliper and a scale and a file then make some parts
[23:41:09] <Gamma-x> BigJohnT, already have those
[23:41:16] <archivist> manual lathe is useful
[23:41:20] <Gamma-x> i actualy have everything haha
[23:41:39] <BigJohnT> then make some parts
[23:41:43] <archivist> hmm pebkac
[23:41:44] <Gamma-x> i bought a vice, parralels, bolt and clamp set, all indicators.
[23:41:49] <Gamma-x> micromers etc lol
[23:42:06] <Gamma-x> well how the hell do u make a circile on a manual mill
[23:42:09] <chr0n1c> you don't need anything but a vise and a file
[23:42:21] <BigJohnT> did you buy a file?
[23:42:44] <Gamma-x> i have one but thats a lot of metal haha ur bringin it back to the stone age lol
[23:42:44] <BigJohnT> boring head Gamma-x
[23:42:44] <chr0n1c> doh!
[23:42:55] <Gamma-x> BigJohnT, u go huntin this year yet?
[23:43:00] <chr0n1c> the stone age built some amazing things ;)
[23:43:21] <Gamma-x> i wonder if BigJohnT learned about the wheel yet
[23:43:27] <BigJohnT> if you don't understand the stone age you will be lost in the space age
[23:43:37] <Gamma-x> ohhhhh snap!
[23:43:38] <Gamma-x> lol
[23:43:49] <BigJohnT> no, I might wait till black powder season
[23:43:56] <Gamma-x> BigJohnT, if i visit my family in leavenworth u should come pick me up and bring me huntin!
[23:44:08] <BigJohnT> * BigJohnT can run a cnc mill :)
[23:44:21] <Gamma-x> ouch
[23:44:30] <BigJohnT> I started with a file and vise
[23:44:44] <BigJohnT> and a lump of charcoal
[23:44:49] <Gamma-x> * Gamma-x has removed asbestos for a living!
[23:45:15] <Gamma-x> that has nutin to do with anything but i thought id add it in lol
[23:45:31] <BigJohnT> know what the charcoal is for?
[23:45:46] <Gamma-x> put it in ur ass?
[23:45:49] <Gamma-x> ohhh haha im jsut kidding
[23:45:52] <chr0n1c> i think i milled some asbestos brake pads from the 60's the other day to make them fit on an old tow truck?
[23:45:53] <BigJohnT> I'm not
[23:46:13] <Gamma-x> to smelt ur aluminum?
[23:46:29] <BigJohnT> you put the lump of charcoal in with the parts in a stainless steel sheet bag then put them in the heat treat oven
[23:46:40] <BigJohnT> the charcoal burns up the O2
[23:46:45] <Gamma-x> carbonizing?
[23:46:45] <BigJohnT> in the bag
[23:46:49] <BigJohnT> no
[23:47:53] <BigJohnT> a long time ago you used ground up dog bones for surface hardening
[23:48:13] <chr0n1c> i hope the dogs dies of natural causes
[23:50:32] <BigJohnT> the point is Gamma-x you need to learn the basics first...
[23:51:21] <BigJohnT> look online for machinest test or similar and try and make the part with just hand tools
[23:51:39] <BigJohnT> * BigJohnT wanders off