Back
[00:02:28] <ziph> The family member that makes the soap on that site does quite well at one despite the same soap being available at the local pharmacy and supermarket. :)
[00:03:38] <inflex> that's good - business is tough in a lot of places
[00:05:04] <ziph> By well I mean better than nothing. ;)
[00:05:21] <ziph> The normal sales have dropped off, which prompted trying the markets.
[00:05:34] <inflex> markets are a good place
[00:05:38] <inflex> (for that sort of thing)
[00:05:46] <Casper> hmmm
[00:05:54] <Casper> pspice really isn't good for smps
[00:05:57] <Casper> not at all...
[00:06:07] <ziph> Husbands get dragged along to them too though. :)
[00:06:48] <eadthem> use LTspice if your doing SMPS
[00:06:51] <eadthem> off to bed cya
[00:07:14] <Casper> I made roman black's smps in pspice since I have it installed....
[00:07:32] <Casper> ... well, it's not oscillating, it's a linear regulator based on pspice!
[00:08:58] <inflex> hahah
[00:10:13] <Casper> based on it, the start up is 100% linear
[00:10:36] <Casper> hmmm maybe it's just too fast...
[00:10:41] <inflex> do multivibrators work in spice?
[00:10:46] <inflex> (the dual transistor-crossover types...
[00:10:52] <inflex> since I've always found you have to make one cap or transistor slightly different to the otehrs to make it oscillate
[00:11:34] <Casper> dunnot, but this seems to be a step issue, where the oscillation is too fast, causing pspice to aproximate it
[00:12:05] <inflex> ah yep, transient analysis is quirky at best.... QUCS fails epically with it
[00:12:33] <inflex> QUCS is like "OMG, the state rapidly cha.... *BLEAH I FAILZ*"
[00:12:39] <Casper> trying 0.01us
[00:13:41] <inflex> mmm... given that they oscillate at ~30kHz, should you no try rather more like 10us?
[00:13:51] <Casper> nope
[00:13:56] <inflex> o
[00:13:58] <inflex> ok
[00:13:59] <Casper> 1us = fail
[00:14:10] <Casper> 0.1 seemed to be ok, but seems like not
[00:14:39] <Casper> 0.01 still show linear at start up with 0.1 ohms for RS
[00:14:43] <Casper> for 100 ohms load
[00:23:01] <Casper> the more I look into the circuit, the more I think pspice might actually be right
[00:25:01] <Casper> inflex: I think pspice is actually simulating it correctly
[00:27:32] <Casper> inflex: can you tell me if my analysis of what happend is right? CLoad is lower, so C2 turn on, which partially turn on, raising the voltage at point A, which raise the base of Q2 via C2, which turn on Q1 more, turn Q2 more... but then the base lower of Q2, making Q1 possibly only partially turn on and fall in a linear regulation....
[00:27:58] <Casper> so unless the load actually vary, there is a risk that the circuit mess up and go linear
[00:32:41] <rue_house> abcminiuser, wow, your audio out is pretty impressive, I cant find a single cut corner
[00:32:53] <rue_house> I thought it was only 8 bit?
[00:33:00] <rue_house> .BitResolution = 16,
[00:33:09] <rue_house> I wonder if 12 is valid
[00:43:22] <inflex> Casper: should be a way to make the circuit self-start more reliably
[00:44:36] <rue_house> use a TL494 :)
[00:45:15] <rue_house> hahah he gets them as 16 and squishes them down to 8, coool!
[00:48:13] <inflex> ?
[00:48:37] <Casper> inflex: yeah, but what I found is that if the current excede what the transistor can give then it will mandatory go linear and self destruct
[00:49:07] <Casper> at power up, it does get overloaded, but once things get stabilised (1-2ms later) then it's switching
[00:49:29] <inflex> Casper: such are the perils of 'simpler' circuits
[00:49:41] <Casper> yeah
[00:49:42] <inflex> guess that's why a lot of us just end up using controller chips
[00:49:52] * inflex is using a LT1935 in his latest design
[00:49:56] <inflex> bloody expensive though :(
[00:50:14] <ziph> That's LT's new slogan.
[00:50:35] <inflex> heh, I'll grant you, their gear is often fairly expensive - but it is good
[00:50:44] <Casper> the problem here is that I'ld have to make a digikey order just for those, but I guess I will, because anyway 2 out of 3 smps that I'ld need I can't do it with black one
[00:51:05] <ziph> Casper: Did you post a schematic somewhere?
[00:51:22] <inflex> Casper: if you were after the LT1935, there's a cheaper TS1935 "clone"
[00:51:43] <inflex> but I'm dubious about the clone because it looks like it can't do Vout < (Vin +3.0V)
[00:52:07] <Casper> ziph:
http://www.romanblack.com/smps/a04.htm
[00:53:06] <Casper> inflex: do you know if there is a big efficiency penality for buck boost instead of buck or boost?
[00:53:35] <inflex> Casper: not really a severe one afaik, just the extra switch loss I think
[00:53:47] <inflex> (I wish there were more SM controllers that let you use your own external MOSFET
[00:54:28] <inflex> eg, the LT1935 would make a lovely controller but for the fact that it has its own internal MOSFET, I suppose I could jury rig my own in some how
[00:55:12] <Casper> yeah I was annoyed by that too... when I last checked... those that allowed the use of external fet had way too many external parts and missing formula... those who had good formula and good explanation on how to build... had internal fet with way too low current...
[00:55:21] <inflex> Casper: precisely
[00:55:44] <Casper> and the controller that do current limiting are even harder to find
[00:57:30] * Valen couldn't find an IC to do a good enough job for his bigass buck converter ;->
[00:57:49] <Valen> 8A and easily variable voltage
[00:57:54] <inflex> Valen: ?
[00:57:55] <Valen> ~4V
[00:58:09] <Valen> doing it with a fet driver and some fets now
[00:58:36] <Valen> synchronous buck at 250khz
[00:58:40] <Casper> sometime I wonder if using an AVR wouln't be easier..
[00:58:47] <ziph> Umm, I'm not even sure how you'd linearise that for stability analysis.
[00:58:48] <Valen> thats what i'm doing lol
[00:58:49] <inflex> Casper: ja, that's probably true, use the comparator
[00:59:00] <Casper> but I don't think the avr would respond fast enought to load change
[00:59:01] <inflex> hell, a Tiny10 might even do it
[00:59:02] <Valen> i want current controlled output for a LED
[00:59:13] <Valen> not at 250khz inflex
[00:59:22] <inflex> oh sure, not at 250kHz
[00:59:32] <Valen> i need to minimise inductor size too
[00:59:34] <inflex> T25 migh though with its 64MHz PWM
[00:59:47] <Valen> that may be a little too fast ;-P
[01:00:07] <Valen> "when ever i turn my torch on all the radios within 4km make static"
[01:00:11] <inflex> well, your max 8-bit PWM speed would still be 250kHz
[01:00:37] <inflex> and the AVR itself runs at 8MHz still
[01:01:01] <Valen> so whats the 64mhz part?
[01:01:19] <inflex> x8 PLL that the T25/45/85 have
[01:01:58] <Valen> i think thats what i'm using in the atmega32u4 i'm using
[01:02:15] <Valen> for the 250khz pwm
[01:02:20] <Casper> how does synchronious smps work? just time based?
[01:02:27] <inflex> the Tx5's have been build with PWM switching in mind, there's even a dead-band parameter to
[01:02:28] <inflex> too
[01:02:29] <Casper> as a percentage of Ton?
[01:02:39] <Valen> u4 has that too
[01:02:46] <Casper> or have a voltage/current feedback?
[01:03:50] <inflex> If you drop to only 64 levels you could get a nice 1MHz SMPS
[01:03:57] <inflex> but I'm not sure how well that'll work
[01:04:19] <Valen> i was thinking of doing that and then dithering it around the value
[01:06:05] <Valen> question is how to measure the current in my loop
[01:06:33] <inflex> just not sure if you can effectively do it - because you have to kludge the TCNT0 value 192 at each start - and that'll suck up a lot of cycles relative to the 64MHz speed
[01:06:33] <Valen> i was thinking of just measuring the voltage across the LED and using a lookup table for voltage vs current vs temp
[01:07:00] <Valen> i don't think it'd even really be possible lol
[01:07:15] <Valen> at 1mhz it'd need to do it every 16 instructions
[01:07:39] <inflex> mmm... maybe you could do it with the output compare
[01:07:40] <Valen> unless you can set the top value of the counter
[01:07:56] <Casper> Valen: better use a feedback resistor
[01:08:09] <Valen> like the normal timer i think you can set either the top or bottom value?
[01:08:58] <inflex> not me speciality alas
[01:15:27] <Casper> does anyone have a good documentation on smps? I understand the basis but can't design one... and the formula seems to be conflicting...
[01:15:49] <rue_bed> yes
[01:15:58] <rue_bed> dont know if its in pdf tho
[01:16:32] <rue_bed> the BEST stuff was done by motorola, but they were sold and bought 3 times and I dont know if the program you could get the docs thru even exists anyhmore
[01:16:44] <rue_bed> but all the documents only tell half the story
[01:16:56] <rue_bed> and everyone phrases their formulas different
[01:17:24] <rue_bed> you have to be carefull, alot of the formulas are based on assumption of 50% duty cycles
[01:17:46] <rue_bed> Casper, the smps isn't the hard part, making the iductor is
[01:18:07] <rue_bed> but I'm in bed, if you ask leter I'll see what pdf's I have
[01:18:10] <Valen> eh buy one
[01:18:35] <rue_bed> if you really want, I'll digitize the special book I made
[01:18:59] <rue_bed> it has the REAL formulas, with ALL the drops
[01:19:18] <Casper> I'll think about it, now it's bed time
[01:19:19] <Casper> niter
[01:19:30] <rue_bed> even if you buy the inductors, their still trickey
[01:19:37] <rue_bed> gnight
[01:19:38] <Casper> if you have something to tell, drop me a line in pm
[01:19:50] <rue_bed> I dont know what I have in pdf
[01:20:14] <rue_bed> I had to write my own book to get the picture straight and clear
[01:20:17] <Casper> and yeah, I'ld buy... on next digikey order IF I can find some good formula that don't differ more than 50% of their example...
[01:21:00] <Casper> because... most of the damn datasheet that give the formula AND an example... the example do not follow the formula at all
[01:21:21] <Casper> and of course, they don't tell how to round.. up, down, or just "whatever you want"
[01:29:41] <inflex> sometimes it's just easier to cut-and-try
[02:10:36] <Triplell> can you connect more then 1 device to the tx/rx pins of the atmega328?
[02:13:14] <Valen> you could
[02:13:20] <Valen> in theory
[02:13:31] <Landon> yay timeslots!
[02:13:45] <Triplell> I want to connect both usb and xbee to my atmega328, but both require the rx/tx pins
[02:14:26] <Valen> not a good plan
[02:14:41] <Valen> you are probably better off just bit bashing whichever one you use less
[02:15:09] <Triplell> or find a chip that has two uarts?
[02:15:49] <Valen> or that
[02:15:52] <Valen> or use spi
[02:16:04] <Valen> or use a chip with hardware usb and use lufa
[02:16:06] <Triplell> xbee with spi?
[02:16:16] <Valen> dunno
[02:16:27] <Triplell> haha, thanks. I will look into it
[02:18:00] <Valen> otherwise you have issues if both try to tx at the same time oe other such
[02:18:04] <Valen> or
[02:18:29] <Triplell> What if I used a double dpdt switch?
[02:18:35] <Triplell> no dpdt*
[02:18:56] <Landon> you have that luxury? :)
[02:18:58] <Valen> if you don't need to use both of them at the same time you could get away with it
[02:19:25] * Landon was thinking something more along hte lines of a multiplexer if you aren't limited by pins
[02:19:57] <Triplell> and then use one of the ports as a select?
[02:20:32] <Landon> yep
[02:21:00] <Triplell> Only problem is xbee and usb are going to be my only i/o...so it's kind of hard to enable/disable your own device
[02:21:19] <Valen> something like that
[02:21:39] <Triplell> meh, I'll just put in a dpdt
[02:21:45] <Triplell> I don't need to use them both at the same time
[02:22:26] <Landon> yeah, my idea would be a bit of a pain depending on application
[02:22:53] <Triplell> basically it's a jpeg trigger that is trigger remotely either by xbee or usb
[02:23:40] <Landon> meh, just got an email inviting me to take an "exit survey" for college... can win a $25 gift card
[02:23:45] * Landon wouldn't mind winning some tuition back
[02:24:03] <Landon> they sure increased it enough while I was here
[02:24:49] <Landon> Triplell: sounds like fun
[02:29:39] <inflex> Landon: sounds like a scam
[02:30:30] <Landon> heh, well it links straight to the schools survey portal
[02:31:15] * Landon would expect a scam to be a little more tempting too :(
[02:36:04] <Valen> see if they have a permit for the "game of chance" and report them to the feds!
[02:36:14] <Landon> ha :D
[02:36:41] <Triplell> could there be issues with the tx/rx signal going through the switch?
[02:36:51] <Valen> probably not
[02:37:02] <Valen> not at usual serial speeds
[05:32:08] <karlp> Triplell: if you want an "xbee" with spi, you can use the mrf24j40ma modules from microchip.
[05:32:12] <karlp> about half the price too
[05:32:23] <karlp> if you're just using them as data transfer.
[05:33:36] * RikusW thought cracking AES might not be that hard, that is until I started looking at mixcolumns....
[06:18:30] <scuzzy> hey RikusW
[06:18:31] <scuzzy> how you doing?
[06:21:06] <abcminiuser> Hrmph
[06:21:15] <abcminiuser> My robot has a main 5V switchmode power supply
[06:21:21] <abcminiuser> And a secondary 3.3V LDO chained off that
[06:21:29] <RikusW> hi scuzzy
[06:21:32] <abcminiuser> I see 80mV ripple on the 5V, and 120mV on the LDO
[06:21:42] <RikusW> I'm doing well, you ?
[06:21:46] <abcminiuser> How is that possible? Shouldn't the LDO ripple be the same or better than the 5V on which it's based?
[06:22:36] <RikusW> abcminiuser: maybe the ldo somehow amplifies it ?
[06:23:14] <scuzzy> RikusW: yeah, not that bad, just suffering from brain rot
[06:23:20] <scuzzy> work is... not very trimulating
[06:23:23] <scuzzy> stimulating
[06:23:33] <RikusW> scuzzy: Have you ever programmed encryption stuff ?
[06:23:46] <scuzzy> RikusW: many years ago, wrote some RSA stuff
[06:23:47] * RikusW is looking at how AES works
[06:23:47] <scuzzy> why?
[06:23:54] <scuzzy> AES is amazingly sweet
[06:23:56] <scuzzy> and not too complex
[06:24:04] <scuzzy> however, practically impossible to crack
[06:24:14] <ziph> abcminiuser: What frequency?
[06:24:20] <scuzzy> I've seen a few AVR implementations of it
[06:24:24] <RikusW> would've been easy to simplify, was it not for mixcolumns...
[06:24:33] <RikusW> like apnote AVR231
[06:24:52] <scuzzy> RikusW: I only have a basic idea about AES tho
[06:24:55] <abcminiuser> ziph, of the switchmode?
[06:25:06] <scuzzy> never tried to write anything myself
[06:25:06] <abcminiuser> It says 150KHz IIRC in the datasheet
[06:25:13] <ziph> abcminiuser: Of the ripple on the LDO; is it the same as the SMPS ripple?
[06:25:15] <scuzzy> I know it's a symetric key system
[06:25:23] <abcminiuser> Oh, damn, didn't check
[06:25:26] <ziph> abcminiuser: Typical ripple rejection is 40-60dB, so your LDO should kill it.
[06:25:28] <abcminiuser> That was stupid :P
[06:25:44] <abcminiuser> That's what I would have thought, I think my CRO is lying to me
[06:26:02] <abcminiuser> Given the cheap rigol doesn't give a flat zero out even with GND'd inputs
[06:26:03] <abcminiuser> :P
[06:26:05] <ziph> It could be really bad common mode noise going in to the CRO.
[06:26:19] <ziph> Oh, your grounded input touched on the supply is showing it?
[06:26:55] <abcminiuser> I mean disconnecting the CRO completely, shorting the CRO channel 1 to it's GND still shows ~40mV of P-P ripple
[06:27:44] <ziph> You've probably got RF floating around then that is being picked up by the probe loop.
[06:28:17] <ziph> Or magnetic coupling into the loop, but that should only be present when the loop is near one of the inductors.
[06:29:35] <ziph> Oh, and if you have a Wacom tablet, keep it well away from the CRO. They're massive flat magnetic radiators. :)
[06:30:51] <abcminiuser> No tablet here
[06:30:59] <abcminiuser> And nothing else but my laptop, perhaps that's injecting noise
[06:31:15] <abcminiuser> I'll re-do it with the expensive CRO at Uni tomorrow or so I think
[06:31:58] <ziph> Make sure the probe gain matches the setting in the CRO too, you might just be seeing 4mV. :)
[06:32:21] <abcminiuser> Yes, checked that :)
[06:32:29] <abcminiuser> Goddam this thesis is going to cost a bomb to print
[06:32:47] <abcminiuser> It's not done and already 78 pages (lots of pictures, etc.)
[06:32:52] <ziph> Oh, and when checking an unknown frequency, wind up the speed to make sure it isn't being alised to a lower frequency.
[06:32:59] <ziph> bbl
[06:54:07] <inflex> lo folks
[06:54:24] * inflex is a bit down and out today - no exciting new PCBs... and I destroyed my beloved DLG (discus launch glider)
[07:17:09] <Valen> abcminiuser: take a look at getting it printed with like vistaprint or perhaps some of the online publishers
[07:17:24] <Valen> i think they can do books for a few $
[07:19:15] <Valen> how many copies do you need done?
[07:19:19] <abcminiuser> Valen, HOLY CRAP NO
[07:19:33] <abcminiuser> VistaPrint is one of the scammiest organisations known to mankind
[07:19:39] <Valen> well there is that
[07:19:40] <abcminiuser> (Minus Scientology)
[07:19:50] <Valen> but they do actually print stuff for not outrageous sums
[07:19:56] <abcminiuser> Officeworks will print and bind, 9 cents/B&W page
[07:20:06] <inflex> mmm... I see no AVR talk here
[07:20:09] <abcminiuser> Plus binding - around 100 pages, three copies
[07:20:32] <Valen> i can probably run that out at a client of mine if your happy with the ring binder stuff?
[07:20:44] <Valen> colour if you want it lol
[07:20:47] * inflex wondersif he's going to have success in keeping power consumption on his "soft switch" configuration low enough
[07:20:54] <abcminiuser> ^ They do print things, but they a) try to sing you up to monthly fees and b) advertise super cheap or even "free" cards, then hit you with a HUEG shipping fee right at the end
[07:20:57] <Valen> that black plastic thingo
[07:21:14] <abcminiuser> Valen, it's fine, I'm time limited so I'll just pay the $50 at officeworks
[07:21:24] <Valen> no worries ;->
[07:21:40] <abcminiuser> For a little more I could by a new laser printer :P
[07:22:45] <mrfrenzy> the key to ordering from vistaprint is using a throw-away mail and phone
[07:22:48] <inflex> mmm... maybe I'll run on the 128kHz internal osc on the T25
[07:22:51] <Valen> probably only can print 10 pages before needing $800 in new carts
[07:23:33] <inflex> $99 for a laser printer with 1500 page "test" toner
[07:23:49] <inflex> but it's not duplex... and you really probably want duplex...right?
[07:23:57] <inflex> or do they forbid that for thesis now?
[07:24:44] <mrfrenzy> you have to look at the specifications of your school
[07:26:01] <mrfrenzy> and if it's only 70 pages and on a budget you could always do manual duplex
[07:26:07] <mrfrenzy> that's how we printed thesis in the old days ;)
[07:26:23] <inflex> eww... I hate manual duplex... there's almost always a page skip and everything's fubar'd
[07:26:29] <Valen> how many times did you !@#$ it up the first time
[07:26:33] <Valen> upside down and back to front
[07:26:34] <mrfrenzy> then you have a crap printer
[07:26:40] <mrfrenzy> any decent laser will not skip pages
[07:26:43] <inflex> hhahahahah
[07:26:46] <inflex> humans make errors
[07:26:52] <mrfrenzy> yes ofcourse
[07:26:55] <inflex> anyhow, I prefer using something like a reliable HP2200D
[07:26:55] <ziph> abcminiuser: Your uni guild doesn't do printing?
[07:27:06] <mrfrenzy> first thing I do with a new printer is figure out which way to insert the papers
[07:27:16] <mrfrenzy> then I take a black marker and paint an arrow
[07:27:20] <inflex> don'tforget toothough, paper behaves a bit different after it's printed and restacked
[07:27:20] <Valen> one of my clients just got a new kyocera
[07:27:27] <abcminiuser> ziph, they sure do, also for 9 cents/page
[07:27:34] <ziph> Ah.
[07:27:46] <Valen> i think the only thing it doesn't do is give you money for the cab ride home after
[07:27:53] <ziph> Why do you even need printed copies? :)
[07:28:11] <Valen> 80 A4 pages, i'd rather it in dead tree form
[07:28:32] <inflex> some things are nicer in PDFs... like Atmel AVR datasheets :D
[07:28:56] <mrfrenzy> 9 cents per page is not too expensiver. toner is like 3 cents, paper 1 cent and work 5 cents
[07:29:00] <inflex> I do like how they've tied datasheets and stuff into Eagle with Element14 now
[07:32:35] <abcminiuser> ziph, Uni likes physical everything
[07:32:58] <abcminiuser> For archival purposes, and because our professors are fossils
[07:33:11] <abcminiuser> Decent fossils in some cases, my supervisor used to work for NASA
[07:33:11] <ziph> I think my Uni started accepting CS assignments electronically while I was there.
[07:33:21] <ziph> You'd upload everything to a web application..
[07:33:26] <abcminiuser> ziph, ours does for CS, not for EE
[07:33:30] <abcminiuser> Inexplicably
[07:33:32] <ziph> And it was nice because you could reupload everything too.
[07:33:38] <karlp> my these had to be submitted in pdf and 2xhardcopy
[07:33:40] <Valen> one of my friends used to get the lecturer to sign a reciept when she handed her work in
[07:33:41] <karlp> that was EE
[07:33:54] <abcminiuser> I'm off to bed, it's getting late
[07:33:57] <Valen> because they kept loosing stuff and blaming the students
[07:33:58] <Valen> same
[07:34:00] <Valen> nighty night
[07:34:04] <abcminiuser> Need to be bright eyed for another fun Thesis writing day
[07:34:11] <ziph> abcminiuser: Goodnight, hope there's already someone in there. ;)
[07:34:17] * Valen wrote 40 pages of report in 8 hours
[07:34:22] <Valen> got 87 for it ;->
[07:34:25] <abcminiuser> Not today :(
[07:34:30] <abcminiuser> Valen, holy crap
[07:34:43] <Valen> google image search is the bomb ;->
[07:34:46] <abcminiuser> I've only got 78 pages (~55 pages of text) and it's taken me ages
[07:35:03] <Valen> of course the pictures need to be relevant but still
[07:35:08] <ziph> abcminiuser: Dragon NaturallySpeaking works for that kind of thing. :)
[07:35:19] <Valen> digital satellite communications, ftw
[07:35:52] <abcminiuser> ziph, I find myself terrible at dictation, it breaks my thought process
[07:36:07] <Valen> I once gave a 40 minute presentation on the japanese remote manipulator arm based on 20 minutes of finding photos of it on the net, the fact all the text was in japanese meant i could get away with alot
[07:36:18] <ziph> abcminiuser: Dictating to a computer is different.
[07:36:26] <ziph> You can sit there and pause and it doesn't care. :)
[07:36:30] <abcminiuser> Also I don't want to turn in a paper with the words "As is shown in figure 2, an hey mike how are you year not bad see you around"
[07:36:43] <Valen> i'm with abcminiuser technical writing is not suited to dictation
[07:37:03] <abcminiuser> I'm not a good speaker :P
[07:37:04] <ziph> abcminiuser: Yeah, Dragon doesn't include expletives. I was surprised until I realised why it wouldn't. ;)
[07:37:23] <ziph> Pft, it improves technical writing. :)
[07:37:25] <abcminiuser> ziph, I wonder what happens if you shove some Gordon Ramsay into it then?
[07:37:28] <abcminiuser> Blank pages?
[07:37:40] <ziph> Because you can't write nonsensical phrases that would never be spoken.
[07:38:06] <abcminiuser> Anyway, off to bed
[07:38:07] <abcminiuser> Night all
[07:38:14] <ziph> The writing has to go through the part of your brain that makes sure your mouth isn't just spitting out nonsense. :)
[07:38:26] <inflex> night
[10:05:51] <soul-d> i think ill stick to uv lighting instead of toner transfer
[10:05:53] <soul-d> http://i.imgur.com/JcOc4.jpg
[10:06:57] <karlp> soul-d: I believe that twas the channel consensus here a week or two ago when that last came up :)
[10:10:27] <soul-d> you guy's predrill holes thats the last problem i have drilling holes might snag tiny pads
[10:11:26] * karlp shrugs. I do neither. :) But I believe that, of those two options, most people preferred UV lighting
[10:48:11] <Casper> inflex: finally I think the problem for the smps is that roman black tuned the values of the resistor a bit too borderline, causing the switch transistor to barelly saturate... so if the load is slightly more than the saturation point, it go linear. I toyed with the resistors and I doubled the current output and it go well in switching now with just under 2W wasted in the switch
[11:56:52] <spow> Hi. I want to measure the time an A/D conversion takes. Is it possible to set an interrupt on the 'conversion finished' bit of the ADC ?
[11:57:32] <karlp> of course
[11:57:38] <karlp> have you read the datasheet on the adc?
[11:57:58] <mrfrenzy> it requires reading a few times, taking notes ;)
[11:58:24] <mrfrenzy> but yes, everything is possible. It is quite normal to have an interrupt on conversion finished, then you launch whatever code usese the read value
[11:58:32] <ftc> those datasheets can be downright incomprehensible
[11:59:46] <spow> yes yes I read the datasheet. I mean is it the proper solution ? Won't the interrupt cause many cycles being read wrongly ?
[12:00:21] <spow> or maybe the interrupt is low in cycles ?
[12:03:06] <ftc> it should only take the number of cycles in your interrupt handler plus a little bit of overhead for the other things the cpu needs to do
[12:04:36] <ftc> given how much stuff relies on interrupts I can't imagine it is too terribly heavy
[12:04:40] <mrfrenzy> the interrupt handler would only need to stop your timer
[12:05:00] <spow> so 4 clock cycles for the handler ?
[12:05:45] <district> there's jitter in the interrupt latency depending on what the currently executing instruction is, but the maximum is still pretty small
[12:06:20] <karlp> doesn't the datasheet alsos imply define how long an ADC conversion takes?
[12:06:33] <ftc> i don't have the data sheet for an avr chip on hand but i seem to recall that the adc max sample rate was orders of magnitude smaller than the cpu clock
[12:07:01] <spow> okay i'll measure it with the interrupt and then decide if the handler is negligible or not
[12:07:12] <spow> karlp: it's my homework
[14:44:48] <ripthejacker> Hey all, anyone could help me a bit for a design im working on?
[14:46:28] <_Shurik_> like, for free??
[14:49:24] <ripthejacker> Why of course.
[14:51:37] <_Shurik_> Well, I'm not very experienced, but there are lots of people who are
[14:51:43] <_Shurik_> what you trying to do there?
[14:53:33] <ripthejacker> well firstly just had a basic question, are there really any requirements for a zigbee module to be put with a AT chip aside that itneeds the UART protocol pins...i read somewhere memory may become an issue if using too low of ram in a chip or something
[14:55:42] <_Shurik_> I actually wouldn't know that one. Somebody might wake up in a minute with an answer :)
[15:09:06] <karlp> ripthejacker: memory?
[15:13:34] <ripthejacker> yeah I first thought aslong as the AVR chip had the UART pins it would be fine
[15:13:54] <ripthejacker> which i dont understand since its just sending data to the xbee
[15:14:07] <ripthejacker> but yeah, someone said somethign about lower memory chips may not work
[15:14:27] <ripthejacker> on the other hand, ive seen zigbee used with both atmega8 and attiny2313 so i thought there wouldnt be any memory issue
[15:22:36] <soul-d> http://imgur.com/a/9T26z still need more smd soldering practice though
[15:24:19] <JanneP> toner transfer PCB?
[15:24:23] <JanneP> looks pretty nice
[15:25:10] <soul-d> no uv
[15:25:41] <JanneP> okey
[15:25:51] <JanneP> SMT without solder mask is always s bit difficult
[15:26:09] <soul-d> still need to fetch chalk paper instead of transparant
[15:26:16] <JanneP> i've found it's easiest if I first coat the tracs with flux
[15:26:18] <soul-d> if thats how it's called in english :P
[15:26:43] <soul-d> yeah used enough of that this time last time i used a blob of tin
[15:27:02] <JanneP> no idea, english's not my primary language :P
[15:27:09] <JanneP> what wrong with using transparencies?
[15:27:36] <soul-d> toner difrences and bit less dense then paper
[16:04:15] <karlp> ripthejacker: I think they might have been talking about how much of the zigbee stack you want to use
[16:04:40] <karlp> there's quite a bit of difference between implementing a zigbee router, coordinator or node, vs using 802.15.4 radio modules
[16:05:10] <karlp> I have happily used a xbee with an attiny45, and software serial for the uart, but that's not zigbee
[16:13:19] <ripthejacker> karlp : basically you were able to implement an Xbee without using UART pins?
[16:14:23] <Casper> software uart does work, just more of an annoyance code wise
[16:15:30] <ripthejacker> Hm I think ill stick with a hardware UART/zigbee module since im a beginner
[16:17:13] <Casper> hardware is always easier than software
[16:17:31] <Casper> specially when you need timing sensitive stuff
[16:20:49] <karlp> I was doing tx-only, 802.15.4, using an xbee, yes.
[16:20:53] <karlp> but that's not zibee
[16:21:11] <karlp> and you'll only confuse people if you talk about using a uart/xbee combo as being zigbee
[16:41:50] <inflex> Casper: congrats on the solution for the SMPS
[16:44:25] <Casper> inflex: is there an eay way to guestimate the inductor value ?
[17:00:07] <inflex> nfi
[19:21:13] <inflex> why would ppl still use the atmega8 over the mega88
[19:21:38] <Tom_itx> stupid
[19:22:04] <Tom_itx> or overstocked with them
[19:22:52] <inflex> *nod* that's what I'm thinking
[19:23:39] <inflex> just noticing all the speed conrollers coming outof China's cheap hobby stores have previously had the mega8L in them... but now they've moved to SiLabs chips, so I'm thinking someone bought a few hundred thousand of the mega8/8L's and just used them up
[19:23:48] <inflex> the SiLabs chip is likely a lot cheaper
[19:24:41] <Casper> or didn't had access to them
[19:24:45] <Casper> or didn't knew about them
[19:25:04] <Tom_itx> highly doubtful
[19:26:23] <inflex> given that the designs of these controllers are all very very similar, I'm thinking it's just one or two factories
[19:26:41] <Casper> or copycat
[19:27:06] <inflex> true, plenty of that too
[19:28:01] <Casper> you know, the roman black smps... probably would fail to work proprelly with tl431 instead of the zener
[19:29:01] <inflex> how come?
[19:29:14] <Casper> I'm pretty sure that the crappiness of the zener help... specially the "high" impedance
[19:29:40] <Casper> when the switch turn on, it also push some current there, raising a bit the voltage
[19:30:02] <Casper> not for long, but probably enought to help a bit
[19:30:20] <Casper> the tl431 have a very low impedance, so would probably just "eat" that power
[19:30:32] <Casper> so no small bump
[19:37:58] <inflex> right, so you put a cap on it to induce the startup bump :D
[19:42:23] <Casper> I wish I knew more about smps... I'ld like to be able to mod it up for higher output current and possibly higher efficiency
[19:42:45] <Casper> sadly... me don't know much about smps... the course got quite... what course....
[19:43:08] <inflex> well, you could always just ignore all the discrete stuff and use the T25
[19:43:15] <inflex> cheaper/simpler really :\
[19:43:35] <Casper> still don't know the math to select the cap and inductor
[19:43:55] <Casper> all I know is: if frequency goes up, inductance must go down
[19:43:58] <inflex> pick something from a datasheet using between 250~1MHz
[19:44:21] <inflex> eg, find a controller with comparable switching speed and current capability
[19:44:31] <vectory_> rather know the math );
[19:44:40] <Casper> last time I dug into datasheet... I couln't find a single one that the formula was giving something close to their example
[19:45:08] <Casper> sometime the difference was a matter of 2 or 3 times bigger/smaller
[19:46:23] <vectory_> is it the proportion from inductor to cap that counts?
[19:48:15] <Casper> afaik, no
[19:49:01] <Casper> the inductor is sized based on Vin and Vout and speed and current... the capacitor is sized for the ripple wanted
[19:52:41] <ziph> LT's SMPS application notes aren't any help?
[19:53:18] <Casper> I'ld need to check again
[19:56:39] <karlp> Casper: read the MCP1640 datasheet, it has some good stuff
[19:56:54] <karlp> MCP1623 does too, they're basically the same, just different current ratings
[19:58:48] <rue_house> !thislog
[19:58:48] <tobbor> This one:
http://rueshouse.dyndns.org:82/~ircjunk/irclogs/html/%23avr-2011-11-15.html
[19:59:18] <rue_house> last entry at 13:30
[19:59:23] <rue_house> zlog,
[20:01:09] <Casper> so... mother ask me: "Do you have a trick to make this door stay open?" "it's your stuff that you put on it that close it up" "no, it never stayed open look!" *remove everything behind* *me open the door* "see, it was your stuff"
[20:01:27] <rue_house> I see like 12 hours down
[20:03:03] <inflex> """Well, who cares, I for one am glad they've switched to SiLabs, anything is better than avr these days. The only reason they still exist is probably arduino. """
[20:03:09] <inflex> gawd, I want to punch people like that in the face
[20:03:26] <inflex> Reeks of elitism to me - "Boo, AVR is popular now, I have to hate it!"
[20:04:24] <ziph> I hate AVR's too.
[20:04:41] <inflex> O_o
[20:04:45] <ziph> But that's because of cost and piss poor reliability of availability. ;)
[20:04:55] <inflex> indeed, that'd be the two things I dislike as well
[20:15:55] * Casper wonders if he have some good pchan fet...
[20:40:33] <rue_house> microchips policies alienated their products from the people who would have made it popular
[20:41:21] <karlp> masochists will always exist
[20:41:37] <rue_house> if there had been a means to program the new ones via parallel port when they came out, and there had been a FREE C compiler, avrs prolly woulnd't have had the chance
[20:41:42] <ziph> What policies?
[20:42:00] <ziph> Keep in mind the PIC architecture is also junky.
[20:42:04] <inflex> well, their initial policy was positive for the PIC
[20:42:07] <rue_house> "we shall make non-backwards compatible new versions of our old chips"
[20:42:09] <inflex> eg, easy to access free samples
[20:42:29] <inflex> and yes, the old PIC architecture was a major PITA
[20:42:52] <rue_house> and , maybe not their fault, but keeping all the programmer hardware as unbuildable by hobbists
[20:42:55] <inflex> OOoooh, I've got no spare registers, you better shuffle your shit all around the place, don't worry, we can do it at 40MHz!
[20:43:14] <rue_house> yea, and not to mention paged memory
[20:43:23] <inflex> that was a feature ;)
[20:43:24] <rue_house> and clock dividers of 4
[20:43:39] <inflex> What are the new PICs like though?
[20:43:44] <rue_house> dont care
[20:43:55] <inflex> that's my view too - but I was merely curious
[20:44:00] <rue_house> still no parallel port programmer schematics that I'm aware of
[20:44:41] <inflex> damn, how will they beat the 3-resistor //port programmer
[20:44:46] <inflex> butterflies!
[20:44:59] <rue_house> do they give otu a free C programmer now?
[20:45:11] <inflex> Anyhow, I hate people who shoot out with blanket statements like that before and then you find out later just how insanely specific their real complaint is
[20:45:23] <rue_house> did atmel sponsor the gcc port for avr?
[20:45:24] <inflex> rue_house: they probably have to, probably with a 4K limit or something?
[20:45:35] <inflex> rue_house: yes, afaik Atmel did invest in to gcc to some level
[20:45:41] <Tom_itx> cpld is where it's all at now
[20:45:41] <rue_house> yea
[20:45:49] <Tom_itx> drop all your avr's in the shit can
[20:45:59] <rue_house> hah
[20:46:28] <inflex> Tom_itx: and that can would be coincidently located at your house?
[20:46:37] <Tom_itx> absolutely
[20:46:46] <inflex> http://www.rcgroups.com/forums/showpost.php?p=19876656&postcount=393 <=- my response to the person who hated AVRs
[20:47:48] <Tom_itx> sensored?
[20:47:59] <inflex> ?
[20:48:42] <rue_house> sensord is the i2c system I use for logging temps with my boiler and house
[20:49:04] <inflex> oh yeah, I2C daemon for linux
[20:49:11] <rue_house> I C an E where I C no E B 4
[20:50:27] <inflex> I shouldn't bother replying to people like that
[20:50:43] <ziph> inflex: Yeah, haven't you been around long enough to know that? :)
[20:51:13] <inflex> ziph: I have moments of weakness to pretentious people.
[20:52:08] <inflex> I do want to say "Oh do just shut the fuck up you pretentious misguided myopic arsehole" ... but instead I waste an hour playing "Change the goal posts" with them for no good reason.
[20:52:52] <ziph> :)
[20:53:02] <Tom_itx> bored?
[20:53:29] <inflex> I've seen this person before, surprised they didn't start slapping down the "I worked with a Fortune-100 design group and we all agree AVRs are SHIT... and we prefer [insert some hideously expensive exotic uC that requires a billion support components and a $20k programmer]"
[20:54:48] <inflex> Tom_itx: I must be :(
[20:54:56] <inflex> also want more coffee, and food... and money
[20:55:22] <inflex> it's a very dangerous game to play with people like that... becuase if you stuff up (and it happens) you end up looking like a mega jerk to everyone.
[20:56:13] <Tom_itx> they start out thinking they hold the upper hand
[20:57:01] <inflex> yes, "This is my herd of sheep"
[20:57:39] <inflex> but yes, bamboozle people with enough of the right talk and you'll have a following
[20:57:51] <dehuman> hahah are you in a flame war with timecop on rcgroups?
[20:57:52] <dehuman> lulz
[20:57:58] <inflex> Not specifically because the followers are "stupid", rather because they made the mistake of actually beliving someone on faith.
[20:58:08] <inflex> dehuman: bah, not a flame war
[20:58:17] <inflex> dehuman: just getting him to be more -specific-
[20:58:43] <inflex> dehuman: am thinking it's not the first time either, I swear I've smelt him before.
[20:58:58] <dehuman> probably
[20:59:06] <dehuman> he hangs on in #electronics@efnet
[20:59:22] <dehuman> also affiliated with trolls, GNAA etc etc
[20:59:29] <inflex> oh good, I should go over there and shove him
[21:00:12] <inflex> You know, none of this would have happened if he had simply said "I am glad that they've moved to SiLogic because it'll be easier to adapt the system to our needs"
[21:00:23] <inflex> as opposed to "... anything is better than AVR"
[21:00:38] <inflex> but I suppose the former requires too much thinking and doesn't score enough hit points.
[21:00:57] <Tom_itx> it's not about facts
[21:01:02] <Tom_itx> it's all about perception
[21:01:04] <inflex> Tom_itx: indeed.
[21:01:31] <dehuman> pay no mind to timecop
[21:01:32] <dehuman> no one does
[21:01:40] <inflex> Still I added the twist-back at the end; he can choose to be happy that I agreed, but it invalidates his original statement.
[21:01:44] <Tom_itx> is he another flyback?
[21:01:47] <dehuman> when he isn't trashing avr's
[21:01:55] <dehuman> he is extolling the virtues of microsoft over *
[21:02:08] <inflex> oh man, I think that's how I've sniffed him before.
[21:02:10] <inflex> (MS)
[21:02:40] <inflex> in a way I'm waiting for the personal attack to come next...
[21:02:40] <ziph> Microsoft is better for electronics work.
[21:02:58] <inflex> here comes Ziph with his anti-Eagle tyrade ... *runs*
[21:02:59] <ziph> I don't know many professional EE's that don't just use 100% Windows.
[21:03:07] <inflex> ziph: well, I'm one.
[21:03:36] <ziph> inflex: This is bait training, you just failed the first test. ;)
[21:03:36] <inflex> we're growing in numbers again, after the great purging between 1995~2005
[21:03:57] <inflex> ziph: bah - I'm back to being a poop scooper
[21:04:26] <inflex> not even a padiwan :(
[21:05:20] <ziph> #arduino is fairly hilarious that way.
[21:09:15] <inflex> http://www.rcgroups.com/forums/showpost.php?p=19876893&postcount=394 <= heh, he came back, a bit flustered.
[21:09:23] <inflex> Impressed though he did avoid personal insults.
[21:09:33] <inflex> I did reply alas...
[21:11:11] <dehuman> speak of the devil
[21:11:44] <inflex> oh?
[21:12:28] <inflex> He's in here now?
[21:23:06] <ziph> inflex: I think if I was going to bother with something requiring commercial toolchains like the SiLabs I wouldn't bother with 8 bit.
[21:23:21] <ziph> inflex: I'd much prefer an MSP430 or something similar.
[21:26:30] <inflex> yes, actualy I mentioned that in the thread
[21:26:41] <inflex> oh, no I didn't - I mentioned it here
[21:26:53] <inflex> anyhow, he's trying hard to reignite the trolling, after saying "End of story".
[21:32:46] <ziph> inflex: Tell him he's wrong because you've worked at Intel since 1981.
[21:34:35] <inflex> hah
[21:34:42] <inflex> I'll change that to 1781
[21:34:54] <Tom_itx> zilog
[21:34:58] <ziph> Yeah, I actually meant to say 1961, which would be before it was formed.
[21:35:37] <inflex> wonder how he's going popping that zit on his face... might get it done before he has to go to bed.
[21:35:57] <inflex> okay, that was just uncalled for
[21:37:18] <Tom_itx> leave it to you to take it over the top
[21:37:30] <Tom_itx> gawd.
[21:37:32] <Tom_itx> :)
[21:38:58] <inflex> whaaa... the zit thing?
[21:39:04] <Tom_itx> btw, have you had a chance to use the programmers?
[21:39:07] <ziph> Post his home address and telephone number to some Arduino forum. ;)
[21:39:19] <inflex> Tom_itx: yours?yep - use them all the time
[21:39:25] <Tom_itx> the 12v thing?
[21:39:36] <inflex> Tom_itx: haven't used the HV yet... but I did need the HV for a mega88 the other night... but sadly did not have :(
[21:39:49] <inflex> the HV thing will come in use very soon though, soon as this new batch of PCBs arrive
[21:40:23] <Tom_itx> you can use the switch on the 12v board to go back and forth
[21:41:57] <inflex> yep, saw that
[21:51:27] <matt5> hey guys, for the ldm avr instruction, is it a requirement that the Z pointer _always_ be used?
[21:57:18] <inflex> wait, wth... what is "LDM" ?
[21:57:20] <inflex> which chip?
[21:57:44] <inflex> I know LD, LDI, LDS, LDD ...
[21:59:15] <inflex> Ah, so I see now that timecop (come on now, don't be shy) decides to look here
[22:01:39] <matt5> timecop?
[22:01:59] <matt5> anyway, I think it's a load program from memory
[22:02:05] <matt5> loading from flash
[22:02:12] <matt5> I've an atmega32 board
[22:04:30] <matt5> what's ldd though? I use lds to load from sram
[22:09:55] <inflex> Well Timecop, it's a bit of a pity you cannot come out here and actually talk to me.
[22:10:33] <Tom_itx> oh, we have a lurker?
[22:10:45] <matt5> what the heck at you guys talking about
[22:11:03] <inflex> meh, someone cut-paste one of my statements from here and gave it to Timecop
[22:11:12] <inflex> who subsequently posted it on RCGroups as if I had written it in a thread there.
[22:11:17] <matt5> what/who is timecop?
[22:11:39] <matt5> Nevermind actually :) so, say again - what do you use the ldd instruction for?
[22:13:43] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/pdf/atmel/AVR_instruction_set.pdf
[22:13:55] <Tom_itx> appears to be the same as ld
[22:15:23] <inflex> but with displacement
[22:16:25] <matt5> I'm still a grasshopper in training. can you tell me please when that would be used?
[22:16:55] <inflex> mmm... say you want the N'th item in an array that's located at the address stored in Rd
[22:17:07] <inflex> that'd be my first thought
[23:34:16] <ziph> inflex: Looks like the master baiter gave up.
[23:35:13] <inflex> ziph: I'm sure he's just waiting
[23:35:25] <inflex> and suddenly... maggots!!!! bleah!
[23:35:42] <inflex> hate it when you forget about some cat food on a dish for a few days that you stacked with some other dishes in an obscure location
[23:37:11] <inflex> People like that do seem to tend to keep following you afterwards, looking, seeking, waiting
[23:37:31] <ziph> Well, he's already stalking you in this channel. :)
[23:37:48] <ziph> Do you think the fascination is part sexual? You do post your photograph with your forum comments.
[23:42:42] <inflex> I am admittingly always intrigued by such people
[23:43:16] <inflex> curious as to how their logical processing works. . . likewise with people who believe they have the "right" to everything like movies, stuff they steal, or things they should be allowed to do.
[23:44:16] <ziph> Well, at the root of it is the belief that they aren't bad people, or that they are smart.
[23:44:37] <ziph> e.g. I don't steal, therefore what I'm doing isn't stealing.
[23:46:04] <Landon> ziph: or some that genuinely don't understand how it works
[23:46:49] <Landon> I have several friends that thought clicking through the "don't pirate with this software" disclaimers for whatever download client they had meant that they could freely download whatever they wanted
[23:47:04] <Landon> with no repercussions, because they agreed to the terms
[23:47:42] <ziph> There are generally hints when you're wrong like that.
[23:48:22] <ziph> A normal person would quickly realize that it isn't the case because programmer salaries aren't paid by magic fairies.
[23:48:46] <Landon> one could argue college students are hardly normal people yet though
[23:48:58] <Landon> ;)
[23:50:13] <ziph> They're full of crap if they tell you they believe that.
[23:52:07] <inflex> I have however come to the conclusion that RIAA/MPAA etc are actually just adult versions of a lot of the pirates.
[23:52:41] <ziph> Oh?
[23:52:43] <Landon> heh
[23:52:52] <ziph> I actually think the correct approach was for them to sue individuals.
[23:53:05] <inflex> ziph: I was more thinking, they're both scum ;)
[23:53:42] <Landon> ziph: I agree with that, but I'd draw the line at irreparably harming individuals
[23:53:46] <inflex> they both deprive artists of money (virtual or otherwise)
[23:54:06] <Landon> at least financially, I don't imagine they're putting people to the rack >.>
[23:54:14] <ziph> Landon: That's more a problem with the legal system.
[23:54:36] <inflex> I find it disturbing that they (RIAA/MPAA) can put forward so many john-doe cases
[23:54:49] <inflex> for sure, if they find a legit pirate, go for it
[23:55:04] <inflex> but the whole blanket casting approach is disgusting
[23:56:00] <inflex> I think what really spun me was the /. story the other day about how the german "RIAA" charged licence costs to an all-creative-commons concert
[23:56:27] <inflex> and stuff like the levies on recordable media
