Back
[00:00:43] <Casper> same way?
[00:00:54] <Casper> and why do you complicate your life with binary?
[00:12:46] <biker_> Casper: i dont know,., its easy for me to do binary :p
[00:13:06] <abcminiuser> Hrm
[00:13:12] <biker_> Casper: but what same way?? i tried that way and it didnt work
[00:13:20] <abcminiuser> Is there a nice way to help a guy collect interest in his work?
[00:13:39] <Casper> is the 3 a 16 bits one?
[00:13:47] <biker_> Casper: yea
[00:14:02] <Casper> then you'll have to address it in 16 bits
[00:15:25] <biker_> Casper: like in the other ones (OC1A and OC1B)
[00:15:38] <biker_> but i tried just calling OCR3A and OCR3B
[00:15:42] <biker_> but they didnt work
[00:16:05] <Casper> and you sure the part you're compiling for have the 3?
[00:16:30] <biker_> Casper: i wrote to you waht i have
[00:16:40] <biker_> Casper:
http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&p=899422&sid=1698721a178a1bc696de1d51088e7e4b#899422
[00:16:47] <biker_> i have to go,., ill keep watching the forum
[00:16:50] <biker_> thanks for the help :)
[00:29:51] <abcminiuser> Ok, is this too harsh?
http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&p=899425#899425
[00:32:14] <Casper> not harsh enought
[00:32:18] <Casper> :D
[00:32:34] <Casper> I hate thread bumbers and the alike
[00:34:48] <Casper> abcminiuser: are you good with smps?
[00:35:02] <abcminiuser> Switchmodes?
[00:35:05] <Casper> ya
[00:35:07] <abcminiuser> I've only layed out two :P
[00:35:18] <Casper> do you have some good doc for them?
[00:35:37] <abcminiuser> ^ I like they guy's enthusiasm, but holy crap he's done everything wrong when it comes to trying to gain e-friends, donations, etc
[00:35:58] <Casper> specially what I look for is some tl494 with fet
[00:36:06] <abcminiuser> Casper, we only got told "keep the feedback path as short as possible" and to have a nice big ground plane underneath them
[00:36:09] <Casper> and working circuit
[00:37:32] <Casper> I need to find a place to host a small site, but all free one suck
[00:37:40] <Casper> and my home connection is kinda too slow
[00:37:49] <Casper> my isp offer web host, but only 5MB !
[00:38:55] <abcminiuser> Couple of bucks a month will give you an adequate server
[00:39:01] <abcminiuser> Shared, but workable
[00:39:26] <Casper> I guess tom... today I'll try some crappy fet driver, I hope that 2n3904/06 would do the job...
[00:45:58] <ziph> Casper: What kind of SMPS?
[00:49:30] <Casper> currently I'm in forward topology
[00:50:10] <Casper> I will then, if I get good efficiency and power, move to buck, then boost
[00:50:18] <ziph> So just a small current DC/DC?
[00:50:33] <Casper> 75W for the first
[00:50:58] <Casper> 15-22-> 13.65 current limited
[00:51:36] <ziph> Digging up a few evaluation board datasheets might help you get started, they have the layouts usually.
[00:51:46] <Casper> I have issues with the damn magnetics
[00:52:01] <ziph> Rating issues, or coupling?
[00:52:10] <Casper> and the transistors that sometime refuse to go down enought...
[00:52:34] <Casper> more like what should be a 70+% efficient first was 5% :D
[00:52:45] <ziph> Do they heat up? :)
[00:53:14] <Casper> the transistors, enought to smell the heat in 20 seconds, with some heatsink too
[00:54:09] <Casper> I read that for 12V in 6 turns was a good start
[00:54:28] <Casper> first core I have 8, and it do not transfert enought power
[00:54:45] <Casper> this one is smaller, but I have 6 turns... and all I get is extremelly high current
[00:54:58] <Casper> and almost nothing on the output
[00:56:15] <ziph> That probably means the inductance isn't high enough.
[00:56:59] <Casper> and I keep reading conflicting stuff about how to wind the transfo
[00:57:58] <ziph> What kind of core?
[00:58:09] <Casper> from an old atx psu
[00:58:30] <Casper> which run at about 67kHz, I'm at 20kHz however
[00:58:43] <ziph> Is it a toroid?
[00:58:46] <Casper> no
[00:58:54] <Casper> EI core
[00:59:12] <Casper> http://img155.imageshack.us/img155/7166/img3065eu.jpg ← current bad setup :D
[01:01:39] <Casper> see theBear transistor heatsink? 20 seconds and they are too hot to touch with that transfo
[01:03:02] <Casper> theBear??? woops
[01:03:09] <Casper> look like I hit tab?
[01:04:59] <ziph> Casper: My bet would be that you need to increase the inductances on the transformer (even if you don't change the winding ratio).
[01:05:44] <Casper> ziph: possibly...
[01:06:48] <Casper> original winding: 19Primary secondaries 19Primary both in series, for 38 turns in total
[01:06:50] <Casper> for 340
[01:06:54] <Casper> 340V
[01:07:05] <Casper> that's almost 10 volts/turn
[01:07:20] <ziph> They were using higher frequencies though.
[01:07:30] <Casper> I have 12V, I first tried 1 turn, was total short, 6 turns barelly work
[01:07:36] <Casper> well, 3 times the frequency
[01:08:03] <Casper> I doubt it change much
[01:08:58] <ziph> It will if the majority of the energy is no longer in the midband of the transformer transfer Fn. :_
[01:08:59] <ziph> :)
[01:09:30] <Casper> enought to go down to 1% efficient?
[01:13:52] <ziph> Dunno.
[01:14:24] <ziph> Say your coupling factor was 0 though; then you'd basically have an inductor in series with the transistors.
[01:15:18] <Casper> but bed time, tomorrow I'll try to move from that half fried chip to a tl494
[01:15:24] <Casper> hopefully it will work better
[01:15:55] <Casper> that chip somehow survived a misconnection, power -> Vref out yikes :D
[01:16:11] <Casper> now the vref is about 9V instead of 5 hehe
[01:16:18] <Casper> frequency is also all messed up
[01:16:37] <Casper> bed time, so good night
[02:39:51] <earthshine> o/
[05:02:19] <amee2k> mmh, is there a temperature sensor like LM75 (i.e. with built in over-temperature comparator) but with analog output instead of TWI??
[05:07:34] <karlp> you want it to send you _some_ digital signals, but not others?
[05:12:02] <amee2k> karlp: yes
[05:12:21] <amee2k> whats wrong with that?
[05:16:25] <amee2k> well, i'm really after the analog temperature signal, but i'd only end up adding my own overtemp comparator
[05:16:46] <amee2k> and the protection shutdown is somewhat critical so if the sensor could do it, i would strongly prefer that
[05:29:04] <RikusW> http://uvicrec.blogspot.com/
[05:57:44] <carp3> wow
[06:01:14] <mapee> hi
[06:01:15] <tobbor> mapee! like, totally tell us about the project!
[06:01:24] <Regenschein> hi mapee
[06:10:58] <elektrinis> hi
[06:13:18] <inflex> anyone used a few spare pins on an AVR as a charge-pump doubler/tripler?
[09:18:00] <rue_house> you could
[09:18:26] <rue_house> having a problem?
[09:18:50] <rue_house> amee2k, LM35
[09:19:16] <rue_house> but no alarm, use your own comparitor
[09:20:06] <rue_house> amee2k, maybe you want a lm75 and a dac?
[09:21:42] <amee2k> well, then i can just do the comparison in an MCU, eh?
[09:22:06] <rue_house> if you use a digital with the alarm
[09:22:07] <amee2k> err, wait
[09:22:16] <rue_house> then use the dac to get your reading
[09:22:41] <amee2k> are there SPI to I2C bridges? >_<
[09:22:52] <rue_house> use an spi dac
[09:23:38] <rue_house> er i2c
[09:23:39] <amee2k> mmh, this is getting screwed up >_<
[09:23:50] <rue_house> dont know what your goal is :)
[09:24:17] <rue_house> if its anything like my boiler, I ahve a good ole hard switch bolted to it, overrides the computer
[09:24:29] <amee2k> finding out if there is a cheap temperature sensor IC with both overtemperature warning output and analog output
[09:24:57] <rue_house> no
[09:25:03] <amee2k> apparently there isn't so i'll have to go for plan B anyway
[09:25:18] <rue_house> and the TCN75 is cheaper
[09:27:17] <amee2k> MCP9700T/AT comes to mind
[10:25:30] <landonf> Has atmel basically abandoned 6lowpan?
[10:25:41] <landonf> (Does *anyone* use 6lowpan?)
[10:30:47] <karlp> abandoned in favour of something else?
[10:31:10] <karlp> wirelesshart? rf4ce?
[10:31:23] <karlp> contiki is all 6lowpan isn't it?
[10:38:58] <amee2k> zigpee
[10:40:27] <landonf> what he said
[10:40:54] <landonf> The bitcloud SDK is allll about zigbee, references to 6lowpan seem to be marked 'mature' and haven't received updates in years.
[10:44:48] <landonf> It looks like some people are still pushing 6lowpan forward, eg,
http://tools.ietf.org/html/rfc6282
[10:47:34] <karlp> stm32w promo material still talks about it..
[10:49:46] <landonf> It kind of looks like z-wave is winning for actual consumer home automation products, zigbee is winning with smart meters / utilities, and 6lowpan is the tertiary contender for anyone that doesn't want to deal with the z-wave or zigbee alliances' proprietary mesh stuff.
[10:50:08] <landonf> Getting my head around the options. Can't help but wonder if straight wifi will just leap frog them all at some point.
[10:50:57] <landonf> Looks like stm32w's referencing Contiki
[10:51:08] <mrfrenzy> I prefer x10/plc-based stuff
[10:51:53] <Essobi> didn't zig has some pretty big security drawbacks?
[10:52:23] <karlp> we've been seeing a bit of zwave europe, zigbee north america
[10:52:59] <karlp> (for power utilities that is)
[10:53:50] <karlp> if they can make the wireless chips cheap enough maybe.
[10:54:40] <karlp> there's been some good work in making 802.11 run onbatteries, but it's still lots more current than 802.15.4 in what I've seen at least.
[10:54:55] <heymaster> Hello. Why there's no avr chips with usb support in dip package ?
[10:55:20] <karlp> because the rest of the world's using SMD now :)
[10:55:46] <heymaster> karlp: I will need to learn how to solder SMD :)
[10:56:05] <karlp> it's much the same, except they don't stay in one place as easily when you first put them down.
[10:56:28] <heymaster> karlp: i need to buy adapter which converts SMD to DIP ?
[10:56:34] <mrfrenzy> it's very easy, I always recommend starting with this:
http://store.curiousinventor.com/guides/Surface_Mount_Soldering/101/
[10:56:46] <mrfrenzy> he recommends good tools (that can ofcourse be bought elsewhere)
[10:58:10] <heymaster> i still using breadboard so will need to buy adapter and they a not so cheap.
[10:58:39] <mrfrenzy> which package are you using, and in what country are you? there are plenty of cheap adapters
[11:00:13] <heymaster> mrfrenzy: except ebay. i thinking to buy this adapter
http://www.ebay.com/itm/QFP32-44-64-0-8mm-pitch-dip-2-54-pcb-adapter-Board-/320720084309?pt=LH_DefaultDomain_0&hash=item4aac681d55#ht_2449wt_777
[11:01:10] <heymaster> mrfrenzy: i living in Lithuania. But i can buy from this page:
http://www.farnell.com/
[11:01:49] <mrfrenzy> farnells adapters are expensive
[11:01:53] <mrfrenzy> your best bet is ebay or sparkfun
[11:01:58] <heymaster> mrfrenzy: yes :)
[11:02:09] <mrfrenzy> that was a good price with free shipping
[11:05:03] <karlp> dangerous prototypes sells them too right?
[11:05:12] <karlp> for QFP at least
[11:05:19] <karlp> http://www.seeedstudio.com/depot/index.php?main_page=advanced_search_result&search_in_description=0&keyword=xQFP+breakout&x=20&y=13
[11:05:40] <karlp> no drill holes though...
[11:09:05] <powderhound> Anyone dealt with XMega USB before?
[11:21:04] <landonf> karlp:
http://www.airstash.com/ has been doing interesting stuff with wifi + batteries, new version will run ~7 hours (actively streaming), apparently.
[11:25:58] <karlp> pretty different to 2xAA for 3 years though...
[11:26:29] <karlp> looks like a cool idea though
[11:27:03] <karlp> kinda an inverted eyefi
[11:28:26] <karlp> and a lot more useful
[12:37:21] <RikusW> uncapped about 90% of my m32l :0
[12:37:24] <RikusW> :)
[12:37:54] <RikusW> only the die is still capped under about 0.6mm black stuff
[12:40:46] <RikusW> its incredibly hard.... a knife barely makes a scratch.....
[12:45:40] <scuzzy> how did you uncap it?
[12:45:56] <Steffanx> lol RikusW
[12:46:00] <Steffanx> You tried it :)
[12:47:18] <RikusW> angle grinder ;)
[12:47:29] <RikusW> first I tried acid
[12:47:32] <RikusW> and heatgun
[12:47:43] <RikusW> sulphuric seems to be faster
[12:48:11] <RikusW> it starts cooking then you immediately scrape it with a blade, and a layer comes off
[12:48:36] <RikusW> decided to try an angle grinder on the sides, much faster
[12:48:49] <RikusW> that epoxy has glass in it... quite hard....
[12:49:10] <RikusW> will try the last part above the die with acid...
[12:49:55] <Steffanx> Don't remove the skin from your fingers
[12:50:23] <RikusW> NaOH will do that ;)
[12:58:32] <RikusW> here is what it looks like now:
http://imageshack.us/photo/my-images/812/m32l.jpg/
[12:58:48] <RikusW> android camera's focus is screwed up a bit :(
[12:59:06] <Steffanx> No better camera?
[12:59:33] <RikusW> maybe I should get a better camera app...
[12:59:36] <RikusW> it might help
[13:04:12] <keenerd> RikusW: Do you have a loupe? Any strong magnifier lets you take good macro shots with cheap cameras.
[13:05:12] <RikusW> ah, do have one
[13:05:53] <RikusW> http://imageshack.us/photo/my-images/88/uveprom.jpg/
[13:06:13] <RikusW> this one was taken at 160x I think
[13:06:24] <RikusW> or 100x
[13:06:46] <Steffanx> or 200x
[13:06:53] <Steffanx> or ….
[13:07:13] <RikusW> not sure if I have the 10x or 16x eyepiece in...
[13:07:23] <RikusW> think the bottom one was at 10
[13:07:28] <RikusW> not 4x
[13:07:47] <RikusW> there is a 40x objective too, but it needs to be like 0.5mm from the object....
[13:08:09] <RikusW> Steffanx: seen a uveprom before ?
[13:08:23] <Steffanx> Yes
[13:08:34] <RikusW> under a microscope ?
[13:09:07] <Steffanx> No
[13:09:27] <RikusW> the bright white part in the center is the rom
[13:09:28] <amee2k> neat
[13:09:42] <RikusW> left is the address decoders and on top is the data output circuits
[13:09:51] <RikusW> 64kbit
[13:10:08] <amee2k> yeah, i was wondering about the triangular structures being some kind of row select
[13:15:30] <RikusW> under the microscope I could see the actual bits in the rom, as little dots in a grid
[13:16:21] <mrfrenzy> could you read the bit state?
[13:16:25] <RikusW> it wass quite difficult to get that shot.... holding the camera still enough, I don't have an eyepiece camera...
[13:16:28] <RikusW> nah
[13:16:37] <RikusW> only see it, only just
[13:16:46] <ziph> An MP-E lens can get close to that.
[13:16:51] <mrfrenzy> okay, so need slightly more expensive microscope ;)
[13:17:01] <ziph> http://www.bitplantation.com/transfer/mpe/
[13:17:06] <RikusW> mrfrenzy: you're welcome to read 64kbit manually ;)
[13:17:18] <RikusW> your eyes will be sore.....
[13:17:39] <ziph> The last one is of an EEPROM.
[13:17:40] <mrfrenzy> ofcourse you need to bring it into pc some way
[13:17:43] <ziph> http://www.bitplantation.com/transfer/mpe/IMG_1934.jpg
[13:17:51] <ziph> (You need to zoom to see detail)
[13:17:52] <mrfrenzy> interesting when you have read protected chips
[13:18:16] <RikusW> mrfrenzy: there is no visual way to see it...
[13:18:45] <ziph> I read that one way to tell was visually finding EEPROM gates that weren't in arrays.
[13:19:08] <RikusW> ziph: aren't you talking about masked rom ?
[13:19:15] <ziph> The protection bits would be one or two EEPROM bits sitting in the middle of nowhere.
[13:19:20] <ziph> RikusW: For which?
[13:19:45] <RikusW> oh
[13:20:11] <RikusW> you are talking about lockbits now
[13:20:28] <ziph> This paper was talking about read protection on MCU's like AVR's and PIC's.
[13:20:29] <RikusW> some lockbits are covered by metal now...
[13:20:34] <ziph> Yeah.
[13:20:45] <mrfrenzy> evil
[13:20:51] <ziph> That might be undone soonish.
[13:21:05] <ziph> It makes it hard for forensics to be done on recovered devices.
[13:21:30] <RikusW> making life easier for hackers again ? :)
[13:21:43] <mrfrenzy> I highly doubt they will make it illegal
[13:21:57] <ziph> They won't make it illegal, they'd just lean on the manufacturers.
[13:22:15] <ziph> There's always secure MCU's around for applications that actually need it.
[13:22:20] <landonf> What kind of recovered devices?
[13:22:31] <ziph> And they go to a lot more effort than just putting a bit of fluff around the lock bits.
[13:22:35] <landonf> (… would said forensics need to be performed on ...)
[13:23:13] <ziph> Small electronic devices used in the commission of large scale international crime and fraud.
[13:23:54] <keenerd> Still waiting for thermite-on-die :-(
[13:24:15] <ziph> Nah, bombs aren't all that interesting that way. ;)
[13:24:55] <ziph> Although they might start getting more interesting if someone decides to make them IMEI/RF proximity triggered.
[13:25:06] <ziph> Then you might want the particular IMEI that a recovered device was looking for.
[13:29:57] <Fleck> capacitors in chain (one after other) voltage increases, capacity? stays the same?
[13:30:20] <ziph> You mean in series?
[13:30:25] <mrfrenzy> it gets easier if you use the terms series and parallell
[13:30:30] <RikusW> decreases capacitance
[13:30:40] <mrfrenzy> and think about how voltage is divided amongst resitors connected in series or parallell
[13:30:56] <ziph> Wait a minute, we weren't finished correcting his nomenclature. ;)
[13:31:07] <RikusW> 1/T = 1/C1 + 1/C2 + 1/C3 .....
[13:31:13] <Fleck> yeah in series, sorry
[13:31:22] <RikusW> for series capacitors
[13:33:23] <Fleck> RikusW whats T ?
[13:33:54] <RikusW> Total
[13:34:10] <RikusW> T = 1(1/C1 + 1/C2 + 1/C3)
[13:48:59] <RikusW> here is my m128 breakout
http://imageshack.us/photo/my-images/196/m128.jpg/
[13:51:54] <keenerd> That is dedication.
[13:52:36] <Fleck> any ideas why switching power supply makes high pitch noise (but still works...)
[13:53:28] <Fleck> i would say - somewhere at 10KHz
[13:53:45] <scuzzy> I have to agree, I don't think I could spend that long soldering all the little wires
[13:53:53] <RikusW> keenerd: more than 6 hours of work in there....
[13:54:12] <RikusW> about an hour for one 2x5 header...
[13:54:17] <scuzzy> RikusW: and 100% crazy
[13:54:30] <RikusW> cutting the wires is the most time consuming part...
[13:54:38] <scuzzy> and stripping the right length off them
[13:54:43] <scuzzy> I hate doing it
[13:55:03] <RikusW> and then cutting off the wire by accident :(
[13:55:12] <RikusW> and making a new piece...
[13:55:14] <scuzzy> Haha, that's why you need a stripper
[13:55:30] <RikusW> that wire wrapping wire is tiny...
[13:55:38] <scuzzy> Ahhh
[13:55:42] <scuzzy> wire wrappers
[13:55:48] <scuzzy> I saw one for the first time this weekend
[13:55:50] <scuzzy> neat idea
[13:55:59] <RikusW> its awg #30
[13:56:09] <scuzzy> use a metric measurement please
[13:56:39] <RikusW> 250um
[13:56:45] <scuzzy> 0.25mm
[13:56:56] <RikusW> yes
[13:57:10] <scuzzy> you got a tool I presume?
[13:57:13] <scuzzy> the wrapping tool?
[13:57:20] <RikusW> no
[13:57:28] <scuzzy> you should get one, it'll save you hours
[13:57:47] <RikusW> I cut off a length the use a stanley blade and roll the wires cutting off the plastic...
[13:58:07] <scuzzy> the tools are cheap
[13:58:13] <amee2k> Fleck: typically some part vibrating in the magnetic field somewhere
[13:58:15] <RikusW> the pinheaders is too short...
[13:58:30] <scuzzy> Ahhh
[13:58:47] <keenerd> Why not print a one-off board?
[13:58:51] <amee2k> Fleck: sometimes it goes away with different load conditions or when the power supply is warmed up
[13:58:58] <RikusW> and routing the wires also takes time...
[13:59:26] <keenerd> I dunno, simple point to point should be autoroutable.
[13:59:26] <amee2k> pouring transformer varnish or hot glue on the magnetics sometimes helps too
[13:59:27] <soul-d> RikusW, woulnd roadrunner work for you
[13:59:33] <soul-d> or hows the stuff called
[13:59:39] <RikusW> and the order in which it is soldered also matters... it might be easier one way than the other...
[14:00:02] <mrfrenzy> a stripax can easily strip those wires RikusW
[14:00:09] <RikusW> think I'll make a pcb next time...
[14:00:09] <soul-d> yeah it's called roadrunner signal wire with enmal you can solder trough
[14:00:19] <mrfrenzy> I have done similar when prototyping small smd chips
[14:00:55] <soul-d> that probably is easyer and quicker
[14:01:03] <RikusW> I do have some thin enamelled wire here...
[14:01:23] <soul-d> http://www.rrunner.co.uk/wire/bobbins.htm
[14:01:23] <Fleck> amee2k ok, well, it turns on slowly too, like it has been weekened somehow... led blinks few times and then comes on, and... when under load - noise actually is louder
[14:01:28] <soul-d> try that otherwise
[14:03:07] <RikusW> roadrunner looks nice, thanks
[14:03:11] <amee2k> maybe worn out filter caps increase ripple and make some filter buzz?
[14:03:29] <Fleck> filter caps are in output?
[14:03:36] <Casper> does anyone know of a good mosfet gate driver?
[14:03:57] <amee2k> yeah. input filters on mains don't see enough current to vibrate usually
[14:03:58] <Casper> tried with some 2n3904/06 with 10 ohms... I get shitty waveform...
[14:04:26] <amee2k> mmh, i have some 1.5A drivers from ID, but don't ask me the type number right now :)
[14:04:54] <Casper> ok, what's the part number then? :D
[14:04:55] <Fleck> amee2k one more thing is - that when power supply is in stand by mode, and i disconnect it from main - noise stays there for about 10 - 20 seconds!
[14:05:05] <amee2k> IR4426
[14:05:28] <amee2k> Fleck: then it has some kind of idle circuit. is this a computer PSU?
[14:05:58] <Casper> which of course the local store do not have
[14:06:07] <Fleck> amee2k no
[14:06:27] <Fleck> wireless 3 channel audio amplifier for home theater
[14:06:40] <amee2k> okay
[14:06:41] <Fleck> sound comes trough speakers too :(
[14:07:28] <amee2k> well, big PSUs are quite inefficient at low loads so PSU modules that have a standby mode typically have a second smaller regulator to cover for that
[14:08:06] <amee2k> this one draws only very little current, especially when unloaded so the input filter caps can keep it alive for a good while after pulling the plug
[14:09:01] <amee2k> if its coming through the speakers then something is funny. check the power rails for noise if you can and i'd point at the output filters
[14:09:25] <Fleck> amee2k ok, ill check those big cups at output
[14:09:32] <Fleck> thx amee2k!! :)
[14:09:41] <amee2k> you're welcome :)
[14:10:03] <Fleck> i dont have replacement for them thou :D
[14:10:12] <Fleck> 1000cF, 50V :D
[14:10:29] <Fleck> but can check if they are good
[14:10:30] <jakllsch> centifarad?
[14:10:38] <Fleck> no :D
[14:10:39] <Fleck> mikro
[14:10:41] <RikusW> Fleck: it must be low impedance caps...
[14:10:51] <RikusW> low esr...
[14:10:55] <Fleck> RikusW ok, so?
[14:11:00] <amee2k> yeah, being cheap on PSU filter caps can come back and haunt you
[14:11:13] <RikusW> Fleck: any cap wont work there...
[14:11:20] <RikusW> it needs to be low esr
[14:12:07] <RikusW> low esr ones are a bit more expensive...
[14:12:51] <RikusW> I got 100x 1000uF 25V for ZAR2 each, would be about 0.20 Euro
[14:13:04] <RikusW> low esr
[14:13:20] <amee2k> nice shot :)
[14:13:44] <RikusW> I repair a lot of pc smps's
[14:14:05] <RikusW> probably already used 15 or so
[14:14:05] <amee2k> aren't 25V ones a bit big for most?
[14:14:35] <amee2k> i usually stock 1000u and 2200u for 10V and 16V because pretty much all PSUs use these
[14:14:41] <RikusW> I prefer 25 to 16V
[14:14:44] <RikusW> last longer...
[14:15:09] <mrfrenzy> may I ask which country you are in?
[14:15:12] <amee2k> most of the time i find bigger ones don't fit because the space is pretty cramped
[14:15:18] <mrfrenzy> I have never repaired a ATX PSU
[14:15:21] <RikusW> mrfrenzy: South Africa
[14:15:26] <mrfrenzy> you can buy a new PSU for 1/2 hour work time
[14:15:48] <RikusW> some are easy to repair...
[14:15:54] <amee2k> i can fix a PSU in 1/2 an hour. if i buy a new one it takes 1/2 a week to arrive
[14:16:32] <mrfrenzy> I certainly miss when that was the situation here, psus, tvs, etc got repaired
[14:16:40] <mrfrenzy> now we just throw them away and buy new stuff
[14:17:00] <mrfrenzy> 90% of all electronics repair shops have closed
[14:17:15] <RikusW> mrfrenzy: USA ?
[14:17:20] <mrfrenzy> Sweden
[14:17:30] <amee2k> most of the time i fix stuff because i like it and want to keep it
[14:17:35] <mrfrenzy> USA have even cheaper hardware than us, but also slightly cheaper work
[14:17:48] <amee2k> or because i don't have the time to wait for snailmail to get a replacement here
[14:17:48] <Fleck> Rubycon caps
[14:18:13] <mrfrenzy> don't you have any local distributors?
[14:18:22] <RikusW> Fleck: the code on my Rubycon caps is YXF and 105C
[14:18:35] <amee2k> store prices here are usually 2x or more the online prices
[14:18:44] <Fleck> YK 85C
[14:18:44] <amee2k> that pretty much ruins any price advantage
[14:18:48] <mrfrenzy> well by local I mean they can deliver goods to the next day
[14:19:04] <mrfrenzy> if I order any computer part today I will get it tomorrow
[14:19:06] <Fleck> but measure s 900uF not 1000
[14:19:14] <amee2k> i usually get 105C panasonics and sometimes ruby or nichis
[14:19:20] <RikusW> Fleck: mine measures 900uF too
[14:19:32] <amee2k> Fleck: thats normal. electrolytics have really shitty tolerances usually
[14:19:44] <Fleck> ok soo its ok?
[14:19:45] <amee2k> something like +20%/-10% isn't uncommon
[14:19:52] <RikusW> Fleck: read some datasheets on it to see what YXF and YK means
[14:20:08] <amee2k> where you can ignore the +20% part because 90% are in the -10% half ;)
[14:20:18] <RikusW> amee2k: I'd prefer it to be +20% ;)
[14:20:39] <amee2k> RikusW: china prefers otherwise
[14:20:47] <RikusW> seems so
[14:21:49] <RikusW> Fleck: unfortunately I can't email you a few caps ;(
[14:22:01] <Fleck> :(
[14:22:18] <amee2k> sometimes i'm actually grateful for electrolytics having shit service life when they're undersized by a factor of 3
[14:22:56] <amee2k> this way china doesn't need to develop time bombs for their equipment that are actually hard to fix
[14:23:27] <RikusW> replacing caps are easy too :)
[14:24:40] <amee2k> after the third set of caps my good old PC PSU blew the switching transistor in the PFC module
[14:25:44] <amee2k> it went 3-way short and blew the controller IC which would cost ~28EUR to import here
[14:26:30] <Fleck> second cap 906uF :D
[14:26:44] <amee2k> Fleck: does your meter show ESR too?
[14:26:53] <amee2k> because thats the important spec for filter use
[14:26:54] <Fleck> i dont think so :(
[14:28:40] <Fleck> is there any way i can measure it?
[14:29:02] <amee2k> well, better capacitance meters show it
[14:29:27] <Fleck> naah, its cheapest
[14:29:28] <Fleck> http://www.aslanllc.biz/index.php?main_page=product_info&cPath=6&products_id=40
[14:29:28] <RikusW> Fleck: have a look at what YK means...
[14:29:33] <amee2k> there are also various homebrew circuits to measure it that you could build. but thats probably only worth it if you need it more than once
[14:29:39] <Fleck> RikusW and?
[14:29:50] <amee2k> when i suspect caps, i usually just replace them unseen
[14:30:15] <Fleck> amee2k i would do that too... but i dont have replacement :D
[14:30:59] <RikusW> Fleck: so the YK ones come from the psu ?
[14:31:19] <Fleck> yes
[14:31:25] <Fleck> ampl psu
[14:35:44] <RikusW> http://www.rubycon.co.jp/en/catalog/e_pdfs/old/aluminum/e_YK.pdf
[14:36:37] <Fleck> RikusW soo? ;)
[14:37:18] <Fleck> i cant see, how this can help me
[14:37:49] <RikusW> http://www.rubycon.co.jp/en/catalog/e_pdfs/aluminum/e_YXF.pdf
[14:38:02] <RikusW> these are low impedance
[14:38:20] <Fleck> So i can get any?
[14:38:32] <Fleck> or better low imp?
[14:40:29] <RikusW> Ask for low impedance when buying
[14:40:48] <amee2k> and take a netbook with you to check the datasheet right on the counter
[14:41:07] <RikusW> postage from me to you will be more than the caps ....
[14:41:09] <amee2k> because my 5$ say the dude on the counter doesn't even know how to spell impedance
[14:41:33] <Fleck> RikusW if only u know how our local store sux and sells cheap stuff :(
[14:41:49] <Fleck> i will ask, and they will look at me with big eyes - whaaat?
[14:41:58] <Fleck> take what we got our LEAVE! :D
[14:42:07] <RikusW> Fleck: ask for rubycon yxf caps then...
[14:42:23] <Fleck> they dont have such caps ;D
[14:42:56] <Fleck> http://web.argus.lv/shop/download/424248/hwindex.htm
[14:43:03] <Fleck> this is what they got!
[14:43:04] <RikusW> ask them to order some ?
[14:43:13] <Fleck> they cant :D
[14:43:22] <amee2k> they can
[14:43:26] <amee2k> doesn't mean they want to ;)
[14:43:30] <Fleck> better will be, if i buy them at ebay myself ;D
[14:46:17] <Fleck> one is Rubycon ZL 220uF 50V
[14:46:18] <amee2k> what country btw?
[14:46:25] <Fleck> 105C
[14:46:29] <Fleck> Latvia
[14:47:19] <amee2k> hmm if you have a student ID or a company name you can use, you might want to take a look at Farnell
[14:47:44] <amee2k> they're kind of the big shorts for parts in the EU or something
[14:48:19] <RikusW> my one local supplier get stuff from farnell too
[14:48:38] <RikusW> sort of a local rep for them
[14:48:48] <amee2k> theres digikey too, but they're mostly a US thing so no way to get shit without a credit card as it seems
[14:49:58] <soul-d> mmm here in netherlands we can still use farnell as normal consumers although not online only trough mail , no such thing for other area's ?
[14:52:23] <soul-d> k seems that link is only in the dutch shopping basket
[14:52:28] <amee2k> registering a company here is like 50EUR. when i'm done studying and they complain, i'll just do that
[14:53:52] <soul-d> that one time only yearly ?
[14:54:05] <amee2k> one time only
[14:54:32] <amee2k> since i'm not making any actual money with that letter box company, i don't think i have to pay anything anually
[14:54:41] <amee2k> and even if its not much
[14:54:53] <RikusW> Fleck: ZL is fine too, high ripple + low imp
[14:55:09] <RikusW> http://www.rubycon.co.jp/en/catalog/e_pdfs/e_SERIES-CHART.pdf
[14:55:14] <RikusW> very nice chart
[14:55:56] <Fleck> those seems to be from low power part...
[15:11:10] <Fleck> RikusW, amee2k now when i replaced all caps i had with cheap ones, it doesn't power on at all, but, when i blow hot air on caps... it turns on after a while :D
[15:12:58] <RikusW> weird
[15:13:59] <Fleck> is it? my dad showed me this trick :D
[15:14:18] <Fleck> when something "tries" to turn on but cant, try hot air on caps
[15:14:30] <Fleck> then probably u will find one that is dead
[15:16:30] <RikusW> or look for a leaky or swollen one
[15:16:32] <sabesto> meh, my AVR Dragon is acting up, keeps connecting/disconnecting when i try to use it, anyone know about this?
[15:17:43] <RikusW> is the target powered ?
[15:18:24] <sabesto> tried with/without target
[15:18:37] <soul-d> faulty usb cable maybe ?
[15:18:41] <RikusW> sabesto: ISP or JTAG ?
[15:18:42] <sabesto> it was working 2 min ago, nothing has changed
[15:18:46] <sabesto> ISP
[15:19:26] <sabesto> will try different USB cable, have tried different port
[15:21:05] <sabesto> same same
[15:23:34] <sabesto> it has always been wonky, when i close the programming tool in AS5 it disconnects and reconnects, cutting power on the VCC pin in the same go
[15:24:31] <RikusW> sabesto: thats normal
[15:24:35] <RikusW> it does that
[15:24:42] <willie_w> someone who has created an I2C slave on a attiny?
[15:26:30] <sabesto> god damned this thing
[15:28:28] <Fleck> RikusW soldered back two ZL caps, works again in the same way - led blinks few times and powers on w/o heating
[15:29:37] <RikusW> Fleck: don't they look swollen at all ?
[15:29:53] <Fleck> nope
[15:30:00] <Fleck> all look just fine
[15:31:14] <RikusW> what kind of supply is it ?
[15:31:21] <Fleck> ?
[15:31:37] <RikusW> pc , printer ? etc ?
[15:31:58] <Fleck> [21:57:56] <Fleck> wireless 3 channel audio amplifier for home theater
[15:32:14] <RikusW> volts amps output ?
[15:34:24] <Fleck> RikusW 60W from mains...
[15:36:11] <amee2k> Fleck: i've never heard of that trick before, and i can't think of any reason why it would help off the top of my head
[15:36:43] <Fleck> dead caps somehow work better when heated
[15:37:20] <amee2k> maybe dried out electrolytic goop starts to work again when it gets warm
[15:37:24] <sabesto> are the ext_power pins on the dragon protected in some way?
[15:37:43] <amee2k> because heat is exactly what electrolytics normally don't like, and what kills them in the end
[15:38:17] <RikusW> sabesto: its connected via a switching regulator...
[15:39:02] <RikusW> I did connect vcc and gnd the wrong way to an avr once.... my dragons and that avr still works...
[15:39:27] <RikusW> the dragon just turned off.
[15:39:36] <sabesto> the firmware update tool says the firmware on the dragon is "unknown"
[15:40:03] <RikusW> have you tried updating it ?
[15:40:22] <sabesto> yeah
[15:40:34] <RikusW> didn't help ?
[15:40:53] <sabesto> i get that disconnected/connected sound all the time
[15:41:12] <sabesto> now it might have worked, at least it says up to date
[15:41:34] <sabesto> yep, now its back
[15:41:42] <sabesto> but there is something wonky about it
[15:41:50] <sabesto> it has been since i bought it
[15:41:54] <RikusW> like ?
[15:42:10] <sabesto> one time it refused to work on 3 different computers
[15:42:17] <sabesto> probably the same thing
[15:42:39] <sabesto> it totally failed to program an STK1000 using jtag
[15:42:52] <sabesto> it succeded in deleting the bootloader though
[15:44:08] <RikusW> have you tried with the new fw ?
[15:44:19] <sabesto> it worked now
[15:44:30] <sabesto> i update nearly every day i use it
[15:44:37] <RikusW> why ?
[15:44:59] <RikusW> using AS4 and AS5 ?
[15:45:22] <sabesto> AS5, i mean, i update every time its outdated
[15:45:26] <sabesto> 22:30:25.024: [ERROR] Firmware on AVR Dragon (00A200021358): upgrade required
[15:45:40] <sabesto> cant do anything without updating either so...
[15:46:09] <RikusW> and it asks that every time ?
[15:46:26] <sabesto> no
[15:47:05] <sabesto> some times it asks that every time i connect it to a new computer (on the same day)
[15:47:16] <RikusW> I patched AS4 so it don't complain about downgrading ;)
[15:47:17] <sabesto> happened at least twice
[15:47:31] <RikusW> http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&t=110877 think this is the right link...
[15:47:45] <RikusW> and it actually works too
[15:48:25] <sabesto> A4/AS5 requiring different firmware?
[15:48:32] <RikusW> yes
[15:48:33] <sabesto> *AS4/AS5
[15:48:42] <RikusW> but the AS5 fw works in AS4
[15:48:45] <sabesto> ok, i'm only using 5
[15:48:47] <RikusW> with my patch....
[15:48:51] <sabesto> cool
[15:49:36] <sabesto> did have to patch some files to get atmega8 to work
[15:49:58] <sabesto> atmel being lazy
[15:49:58] <RikusW> yeah, those xml stuff is incomplete...
[15:50:25] <RikusW> sabesto: using linux too ?
[15:50:32] <mrfrenzy> do you submit your patches so they get included in the next update?
[15:50:56] <sabesto> no, windows (mainly because of AS5, multisim)
[15:51:09] <RikusW> mrfrenzy: don't think atmel will like my hack ;)
[15:51:13] <sabesto> but i have linux servers and osx laptop
[15:51:18] <Fleck> RikusW, amee2k
http://fleck.rullz.lv/psu/IMG_20111206_220002.jpg
[15:51:37] <mrfrenzy> I always try to make my hacks good enough quality so I don't feel ashamed of them ;)
[15:52:04] <Fleck> there is a cooler at the left side upper corner, it turns on for a fraction of second - when this happens - noise gets pitched up then down :D
[15:52:05] <RikusW> mrfrenzy: I totally ignored fw versions with my hack...
[15:52:39] <mrfrenzy> aah, I was mostly thinking about the fix to include support for other atmegas
[15:52:42] <RikusW> mrfrenzy: and its probably against atmel policy, although it seems to work so far
[15:53:34] <RikusW> I added support for some stk500 parts to AS5, don't think atmel will have a problem with that....
[15:54:06] <mrfrenzy> why don't you try and submit it?
[15:54:27] <RikusW> atmel should actually have provided those xml files....
[15:54:44] <RikusW> and its very easy to make one to support a part...
[15:59:37] <Fleck> RikusW, amee2k with magnet i was able to tell - the noise makes small transformer :D
[16:00:44] <amee2k> hehe
[16:01:10] <Fleck> ;)
[16:01:29] <Fleck> amee2k and i think thats the weak part of supply
[16:01:35] <Fleck> for standby etc
[16:01:49] <amee2k> would make sense :)
[16:02:17] <Fleck> i can even turn it off with magnet, i guess thats not a good thing to do
[16:02:30] <Fleck> but i can make the noise almoust go away with magnet
[16:02:53] <amee2k> hmm a magnet...
[16:03:17] <amee2k> does it run normally with the magnet too?
[16:03:44] <Fleck> at some angle/point it turns off (led goes dark...)
[16:03:49] <Fleck> led goes off...
[16:03:55] <amee2k> i have a hunch the extra flux saturates the core and the circuit goes into overload shutdown or something
[16:04:56] <Fleck> yeah
[16:05:02] <Fleck> but why it does this?
[16:05:16] <Fleck> and why it doesn't power ON normaly?
[16:06:21] <Fleck> dead transformer?
[16:06:35] <Fleck> then why my caps made it work even worse
[16:07:04] <amee2k> i don't know why the caps thing went wrong but i'd say the transformer is fine, and the magnet you put on it just saturates the core
[16:07:58] <amee2k> that trips some overload protection in the PSU circuit and effectively shuts it down
[16:08:26] <amee2k> at which point the noise will stop too, obviously
[16:08:37] <Fleck> amee2k
http://fleck.rullz.lv/psu/VID_20111206_235651.m4v
[16:09:05] <Fleck> didnt think that mob will record sound so good! :D
[16:09:08] <sabesto> bah, now the dragon is messing around again
[16:10:31] <amee2k> hmm maybe that question sounds dumb, but did you put the caps on the right way around?
[16:10:43] <Fleck> yeah
[16:11:03] <Fleck> as i said - when heated - blue led did come on! :)
[16:11:25] <amee2k> mmh, well i have no idea about the heating thing
[16:11:44] <amee2k> if it doesn't work at all this way, leave the old caps on until you can get better replacements
[16:12:01] <Fleck> i did allready put them back, video is with old caps
[16:12:09] <amee2k> o.O
[16:12:51] <Fleck> new ones made noise, but, led did not even blink, when started to heat - led starts to blink, then goes one after a while, but clearly - works much worse than with old ones
[16:15:00] <amee2k> and then you put the old caps back and now it is refusing to start up too?
[16:16:11] <Fleck> well it was behaving like this with old cups when i started to look at it
[16:16:24] <Fleck> thats why i tought that some caps are dead
[16:16:35] <amee2k> so its only making the noise now but otherwise works?
[16:17:17] <amee2k> i'd say if it runs at all and there is no visible damage, then the transformer should be fine
[16:17:41] <Fleck> it runs, sound is coming out of it... wireless receiver works, ampls too
[16:17:52] <Fleck> just this noise comes out from speakers too
[16:17:58] <Fleck> makes it kind of useless
[16:18:09] <Fleck> i cant stand that noise for a minute ;D
[16:18:20] <amee2k> transformers are almost impossible to kill by accident... they need huge sustained overload before they go
[16:19:11] <Fleck> yeah well, it was staying in standby mode for years...
[16:19:16] <amee2k> as for the noise coming from the PSU, you could try putting a rag over the transformers to see if that changes the sound. if that dampens it, then its something in the magnetic field that started vibrating
[16:19:21] <Fleck> then i one day noticed that noise and turned it off
[16:20:11] <amee2k> as for the noise coming from the speakers, do you have access to a scope?
[16:20:55] <amee2k> i'd point at power supply ripple, but depending on the circuit it could get picked up anywhere
[16:21:14] <Fleck> i dont have scope at the moment
[16:21:26] <Fleck> but i am thinking of getting one
[16:22:25] <mrfrenzy> you can start by measuring AC and DC in various points in the psu
[16:22:36] <mrfrenzy> after trafo, after rectifier, after caps etc
[16:22:47] <mrfrenzy> assuming you have a multimeter ofc
[16:23:07] <Fleck> i have multimetter :)
[16:23:25] <mrfrenzy> also, try and not kill yourself from the live parts
[16:23:32] <Fleck> ;))
[16:23:38] <Fleck> thx for suggestion :)
[16:23:41] <Fleck> ill try! :)
[16:23:41] <amee2k> dmm in Vac mode + series capacitor? would that work for measuring ripple?
[16:23:50] <amee2k> never tried it though :)
[16:46:46] <Fleck> amee2k without load i can still hear that transformer... but its silent
[16:47:07] <Fleck> when in case - i think it would be noticeable
[17:01:40] <mrfrenzy> what was your results on the AC and DC?
[18:04:34] <winferno> quit
