#avr | Logs for 2012-02-22

Back
[00:50:54] <pingec> "PORTC bit 5 = 1 & DDRC bit 5 = 1" vs. "PORTC bit 5 = 1 & DDRC bit 5 = 0" Is this the same from the point of view of a DS18B20? It should detect a high signal in both scenarios, or not?
[01:09:31] <wollw> pingec: DDR registers determine if the pin is being used for input or output, PORT registers are used to set the pin high or low
[01:10:15] <pingec> so it's the same for the 1wire slave?
[01:11:28] <wollw> I haven't tried it myself but if the DDR bit is set and PORT bit isn't is defined as "output low"
[01:11:30] <Kevin`> pingec: you may need a stronger bus pullup than the one built into the io pin
[01:12:50] <Kevin`> pingec: watch it with a scope if you aren't sure
[01:13:18] <pingec> I'd need to borrow a scope
[01:13:21] <Casper> pingec: if the DDR is 0, then the pin is in input mode. writting to PORT will enable or disable the very weak internal pullup (20-60k, depend on the chip)
[01:13:31] <Casper> 1 wire ask for around 5k pullup
[01:13:52] <pingec> I see so it would not work?
[01:14:02] <Casper> so no, it won't be the same. it might work, it probably won't work. if it work it will surelly be unreliable
[01:14:21] <wollw> pingec: there should be a nice table enumerating the states in the datasheet for the mcu in the IO section
[01:14:33] <pingec> Yes I get that
[01:14:43] <pingec> it's jsut that i've found 2 sample codes
[01:15:05] <pingec> one uses first and the other the second way
[01:15:15] <pingec> But neither work for me :P
[01:15:37] <pingec> I am able to detect presence signal
[01:16:20] <pingec> Now I'm trying to get the match rom command to work
[01:16:32] <pingec> I guess I need a scope
[01:18:12] <Kevin`> you aren't supposed to drive the output high
[01:24:28] <pingec> So do I need an external pullup resistor then?
[01:26:42] <Kevin`> yes
[01:35:47] <RokLobsta> yes you need a scope
[01:35:53] <RokLobsta> so much less hassle
[02:17:15] <pingec> So to setup an external pullup I need to add a 4.7k resistor to connect the chosen data pin to VCC and set this pin to tristate mode?
[02:18:08] <Kevin`> you'll toggle it between tristate and low output
[02:25:46] <Casper> and for some command you will need to high output if your devices are bus powered as some ask more power than what the pullup can give
[02:28:58] <pingec> Hmm
[02:29:13] <pingec> Casper what about <Kevin`> you aren't supposed to drive the output high
[02:29:37] <pingec> It's not bus powered but I'm confused
[02:33:11] <Casper> no you ain't supposed to drive the output high during data communication
[02:33:12] <Casper> but
[02:34:02] <Casper> after some commands (mainly eeprom write) you might need to enable the output high to give more power IF and only IF your stuff is powered by the data line
[02:34:26] <Casper> because those ds1820 use so little power that the VCC pin on them is optional
[02:34:41] <Casper> you can literally have only 2 wires: ground and data
[02:34:43] <Casper> and that's it
[02:35:44] <pingec> ok I see
[02:37:31] <pingec> So if I got it right (focusing on the external pullup scenario with power over VCC pin): that means I only need to initialize the bit in PORTC to 0 and then I just toggle the bit in DDRC when signaling 1/0 and for reading I set DDRC bit to 0 and read PINCx bit ? So I don't really need to touch PORTC after init...
[02:39:37] <ziph> You might want to check timing too to make sure you aren't reading while the edge is still rising.
[02:43:58] <pingec> Yes I will thanks.
[03:31:17] <Kevin`> <- connected to the internet with my usb-fsk bridge :)
[03:32:20] <Kevin`> it actually works reasonably well once some bugs are squashed;p
[03:34:51] <Kevin`> well, except for being slower than dialup of course
[03:50:20] <RokLobsta> ew
[03:50:26] <Kevin`> hmm, it should be able to run a bit faster than that, probably still some bugs left
[03:50:35] <RokLobsta> don't take me back to the early 1990's.
[03:51:00] <Kevin`> it's intended for long range use and telemetry
[03:51:16] <RokLobsta> so you were saying
[03:51:29] <RokLobsta> how robust is your data protocol?
[03:51:43] <RokLobsta> or are you relying on IP to deal with it?
[03:51:52] <Kevin`> fairly robust
[03:52:02] <Kevin`> (aside from the bugs)
[03:52:08] <RokLobsta> what freq?
[03:52:22] <Kevin`> 900mhz ism
[03:52:33] <RokLobsta> ah the bugs, it works all the time except when it doesn't.
[03:52:45] <Kevin`> indeed, it works except when it doesn't :D
[03:52:52] <RokLobsta> any reason you didn't use one of nordic or chipcons radios?
[03:53:42] <RokLobsta> they have some nifty stuff like AES and high bit reates at 900Mhz.
[03:53:44] <Kevin`> I didn't see them having any high power devices, do they?
[03:53:53] <RokLobsta> hrm, i think they
[03:53:57] <RokLobsta> what power do you need?
[03:53:58] <Kevin`> higher bandwidth means lower range
[03:54:16] <RokLobsta> of course, the lower bit rate the more range you get for a given power
[03:54:27] <Kevin`> 500mw
[03:54:47] <RokLobsta> i am sure they had appnotes for strapping an LNA on.
[03:55:04] <RokLobsta> and RF booster.
[03:55:12] <Kevin`> I really don't wanna do my own rf design
[03:55:25] <RokLobsta> no, you don't. how good are you with smith charts
[03:55:36] <Kevin`> well I know what it is.
[03:55:39] <Kevin`> that's about it
[03:55:49] <RokLobsta> me too, we did them in one lecture at uni...
[03:56:05] <RokLobsta> no rf practise since uni. damnit.
[03:56:08] <Kevin`> I don't have any fancy equipment for stuff like that though
[03:56:39] <RokLobsta> what model rxtx did you use?
[03:57:04] * RokLobsta ponders.
[03:57:06] <Kevin`> i'm testing with rfm22b right now
[03:57:19] <Kevin`> didn't even turn the PA on those on, since they are right next to each other
[03:57:54] <RokLobsta> http://www.sparkfun.com/products/10153 ?
[03:58:21] <Kevin`> yes, like that, only one of the 915mhz tuned ones, since 433mhz is less open in this country
[03:58:31] <RokLobsta> right
[03:58:38] <RokLobsta> so are you using fm or GMSK?
[03:59:00] <Kevin`> gfsk is fm too, it's just filtered
[03:59:01] <Kevin`> also, gfsk
[03:59:33] <Kevin`> they don't do msk
[04:01:02] <RokLobsta> fsm
[04:01:21] <Kevin`> well, actually, the PA is enabled, just at 1.2mw
[04:01:55] <Kevin`> flying spaghetti monster?
[04:05:38] <RokLobsta> fsk i mean. interesting none of the chips use multitone as used in the military's LINK11 standard.
[04:06:00] <RokLobsta> LINK11 is sort of like a very narrowband OFDM
[04:06:23] <RokLobsta> i guess putting an FFT into them wouldn't make them so cheap.
[04:06:44] <RokLobsta> what sort of telemetry?
[04:07:43] <Kevin`> control to and data from one of those little flying machines, when I build it
[04:07:57] <Kevin`> that part probably won't use ip/ethernet unless i get it working really well, just the first layer
[04:09:20] <RokLobsta> here's what the military uses. LINK22. http://www.is.northropgrumman.com/products/data_link_processing_and_manag/assets/NG_278-A007-Link22_Guidebook.pdf
[04:09:27] <RokLobsta> sort of like Zigbee for battle.
[04:14:26] <Kevin`> these are actually tunable to the uhf frequencies used there, although there wouldn't be much point in doing that
[04:15:38] <Kevin`> I was thinking of using tdma, but I probably won't have very many nodes, not sure if it will ever be worth it
[04:15:50] <Kevin`> although you can cut power consumption a lot by using it
[04:18:30] <Kevin`> (right now it uses basic csma)
[04:21:44] <pentths> hi all!
[04:22:21] <pentths> 155th anniversary of Hertz today.... thank you google.
[04:23:09] <Kevin`> I wonder if the graphic is saying anything
[04:27:02] <RokLobsta> "Drink your Milo".
[04:27:25] <ziph> Wouldn't it be Ovaltine? :)
[04:39:11] <Valen> mix them both with a dose of quick
[04:40:58] <amee2k> huge what the fuck at google's logo today?
[04:41:14] <RokLobsta> no we do Milo here in .au
[04:41:16] <amee2k> i only see a wavy line
[04:41:56] <pentths> 11:13 < pentths> 155th anniversary of Hertz today.... thank you google.
[04:42:09] <ziph> RokLobsta: We're also not meant to know about Orphan Annie or Jean Shepherd in Australia.
[04:42:18] <amee2k> um.
[04:42:55] <amee2k> the rightmost wave looks mildly phallic...
[04:43:17] <amee2k> >_>
[04:43:59] <RokLobsta> has anyone here ever used an accelerometer from freescale before?
[04:44:24] <RokLobsta> i don't know who jean shepherd is.
[04:44:52] <ziph> Why did you say "Drink your Milo" in reference to a coded message?
[04:45:28] <RokLobsta> sort of slashdot meme referring to some code at cia HQ
[04:46:09] <ziph> The meme would've come from "Drink your Ovaltine" off Orphan Annie.
[04:46:35] <RokLobsta> Ohhh well the circle is complete. I saw that movie once as a kid so i don't rem that.
[04:46:43] <ziph> Which is referenced in a Jean Shepherd story.
[04:46:48] <ziph> Yeah, which was made into that movie.
[04:46:56] <ziph> Well, a bunch of stories of his were made into that movie.
[04:47:00] <ziph> It's very well known in the states.
[04:47:17] <RokLobsta> anyway, if I had my way it'd be "Drink your Aldi branded pils beer"
[04:47:55] <amee2k> yuck, beer >_<
[04:48:10] <RokLobsta> you sound like Mrs Roklobsta
[04:48:35] <RokLobsta> accelerometers anyone?
[04:49:01] <ziph> Try a question, we're fairly clever even if we haven't used a specific device.
[04:49:47] <amee2k> or ask something more common, like if anyone is using 2N2222s
[04:49:58] <RokLobsta> i was wondering if anyone has used a freescale device? I am just trying to get one working here.
[04:50:18] <RokLobsta> tonights AVR coding is powered by Arnotts Ginger Nut biscuits.
[04:50:34] <ziph> Don't do that, you'll get 10 people telling you 2N2222's suck and that you should fish some obscure BJT out of a $3 pink vibrator. ;)
[04:50:35] <amee2k> shameless plug :P
[04:50:51] <amee2k> ziph: LOL
[04:51:03] <RokLobsta> you pay too much for your pink vibrators.
[04:51:15] <amee2k> if enough people are using something, there is always someone who doesn't like it :P
[04:51:30] <amee2k> right. never go over 2.95$
[04:52:03] <RokLobsta> a tenant left a green vibrator in my rental house.
[04:52:12] <RokLobsta> so that was free.
[04:52:29] <amee2k> well, technically i've been using mostly BC337s for a while now anyway
[04:52:48] <RokLobsta> ok, so noone has used an accelerometer before?
[04:53:06] <amee2k> no, the quadcopter crowd is still in #arduino :P
[04:53:18] <ziph> amee2k: Why no FET's? :)
[04:53:33] <RokLobsta> remote contrlling anything like a car of plane or coptor seems didgy using packet radio
[04:53:41] <RokLobsta> my typing sucks
[04:53:52] <amee2k> ziph: only small signal FETs i have here are BS170 and some SMT 7001s
[04:54:40] <ziph> SOT-23 2N7000's are always useful to have around.
[04:54:50] <Kevin`> RokLobsta: actually, at the settings I am using, gfsk=gsmk
[04:54:53] <Kevin`> randomly
[04:55:01] <ziph> I wouldn't exactly call them small signal though. :)
[04:55:04] <amee2k> ziph: well, good luck sticking one of these on a breadboard
[04:55:21] <ziph> They come in TO-92.
[04:55:39] <amee2k> TO-92 counts as small in my book
[04:56:06] * amee2k idly starts throwing dead TO-3 transistors at ziph
[04:57:29] <amee2k> i want some power LEDs in clear top TO-3 packages
[04:57:53] <amee2k> ziph: no, i meant your SOT ones
[04:58:17] <amee2k> also, from what i remember of the datasheet the 170/7000 don't work well with 3.3V anymore
[04:58:24] <ziph> Yeah, I know, but if you want to breadboard you can still get through hole.
[04:59:12] <ziph> Yeah, Vth is fairly high on them.
[05:02:48] <RokLobsta> kevin: what effect will latency and errors have on control of your copters?
[05:03:04] <RokLobsta> will it be like trying to control something on the moon (for latency)
[05:03:34] <amee2k> moon is a quater second one-way, no?
[05:03:45] <amee2k> that sounds awfully long for real time control
[05:04:02] <Kevin`> RokLobsta: there's ecc, so no errors. and I have the option to transmit packets without ACK (default is to retransmit until it gets through) to provide guaranteed latency
[05:04:05] <Kevin`> erm
[05:04:08] <Kevin`> not ecc, crc
[05:04:33] <RokLobsta> should add some FEC or ECC in for control
[05:04:34] <ziph> You can do FEC with CRC.
[05:04:39] <amee2k> WEP had CRC too and claimed it didn't have errors >_>
[05:05:01] <Kevin`> amee2k: I have no 'security' at all, that's different ;p
[05:05:16] <amee2k> hehe
[05:05:25] <ziph> A CRC is just a linear code where all entries in the codebook are cyclic shifts of each other.
[05:05:41] <Kevin`> RokLobsta: do you know of any efficient fec implementations for avr? it would require disabling the hardware packet handling, but it might be worth it in some odd situations
[05:06:12] <RokLobsta> god you are making me stretch back to uni...
[05:06:14] <Kevin`> RokLobsta: what you normally do for this type of control is just transmit the same thing over and over again whether it's received or not
[05:06:46] <Kevin`> (set axis to 0,0.0005,1.5,1.2)
[05:06:46] <Kevin`> (set axis to 0,0.0005,1.5,1.2)
[05:06:47] <Kevin`> (set axis to 0,0.0005,1.5,1.2)
[05:06:48] <Kevin`> (set axis to 0,0.0005,1.5,1.3)
[05:06:48] <Kevin`> (set axis to 0,0.0005,1.5,1.3)
[05:06:51] <Kevin`> (set axis to 0,0.0005,1.5,1.3)
[05:06:53] <Kevin`> etc
[05:07:13] <ziph> amee2k: A low strength binary viterbi decoder or fano convolutional decoder might work.
[05:07:17] <amee2k> so much like old-style remote controls?
[05:07:48] <amee2k> ziph: i think you meant Kevin`, no?
[05:07:59] <ziph> amee2k: Yes.
[05:08:09] <amee2k> :)
[05:08:12] <Kevin`> amee2k: well yeah. if i'm controlling with a set of analog sticks anyway, the control data is going to look like that in any case. once I get to more complex things like "go to gps coordinates x/y/z", latency won't matter anyway
[05:08:27] <RokLobsta> reddsolomon is a classic one. don't know if there is a cheap algorithm
[05:08:41] <ziph> RS is nasty.
[05:09:10] <ziph> Viterbi was one of the first to be used and is simple.
[05:09:13] <Kevin`> amee2k: .. what, exactly, would new-style remote contriols be? ;p
[05:09:13] <amee2k> Kevin`: yeah... at that point an RPC like model would be better for your protocol anyway than state synchronization
[05:09:38] <ziph> Fano uses far more memory but is faster; the difference with Fano is that it's slower the more errors there are.
[05:09:53] <amee2k> something digital and fancy that doesn't work :P
[05:11:44] <RokLobsta> viterbi has a simple state machine.
[05:11:48] <RokLobsta> from memory
[05:13:00] <ziph> The convolutional code has a state machine.
[05:13:09] <ziph> Viterbi uses a trellis.
[05:20:53] <RokLobsta> kevin: i reckon you should eventually look at soe FEC rather than just reply on retransmits
[05:21:23] <RokLobsta> the trick is to find something effective but quick to run on an AVR
[05:21:48] <Kevin`> if you (or anyone here) know of some magic code for it, point me to it :)
[05:24:14] <ziph> FEC will only help bring down the uncorrected error rate for a particular power level.
[05:24:41] <ziph> You might still get a channel that is completely wiped out.
[05:24:42] <Kevin`> hm, seems like dropping the data rate down to 56/112kbps improved reliability quite a bit somehow. I had it at 115/230kbps before because of a logical mistake (effective data rate/raw data rate, from manchester)
[05:24:58] <Kevin`> some of the other modules I have only go up to 115kbps
[05:25:31] <ziph> Reducing baud rates also decrease error rates.
[05:25:34] <Kevin`> ziph: yeah, it would be nice. know of some code for it? i'm not so good with the heavy math stuff
[05:26:13] <ziph> Phil Karn had some about.
[05:26:26] <ziph> Optimised for x86 though.
[05:26:45] <Kevin`> heh, x86 kind of has infinite processing power for this sort of thing
[05:27:05] <ziph> x86 in the dark ages.
[05:27:26] <ziph> It's hard to work with it without having an understanding though, the decoders tend to be tuned to particular parameters.
[05:29:57] <RokLobsta> bleh, my brief undergrad lectures on coding were next to useless
[05:30:00] <Kevin`> 2012-02-22 05:19:58 (2.03 KB/s) - “/dev/null” saved [1048576/1048576] - like 1995 all over again :D
[05:31:40] <Kevin`> no more stalls now though, perfectly consistant
[05:32:03] <Kevin`> once it gets warmer i'll be able to use irc and such from the park down the road :)
[05:32:31] <Kevin`> (or across town, if I set up a tower)
[05:33:20] <RokLobsta> that's so... Ricochet.
[05:33:31] <Kevin`> anyone want the code? it's pretty nasty looking right now
[05:33:55] <RokLobsta> will we still respect you in the morning?
[05:34:05] <Kevin`> well it DOES work.
[05:35:32] <RokLobsta> you'll look odd with your laptop and a big yagi down in the park
[05:36:23] <Kevin`> with a big yagi I could probably make that link with 100mbit wifi
[05:37:05] <ziph> Are you on 433MHz?
[05:37:14] <Kevin`> http://kwzs.be/~kevin/usb-rfm.txt
[05:37:19] <Kevin`> 915mhz, since this is US
[05:37:25] <ziph> Ah.
[05:37:31] <ziph> We get both 433 and 900.
[05:37:50] <ziph> I personally like 433 better because I can use 200W on it. ;)
[05:38:05] <Kevin`> i'd need to go take a test or something easy to be able to do that
[05:38:11] <Kevin`> :/
[05:39:31] <RokLobsta> ziph: you got a license?
[05:39:47] <RokLobsta> vk?
[05:40:07] <ziph> -.-- . ...
[05:41:28] <Kevin`> mm, I forgot to have the bridge table entries expire
[05:41:50] <Kevin`> (for mac address - rf node mapping)
[05:41:57] <RokLobsta> .. ... / .. - / .- / ..-. ..- .-.. .-.. / --- .-. / .-.. .. -- .. - . -.. / .-.. .. -.-. . -. ... . ..--..
[05:42:12] <Kevin`> you know, that's not very efficient since you have a keyboard
[05:43:03] <ziph> And since to read it you have to make the noises in your head.
[05:43:22] <RokLobsta> morse coders beat out speed texters on letterman once.
[05:43:42] <Kevin`> that's because texting is stupid
[05:43:49] <RokLobsta> it is
[05:43:51] <Kevin`> try beating someone with a laptop
[05:44:47] <Kevin`> it would be interesting to try having the module do CW encode/decode. it should be able to
[05:44:59] <Kevin`> complete waste of the hardware, but it would be fun :)
[05:45:20] <ziph> Rok: Is that what they still call the licences?
[05:45:47] <RokLobsta> .. / .... --- -. . ... - .-.. -.-- / -.. --- -. .----. - / -.- -. --- .-- .-.-.-
[05:46:06] <ziph> I'm not reading that one unless you can make it make the sound instead. :)
[05:46:41] <RokLobsta> dit dit dit dit dit dit dah dah oh bugger it.
[05:46:49] <ziph> Heh.
[05:47:20] <RokLobsta> ... -- ... Good ol' nokia
[05:47:33] <ziph> I can't stand nokia's.
[05:47:54] <ziph> You hear morse wafting in from a distance and perk your ears up and then realise it's just "connecting people".
[05:48:02] <RokLobsta> i had a nokie 1610 once that just refused to work on the optus network properly.
[05:49:06] <RokLobsta> now why won't this f'ing accelerometer calibrate?
[05:49:51] <ziph> Because you're in Australia.
[05:49:56] <ziph> It's confused that it is upside down.
[05:50:10] <RokLobsta> oh yeah
[05:50:12] <RokLobsta> duh
[05:51:00] <RokLobsta> at least the TMP75 is wokring. 23.5C in here.
[05:52:05] <RokLobsta> re: hemispheres... it reminds me of the old Sun monitors that had hemipshere designations
[05:52:33] <RokLobsta> probably beacuse of the coriolis effect.
[05:55:05] <ziph> It was because of the different magnetic field.
[05:55:19] <RokLobsta> i was kidding
[06:46:34] <hetii> Hell :)
[06:48:30] <hetii> Q: I want to build sms notification box, maybe someone know with you, is it possible to change sim card withour rebooting the phone ?
[06:56:24] <RikusW> seems unlikely
[06:56:40] <RikusW> I think its better to power it off entirely.
[07:01:40] <RokLobsta> do you have a phone or a module?
[07:03:16] <hetii> i have different old phones
[07:04:15] <hetii> btw is there some avr library for communication with phones?
[07:13:28] <Tom_itx> hetii, what is your purpose for this?
[07:13:48] <hetii> support different caller id
[07:20:25] <Kevin`> hetii: send from a trunk line and you can set random caller id. why do you need to though?
[07:22:50] <specing> Is it normal for an ADC channel to affect a nearby one by ~1V?
[07:23:32] <Kevin`> no
[07:24:08] <Kevin`> maybe you aren't providing enough input power or time fot the cap charge to change
[07:26:18] <specing> Hmm
[07:27:52] <specing> I have the reference set to internal 2.56V and the setup as folows: GND<-->1Mohm<-->ADC0<-->1Mohm<-->input
[07:27:56] <specing> The same for ADC1
[07:29:08] <hetii> Kevin`: what you mean by trunk line ?
[07:31:07] <hetii> i guess when i use phone i`m not able change caller id
[07:31:30] <hetii> its not a VoIP, where i can set such argument
[07:50:34] <RokLobsta> specing: are you allowing enough settling time between ADC mux changes?
[07:51:22] <Kevin`> specing: 1mohm is way too high to use without a buffer
[07:51:27] <Kevin`> Mohm*
[07:51:35] <specing> RokLobsta: The time between the change and starting a new conversion is exactly 1 cycle
[07:52:04] <specing> Kevin`: should I use 10kOhm, 1kOhm or below?
[07:52:34] <Kevin`> specing: 10k will probably work, but even then you may have problems because of the fast switching you are trying to do. use a buffer.
[07:52:54] <specing> define 'buffer'?
[07:53:15] <Kevin`> op-amp unity gain non-inverting
[07:53:56] <Kevin`> specing: http://en.wikipedia.org/wiki/File:Op-Amp_Unity-Gain_Buffer.svg
[07:55:59] <Kevin`> you could of course actually set a gain there too
[07:56:59] <dirty_d> specing, what are you building?
[07:57:19] <specing> Kevin`: Don't have that heh
[07:57:34] <Kevin`> specing: why do you need the 1/2 voltage divider?
[07:57:56] <specing> Kevin`: So I can measure 0-5V on the internal reference
[07:58:50] <Kevin`> specing: if you are measuring the power supply voltage, do it backwards and measure the interernal reverence with the adc set to 0-vcc
[07:59:04] <Kevin`> reference
[07:59:51] <RikusW> clever idea
[08:00:03] <specing> Im building a battery charger
[08:00:05] <RikusW> it will involve a little math ;)
[08:00:29] <specing> And I need one ADC to measure the battery voltage and the other one for current
[08:01:10] <dirty_d> specing, why not just use like 1k resistors as the divider, 10k is probably fine too
[08:01:25] <specing> I can use both 1k and 10k
[08:01:47] <specing> I'll try it in a moment
[08:02:13] <specing> I just need to implement something on the computer-side interface client
[08:04:00] <dirty_d> specing, the input impedance of the adc is 100M
[08:05:07] <dirty_d> the manual recomments an output impedance of what youre measuring of 10K or less
[08:06:00] <Kevin`> specing: btw, if you have a chip with an internal gain amplifier an use it, you can use a higher impedance input
[08:06:11] <Kevin`> or using it in differential mode
[08:07:28] * specing can save graphs now :D
[08:09:58] <Kevin`> is your design nice? I want to build a battery charger+discharger for measuring capacity
[08:10:12] <specing> define nice
[08:10:44] <Kevin`> actually, explain it even if it's not nice
[08:10:59] <specing> It doesent work yet :P
[08:11:06] <keenerd> Kevin`: Doing it right is tricky. See http://elm-chan.org/works/bt/report_e.html
[08:14:34] <Kevin`> interesting design on that, useful :)
[08:26:29] <specing> That 1Kohm divider seems to have cut it
[08:26:39] <specing> But now I have a different problem
[08:26:58] <specing> Not sure how to describe it, too
[08:27:50] <specing> Looks like my trimmer pot somehow got broken
[08:27:59] <specing> sigh
[08:28:49] <specing> And also Im kinda pot-poor
[08:29:20] <specing> ...
[08:29:52] * specing needs an another source of voltages
[08:30:12] <amee2k> i don't think you're supposed to smoke the trimmers, even if they're pot
[08:31:37] <RikusW> oooh ;)
[08:31:42] <RikusW> lol
[08:33:08] <specing> :) no crosstalk
[08:33:24] <specing> amee2k: lol
[08:33:51] <specing> Also my other trimmer looks kinda broken too
[08:34:02] <specing> nonlinear voltage adjustment ...
[08:34:07] <dehuman> i hate trimmers
[08:34:11] <dehuman> i always break em
[08:34:15] <dehuman> fucking bastards
[08:34:33] <dehuman> i'm glad its not just me
[08:34:54] <dehuman> i look at trimmers and i'm like 'i need that, but its just gonna break when i go to trim it and be a bitch to replace'
[08:35:37] <specing> Too bad I only ordered one from farnell ;/
[08:37:06] <amee2k> fixed resistors + software calibration?
[08:37:32] <amee2k> or start stocking 5W wirewound rheostats as trimmers
[08:37:48] <Kevin`> those are even more rare
[08:38:12] <amee2k> but rarely if ever get damaged accidentially during adjustment
[08:38:40] <amee2k> speaking of potis... i need a new pair of 4.7k panel mount pots for my power supply
[08:38:56] <amee2k> they're getting pretty scratchy and one is acting funny when i'm near the lower end
[08:39:29] <Kevin`> I have lots of trimpots that come from random electronics, I wonder why other people don't
[08:40:13] <amee2k> i've got 1k 3/4T trimmers and 10k/100k 23T precision trimmers
[08:40:17] <amee2k> that works for most of my stuff
[08:41:16] <amee2k> if it doesn't then i usually need a funny value that i'd never think about stocking up on anyway
[08:42:52] <specing> Hehe
[08:43:02] <specing> Everyone always needs funny values ;D
[08:43:23] <amee2k> better than funny money
[08:43:26] <amee2k> or funny mushrooms
[08:43:58] <amee2k> speaking of money
[08:44:04] <amee2k> gotta bounce, guys
[08:50:39] <OndraSter> is there wire thinner than the one in FDD cable?
[08:50:50] <rue_house> yes
[08:50:57] <rue_house> "wire wrap wire"
[08:51:06] <OndraSter> is it 0.00nothing mm thick?
[08:51:17] <RikusW> the udma 80 conductor cables
[08:51:22] <OndraSter> http://a6.sphotos.ak.fbcdn.net/hphotos-ak-snc7/422657_2651125242943_1400123387_31937295_1381409349_n.jpg
[08:51:29] <OndraSter> see those small pads
[08:51:33] <OndraSter> that is my target
[08:51:40] <RikusW> my wire wrap wire is awg 30 = 10mil = 250um
[08:51:46] <rue_house> wire wrap wire is smaller than 80c hd cable
[08:52:03] <OndraSter> 10mil is like ... 2 times this
[08:53:04] <rue_house> but the jacket on wire wrap wire is really thin
[08:54:55] <RikusW> and somewhat difficult to strip
[08:55:52] <OndraSter> btw any tips about soldering onto that?
[08:55:59] <OndraSter> except steady hand
[08:56:02] <OndraSter> because I will never have that
[08:56:04] <Kevin`> OndraSter: use flux, reflow
[08:56:39] <OndraSter> reflow to solder wires?
[08:56:39] <OndraSter> hmm
[08:57:01] <Kevin`> yeah, put solder on the wire first, then just heat up the wire and pad with the iron to join them
[08:57:14] <OndraSter> oh
[08:57:15] <OndraSter> yeah
[08:58:02] <OndraSter> fine, 80 wire IDE ripped
[08:58:34] <Kevin`> magnet wire is pretty common in stupidthin sizes
[08:58:47] <RikusW> or buy some real wire wrap wire..
[08:59:05] <RikusW> I heard a hot enough tip will remove the enamel
[08:59:45] <dirty_d> i priced the rc stuff i need for my quadcopter last night, $450....
[09:00:14] <rue_house> heh
[09:00:26] <rue_house> I wonder how much it would cost me to make one
[09:01:05] <Kevin`> dirty_d: what do you mean the "rc stuff"?
[09:01:17] <Kevin`> the radio link? all the parts?
[09:01:26] <dirty_d> motor esc batteries
[09:01:34] <Guest8963> hi guys
[09:01:53] <OndraSter> hey
[09:02:02] <OndraSter> why can't be JTAG like... only one or two wires? :(
[09:02:04] <Guest8963> where can i find the spi notations, i forgot them
[09:02:09] <dirty_d> 4x1.5kW motors, 4x60A esc, 4x5Ah 4s lipo
[09:02:27] <dirty_d> need props too
[09:02:34] <Guest8963> and i dunno exaclty wich is sda, scl, sdd
[09:02:34] <rue_house> wow thats big
[09:02:35] <Guest8963> etc
[09:02:39] <dirty_d> the 16-18" props are expensive
[09:03:02] <Kevin`> Guest8963: sda,scl is notation for i2c, not spi
[09:03:19] <Kevin`> Guest8963: anyway, it's in the datasheet
[09:03:43] <specing> dirty_d: you are crazy :D
[09:04:12] <rue_house> your making a 48" quadcopter and complaining about it costing $450?!?!?!
[09:04:53] <norbi> Kevin`: im using a device that uses i2c and spi on the same ports
[09:05:07] <Kevin`> compatible with both?
[09:05:12] <norbi> Kevin`: yes
[09:05:20] <norbi> Kevin`: it needs to be selected
[09:06:11] <rue_house> iirc they are designed to be able to be put ont he same bus
[09:06:23] <rue_house> never heard of it actually being done tho
[09:06:45] <Kevin`> rue_house: the two protocols are not designed to be compatible
[09:07:02] <rue_house> no, they are designed to be able to be run on the same bus
[09:07:21] <norbi> sdi - d out, sdo - d in, sclk d out, cs - dout(ss)
[09:07:23] <Kevin`> with i2c both lines are bidirectional
[09:07:30] <Kevin`> what's "they"
[09:07:50] <Kevin`> also i2c is open collector instead of push-pull
[09:08:01] <rue_house> this requires a long explination I cant make during breakfast
[09:08:07] <norbi> Kevin`: you need to set a register bit to select or spi or i2c
[09:08:14] <Kevin`> you just have to say what "they" are that are compatible
[09:08:55] <rue_house> electrically i2c and spi (I'm sure it was spi) dont get each other mixed up when on the same bus
[09:09:26] <Kevin`> clock stretch i2c or i2c data output + spi = short
[09:09:58] <dirty_d> rue_house, lol
[09:10:08] <norbi> rue_house: they don't get mixed up, but it can create bus errors, but those errors can be prevented and its specific for each device how to prevent bus error
[09:10:22] <dirty_d> it should be able to lift 20 pounds too
[09:10:24] <dirty_d> maybe more
[09:10:39] <rue_house> scarry
[09:10:47] <dirty_d> maybe i can sell it for $1000 and make another one for free
[09:11:04] <dirty_d> that could carry some pretty heavy equipment
[09:11:13] <norbi> Kevin`: you can use logic chips to prevent errors
[09:12:04] <norbi> MOSI is d out, right, and miso is d in
[09:12:12] <norbi> sck is clock
[09:12:23] <norbi> ss is chipselect d out
[09:12:43] <rue_house> like a cop cam
[09:13:08] <dirty_d> or rockets
[09:13:09] <dirty_d> lol
[09:13:20] <rue_house> sell to protestors for watching cops
[09:13:27] <Kevin`> norbi: mosi = master out, slave in
[09:14:49] <norbi> Kevin`: yea, so it is data out pin of spi
[09:14:51] <norbi> right?
[09:15:37] <Kevin`> norbi: it's the data output pin of the master, you connect it to the data input pin of the slave
[09:15:45] <norbi> Kevin`: i mean data out from the master point of view, it is out pin of mcu
[09:15:54] <norbi> Kevin`: yea, thanks
[09:17:34] <norbi> everybody is saying that spi is the coolest one, but why most people are using i2c?
[09:18:05] <rue_house> if you want a bunch of the same device on i2c it gets bad
[09:18:44] <rue_house> device addresses are, for the most part, hard coded on the chip
[09:20:54] <rue_house> on my hexapod I was going to use i2c out to the controllers
[09:21:15] <rue_house> just using plain serial now
[09:24:35] <Kevin`> norbi: what do you mean most people? there are tons of devices using spi
[09:24:56] <Kevin`> norbi: spi is a lot faster than i2c. you can put multiple devices on an i2c bus with only two pins.
[09:25:49] <norbi> Kevin`: i know, but i just asked why most people are using i2c, there are a lot, who i met and saying that never used spi devices, he prefered ot use i2c
[09:26:37] <Kevin`> you probably just talked to the wrong people
[09:27:21] <Kevin`> both of them are common, but for different tasks
[09:28:09] <Kevin`> in the average consumer gizmo, i2c is used for sensors and pmic, spi is used for communication and storage devices
[09:28:50] <norbi> ah, ok thx
[09:36:38] <OndraSter> http://a5.sphotos.ak.fbcdn.net/hphotos-ak-ash4/409571_2651337088239_1400123387_31937380_909041288_n.jpg
[09:36:41] <OndraSter> soldering skills++
[09:37:22] <Kevin`> that doesn't look stable
[09:37:39] <Kevin`> I would hot glue those to the board, not leave them flying
[09:37:42] <Kevin`> or you'll pull off the pads
[09:44:57] <norbi> did you ever played with adxl345?
[10:16:28] <OndraSter> Kevin`, you just had to say it
[10:16:29] <OndraSter> :(
[10:17:14] <OndraSter> TDO pad is off
[10:17:16] <OndraSter> damnit
[10:17:38] <specing> :D
[10:18:54] <Kevin`> OndraSter: it's happened to me more than once, so i'm rather careful about it now. especially with solid wires like you were using
[10:21:23] <OndraSter> :(
[10:39:50] <OndraSter> hmm
[10:39:56] <OndraSter> wish there was some way to "ignore TDO"
[10:40:02] <OndraSter> and just expect that the data is right
[10:40:34] <Kevin`> um, you can?
[10:40:45] <Kevin`> why would you want to though
[10:41:04] <Kevin`> oh, right.
[10:41:38] <Kevin`> get out your xacto knife and solder a wire to the trace the pad was connected to
[10:42:00] <OndraSter> there is no trace
[10:42:12] <Kevin`> what's it connected to
[10:42:18] <OndraSter> it just disappears
[10:42:26] <Kevin`> there's no hole?
[10:42:31] <OndraSter> nope
[10:42:37] <Kevin`> it can't just disappear
[10:42:45] <Kevin`> microscope
[10:43:32] <OndraSter> first I will ask LG if the phone isn't still in warranty actually :P
[10:43:57] <Kevin`> normally lifting pads voids the warranty
[10:44:10] <OndraSter> I know
[10:44:17] <OndraSter> but there was no "warranty void if removed" sticker
[10:44:27] <OndraSter> there is no way to find that somebody opened it (except the soldered JTAG pins lol)
[10:50:42] <amee2k> CapnKennel!
[10:55:38] <OndraSter> Kevin`, too much thick silkscreen to see the nanotrace
[10:55:54] <OndraSter> I had diff idea
[10:56:04] <OndraSter> each pin is usually protected by 100R resistor usually
[10:56:08] <OndraSter> so if I find it...
[10:56:19] <Kevin`> you can break off BOTH the resistor pads then :D
[10:56:58] <OndraSter> both?
[10:57:06] <OndraSter> :D
[10:57:15] <OndraSter> hmm I found the trace
[10:57:20] <OndraSter> it goes immediatily into another layers
[10:57:25] <OndraSter> about 0.5mm
[10:57:29] <OndraSter> from the pad
[10:59:22] <Kevin`> does the via go all the way through the board?
[10:59:51] <OndraSter> no idea, because there is copper shielding on the other side
[10:59:55] <OndraSter> I think I am going for RMA...
[11:00:01] <OndraSter> at least try it lol
[11:04:03] <specing> lol
[11:04:47] <OndraSter> LG has pissed me in the past
[11:04:49] <OndraSter> so... ye
[11:24:50] <pingec> I am trying to find the formula on how to calculate the temperature out of the value in the DS18B20 register. Anyone know where to look?
[11:25:26] <keenerd> I would hope it is in their data sheet.
[11:25:44] <OndraSter> yeah it is
[11:27:36] <nevdull> pingec: a lot of them use a 2's complement of some 8- or 16-bit value to represent the temp
[11:28:13] <pingec> I must be blind
[11:28:21] <pingec> cause I can't find it in the datasheet
[11:29:03] <nevdull> do a search for 'temperature register' or something like that
[11:29:26] <OndraSter> I think it was raw 2 compilment
[11:29:29] <OndraSter> or was it
[11:31:30] <pingec> I am looking in here: http://datasheets.maxim-ic.com/en/ds/DS18B20.pdf
[11:31:54] <nevdull> look on page 4
[11:32:49] <pingec> there is only a table
[11:32:56] <pingec> no formula
[11:33:22] <nevdull> look at the paragraph before it. it's not a formula. it's just a signed 2 compliment of a 16-bit number
[11:33:43] <OndraSter> last 4 bits = after decimal points
[11:33:52] <OndraSter> before = 2s complement of the number
[11:43:43] <pingec> Ok thanks
[11:51:57] <mobilegamer999> I have an issue with an AtMega16M1, I screwed up the clock fuses so I can no longer set fuses via ISP, but ISP is the only way I can program this chip.
[11:52:43] <OndraSter> only clock? just use another mega to generate it :)
[11:53:11] <mobilegamer999> well I think its the clock, but im not 100% sure
[11:54:27] <mobilegamer999> my setup is that ive got a PCB with a SMD ATMega16M1 with a single ISP connector for programming, and I have an STK600 sitting next to it which I am using for the ISP connection.
[11:56:03] <mobilegamer999> And I accidentally flashed the wrong fuse to it, and now it no longer can be read/written to over the ISP connection, which my guess is because of clock, but im not 100%
[11:56:43] <Kevin`> i'm pondering whether it would be worth adding encryption to my data or not. in all likelyhood, nobody will ever receive it because it's a non-standard data link
[11:57:40] <Kevin`> and encryption is almost as complicated as fec
[11:57:44] <mobilegamer999> So can anybody offer up any suggestions on what i can try?
[11:58:00] <OndraSter> hook up wire to the xtal pin(s)
[11:58:05] <Kevin`> mobilegamer999: provide a clock to it
[11:58:08] <OndraSter> and get there signal
[11:58:52] <Kevin`> mobilegamer999: do you know what value you wrote to the fuses? can you find out?
[11:59:18] <mobilegamer999> i believe that the DWEN fuse was enabled when it shouldn't have been
[11:59:44] <mobilegamer999> and i had a simmilar issue before when using the STK600 itself, but that supports HVPP so i was easily able to flash it back to normal
[12:00:15] <Kevin`> mobilegamer999: you need to use debugwire (or HVSP/HVPP) to disable debugwire
[12:00:34] <mobilegamer999> shuold i provide a clock to it via the SCK pin on the ISP connection, or directly to the XTAL pins on the chip?
[12:00:45] <mobilegamer999> well the only connection i can (easily) make to this chip is ISP
[12:00:47] <Kevin`> mobilegamer999: fortunately debugwire just needs unimpeded access to the reset line, so you should be able to connect it
[12:01:14] <Kevin`> providing a clock to the chip won't help you disable debugwire via isp
[12:02:01] <mobilegamer999> so pretty much my only possible solution is to go grab an AVR Dragon and disbale it?
[12:02:22] <Kevin`> mobilegamer999: that's a reasonable solution. does the atmega16m1 use hvsp or hvpp?
[12:02:42] <mobilegamer999> it does, but i dont believe i can make that connection, because the only header i have available is ISP
[12:03:05] <Kevin`> you could desolder the chip, or just make the connection anyway if it won't interfere with the circuit too badly
[12:03:25] <mobilegamer999> well anything is going to be very hard because its SMD
[12:03:36] <Kevin`> which package?
[12:03:51] <mobilegamer999> TQFP32
[12:04:02] <Kevin`> that's not a bad package at all to work with
[12:04:31] <Kevin`> it has pins. they aren't on the bottom. and they are big (relatively)
[12:04:44] <mobilegamer999> how many pins does HVPP take?
[12:05:01] <Kevin`> um, 12 maybe? look at the datasheet
[12:10:32] <mobilegamer999> the datasheet doesnt seem to mention HVPP
[12:13:46] <Kevin`> mobilegamer999: 'parallel programming'
[12:14:21] <Kevin`> 11 pins, I was close :)
[12:14:30] <Kevin`> erm, no
[12:14:34] <Kevin`> 19?
[12:15:12] <Kevin`> 8 data pins + 10 control pins
[12:17:41] <Kevin`> I suspect it would be easier to remove the chip and attach it to an adapter board than to connect all those wires to the circuit
[12:42:40] <Jan-> hihi
[12:42:44] * Jan- waves
[12:43:01] <OndraSter> hi
[12:43:03] <Jan-> hey RikusW :)
[12:43:16] <RikusW> hi Jan-
[12:49:36] <amee2k> mmh... anyone know a decent BJT that does at least 40V / 800mA / 10W and has a minimum hfe that isn't completely bullshit?
[12:49:56] <amee2k> something >50 @ 500mA would be nice
[12:50:01] <Jan-> It's cold here :/
[12:50:09] * Jan- shivers *toothchatter*
[12:50:20] * amee2k passes his warm orange hi vis winter jacket on to Jan-
[12:50:23] <amee2k> there ya go :)
[12:50:37] <Jan-> ooh, thankies
[12:50:50] <Jan-> In a weird parallel development, my friend Katie gave me one of her old military parkas (she's in the air force)
[12:50:58] <Jan-> it's super warm and you can turn it inside out and it's safety orange
[12:51:10] <amee2k> wool :)
[12:51:13] <amee2k> cool* even
[12:51:13] * Jan- puts it on
[12:51:30] <amee2k> does it have reflective stripes too?
[12:51:35] <Jan-> I don't think so
[12:51:37] * amee2k <3 reflective stripes
[12:51:38] <Jan-> It's blue on the outside
[12:51:46] <Jan-> sort of air force grey/blue
[12:51:46] <amee2k> meh
[12:52:44] <Jan-> I know she had a problem finding a camelbak that DIDN't have reflective piping on it
[12:52:55] <Jan-> she had to get a desert colored one
[12:52:59] <Jan-> but they all had reflective parts
[12:53:07] <Jan-> which sort of destroys the point of, you know, camouflage
[12:53:12] <amee2k> lol
[12:53:31] <amee2k> well, there always is duct tape :)
[12:53:51] <amee2k> but yeah... getting seen in a military setting gets you a bullet in your butt
[12:53:53] <Jan-> The thing is, when she got to where she was going, it was a huge airbase in the middle of the desert
[12:54:06] <Jan-> so basically the military's decision making process was this:
[12:54:19] <Jan-> Our one-mile-square airbase will be noticed if someone wears a camelbak with eighth-inch reflective piping on it.
[12:54:44] <amee2k> in an industrial setting, getting seen at the right moment can save your butt, which in turn saves a lot of people for a lot of paperwork, which makes it pretty cool
[12:56:10] <amee2k> my good old BD139s aren't cutting it and parametric search is pretty useless for finding *minimum* specs. noone wants to advertise the worst case
[12:56:59] <amee2k> but my biasing is a current source so i kind of need to size it for the worst case gain that can happen
[12:58:09] <Jan-> um, er
[12:58:18] <Jan-> can you say that again but using a major european language
[12:58:34] <amee2k> lol
[12:58:40] <amee2k> which part of it?
[12:59:32] <amee2k> i'm trying to find a power NPN transistor that doesn't have really shitty top end gain :(
[13:00:37] <amee2k> the problem is that parametric search engines usually advertise the typical or maximum gain
[13:00:40] <bram1> amee2k: take a look at the PBSS4032NZ
[13:00:59] <amee2k> BD139 is advertised as "hfe up to 250!" ... ha. turns out they only guarantee a minimum of 25 @ 500mA
[13:03:08] <amee2k> bram1: wow, thats a lot of gain for 6 amps. do they make these in a real package too?
[13:03:25] <amee2k> the NXP datasheet only has some flimsy SOT thingy
[13:03:46] <bram1> amee2k: this is the big sot package, it is easy to solder by hand
[13:04:35] <amee2k> but it isn't very easy to get heat out of
[13:04:45] <bram1> Rcesat is 62.5mR
[13:05:25] <bram1> at 4A that is 1W
[13:05:31] <amee2k> rated for only 2W and the datasheet says only really useful for 700mW tops on a normal board
[13:05:51] <bram1> with Rth of 75K/W at 6cm2
[13:06:13] <amee2k> even 2W on the aluminium core board is too low
[13:06:31] <bram1> how much current do you need?
[13:06:42] <amee2k> 5W absolute minimum, better rated for 10 or more
[13:07:08] <bram1> why? the Rcesat is only 62.5mR, so dissipation is very low
[13:07:09] <amee2k> only 1A or so, but at least 5V of drop
[13:07:33] <amee2k> what does some resistance have to do with the heatsink?
[13:07:42] <bram1> everything
[13:07:53] <bram1> I2 * R = P
[13:08:09] <amee2k> transistors aren't resistors
[13:08:33] <bram1> if you put them in saturation, they are
[13:08:39] <amee2k> well, i don't
[13:08:43] <amee2k> and i never said i did
[13:09:01] <bram1> ok, then this transistor is no good for you
[13:09:08] <amee2k> :P
[13:09:40] <amee2k> thats why i said "at least 40V / 800mA / 10W" ;)
[13:09:48] <bram1> you can do pwm and put a C to smooth it
[13:10:35] <bram1> ah, it was outside my screen when I....
[13:10:35] <amee2k> well, its class B, not class D
[13:10:42] <amee2k> oh, okay
[13:10:57] <bram1> excuse me for injecting sh*t solutions
[13:11:54] <amee2k> i'm trying to eliminate a predriver on a regulator to get the compliance voltage down... already tempted to turn it upside down and use a PNP for it
[13:12:46] <amee2k> fuck. and now stores are closed too >_<
[13:12:52] * amee2k desires pizza
[13:13:07] <asteve> pizza?!
[13:13:12] <asteve> stores closed? where are you?
[13:13:23] <amee2k> in a country other than the US
[13:14:02] <asteve> i'm not sure what the US has to do with where you are based on the statement you just made stating that you're not in it
[13:14:11] <amee2k> cbf to order tonight, i'll just raid the fridge i guess
[13:14:13] <asteve> but, pizza places stay open late everywhere in the world
[13:14:43] <amee2k> well, some are still open
[13:15:13] <bram1> amee2k: how about the 2n6045?
[13:15:26] <amee2k> but they're the most expensive ones and unless everything else has closed i don't usually order from them
[13:15:56] <amee2k> ooh
[13:15:56] <Jan-> does anyone want a 75-ohm MUSA patch bay in rack mount?
[13:16:06] * amee2k <3 TO-220
[13:16:07] <Jan-> you can have it for the price of postage
[13:16:10] <Jan-> links included
[13:16:50] <amee2k> bram1: are you sure thats not a darlington??
[13:17:00] <amee2k> B-E drop is really shitty
[13:17:16] <amee2k> 2.8V... even the predriver i'm trying to eliminate doesn't drop more than a volt
[13:18:05] <amee2k> and hfe 1000 @ 3A is creepy
[13:18:25] <bram1> it has the specs of a darlington
[13:19:31] <amee2k> i love how the ON Semi one on farnell is listed as "hfe 20" and the Multicomp for 1/5th the price is listed as "hfe 20000"
[13:21:29] <amee2k> i don't need HUGE gain... the BD139 with >25 is really only short by a factor of 2 or 3 or so
[13:33:21] <Kre10s_> hmm. ok. so I caught an interrupt, and am in the interrupt handler routine. When I return, the previous code that was running will be executed.
[13:33:45] <Kre10s_> whats the simplest way I can jump to a while(1); after the interrupt?
[13:34:59] <amee2k> put it into the ISR?
[13:35:07] <Tom_itx> the main loop usually runs in a while(1)
[13:35:12] <amee2k> or define what you mean by "jump"
[13:36:11] <amee2k> the ISR will naturally return only to the place where it left off. you *can* modify the return address, but that can get mildly awkward if you're using C
[14:16:50] <pthomas> what do you call the units of a raw adc conversion?
[14:17:16] <pthomas> raw sounds funny 2047 raw/Volt
[14:17:20] <pthomas> ?
[14:22:14] <Landon> bits?
[14:22:53] <raek> units/volt?
[14:24:44] <pthomas> bits sounds like the number bits wide the result is i.e. 12-bit, I also thought about counts, 2047 counts/Volt?
[14:27:50] <OndraSter> steps maybe?
[14:28:38] <pthomas> yeah that has a certain ring to it
[14:34:13] <amee2k> pthomas: well, it really is a dimensionless quantity
[14:35:21] <amee2k> signal / refence * 2^n ... units of signal and reference is volt which cancels and leaves the result dimensionless
[14:46:27] <pthomas> OK, I get it, ref gives me the units I need
[14:47:28] * Jan- eats chocolate chips
[14:53:48] <Steffann> Congratz
[14:55:44] * specing eats Jan-
[14:55:54] <Jan-> HEY
[14:55:57] <Jan-> let me out!
[14:56:06] * Jan- kicks specing's insides
[14:56:14] * specing farts
[14:58:46] <Steffann> Bah
[15:04:41] <RikusW> lol
[15:04:52] <RikusW> seems someone is bored
[15:11:19] <Tom_itx> trolls
[15:15:38] <Steffann> Jan- is just our Mrs. Offtopic
[15:17:57] <OndraSter> hmm that phone I was trying to JTAG
[15:18:03] <OndraSter> it is still under warranty, Kevin`
[15:18:07] <OndraSter> I think I am going to try RMA it
[15:18:07] <OndraSter> lol
[15:20:03] <Kevin`> that's so bad =p
[15:20:38] <Kevin`> why do you need jtag, btw? normally you can get enough debugging output via the uart and/or usb interface to get by
[15:27:08] <specing> is it just me or are all women on IRC trolls?
[15:27:26] <Jan-> *sigh* Phil wants to buy a 200 pound video projector
[15:27:28] <Jan-> he's insane
[15:27:35] <specing> there is another one on #electronics meowing randomly
[15:27:35] <Jan-> I don't want a piece of AV equipment that weighs nearly twice as much as me
[15:27:47] * Jan- lights specing on fire
[15:27:50] <norbi> guys, do you know isp mkii ?
[15:27:51] <Jan-> I'm not a troll!
[15:28:02] <specing> Fine. You are just insane.
[15:28:08] <Jan-> We aim to please :)
[15:28:53] <Jan-> OK so he wants to get this old-style, three-gun, CRT-based video projector.
[15:29:07] <Jan-> He claims it'll get us up-to-the-minute picture quality at a fraction of the price.
[15:29:18] * Jan- isn't so sure she wants a 200lb lump of 80s technology hanging from the ceiling
[15:29:48] <specing> Jan-: Why do you need a video projector anyway? Aren't you like...blind?
[15:30:00] <Jan-> I am, Phil isn't
[15:30:20] <specing> And who the hell is phil?
[15:30:33] <Jan-> My long-term, live-in, carnal-sin male friend.
[15:30:41] <specing> A WHAT?!
[15:30:52] <Jan-> Well, I'd say "boyfriend", but it makes it sound like we're 16
[15:31:40] <specing> Ahm
[15:31:53] <Jan-> whaaaaat
[15:31:55] <Jan-> aaaa
[15:32:53] <specing> And Im not sure you'll get "up-to-the-whatever" quality with something from the 320x280 era...
[15:33:29] <Jan-> apparently they do HD
[15:33:57] <specing> HAhahahaa
[15:34:14] <Jan-> what
[15:34:20] <specing> HD my ass.
[15:34:37] <Jan-> OK I'm wrong they're not from the 80s, they're from the mid 90s
[15:34:59] <Jan-> wbat about your ass
[15:35:50] <norbi> how is that possible that isp cannot connect to the board in programming mode and the isp is working and the board is fine too..?
[15:35:52] <specing> My ass thinks this debate is pointless
[15:36:16] <Jan-> well, you are talking to a blind person about movie projectors, yes.
[15:36:26] <Jan-> but that's OK, that sort of thing happens all the time :)
[15:36:29] <specing> >_>
[15:37:15] <norbi> anybody?!
[15:37:41] <specing> norbi: the connection is broken somehow
[15:38:02] <norbi> specing: it is somehow, im able to connect with jtag, but not with isp
[15:38:39] <specing> maybe you've disabled ISP?
[15:38:45] <norbi> specing: there are times when isp works and when it doesnt, first thing was trying debuging the problem from bottom up
[15:38:54] <norbi> cable is fine
[15:39:04] <norbi> checked with multimeter
[15:39:29] <norbi> specing: disabled isp?
[15:39:58] <specing> FUSEs...
[15:40:07] <specing> Its always the fuses...
[15:40:26] <norbi> specing: but how its that i can connect with jtag and programming it fine
[15:40:50] <norbi> specing: what part of fuses can be affected, i never change the fuses
[15:44:06] <specing> "times when isp works and when it doesnt" -- contact issue?
[15:48:01] <norbi> specing: tried to debug from bottom up, but cable is fine
[15:50:06] <norbi> specing: what should i check for at fuses?
[15:50:56] <amee2k> http://ve504.cugnet.net/~amee2k/sandbox/2012-02-22-222820.jpg << can anyone show me a datasheet for this transistor??
[15:51:20] <amee2k> (hint: it is not the datasheet for the 100W TO-3 transistor that goes by the same type number)
[15:52:39] <amee2k> or at least identify the manufacturer from that logo?
[15:53:22] * amee2k once knew a website with like 200 obscure brand decals, but lost the link :(
[15:55:32] <urief> amee2k: what is it in ?
[15:55:50] <amee2k> an electronic fluo ballast board
[15:56:31] <amee2k> i want to reverse engineer it because i'm suspecting some kind of design flaw that causes excessive filament wear
[15:56:51] <norbi> amee2k: ? http://www.datasheetarchive.com/BU102-datasheet.html
[15:57:10] <urief> amee2k: its a switching application, just pick any high voltage transistor.
[15:57:20] <urief> they're all more or less the same.
[15:57:30] <amee2k> i'm trying to identify it, not replace it
[15:57:39] <amee2k> replacing random parts won't fix a design flaw
[15:58:01] <keenerd> It will if you can replace them fast enough.
[15:58:23] <keenerd> (See Apple's genius bar.)
[15:58:27] <amee2k> norbi: did you ever actually look on that link?
[15:58:51] <amee2k> keenerd: how does replacing transistors fix excessive filament wear?
[15:59:31] <urief> it will be some random chinese low cost manufacturer
[15:59:37] <keenerd> Bah. Sorry, it sounds sounds like a very tricky problem.
[15:59:42] <urief> even if you identify the part there may be no datasheet
[16:00:07] <amee2k> yeah, except for the TO-3 power transistor i can't find anything at all
[16:00:32] <amee2k> only obscure scanned databook pages that i didn't inspect too closely
[16:00:56] <amee2k> looks like BU series transistors have been around for several decades
[16:01:09] <urief> amee2k: is it JD or IAD on the transistor ?
[16:01:54] <amee2k> it looks like a "J" and a mirrored italic "C" that could be a stylized "JD"
[16:02:29] <amee2k> with the vertical bars on the D and the J combined
[16:02:42] <keenerd> Any chance of a picture? If you have a loupe you can use it to coax 2cm macro shots from a normal camera.
[16:03:22] <amee2k> http://ve504.cugnet.net/~amee2k/sandbox/2012-02-22-222820.jpg << this isn't macro enough?
[16:03:30] <urief> no
[16:03:42] <amee2k> hmm lemme see if i can get a better shot
[16:04:03] <keenerd> Heh. No, that is a good shot.
[16:05:58] <keenerd> http://www.jdsemi.cn/en/getfile.asp?id=40
[16:06:17] <keenerd> First google hit for jd bu 102 transistor.
[16:09:54] <keenerd> I hope that makes up for me acting like an idiot? :-)
[16:12:43] * Jan- is sleepy
[16:12:51] <amee2k> http://ve504.cugnet.net/~amee2k/sandbox/2012-02-22-230118.jpg http://ve504.cugnet.net/~amee2k/sandbox/2012-02-22-230209.jpg << these better?
[16:13:13] <urief> keenerd: way to make the rest of us look bad.
[16:13:24] <amee2k> keenerd: ooh, that sounds like bullseye. thanks \o/
[16:13:44] <Kevin`> amee2k: nice camera
[16:13:50] * Jan- falls asleep with her head on amee2k's shoulder
[16:14:04] <amee2k> 0.0
[16:14:12] * urief watches the spreading stain of drool on amee2k's overalls
[16:14:15] * Jan- dribbles, as she often does when asleep
[16:14:27] <Jan-> Shhhgggllllllllgghhhh
[16:14:31] <Jan-> *spit bubbles*
[16:14:41] <amee2k> Kevin`: one of these cheap microscopic USB webcams
[16:14:56] <amee2k> urief: bib overalls or jumpsuit?
[16:15:28] <Kevin`> amee2k: ah, that would explain it :)
[16:15:54] <Kevin`> taking pictures of stuff like that with a normal camera is rather difficult
[16:15:58] <urief> amee2k: do they make Hi-Vis jumpsuits ?
[16:16:02] * Jan- snores loudly
[16:16:10] <amee2k> urief: of course \o/
[16:16:15] <urief> :D
[16:16:30] <amee2k> i think they make pretty much everything in hi vis these days
[16:16:37] <amee2k> maybe short of underpants
[16:17:20] <amee2k> Kevin`: Conrad had one on display in their store like a year and a half ago for 150... i thought like mmh nice toy. would be handy for identifying tiny SMT crap or making pics of IC markings. but not for that price
[16:17:46] <urief> you can easily DIY a macro lens.
[16:18:18] <amee2k> and a few weeks ago i my dad noticed it for 49.something in an advertisement flyer... i must have mentioned it sometime because he told me
[16:19:54] <asteve> amee2k: what's the fastest way to toggle an led pin on and off after entering an ISR
[16:21:16] * Kre10s_ probs Jan with a pretzel stick.
[16:21:39] <amee2k> asteve: gcc is clever enough to reduce "PORTx |= (1<<n);" and "PORTx &= ~(1<<n);" to the IO bit set/reset instructions
[16:21:51] <Jan-> Gmrhrzlwhassamatter?
[16:22:09] <amee2k> i *think* they're two cycles each which should be the fastest you can get
[16:22:48] <specing> One cycle.
[16:22:55] <amee2k> hmm even on IO locations?
[16:23:23] <amee2k> i thought pretty much everything memory/IO related is two cycles minimum
[16:23:41] <specing> sbi/cbi is single sycle
[16:23:45] <specing> cycle*
[16:23:54] <amee2k> nice :)
[16:27:03] <asteve> sbi = ?
[16:27:32] <amee2k> its an avr assembler instruction
[16:27:35] <amee2k> Set Bit hm... IO or something
[16:27:43] <asteve> i know, i was hoping for you to explain it's initials :)
[16:27:49] <asteve> its*
[16:28:19] <amee2k> "SBI – Set Bit in I/O Register
[16:28:21] <amee2k> "
[16:28:46] <amee2k> thats what the ISA manual says
[16:29:59] * Jan- is aghast
[16:30:08] <amee2k> is that edible?
[16:30:10] <Jan-> I bought some hexagonal boxes on ebay
[16:30:14] <Jan-> and they're octagonal
[16:30:26] <Jan-> and the seller says "It's the same thing"
[16:30:33] <amee2k> ...
[16:30:41] <amee2k> negative feedback
[16:30:48] <amee2k> resell them yourself
[16:30:52] <Jan-> doesn't know six from eight
[16:30:56] <Jan-> that's some feedback right there
[16:31:22] <amee2k> "goods not as advertised; refused to accept return / refund"
[16:31:28] <mrfrenzy_> I've never had to leave negative feedback. I just open a paypal case and the seller promptly makes everything right
[16:31:50] <Jan-> I left neutral feedback once
[16:31:51] <amee2k> i gave neutral once when ordering from china
[16:31:58] <Jan-> yeah ditto
[16:32:05] <Jan-> I ordered some side-release buckles for my backpack
[16:32:32] <amee2k> they didn't have the stuff they were selling and kept moaning at me for two weeks how their wholesale contact is leaving them in the wind
[16:32:51] <amee2k> because noone is moving their asses due to some local holiday
[16:33:10] <mrfrenzy_> good ebay sellers close their store during new years
[16:33:38] <amee2k> it was hmm summer two years ago i think
[16:34:10] <mrfrenzy_> aha so not new year
[16:34:14] <amee2k> selling stuff you don't have and only ordering it yourself when the customers are waiting isn't exactly what i'd call good management
[16:34:14] <mrfrenzy_> I think that is about right now
[16:35:24] <amee2k> i was happy that i could gather that they're blaming their wholesale contact for the holdup from the emails
[16:35:41] <asteve> amee2k: that's the way the majority of the cheapest online retailers do business
[16:35:49] <amee2k> if they didn't use google translator for their emails, they should start using it
[16:35:59] <asteve> i ordered 1k in exhaust parts for my bike and it took 1.3 months to get ther
[16:37:00] <amee2k> after i left neutral feedback for that, they kept moaning by email for another two weeks that i should revise it and how miserable their lives are going to be now
[16:39:45] <amee2k> asteve: well, contrary to popular belief, just because some dude in china does it for money doesn't mean it is a good idea
[16:40:25] <mrfrenzy_> that's what the feedback system is for
[16:40:37] <mrfrenzy_> some guys might be willing to wait a month for the cheaper price
[16:41:10] <asteve> it's not a problem if they tell you that they do not have it in a stock and will have to order it
[16:41:36] <amee2k> what asteve just said. but they only told me by email a week after i placed the order
[16:48:39] <Kevin`> post from the park :D
[16:51:07] <Jan-> A friend of mine in the US, who's a bit of a h4x0r, has bought a, er, "Nook"
[16:51:11] <Jan-> what's a "Nook"?
[16:51:55] <Jan-> it's tough to google
[16:51:57] <keenerd> Ah, the Nookie Book.
[16:54:14] <Kevin`> I have a nook touch
[16:54:22] <Kevin`> it's an e-reader
[16:54:45] <Kevin`> hacked to run androi and linux appsof course
[16:55:25] <mrfrenzy_> popular hackaday-topic
[16:59:06] <amee2k> if i wanted to buy an ebook reader, then i'd buy it for reading ebooks
[16:59:32] <Kevin`> I use it for reading ebooks, it's nice for that
[17:00:07] <amee2k> and the only features i really care about for that is a) SD card slot b) can display PDFs with no DRM crap and c) long battery life
[17:00:07] <Kevin`> it's nice to be able to do stuff like run a proper comic viewer on eink, check email, play board games, etc, too though
[17:00:39] <amee2k> where a) and b) are roughly equally important and c) somewhat less important. if the battery is servicable then it isn't too important but nice to have
[17:01:11] <keenerd> amee2k: I'd recommend the wikireader, but it will be a while before I have a PDF viewer on there.
[17:01:17] <amee2k> because the first thing i'd do is put my whole datasheet and appnote collection on an SD card so i can carry it around
[17:01:29] <amee2k> "wikireader"?
[17:01:52] <Kevin`> amee2k: wikireader is a VERY minimal ereader device
[17:02:10] <keenerd> $30 though and the most fun embedded programming toy I've gotten.
[17:02:10] * amee2k googles...
[17:02:42] <amee2k> oh, that thing is tiny :(
[17:02:42] <Kevin`> don't expect to run linux or equivalent on it :)
[17:02:53] <Jan-> I hate touchscreen devices :(
[17:03:05] <keenerd> Its got a 32 bit cpu and 32 megs of ram, it could run linux.
[17:03:31] <amee2k> can the display in that thing even display any useful PDF content?
[17:03:42] <keenerd> Jan-: Yes, they are a failure of accessability in general. I do hope to have a keyboard hooked up through he serial port soon.
[17:04:09] <Jan-> sorry I wasn't talking about your particular device
[17:04:12] <Jan-> just general griping
[17:04:30] <Jan-> although jan tips her hat at your accurate assessment of my gripe :)
[17:04:37] <amee2k> keenerd: MASSIVE ++ for using generic batteries!
[17:04:55] <keenerd> amee2k: Anyway, I've got an image viewer that can handle 6000px wide images. A far far ways away from PDF viewer, I just wanted to put topo maps on the device.
[17:06:34] <amee2k> ouch
[17:06:39] <amee2k> 240x204 pixels
[17:06:44] <keenerd> Yeah.
[17:07:03] <keenerd> I am trying tosee if I can at least fake grayscale by blinking them fast enough :-(
[17:07:06] <amee2k> if i resize a PDF to be 204 pixels wide it looks completely unreadable
[17:07:24] <Jan-> just so you know
[17:07:26] <amee2k> err 240 wide but still
[17:07:30] <Jan-> I really hate touch screen devices
[17:07:38] <keenerd> amee2k: Why would you do that when you could zoom in to 1000 dpi with my viewer?
[17:07:39] <Jan-> my gym has replaced all its treadmills
[17:07:51] <Jan-> and instead of having buttons on them they now have a featureless piece of plastic where the buttons use to be.
[17:07:56] * Jan- goes postal with a fire ax
[17:07:59] <Jan-> HELPFUL MUCH
[17:08:32] <keenerd> amee2k: I read full sized topographical maps on the little 200px screen. A datasheet would be nothing forit to handle :-)
[17:09:02] <specing> Any PWM experts around?
[17:09:28] <keenerd> Admitedly since you can't draw on the map or take bearing it is kind of useless for now.
[17:09:29] <specing> Specificaly: How do I ensure equal voltage across different loads?
[17:10:09] <amee2k> PWM doesn't control voltage
[17:10:11] <keenerd> Equal average voltage or equal peak waveform voltage?
[17:10:41] <amee2k> PWM controls duty cycle, which controls *average* voltage, and usually more or less average current or power
[17:11:02] <specing> Im MOSFET'ing
[17:11:14] <keenerd> I love that dance.
[17:11:19] <specing> :S
[17:12:08] <amee2k> keenerd: thats a pretty cool reader but not really what i had in mind at all
[17:12:29] <Kevin`> I do recommend the nook touch
[17:12:54] <keenerd> amee2k: Get one for fun, it is so easy to hack on ;-)
[17:12:56] <amee2k> i'd want something like 10-12" or so, which is still reasonably compact but can display A4 in full width (when holding it as landscape)
[17:13:11] <amee2k> how much do they cost now?
[17:13:35] <Jan-> *sigh* how long will it be before NOBODY makes a phone with BUTTONS.
[17:13:45] <keenerd> $30 in the US, dunno about over your way.
[17:13:58] <amee2k> last year i said when i see one that reliably meets my first two criteria and cost less than 50EUR, i'm just gonna buy it
[17:13:59] <specing> Jan-: Not until Nokia goes bankrupt
[17:14:13] * Jan- less-than-three blackberry
[17:14:17] <Jan-> I had a nokia
[17:14:25] <Jan-> but people kept stealing it as it was obviously "expensive"
[17:14:51] <mrfrenzy_> go to a less expensive gym
[17:14:56] <Jan-> not at the gym.
[17:14:58] <mrfrenzy_> mine still has no touchscreens
[17:15:08] <amee2k> nook touch == Nook Simple Touch ?
[17:15:12] <Kevin`> amee2k: yeah
[17:15:16] <mrfrenzy_> regarding the treadmill, not the phone thefts ;)
[17:15:24] <keenerd> amee2k: I was looking for anything cheap with a reflective screen. Open hardware, serial port and standard batteries were bonuses.
[17:15:46] <Jan-> mrfrenzy_: bleargh
[17:16:31] <amee2k> keenerd: i see. since i don't want it for hacking primarily open hardware isn't much of a priority for me
[17:18:31] <keenerd> specing: If you want it to be accurate, dedicating an ADC to feedback is pretty much the only way. Or using a high power buffer op amp.
[17:20:29] <Kevin`> on man, i'm gonna have to find how this was done and try it http://techcrunch.com/2012/02/22/hack-makes-nook-touch-e-ink-display-almost-as-responsive-as-lcd/
[17:22:16] <keenerd> Holy cow that page was 2.5MB. 12 minutes for the poor person still on dialup!
[17:26:27] <specing> keenerd: How hot are mosfets supposed to be at 200 mA of current switched at 2V?
[17:26:42] <specing> Mine gets untouchable after 10s...
[17:26:57] <specing> In I-Pak packaging
[17:27:10] <Kevin`> specing: it shouldn't get hot at all if you are switching it right. if you keep it all on or all off there won't be much power dissapated in it
[17:27:22] <Kevin`> what's important is switching it with fairly high drive strength
[17:27:36] <specing> Kevin`: explain
[17:27:49] <Kevin`> the gate of a mosfet behaves like a capacitor
[17:27:55] <specing> yes
[17:27:58] <Kevin`> and of course the device itself is voltage controlled
[17:28:20] <Kevin`> while it's transitioning from ~0r to ~BIGNUM-r, it will dissipate some power
[17:28:27] <Kevin`> so you want that to go as fast as possible
[17:28:58] <specing> And how do I do that?
[17:29:10] <specing> Im driving it with 5V from my mega8
[17:29:11] <Kevin`> and of course you want to drive it with a digital signal, not a varying voltage, unless you specifically DO want it to dissipate power
[17:29:25] <specing> the switched voltage is 5V, too
[17:29:49] <Kevin`> specing: check the specs for your mosfet for what the resistance will be at 5v
[17:29:54] <specing> I've set the internal PWM to 8-bit FASTPWM
[17:30:01] <Kevin`> specing: .. or just turn it on and leave it on and see if it gets hot
[17:31:50] <OndraSter> why are they shoving up our heads "using BJTs as switchesL
[17:31:51] <OndraSter> "
[17:32:01] <OndraSter> wouldn't be more logical to use FETs as switches...
[17:32:02] <Kevin`> specing: you are using a logic-level mosfet, right? some general purpose ones switch on at a rather high voltage (9v or so)
[17:32:27] <keenerd> OndraSter: Are you in school? Most of them do that, since no one has heard of a transistor before.
[17:32:36] <OndraSter> haha
[17:32:53] <specing> keenerd: On-Resistance is 0.15 ohm
[17:32:55] <Kevin`> bjts have their place
[17:33:14] <Kevin`> but definitely overrecommended where mosfets would be better
[17:33:31] <OndraSter> weren't 74LS based on BJTs?
[17:33:36] <Kevin`> specing: what gate voltage is that on-resistance for?
[17:33:52] <specing> -10V
[17:34:23] <OndraSter> that's 10V difference compard to DRAIN
[17:34:29] <OndraSter> what is the DS voltage?
[17:34:32] <OndraSter> on yours
[17:34:45] <OndraSter> and -Ugs you are driving it with
[17:34:51] <OndraSter> -Vgs*
[17:34:57] <Kevin`> it could be -v if it's a p-channel mosfet
[17:35:24] <Kevin`> (would be strange to specify it otherwise)
[17:35:24] <specing> Ugh
[17:35:31] <specing> S is at 5V
[17:35:37] <specing> G is at 5V
[17:35:48] <specing> and I want D to be at 2V
[17:36:08] <specing> (after the inductor and capacitor)
[17:36:24] <OndraSter> I am too sleepy, can't keep my eyes open. gn
[17:36:43] <keenerd> Datasheet link?
[17:36:53] <Kevin`> where you use the bjts is to make an amplifier to drive the mosfet at higher voltage+power :)
[17:36:55] <specing> FINALLY! :D
[17:37:26] <specing> http://www.fairchildsemi.com/ds/FQ/FQU11P06.pdf
[17:37:30] <keenerd> specing: I won't do anything with it, I've got to go. It is just painful watching everyone else try to tickle out details.
[17:38:03] <specing> Yes, especially from someone who is not a Ph.D in electronics :D
[17:39:46] <keenerd> I can tell you you need to be interested in fig 1.
[17:40:47] <keenerd> Maybe fig 11
[17:41:34] <Kevin`> yeah figure 1 looks useful
[17:41:36] <specing> Bonus points for nonlinear graphs!
[17:46:30] <sw33tlull4by> hey there I have a huge problem. It is not directly related to an AVR-Chip but to a Display controlled, this is the datasheet of the display http://www.display-elektronik.de/DEM16216SYH-LY.PDF and this is my code it is supposed to write the letter a to the display, but it fails, and I do not know why. I recently added tons of delays but it still does not work, what am i making wrong?
[17:46:45] <sw33tlull4by> http://pastebin.com/Nzcdu8N8 sry forgot the code
[17:55:03] <amee2k> http://omploader.org/vY3Ywbw/fluo-ballast.png << so far so straight-forward. thats the ballast... grey line denotes magnetic coupling. am i correct in interpreting that the "PV" thing is a MOV?
[18:03:29] <amee2k> the problem is that the tubes for some reason only hold 8-12 months and then fail with one filament open.
[18:03:29] <amee2k> the lamp is only turned on like once a day or so on average, then left running for at least several hours at once. i don't see how the filaments could be wearing out this fast
[18:03:29] <amee2k> unless there is something wrong... either defective, or a design flaw. any ideas where to start about this?
[18:03:29] <amee2k> i'm suspecting the circuit that connects the two filaments... the thing that i think is a MOV seems to disengage the filament heating once the tube voltage drops low enough
[18:04:57] <amee2k> the PTC reads 1k but the "PV" thing reads only 10 ohms cold. does that sound suspicious?
[19:15:21] <amstan> so i managed to program my attiny44 with an low-freq xtal clock and the divide by 8 bit set as well, so i think it's running at 4khz right now
[19:15:27] <amstan> i can't seem to get isp working anymore
[19:15:33] <amstan> is there a way I can save it?
[19:16:02] <amstan> i tried avrdude -B 1024, but it seems to do the same thing
[19:16:27] <amstan> http://pastebin.com/eqie6QRr
[19:17:51] <Tom_itx> add the correct clock source and see if it works
[19:18:07] <Tom_itx> -B 32 is the slowest iirc
[19:18:14] <Tom_itx> errr... i forget
[19:19:01] <amstan> Tom_itx: i have it connected already
[19:19:18] <amstan> scope confirms that it's oscillating properly at 32khz
[19:19:33] <amstan> i just forgot to unset that divice by 8 bit
[19:19:38] <amstan> divide*
[19:19:52] <Tom_itx> do you have a faster crystal you can sub temporarily
[19:20:50] <Tom_itx> ie 1Mhz
[19:21:48] <Tom_itx> i don't think the ckdiv8 affects programming
[19:22:18] <amstan> Tom_itx: what were you saying about -B 32? why is that the lowest?
[19:23:11] <Tom_itx> In general, the clock frequency should be at least 4 times larger than the target clock frequency. Try "-B 32" if you're having clocking issues, that should handle even 128khz clocks.
[19:23:43] <amstan> yes, but i'm using a 32k clock
[19:23:48] <amstan> with a 8/ divider
[19:23:51] <amstan> so it's 4khz
[19:23:53] <Tom_itx> change it
[19:23:57] <amstan> i can't
[19:23:59] <amstan> i already did the fuses
[19:24:04] <Tom_itx> does ckdiv8 affect programming?
[19:24:07] <amstan> idk
[19:24:45] <Tom_itx> you don't have a different crystal you can sub?
[19:25:24] <Kevin`> I suspect clkdiv8 DOES affect programming. otherwise it would be out of spec to program the chip with isp at lower voltages
[19:36:56] <amstan> sigh
[19:37:01] <amstan> i don't think the program works on it anymore
[19:38:40] <Casper> Kevin`: also it would also be out of specs to program a chip with a 64MHz clock
[20:08:03] <learningc> Casper: did you get your PSU working properly now?
[20:16:24] <Casper> learningc: didn't toyed with it yet
[21:55:28] <learningc> in eagle, how do I connect 2 points in schematic without using wires. I have seen some sort of labeling on people's schematics, but not sure if this will connect 2 points on the board layout
[21:55:41] <Tom_itx> buss
[21:55:42] <Tom_itx> or net
[21:55:54] <Tom_itx> use net and name them the same
[21:56:04] <Tom_itx> they will be connected
[21:56:15] <learningc> ah ok
[21:56:26] <Tom_itx> just make a short net off the pin
[21:56:29] <Tom_itx> and label it
[21:56:46] <Tom_itx> btw, don't route with wires
[21:56:54] <Tom_itx> they won't snap to a point like net does
[21:56:58] <learningc> so I leave the short net "hanging off the air"?
[21:57:12] <Tom_itx> you will have unconnected pins where you think they should be connected
[21:57:16] <Tom_itx> yes
[21:57:30] <Tom_itx> as long as the names are the same, they will connect
[21:57:38] <learningc> how do I label properly?
[21:57:45] <Tom_itx> name
[21:58:32] <Tom_itx> then the 'ABC' icon below net will display the name for you
[21:59:04] <learningc> ah ok, thanks
[22:00:29] <learningc> I have seen some sort of labelling enclosed in a sort of box with arrow end on one side, is there such thing in eagle?
[22:01:10] <Tom_itx> its in a lib
[22:01:12] <Tom_itx> like a flag?
[22:01:25] <Tom_itx> i forget where those are, i don't use em much
[22:01:41] <Tom_itx> you don't like the 'abc' icon labelling?
[22:03:07] <learningc> well, if the wires ends with a box, it will be easier to see
[22:04:23] <learningc> but does it work the same?
[22:04:34] <Tom_itx> yes
[22:05:08] <learningc> I mean, if I label the boxes the same name, the wires will be connected together?
[22:05:17] <Tom_itx> no
[22:05:49] <learningc> then what are they used for if they cannot connect same names?
[22:06:10] <Tom_itx> they are just a lib component, like i said, i don't use them
[22:06:26] <Tom_itx> they may but i'm not aware that they connect
[22:06:38] <learningc> so they don't have a pin attached to them?
[22:07:07] <Tom_itx> i don't know you'll have to find them in the lib and see
[22:07:24] <learningc> ok, thanks anyway
[22:08:03] <learningc> I try to get used to the abc labeling then
[22:08:20] <Tom_itx> those change when you rename the net
[22:08:35] <learningc> since the labeling is the name of the wire as well
[22:09:25] <learningc> and by the way, why does "wire" not connect wires together?
[22:09:49] <Tom_itx> if you hit the end of the pin it will
[22:09:57] <Tom_itx> but net is designed to snap to it
[22:10:48] <learningc> so, I can still use "wire" to connect my device pins together?
[22:11:03] <Tom_itx> if you want, but i wouldn't
[22:11:23] <Tom_itx> you will get to the board and likely find some unconnected pins
[22:13:03] <learningc> ok, thanks