Back
[00:13:13] <sparrW> I want to control some digital potentiometers from a uC. Do such exist that I could chain them together and send/latch data like shift registers? Do I need to use SPI and chip select lines, or something similar?
[00:14:19] <rue_mohr> you can put chips on pots, do you want ot change their value pysically, they make motorized pots
[00:15:25] <sparrW> don't need physical knobs
[00:15:29] <sparrW> I'm replacing physical knobs
[00:18:56] <yardleydobon> Can someone link me to a datasheet describing the procedure to reset the i2c bus? I've seen it before but I can't find it now.
[00:19:53] <Casper> yardleydobon: application note is a more apropriate term
[00:20:13] <Casper> datasheet refer to a part specs
[00:21:19] <rue_mohr> there are spi digital potentiometers
[00:32:43] <yardleydobon> http://www.atmel.com/dyn/resources/prod_documents/doc8720.pdf
[00:32:46] <yardleydobon> found it
[00:32:56] <yardleydobon> and it is a datasheet
[00:33:26] <yardleydobon> don't you have anything better to do than incorrectly correct people over minor things?
[00:34:02] <rue_mohr> its a 24c512
[00:34:11] <rue_mohr> k, whats the big deal?
[00:35:08] <ziph> yardleydobon: The first letter of the first word of a sentence should be capitalised.
[00:35:18] <rue_mohr> ?
[00:37:00] <ziph> I'm correcting him for something minor. :)
[00:38:56] <ziph> And for a non-minor correction, while that might be how you reset that particular device, the I2C bus specification doesn't have a reset, unless you're talking about the reset of an I2C receiver when it sees a START.
[00:40:38] <rue_bed> are you making a bit-banged i2c library?
[00:41:27] <rue_bed> I tried once, I couldn't organize my state machines right
[00:43:12] <yardleydobon> No, I want to read samples continuously from a nunchuk but after some length of time I get a bus error. I want to see if I can reset the bus to a good state by clocking out the data that remains in the nunchuk.
[00:43:59] <ziph> The bus itself should reset after a START.
[00:44:36] <ziph> If the device is still messed up then you might have to do something device specific.
[00:45:25] <ziph> I take it if you issue a new START/address after the bus error it doesn't work?
[00:45:37] <yardleydobon> I can reset it by gating power to it but that takes milliseconds. I think trying this procedure is worth a shot.
[00:45:56] <yardleydobon> It's holding SDA low.
[00:46:43] <ziph> Permanently?
[00:47:25] <yardleydobon> Yes
[00:49:17] <ziph> Is there some I2C register to reset it?
[00:49:42] <rue_bed> it screw up after 512 samples?
[00:50:28] <rue_bed> yard, you mean you can clock all you want and the sda stays low?
[00:51:04] <rue_bed> iirc isn't that an 'I'm busy' ?
[00:51:18] <yardleydobon> No, I haven't tried this procedure yet. I was just trying to find it.
[00:51:19] <ziph> No, holding SCK low is "I'm busy".
[00:51:35] <rue_bed> hmm, its been a while
[00:51:39] <yardleydobon> Yes that's clock stretching.
[00:52:18] <ziph> Yeah, that's why I prefer SPI, it doesn't take a day to reload your brain with the queer details. ;)
[00:53:27] <yardleydobon> I need to do more investigation before I try to get any more specific help. Happy Halloween!
[01:22:04] <h4x0r`> people, the uc3 avr's.. 33mhz/66mhz, usb, ethernet, hmmm...
[01:22:37] <h4x0r`> how different would it be to write code for them, as opposed to the 8bit architectures..
[01:53:20] <ziph> Hmmm.
[01:53:27] <ziph> 160 components to solder on to this board.
[01:57:22] <inflex> quit complaining
[01:57:24] <inflex> :p
[01:59:59] <ziph> :)
[02:00:21] <ziph> The problem is it will take me longer to hand solder them than it would to drive to the prototyping place and have them paste them on. :)
[02:00:56] <h4x0r`> lol
[02:02:00] <ziph> inflex: The v-grooving was a bit shallow.
[02:03:47] <inflex> ugh, speaking of which, my PCBs STILL haven't arrived
[02:04:01] <inflex> they've been spinning around at the depot in Syd since the 26th !!!!!
[02:05:14] <h4x0r`> hrrm.. whats this design for ?
[02:07:38] <ziph> Which design?
[02:36:51] <sedres> I've DB-25 based cable -
http://imagebin.org/181697 - with DB-25 male connector at one end and other end with two connector, one of which is DB-25 female -
http://imagebin.org/181699 . Can I ise this cable's DB-25 connectors as parallel programmig ISP, parport?
[03:42:04] <theBear> those pictures won't load.... the 3rd connector is a db-9 right ? that would mean it's an old serial/modem cable and only has just enough wires for 9 pin serial, so it wouldn't work for a parallel isp cable
[04:04:50] <h4x0r`> sub freaks
[04:17:14] <amee2k> rue_bed: yeah, seen the pics
[04:17:39] <amee2k> sorry, i was asleep before, which probably means you are now :P
[04:28:13] * doublebeta hopes that nick is purely sarcastic.
[04:31:04] <h4x0r`> its a badge of honour
[04:55:35] <ziph> wtf
[04:55:52] <ziph> A bug tracker configured to just automatically close bugs if they aren't dealt with in 14 days.
[04:56:53] <h4x0r`> lol
[04:57:03] <h4x0r`> which distro? or what application ?
[04:57:13] <ziph> PdfSharp
[05:03:17] <h4x0r`> oops
[05:03:46] <amee2k> ziph: i tried a lot of stuff on the xtal tester. looks like it works for 4-20MHz ones
[05:03:50] <h4x0r`> Portland State Aerospace Society - <www.psas.pdx.edu>
[05:04:03] <h4x0r`> they have a nice library of eagle parts
[05:04:05] <amee2k> none of the watch crystals seem to work on it though
[05:04:15] <amee2k> and my only 24MHz crystal doesn't either
[05:04:52] <ziph> amee2k: Hmm, interesting. You could fiddle with the value of the top C a bit.
[05:05:22] <amee2k> i see, what range of values would be interresting for this?
[05:06:18] <ziph> amee2k: Hmmmm, you could try 1000pF for the top again.
[05:06:40] <amee2k> think i'll socket the top cap in the final version then
[05:07:20] <ziph> amee2k: It'll be because the tank Q is too low to produce an overall gain > 1.
[05:08:06] <amee2k> what i still can't get through my head is how a tank alone can have gain >1
[05:08:30] <ziph> amee2k: A perfect tank has infinite gain. :)
[05:08:57] <amee2k> then i have a different concept there
[05:09:14] <amee2k> in my understanding, a perfect tank has no loss and so has a gain of 1
[05:10:15] <amee2k> so what am i missing
[05:11:43] <ziph> It's the particular arrangement it is in.
[05:12:11] <ziph> One of the limits of Q in oscillator design is that you start destroying components or overbiasing varactors when the Q gets too high.
[05:12:47] <ziph> I'll dig up the analysis of it I did a while back if I get a chance later on.
[05:12:55] <amee2k> yeah, because the oscillator keeps feeding more power into the tank than can get out?
[05:14:17] <ziph> I'm not sure that's exactly how you'd look at it. The oscillator voltages tend to be limited because of non-linearity (hitting the limits of the amplifier at the worst case).
[05:14:31] <ziph> The voltages within the feedback network can be much higher though.
[05:16:03] <amee2k> my high level view of a tank is a frequency-dependant attenuator. in a feedback loop, that starts to oscillate at the frequency of least attenuation
[05:16:47] <amee2k> for lower level considerations, i always think of the LC equivalent circuit that keeps ringing. the more power a driver feeds into it, the higher the amplitude
[05:17:19] <amee2k> and unless you vent it into an antenna or other signal sink, it'll only loses power through component losses
[05:18:37] <ziph> amee2k: It oscillates at 180 degree loop phase shifts where the loop gain is >1.
[05:18:57] <ziph> amee2k: With most feedback networks the gain of the feedback is < 1 and the amplifier is > 1.
[05:19:36] <ziph> amee2k: However with an emitter follower amplifier the voltage gain of the amplifier is < 1.
[05:19:51] <ziph> amee2k: Therefore the feedback network must provide the actual loop gain. :)
[05:23:23] <ziph> amee2k: Another example might help; if you do: AC Voltage Source --- L -- x -- C ---- GND
[05:23:37] <ziph> amee2k: The voltage across the L C circuit is just the AC Voltage.
[05:23:45] <amee2k> mmh, wait so we're talking voltage gain here, not power gain?
[05:24:54] <ziph> amee2k: But the voltage between the two (where the x is) is infinite at resonance (and typically much higher than the AC V).
[05:26:48] <amee2k> hmm i see... i think. if you find that analysis paper i'd love to read it :)
[05:27:11] <ziph> amee2k: It wasn't a paper, I did it myself.
[05:27:30] <amee2k> well, pdf then
[05:28:02] <amee2k> are there oscillator circuits where the feedback part is less critical?
[05:28:23] <amee2k> for a tester this should work with as wide a range of xtals as possible
[05:29:25] <ziph> Yeah, there are a bunch of other topologies with gain in the amplifier.
[05:31:15] <amee2k> which ones, for example?
[05:31:37] <ziph> Practically all of them. ;)
[05:31:43] <amee2k> lol
[05:31:47] <ziph> I don't know a nice simple one to recommend.
[05:32:18] <ziph> If you didn't want to delve in to network analysis you could play with an AC simulation in Spice to tweak values.
[05:32:44] <ziph> Just with an AC voltage source, the two caps, and a crystal.
[05:33:00] <amee2k> i wouldn't know where to start analysing that oscillator circuit anyway, i'm not /that/ good :P
[05:33:11] <ziph> You just need to analyse the feedback part.
[05:33:26] <amee2k> never had to do something like this on paper before
[05:59:45] <amee2k> http://electroschematics.com/942/crystal-tester/ << hahaa... thats the same circuit as rue had
[06:01:23] <inflex> fnarrr... don't you hate it when you can't do a project on an 8-pin chip :(
[06:01:28] <inflex> makes you wish for a 10 pin one :(
[06:01:36] <amee2k> hehe
[06:01:42] <amee2k> 14 pin is too big?
[06:03:58] <inflex> you thinking the Tiny24 ?
[06:04:08] <amee2k> for example
[06:04:15] <amee2k> depends on what you want, i suppose
[06:04:49] <amee2k> i just use mega8 for most of my stuff. thats got enough pins for lots of stuff >_>
[06:05:14] <inflex> yeah, that's where I think I'll end up... just a bit of a PITA given that the M48/88 is slightly physically larger than I wanted in TQFP
[06:05:52] <amee2k> extreme miniaturization fetish?
[06:05:58] <h4x0r`> emf
[06:06:07] <h4x0r`> smdfreak
[06:06:13] <h4x0r`> like moi
[06:07:03] <amee2k> size is a consideration but if i have to make a trade-off i tend to favour easy construction
[06:08:02] <h4x0r`> idont blame u
[06:08:08] <inflex> SMD pick & place tends to be fine so long as it's over 2x2mm
[06:08:44] <amee2k> i really hate packages that have the pins on the underside like qfn that seems to get popular for all kinds of shit
[06:08:44] <h4x0r`> soic8 covers that doesn't it?
[06:08:49] <h4x0r`> only just iirc
[06:09:06] <amee2k> soic is okay, and qfp variants are about my lower limit
[06:09:27] <amee2k> smallest i've done before was 0805 with a hand iron >_<
[06:09:41] <h4x0r`> nice hehehe
[06:09:43] <inflex> MSSOP is fine for me
[06:09:50] <inflex> below that is painful though, I admit
[06:10:01] <h4x0r`> no doubt
[06:10:18] <h4x0r`> havent done much of that, personally fearing tqfp
[06:10:25] <h4x0r`> tinning the pins?
[06:10:29] <h4x0r`> fear.
[06:10:42] <amee2k> if its less than 1mm pin spacing it tends to be a pita
[06:10:55] <amee2k> pin pitch*
[06:11:11] <amee2k> if its less than 0.5mm pitch it can have The Finger(tm)
[06:11:13] <inflex> TQFP is fine
[06:11:28] <inflex> esp anything with less than 32-pin TQFP
[06:11:41] <h4x0r`> inflex, whats the procedure you use for tqfp
[06:11:48] <amee2k> h4x0r`: 0805 has a way of having the exact same pad spacing as a TO92 with straight leads has lead spacing
[06:11:50] <h4x0r`> caps*
[06:12:08] <h4x0r`> k
[06:12:22] <amee2k> i make tiny current sources for ballasting LEDs by soldering a 68R resistor between pins 1 and 2 of an LM317L
[06:12:31] <amee2k> then cutting the middle pin off
[06:13:07] <amee2k> the outter two pins are .1" apart so it nicely fits on protoboard
[06:13:38] <inflex> h4x0r`: nothing exotic... for manual soldering, just tin two legs down, then blob-flood-and-knock, or, do as many pins as you can with the solder-ball method, then suck up the rest with wick
[06:13:50] <amee2k> 11:40 < amee2k>
http://ompldr.org/vNXN4ZA/cimg4989b.jpg << "one and a half parts led constant current source"
[06:14:01] <inflex> personally I find the blob-and-knock method works great, so long as you do them as the first parts ont he board
[06:14:06] <inflex> else, really, I just use reflow :)
[06:14:32] <h4x0r`> inflex, i understand that
[06:14:52] <h4x0r`> tqfp needs tinning first tho before reflow anyway isnt that right ?
[06:15:53] <inflex> no, never had to, just put the paste down via the stencil and place the part on... into the oven.
[06:16:15] <amee2k> i usually tin both pins and pads, then gently head it with hot air until it starts to flow and sucks the part down
[06:16:41] <h4x0r`> cool
[06:16:59] <amee2k> i don't populate full boards though, only rework or a few parts in an otherwise PTH design
[06:17:04] <inflex> amee2k: oh sure, if you're doing a hot-air workstation reflow
[06:17:18] * inflex tries as much as possible to do everything SMD, saves a lot of time
[06:17:47] <amee2k> i would too if i were doing more stuff with printed boards
[06:18:01] <amee2k> leaded parts are better for ghetto construction
[06:21:42] <amee2k> i need to get my shitty computer fixed >_<
[06:22:17] <amee2k> poll: debian 6 or ubuntu 11.10 for my desktop?
[06:23:05] <mrfrenzy> ubuntu. debian is more aimed at servers with it's slow release cycle
[06:24:28] <inflex> meh, either work
[06:24:33] <inflex> I've got U11.10 on
[06:24:42] <inflex> and I'm using a D5 server (via Xen)
[06:24:53] <amee2k> yeah, thinking ubuntu too but its been pissing me off ever since 9.10 or so :/
[06:25:30] <amee2k> and i'm certainly not going to use that unity accident
[06:25:40] <mrfrenzy> ofcourse you can run debian. just be prepared to pull in some stuff from unstable or backports
[06:26:16] * inflex does U11.10 + Fluxbox
[06:26:20] <amee2k> any alternatives? considering i liked pre-9.x ubuntu?
[06:26:34] <inflex> going with FB means I don't have to deal with any of that CPU/memory whoring overkill glossy UI
[06:27:19] <amee2k> i like having a fully featured DE, i just don't like that all of them these days come with lube, err gloss overkill
[06:27:42] <amee2k> if you get my drift :P
[06:28:16] <amee2k> inflex: which reminds me, how well does fluxbox handle multihead?
[06:38:17] <RikusW> h4x0r`: I use a wide tip (4.5mm Plato EW-103) and liquid flux for soldering tqfp
[07:00:07] <h4x0r`> interesting ;)
[07:00:32] <RikusW> I soldered 120 32u2's using that tip
[07:01:16] <Tom_itx> that's all?
[07:01:17] <RikusW> it works very well, it sucks up solder bridges too ( the 90 degree corner)
[07:01:36] <Tom_itx> :)
[07:01:47] <RikusW> Tom_itx: you did some more of yours ?
[07:01:54] <RikusW> reflowed I guess ?
[07:02:01] <Tom_itx> i reflow now yes
[07:02:19] <Tom_itx> test and clean up spots with the iron
[07:02:51] <RikusW> how many 32u2's have you soldered to date ?
[07:03:00] <Tom_itx> i did a few hundred before by hand though
[07:03:03] <Tom_itx> no idea
[07:04:12] <Tom_itx> you really should consider that if you're making alot
[07:04:57] <RikusW> I have to sell my boards first before considering making anymore...
[07:05:18] <Tom_itx> that's the jtag thing right?
[07:05:21] <amee2k> ziph: what i just discovered as rather peculiar is that xtals >10MHz seem to make the test oscillator work at half the rated frequency of the crystal
[07:05:25] <amee2k> is that normal?
[07:06:12] <RikusW> Tom_itx: its a combo STK500 / JTAGICE mkI / UART / bootloader for custom app , all which can be loaded at the same time
[07:06:41] <RikusW> HVPP does need an external 12V circuit
[07:06:51] <Tom_itx> right
[07:07:15] <RikusW> the modules are loaded high next to the bootloader
[07:08:05] <RikusW> will probably add dW support next, but will have to use a custom protocol.... jtag mkII is too complicated and resource hungry....
[07:09:36] <amee2k> i put my scope as frequency counter between ground and the emitter of the oscillator transistor to measure it
[07:32:47] * amee2k . o O ( where in fuck's name did my transistor junk box go >-< )
[07:32:52] <amee2k> >_<*
[07:33:39] <h4x0r`> rofl
[07:33:48] <amee2k> what.
[07:35:56] <h4x0r`> LOL
[07:36:10] <amee2k> o.O
[07:36:29] <amee2k> do we get to share the punchline?
[07:38:58] <amee2k> how small do they make GPS trackers these days again?
[07:39:26] <h4x0r`> 3cm
[07:39:32] <h4x0r`> give or take
[07:39:36] <amee2k> i want some
[07:39:38] <h4x0r`> too noisy otherwise
[07:39:41] <h4x0r`> lul
[07:39:47] <amee2k> junk boxy boxy boxy boxy come to papa... >_<
[07:39:57] <h4x0r`> LULZ
[07:41:03] <amee2k> i had it on my desk from last night... this morning i got up... cleaned up some random stuff... must have put them somewhere away without thinking about it
[07:41:55] <amee2k> from the size of the stack of known boxes, i must be missing two... bipolar transistors and hm... not sure which other one
[08:04:21] <h4x0r`> a question
[08:04:32] <h4x0r`> avr (5v) and 2.8v ic
[08:04:45] <h4x0r`> avr tx 5v, ic tx 2.8v cmos.
[08:04:58] <h4x0r`> is it viable to do the following..
[08:05:36] <h4x0r`> avr Tx (5v) -> ft232rl (2.8v CMOS) -> ic Tx
[08:07:34] <moe3> i have a question about ADC in AVRs
[08:08:30] <RikusW> h4x0r`: seems like you need a level translator
[08:08:31] <moe3> when using normal 10bit ADC. I have a resolution of 5V/1024=0,00488
[08:09:14] <h4x0r`> RikusW, yep, but im reading the ft232rl datasheet to see if its capable of stepping in *as one*
[08:09:19] <amee2k> h4x0r`: i have no idea what that last line meant, but for interfacing different logic levels, you need a translator like rikus said
[08:09:20] <moe3> but when setting AREF to for example 3,5V. Will the resolution be 3,5v/1024=0,00341V ??
[08:09:52] <amee2k> if the sending side is on the higher voltage side, thats as simple as a voltage divider
[08:10:00] <karlp> moe3: yes.
[08:10:01] <Tom_itx> h4x0r` how many bits?
[08:10:27] <amee2k> moe3: yes, the ADC just linearly divides the reference voltage
[08:10:41] <Tom_itx> 2.8v i think is too low to be reliable for 5v high logic
[08:11:17] <amee2k> i vaguely remember someone had a cool level translator with a single small signal mosfet
[08:11:25] <amee2k> that worked for both directions
[08:11:35] <h4x0r`> hmm
[08:11:43] <Tom_itx> takes 2 i think
[08:12:28] <amee2k> http://static.electro-tech-online.com/imgcache/4056-ST-5.jpg
[08:12:31] <amee2k> i think thats the one
[08:12:58] <RikusW> yes
[08:13:00] <h4x0r`> Tom_itx, hmm...
[08:13:03] <Tom_itx> http://search.digikey.com/us/en/products/GTL2002DP,118/568-1870-1-ND/805395
[08:13:06] <Tom_itx> 2 bit
[08:13:12] <Tom_itx> good to 1v
[08:13:17] <RikusW> though it was 2N7002
[08:13:20] <h4x0r`> jeez, looks nice :p
[08:13:46] <amee2k> RikusW: i think pretty much any logic level fet would do
[08:13:55] <RikusW> then there is the one used by STK500, the schematics is available somewhere...
[08:14:09] <amee2k> it becomes an open drain output in one direction, and a source follower in the other
[08:14:09] <RikusW> amee2k: yes
[08:14:11] <h4x0r`> although the idea was to get avr to speak to ft232rl tx running at 2.8v cmos, so it can pass the signal straight onto the ic..
[08:14:19] <h4x0r`> in theory, feels respectable...
[08:14:34] <RikusW> amee2k: I even used normal bipolar and it seems to work
[08:14:45] <amee2k> nice
[08:14:46] <RikusW> just put a 3k resistor on base
[08:14:53] <h4x0r`> (the avr will snd the signal to ft232 at its usual 5v tho..)
[08:14:55] <amee2k> yeah, makes sense
[08:15:05] <h4x0r`> hm
[08:16:43] <amee2k> i didn't check but isn't an ft232 pretty much the most expensive choice of level translator there is?
[08:17:10] <Tom_itx> probably
[08:17:15] <h4x0r`> probably
[08:17:20] <Tom_itx> either the mosfet or my selection would be a better choice
[08:17:47] <h4x0r`> oh cool i didnt see it
[08:18:02] <Tom_itx> made in anything from 2 to xxx bits
[08:19:10] <RikusW> GTLxxxx is a lot of mosfets in a chip.... like amee2k's link
[08:19:35] <Tom_itx> yup
[08:20:19] <Tom_itx> if you need more than a few, cheaper because you don't need all the resistors and takes less board space
[08:20:27] <RikusW> seems loose mosfets might be cheaper too
[08:20:43] <Tom_itx> add up the board space too
[08:20:53] <RikusW> hmm
[08:20:59] <amee2k> well, what i meant was as long as i have anything else i wouldn't waste a slow and expensive 232 chip on translating logic levels
[08:21:08] <Tom_itx> cheapest i've found that is $5 sq in in small quantity
[08:22:18] <moe3> adc:when setting AREF to for example 3,5V. Will the resolution be 3,5v/1024=0,00341V ??
[08:22:43] <wazzah> .
[08:22:55] <Tom_itx> moe3 i believe so
[08:23:13] <amee2k> dejavu= o.O
[08:23:18] <h4x0r`> .
[08:23:26] <h4x0r`> gents
[08:23:44] <amee2k> in the next room
[08:25:27] <h4x0r`> Tom_itx, true, its the safer, perhaps more proper way to go.. quite happy to see an soic8 level shifter :)
[08:26:32] <inflex> any of the pin-compat mega series chips now support differential ADC ?
[08:26:57] <inflex> (but more importantly, with Gain!!!!
[08:27:04] <rue_bed> unidirectional is easy, 5V tollerant 3.3V chip
[08:27:37] <rue_bed> the m32 has diff with gain
[08:27:54] <rue_bed> pin compat with what?
[08:29:16] <inflex> m48/88/168
[08:29:38] <rue_bed> dunno about those
[08:29:54] <amee2k> i don't think so
[08:30:16] <amee2k> within the same series, the only difference is usually just amount of memory
[08:30:33] <rue_bed> he's looking of rone pin compat with ...
[08:30:36] <amee2k> which is something between small and ridiculously small anyway
[08:31:24] <amee2k> oh, okay
[08:35:14] <inflex> oh well, no matter
[08:35:25] <inflex> it just limits me on the precision of the ADC
[08:38:34] <inflex> hate that the AVR ADC's external Vref cannot be below 1V
[08:38:38] <inflex> anyone know the technical reason why?
[08:39:44] <inflex> Would have been very nice to use a 0.1V reference instead of 1V
[08:58:55] <h4x0r`> a little upset with this datasheet, ive got an ic needing to accept 2.8v cmos levels
[08:59:12] <h4x0r`> and check the specs on the pins for current
[08:59:14] <h4x0r`> Output Current 1mA
[08:59:14] <h4x0r`> Input Current 1uA
[08:59:48] <h4x0r`> how am i supposed to negotiate a 2.8v/1uA feed?
[09:00:02] <Tom_itx> tread lightly
[09:00:30] <h4x0r`> i dont understand the statement Tom_itx
[09:00:31] <inflex> drop a few 1Mohm resistors in :)
[09:00:34] <amee2k> i suspect input current doesn't mean "output pin sink current capability" but "input pin bias current"
[09:00:42] <Tom_itx> it should only use the current it needs
[09:00:47] * inflex goes silly and actually asks on AVRFreaks about the Vref question
[09:00:54] <Tom_itx> you should only power something on it's output it can handle
[09:01:24] <h4x0r`> Tom_itx, i get the reference to current draw
[09:01:25] <Tom_itx> later.
[09:01:40] <h4x0r`> :)
[09:02:00] <h4x0r`> ah np
[09:02:29] <h4x0r`> the putput is rx, so it'll be fed back to the controller
[09:02:45] <h4x0r`> hungreh
[09:08:24] * inflex wonders how strict this 1.0V minimum thing is... perhaps tomorrow I should just rig up someting to push 500mV in on Vref and see if the ADC works as required... though a bit concerned about doing stuff out-of-spec
[09:08:46] <h4x0r`> fair comment :p
[09:16:16] <amee2k> http://img528.imageshack.us/img528/7030/fusereplacement.jpg
[09:16:16] <amee2k> ...
[09:18:05] <h4x0r`> lol ;p
[09:19:55] <Kevin`> a guide for how much tinfoil to use for specific ratings would be useful sometimes
[09:20:48] <Kevin`> those must have been fun to test
[09:21:01] <inflex> love it being "slow blow"
[09:21:44] <inflex> mmm... the USA must be waking up ... getting a heap of spam surging
[09:23:08] <_Shurik_> heyo heyo
[09:23:36] <h4x0r`> inflex, hehehehe
[09:27:50] <inflex> hrmppfff.... it's annoying when the datasheet tells you "You shall not pass!" and is actually correct.
[09:28:31] <_Shurik_> Hey guys, anyone programmed AtTiny13a via DAPA?
[09:30:42] <rue_mohr> http://pastebin.com/UiemLs5g <-- could someone message sedres and tell him what kinda programmer he has?
[09:35:34] <amee2k> wow, ubuntu 11.10 isn't detecting my second monitor
[09:35:47] <amee2k> no dual head support is embarrassing
[09:38:31] <inflex> hrmm
[09:38:37] <inflex> strange, I'm running dual-head here on U11.10
[09:39:03] <inflex> amee2k: what gfx card? I was running it out of this cheap ATi card.
[09:39:18] <amee2k> nvidia gf8600gts
[09:39:33] <amee2k> it has installed the nvidia driver on its own, but it doesn't detect the second monitor
[09:39:56] <inflex> mm.. no idea what to suggest - if it's running the nVidia driver then I guess all you can do is fiddle with it
[09:40:37] <amee2k> i wish, i haven't found nvidia's maintenance tool yet
[09:42:12] <inflex> ah okay, no idea what it is... best of luck *laugh*
[09:42:29] <inflex> times like that I pop open the terminal and go "nv[tab tab]
[09:42:51] <amee2k> i haven't found a terminal yet either
[09:42:56] <inflex> O_o
[09:43:01] <inflex> ctrl-alt-T
[09:43:22] <amee2k> mmh, that worked
[09:43:36] <inflex> np
[09:43:51] <amee2k> looks like a kinda ugly xterm or something though
[09:44:05] * amee2k <3 gnome-terminal from v2
[09:45:24] <amee2k> nope, nvidia-settings only detects one monitor either
[09:56:08] <amee2k> this is pointless if i can't get the dualhead thing to work
[10:02:02] <inflex> Check the forums - I'm sure it'll be up there with a solution
[10:02:05] <inflex> esp if it worked before
[10:06:39] <amee2k> inflex: some time ago i got rid of the general notion that things which worked before still generally do so in later versions
[10:07:07] <amee2k> three years ago my laptop's graphics drivers didn't crash when resuming from suspend or hibernation
[10:07:35] <amee2k> last year, the x server on xubuntu used to have an xorg.conf and actually honor its contents
[10:07:51] <amee2k> three days ago i used to have functional dualhead
[10:08:01] <ziph> inflex: I now have an application that loads up a PCB, creates a simple overlay, and lets me interactively highlight all components of a certain value/type for soldering. :)
[10:09:25] <amee2k> inflex: i've been searching for a while, but other than some generic blah on dualhead nothing is coming up
[10:10:12] <amee2k> looks like its going to be deb6 after all then
[10:11:18] <amee2k> another thing that keeps annoying me is that none of the popular color schemes indicate which window is focused
[10:11:47] <amee2k> other than maybe a slight shade change in the title font color
[11:05:14] <sedres> For using printer port in Ubuntu linux, is there any driver to be installed? I get proper reading for Vcc and other pins when in Windows but in Ubuntu, I get 0.08V only for Vcc
[11:06:57] <Kevin`> sedres: a printer port doesn't have a power supply pin. if you are using one of the data pins for power, ensure your application is setting it to be high
[11:14:23] <sedres> Kevin`: Well, I measured the potential difference between Vcc and GND pin and found 3.80V when in Windows and 0.08V when in linux. Do I need to power the AVR chip when using parport?
[11:15:18] <Kevin`> sedres: a parallel port does not provide power. what do you have vcc connected to anyway? it's possible your chip was getting power through the esd protection diodes on it's data pins
[11:16:11] <sedres> Kevin`: Yes, possible. Vcc is connected through ISP to AVR chip?
[11:16:29] <karlp> you actually have a parallel port, and you actually want to use it?
[11:16:31] <Kevin`> sedres: isp from what? random pins on the parallel port?
[11:16:39] <sedres> no
[11:17:35] <sedres> Your questions seems like I've limited knowledge about it, let me check the pin out of printer ports ...
[11:17:45] <Kevin`> anyway, yeah, be aware of the data pin thing. in a production device that will have it's data pins connected to stuff that's regularly higher than vcc, use a buffer chip
[11:19:17] <sedres> Kevin`: Could you link me to any reference for the same? which will detail out the types etc
[11:21:11] <Kevin`> http://www.google.com/search?q=buffer+logic+gate
[11:21:22] <Kevin`> you might want a level translator in some cases though
[11:21:37] <sedres> ok
[11:22:05] <sedres> That's not what I asked but I'll find out. Thanks
[11:23:00] <Kevin`> i'm curious why you are using a printer port without knowing about it?
[11:24:07] <sedres> I've the basic understanding of it for using it for trnasferring hex. I followed a schematic to make a parport which I could understand ...
[11:24:24] <Kevin`> any reason you aren't using a serial port instead?
[11:25:01] <Kevin`> really easy data transfer with those, and only one pin per direction
[11:25:08] <sedres> Serial port voltages are incompatible with what is required for in-system progarmming afaik
[11:25:25] <Kevin`> oh, you are using it for programming?
[11:25:39] <sedres> for ISP
[11:26:05] <Kevin`> the programmer software will probably set the port to provide power to whatever pin it told you to connect vcc too
[11:26:20] <Kevin`> (not power exactly, but a "high" signal level)
[11:29:32] <sedres> Yeah, you were right about no power supply pin from printer ports. I confused Vcc, GND pins from ISP connections of (MISO, MOSI, SCK, RST, VCC & GND).. I observed that when using usbasp, it acts as a power source hence thought so ...
[11:42:38] <amee2k> wow
[11:42:41] <amee2k> that was mature
[11:42:46] * amee2k discretely points at #ubuntu
[11:49:16] <Kevin`> ?
[11:53:11] <amee2k> Kevin`: i was talking to a dude about that the unity thing is a dick move and not intuitive or anything at all
[11:53:50] <amee2k> and some op came in and kicked us both because our discussion was not support related or something
[11:54:46] <amee2k> no wonder ubuntu is getting top scores for usability. everyone disagreeing is removed from the statistics :P
[11:54:59] <Kevin`> it could be just that, the ubuntu channel is probably the busiest one on freenode. do they have a seperate social channel?
[11:55:48] <karlp> amee2k: ubuntu are pretty sick of hearing "unity is a dick move"
[11:56:02] <karlp> either move on, or try it out as if it was brand new.
[11:56:32] <karlp> (I did both, moved to deb 6 on my machine, left ubu11.10 on the tv computer)
[11:56:48] <JanneP> I installed the new ubuntu a few whiles ago, my first linux
[11:57:00] <JanneP> with no previous experience of linux or ubuntus the unity feels quite ok
[11:57:14] <JanneP> with some quirks, but you get used to them quickly
[11:57:23] <amee2k> JanneP: that you're a newcomer is probably an advantage there
[11:57:41] <amee2k> you get to learn everything from scratch and are fine this way
[11:58:05] <amee2k> karlp: probably got something to do with the fact that they're hearing that a lot. i wonder why. they don't apparently
[11:58:19] <JanneP> yeah I've read a lot of the previous users bashing unity vs. the old gnome
[11:58:43] <amee2k> this is my first afternoon trying it myself and i find it fairly awkward
[11:59:06] <amee2k> if i didn't suspect most of that was intentional, i'd describe it as buggy, really
[11:59:53] <amee2k> no dualhead anymore, the funny icon bar thing on the left keeps disappearing on me, you can't un-maximize windows anymore
[12:00:12] <Kevin`> I haven't really tried any desktop version past the last lts
[12:00:27] <Kevin`> because I just don't like upgrading that often ;p
[12:01:43] <amee2k> my system drive died last weekend so a reinstall was in order
[12:02:13] <amee2k> torn between ubuntu and deb6 and decided to give the new ubuntu a shot
[12:03:51] <h4x0r`> amee2k, DIR is connected to vcc-A, in which direction does the flow go by default in page 1of
http://www.farnell.com/datasheets/1388651.pdf
[12:04:56] <amee2k> "flow"?
[12:06:06] <h4x0r`> yep its a transciever
[12:06:22] <amee2k> flow of what?
[12:06:29] <h4x0r`> so it accepts a differeing voltage on either side
[12:06:47] <h4x0r`> goes in at 5v, comes out at 2.5v for example.
[12:07:00] <amee2k> yeah
[12:07:08] <h4x0r`> which way the flow of data goes is controlled by the DIR pin.
[12:07:35] <h4x0r`> which is connected to vcc-A, so which way is the direction going lol
[12:08:08] <amee2k> function table is on page 2
[12:08:33] <h4x0r`> yeah but I dont get that table.
[12:08:35] <h4x0r`> lol
[12:08:41] <amee2k> low means port b -> port a
[12:09:00] <amee2k> it also has an equivalent logic diagram
[12:09:22] <amee2k> dir low means the buffer sending out of port a is enabled
[12:09:42] <h4x0r`> what defines dir low
[12:09:47] <h4x0r`> the lower voltage?
[12:10:02] <h4x0r`> or whatever vcc is connected to dir
[12:10:04] <h4x0r`> confused
[12:10:06] <amee2k> you're supposed to tie the dir pin to a voltage level
[12:10:13] <h4x0r`> yes
[12:10:20] <h4x0r`> so if i tie dir to A
[12:10:22] <amee2k> from the arrangement i'd say the DIR levels are the same as for port A
[12:10:27] <h4x0r`> then the direction is out via B ?
[12:10:56] <amee2k> you tie DIR either low or high
[12:11:05] <h4x0r`> oh.
[12:11:11] <h4x0r`> oh.
[12:11:11] <amee2k> that pin is exposed to allow an MCU to switch the direction
[12:11:26] <h4x0r`> got it mate
[12:11:29] <amee2k> so it can change a pin from input to output without rewiring it
[12:12:08] <h4x0r`> so if I gnd the pin the flow is b data to a bus
[12:12:11] <h4x0r`> yes ?
[12:12:41] <amee2k> yeah, i#d say so
[12:13:07] <h4x0r`> or if I tell the mcu to pull DIR low then b data to a bus? . kewl :D
[12:13:25] <h4x0r`> yep sounds right, they didnt really say much about L and H grr
[12:14:58] <amee2k> well, low is generally the same level as the ground pin
[12:15:12] <h4x0r`> oh yep
[12:15:13] <amee2k> and high is something else :)
[12:15:14] <h4x0r`> i see it now
[12:15:23] <h4x0r`> The logic levels of the direction-control (DIR) input activate either the B-port outputs or the A-port outputs.
[12:15:29] <h4x0r`> 0/1
[12:15:51] <h4x0r`> L = 0 and H = 1
[12:16:07] <amee2k> yep
[12:16:18] <h4x0r`> awesum mate :)
[12:18:50] <h4x0r`> amee2k
[12:19:33] <h4x0r`> is it plausible to run both tx and rx through the same transciever
[12:19:53] <amee2k> sometimes
[12:20:00] <h4x0r`> hmm
[12:20:21] <h4x0r`> i would need to pulldown/pullup when sending/receiving.
[12:20:24] <h4x0r`> make sense?
[12:20:25] <amee2k> one side being driven high and the other driven low makes as little sense as without a transciever inbetween
[12:20:39] <h4x0r`> but the traces would run into each other.. herm.
[12:21:01] <amee2k> i think what you mean is a bidirectional level shifter, that figures out which side is the input and which the output on its own
[12:21:22] <h4x0r`> wow :D
[12:22:39] <h4x0r`> if only farnell had one for such low voltage levels.
[12:22:43] * grummund jumps straight in without reading the scrollback ;)
[12:22:44] <h4x0r`> 2.8v/5v
[12:22:53] <amee2k> brb, gotta take care of something
[12:22:54] <h4x0r`> wazzup grummund :D
[12:22:56] <grummund> you might be interested in this...
[12:22:58] <grummund> http://www.daycounter.com/Circuits/Level-Translators/Level-Translators.phtml
[12:22:58] <h4x0r`> np, thanks mate
[12:24:26] <h4x0r`> grummund, nice, thankyou
[12:25:17] <h4x0r`> a fancy transciever could do it.. but i needed to read up using fets for shifting aswell, so cool find & ty much
[12:44:02] <amee2k> i think ubuntu just died for me
[12:45:05] <specing> ubuntu is dying all the time, everywhere
[12:45:14] <amee2k> i know
[12:45:34] <amee2k> but you're not even allowed to say that out loud without getting funny looks like you're a social reject
[12:45:56] <amee2k> and that attitude right now was like the cherry on top
[12:47:13] <amee2k> if you say some window manage in debian sucks, noone gives a fuck. say unity is a dick move and you're a heretic
[12:49:00] <amee2k> wondering if i should give mint a shot now
[12:51:36] <specing> amee2k: once I went to a local CS faculty... 3/4 of computers in the lobby were crashed
[12:51:42] <specing> amee2k: ubuntu there
[12:51:58] <amee2k> wow
[12:52:09] <specing> Some were fork bombed, some nuked by an unknown method
[12:52:20] <specing> but still dead
[12:52:22] <amee2k> ...
[12:52:41] <amee2k> thats ubuntu users for ya
[12:52:59] <specing> I can imagine there is a poor humanoid out there constantly rebooting those machines
[12:53:24] <amee2k> 555 on the reset button comes to mind
[12:53:57] <specing> :D
[12:54:27] <specing> maybe they even have a hardware watchdog there
[12:55:50] <ziph> Niiice, Digikey. shipping a 0R instead of a 1nF
[13:00:13] <amee2k> only missed by one :P
[13:13:05] <Kevin`> ziph: a little superconducting rod is kind of cool, though
[13:13:51] <ziph> Kevin`: It's a 5%, too, so half of them are negative resistance. ;)
[13:14:18] <Kevin`> nice
[13:17:16] <tosmo> is there a complete 'fusemap'? the documentation explains the meanings of varius fuse bits, but i didn't find the complete 'layout' where those bits lie in the (i think) total 2 bytes. when writing fuses with avrdude i need to specify the bytes rather than single named bits
[13:18:13] <tosmo> what i'm looking for is basicly a sketch similar to those of the timer or adc control registers, just for fuses
[13:18:16] <LoRez> yes, look for the fuses section of the datasheet.
[13:18:18] <karlp> tosmo:
http://www.engbedded.com/fusecalc/
[13:18:29] <karlp> but the maps are in the datasheet
[13:18:52] <tosmo> hm didn't find it i'll have another look
[13:19:50] <Casper> tosmo: memory -> fuse
[13:19:52] <Casper> and it's 3 bytes
[13:19:58] <Casper> and not all bits are used
[13:23:37] <tosmo> Casper: thnanks, found it
[15:16:11] <magnus_> in a system with no function recursion, would it be true to say that the sum of all function frame sizes is an upper bound of stack usage?
[15:17:07] <Tom_itx> what about memvars
[15:17:45] <magnus_> memvars?
[15:18:13] <Tom_itx> code segment and data segment
[15:18:16] <Tom_itx> are they combined?
[15:18:26] <magnus_> they are not in the stack, are they?
[15:18:30] <Tom_itx> oh
[15:18:34] <Tom_itx> probably not
[15:18:59] <magnus_> im just trying to figure out if this can be true, because the sum of my function frame sizes right now is only 97 bytes
[15:19:10] <magnus_> which is very little, given the amount of functions and code I have
[15:19:56] <magnus_> many of my functions have 0 frame size, which I guess is possible if there are enough available cpu registers at the point when it is invoked
[15:20:35] <magnus_> obviously, for a function that declares a larger variable, such as "char mystr[50]", i will see (correct) frame size of 50 bytes
[15:26:12] <magnus_> however, to contradict this i have painted my entire stack with a known value during initialization. i can then calculate the maximum stack size by checking how much of this paint is left at some point in execution. it always exceeds 97 bytes
[15:26:17] <magnus_> which i find weird
[15:26:33] <magnus_> (and no, im not using any heap allocations whatsoever in my application)
[15:36:46] <Casper> avr-gcc always use a few bytes for the function
[15:36:54] <Casper> around 6 bytes
[15:37:05] <Casper> which it free when it leave it
[16:05:15] <magnus_> Casper, how can i tell how the stack is used by gcc except for variables?
[16:06:29] <Casper> no official way because it depend on the code flow and conditions and all
[16:06:57] <magnus_> so, there is an unspecified amount of data put on the stack for each function call?
[16:07:09] <Casper> it depend on the function content
[16:07:21] <magnus_> depend in what way?
[16:07:25] <RikusW> gcc have 2 stacks
[16:07:39] <RikusW> the Y pointer is used for parameters
[16:08:01] <RikusW> return address on the normal stack
[16:08:20] <RikusW> and r24 r25 is usually used for the first parameter
[16:08:38] <Casper> gcc try to optimise, if it think that some register should be "flushed" to leave space for more usefull data, it save, then reuse, then reload when it exit
[16:09:05] <RikusW> push and pop..
[16:09:22] <Casper> but last time I checked it was around 6 bytes
[16:09:25] <Casper> ya
[16:10:04] <RikusW> avr-libc docs decribe register usage, stack too I think
[16:10:16] <magnus_> ok, will check that. thanks
[16:15:42] * Casper wonders if he have some 6000+rpm motors...
[16:26:52] <ys0> oi
[18:56:35] <rue_mohr> did anyone msg sedres with his programmer help?
[19:28:52] <pld> oh flippen heck... hate it when I cannot talk on most IRC cahnnels
[19:29:00] <pld> (because I'm not auth'd)
[19:29:25] <pld> brb
[19:29:27] <rue_mohr> did we do that here?
[19:29:48] <inflex-> that's better
[19:30:15] <h4x0r`> dudes, lil mosfet q, when shifting from 5v to 2.8v using a mosfet, which voltage should be used at the gate
[19:30:39] <h4x0r`> 2.8?
[19:30:41] <inflex-> h4x0r`, normally the gate is on one side of the shift
[19:30:49] <inflex-> oooh, I see whatyou're saying
[19:31:00] <inflex-> just go look at the spunfark version
[19:31:27] <h4x0r`> inflex-, sparkfun version of?
[19:32:04] <inflex-> mosfet level shifter
[19:32:29] <inflex-> http://www.sparkfun.com/products/8745
[19:32:35] <inflex-> check their schem
[19:36:08] <h4x0r`> its a nice schem, but i dont know which is gate drain and source is ground i think lol..
http://www.sparkfun.com/datasheets/BreakoutBoards/Level-Converter-v10.pdf
[19:36:44] <inflex-> gate is the one going to LV
[19:36:46] <h4x0r`> nice converter too :)
[19:36:49] <h4x0r`> hrm
[19:37:22] <inflex-> source is TX_LV, drain is TX_HV
[19:37:28] <inflex-> they're N-ch MOSFETs
[19:38:10] <inflex-> it's a cute configuration really, a nice example of manipulating inherent characteristics of a device in a new way
[19:38:21] <inflex-> like how the body-diode even plays an important part too
[19:38:41] <h4x0r`> ya i really like it, i just dont know how im getting the 2.8v from LV, do I need to supply it with a 2.8v power source?
[19:39:00] <inflex-> yes
[19:39:51] <h4x0r`> ok i think i see it
[19:40:16] <h4x0r`> LV = 2.8v supply to the 2.8v tx line, concept win?
[21:51:12] <TeknoJuce> any one alive?
[21:51:50] <Casper> hello, we are not available for the moment. Please leave your message after the beep
[21:52:04] <TeknoJuce> damn ghosts!
[21:52:37] <Casper> <old never ending music from tape based machine>
[21:53:22] <TeknoJuce> looking for a method to detect 5 swinging players on a fooseball table and which one connects with the ball any ideas?
[21:53:34] <TeknoJuce> thinking something with embedded magnets or something not sure
[21:54:19] <Casper> magnet could be a way, but... consider that the position need to be accurate
[21:55:06] <TeknoJuce> have you seen anything like this done before or anything that spings to mind with a simtech process?>
[21:57:41] <TeknoJuce> hmm wonder if one of those touch grids for the entire board would be something to look at think I saw something like that
[22:08:04] <Casper> what do you want to do exactly? just know when someone swing? or when it kick the ball?
[22:19:55] <TeknoJuce> yes
[22:20:29] <Casper> A or B can not be answered by "yes"
[22:20:48] <TeknoJuce> but if I put contact sensors in the players not sure if that will go over very well if they spin the polls with wires going though them
[22:21:05] <TeknoJuce> *pole
[22:21:29] <TeknoJuce> ok when some one kicks the ball which exact man on the bar of players
[22:24:11] <TeknoJuce> also the poles pull out and in so the wire thing is really complexing to my brain how would that get mounted with not a lot of wire slack all over the place to break off
[22:28:38] <Casper> slip rings
[22:28:49] <Casper> but good luck...
[22:29:03] <TeknoJuce> haha oh sure just give up!
[22:30:20] <Casper> the easiest way I can think off is the hardest one :D
[22:30:28] <Casper> battery powered wireless :D
[22:32:04] <TeknoJuce> yeah thats a lot of batteries
[22:33:34] <TeknoJuce> guessing powercast would be pricey as well
[22:37:55] <TeknoJuce> only 38$ :S
http://ca.mouser.com/Search/ProductDetail.aspx?qs=kw8vCQu8LVcM8eL3CmnLwQ%3d%3d
[22:47:15] <Casper> hmmm icecream... movie time, bbt! (be back tomorrow)
[22:47:55] <h4x0r`> npcut
[22:48:00] <h4x0r`> np c u then
[22:48:01] <h4x0r`> ()
