Back
[01:15:51] <inflex> hmm... time to look for a serial-USB converter using an AVR
[01:15:57] <inflex> or a very cheap clone
[01:34:26] <inflex> bleh, might just go with a prefabbed chpi
[01:36:29] <inflex> can you source things like the PL2003 chips?
[03:54:38] <amee2k> inflex: can't find it on my end, whats that one do?
[03:59:32] <h4x0r`> guys, trying to work out why these 1k resistors go here,
http://i.imgur.com/R8ya6.jpg
[03:59:43] <h4x0r`> just brushing up
[04:00:16] <amee2k> R2 and R5 are base resistors to limit drive current
[04:00:30] <amee2k> R1 and R4 are pullups to provide the output high level
[04:01:02] <h4x0r`> ok right
[04:01:09] <h4x0r`> can you explain a bit mroe?
[04:01:36] <h4x0r`> r2 & r5 limit the forward drive current?
[04:02:13] <amee2k> well, a BJTs base-emitter junction electrically behaves like a diodes, so driving it with a voltage source will yield very unstable drive current and usually kill the transistor due to overcurrent
[04:02:31] <h4x0r`> HMM
[04:02:34] <amee2k> so the base resistor is used for current limiting, much like the series resistor for LEDs
[04:02:45] <h4x0r`> ok but why the value 1k?
[04:02:52] <h4x0r`> its just like, to tighten the current ?
[04:03:28] <amee2k> its convenient, widely available and probably in the junk boxes of most people, and yields sufficient drive current
[04:04:28] <amee2k> pretty much anything from >10k down to 470 or even less will work here. all you need is enough base current to make the transistor pull the output low
[04:05:15] <amee2k> the pullup is used to make the output settle on the desired high voltage level when the transistor is not conducting (i.e. turned off)
[04:05:48] <amee2k> when the transistor turns on, it conducts current to ground and pulls the output down to ~0.3V, which is low enough to count as low
[04:05:52] <h4x0r`> lol the reason it dint make sense was i didnt even look at the transistor, i didnt realise it was one.
[04:06:03] <amee2k> 0.0
[04:06:18] <h4x0r`> toomany distractions here atm ol
[04:06:25] <amee2k> hehe
[04:07:20] <amee2k> my rule of thumb is, everything within a diode drop of ground counts as low. everything within two diode drops of the supply voltage counts as high. everything inbetween is probably high but may be anything. everything outside the rails is a bad idea
[04:07:58] <h4x0r`> lol
[04:09:06] <amee2k> if the logic output driving these level shifters can pull the output low, you can eliminate R3 and R6... these are there to make sure they settle low when the inputs are floating
[04:09:46] <amee2k> to prevent leakage currents or other kinds of interference from causing random fluctuations
[04:10:08] * Valen adds a distraction http://www.google.com/search?btnG=1&pws=0&q=do+a+barrel+roll
[04:10:42] <amee2k> i think everyone has that link on their ad/spam blockers by now :P
[04:14:15] <inflex> amee2k: np, I found suppliers for the chip, it was the PL2303 I was thinking of
[04:14:34] <amee2k> oh, okay
[04:14:40] <amee2k> i was thinking TDA2003 :)
[04:15:44] <amee2k> personally, i find the pricing on FT232 chips fairly offensive... makes me want to make a clone with the cheapest USB AVR lol
[04:17:49] <inflex> indeed... $8+ for the FT232 is just bonkers
[04:17:56] <inflex> it's a great chip... but not _THAT_ great
[04:18:08] <amee2k> yeah
[04:18:28] <amee2k> but then, i think max232s are ridiclously expensive for what they do as well
[04:18:53] <inflex> *nod* I use the HIN202's instead
[04:18:58] <inflex> vastly cheaper, more what they should be
[04:20:15] <amee2k> i've been wondering if some clever wiring on a pair of NE556 would work for a max232 replacement if you only need one driver in each direction
[04:21:09] <amee2k> one 556 provides the receiver and drives the charge pump, the other is powered off the charge pump output to do tx and the negative voltage
[04:22:33] <amee2k> inflex: i like your HIN202 :)
[04:22:34] <doublebeta> that'd be a nice hack
[04:24:41] <inflex> amee2k: yes, it's a nice solid replacement for a much better price
[04:25:57] <h4x0r`> amee2k,so those resistors of 1kohm are basically standard issue pullups to control the impedence of the circuit during modes of operation
[04:26:41] <h4x0r`> i didnt realise that was a fet
[04:28:27] <amee2k> it isn't, its a BJT ;)
[04:28:38] <h4x0r`> hm
[04:30:16] <h4x0r`> its still a fet lol.
[04:32:05] <amee2k> if you use a FET you can lose the base resistor. FET gates are high impedance by nature
[04:32:31] <amee2k> make sure the FETs are "logic level" type so the drive output can turn them on completely
[04:33:15] <h4x0r`> :)
[04:37:28] <amee2k> my favourite small signal FET is the BS170, which works fine for 5V logic. should be fine for 3.3V too according to the datasheet, but its Vth is a bit high for 2.8V
[04:38:03] <Valen> I use AVR's with USB and lufa ;->
[04:38:15] <h4x0r`> voltage divider is looking awesome right about now.. lol
[04:38:40] <amee2k> if there is any chance the gate of the FET might go floating, you need the 100k puldown though. otherwise the gate WILL do whatever it wants
[04:38:50] <h4x0r`> ya
[04:38:53] <Valen> fets are made with fire
[04:38:59] <h4x0r`> i dont want a circuit that floats
[04:39:00] <Valen> most electronics is made with smoke
[04:39:06] <amee2k> lol
[04:39:07] <h4x0r`> ah
[04:39:09] <h4x0r`> lol
[04:39:11] <Valen> FET = Fire Emitting Transistor
[04:39:21] <Valen> a floating fet on a power circuit = fire
[04:39:21] <h4x0r`> ofc lol
[04:39:25] <amee2k> i had some FETs in an audio amp power stage throw sparks all over the board once :P
[04:39:30] <h4x0r`> Valen, rgr :p
[04:40:01] <Valen> I had them shine through a fiberglass panel in a "spontaneous meso-plasma event"
[04:42:03] <inflex> hiya Valen
[04:42:13] <Valen> howdy howdy
[04:42:14] * inflex put in -another- PCB today, but this one I had been umming and erring over for a bit
[04:42:46] <inflex> it's a 15th revision of my LOM... but it's one that lets me put in the 4 x 1206 array, 1206 + 2 x 2512, or TO220-4 references
[04:43:03] <inflex> oh and the cabling is rearranged to minimise cable cross-over and subsequent cross-talk effects
[04:43:44] <Valen> this is where milling PCB's come in handy lol
[04:43:48] <inflex> really hoping that I don't have to change any more... it's expensive at $30/run
[04:43:57] <Valen> whats shipping?
[04:45:40] <inflex> $5
[04:45:43] <Valen> nice
[04:45:45] <Valen> seed?
[04:45:53] <Valen> also what time
[04:46:00] <Valen> i have some super tiny IC's i'd like to use
[04:46:27] <inflex> itead
[04:46:30] <inflex> 2wks
[04:46:50] <inflex> http://dxp.me/i/LOM15.png <=-obviously there's a lot going on on the bottom layer too
[04:48:14] <inflex> just so many aspects to this whole damned design that keep needing tweaks
[04:48:33] <inflex> I should be happy with how things are, but as you increase the precision you see too much of the wandering from the other parts
[04:48:57] <inflex> eg, at 1milliohm res I'd not see the tempco... but at 10micro ohms you really REALLY see it
[04:49:57] * inflex still thinks he should ahve a shot at making headphone amps in fancy cases
[04:50:19] <inflex> not because I care about doing audio (I suck at listening because of my hearing) but simply because it's such a flushed-with-money-gullible market
[04:50:37] * Valen is thinking the same things
[04:51:20] <inflex> wonder if we could do a co-op build and promo
[04:51:37] <inflex> eg, two separate products, two brands...
[04:52:25] <Valen> what you thinking for co-op promo?
[04:52:59] <inflex> *shrug* pick a product that compliments the other
[04:53:07] <inflex> preamp and headphone amp?
[04:53:09] <inflex> I dunno
[04:53:11] <Valen> you do battery, we do fixed?
[04:53:34] <inflex> mmm... I was looking at the battery thing... can't see it going down nicely without using at least 5+5 :(
[04:53:42] <Valen> 5+5?
[04:53:44] <inflex> would prefer 9+9
[04:53:45] <inflex> volts
[04:54:01] <Valen> thats quite a bit of volts
[04:54:12] <inflex> ja, but opamps love that stuff
[04:54:13] <Valen> for pre/headphone do you need rails that big?
[04:54:38] <inflex> I doubt you'd get away with anything less than 5+%
[04:54:39] <inflex> 5+%
[04:54:42] <inflex> oh ffs
[04:56:44] <inflex> http://diyaudioprojects.blogspot.com/2007/08/grado-ra1-headphone-amplifier.html <=- in a way, I also want to do it just to see if the market really is like that
[04:58:46] <inflex> http://store.tubedepot.com/diy-ek-tu-870r.html <=- this stuff cracks me up a lot too though... you've got a lossy audio, being pushed through a DAC, then through a small class AB or similar... then into a tube amp
[04:59:54] <inflex> I also love that they still use pin-through parts so much...lots of exposed component-leads there for picking up noise... and lots of spiky solder terminations for creating coronas
[05:00:07] <soul-d> :)
[05:00:14] <Valen> personally I don't see why you cant fake all that tube stuff
[05:00:32] <Valen> decent DSP and a "hi fidelity" amplifier
[05:00:34] <soul-d> those are expensive kits
[05:00:35] <inflex> heh, get some neon pigtails and say they're microtubes
[05:01:04] <Valen> cant just use those 5W bulbs in microwaves with some stainless mesh around em?
[05:01:36] <CapnKernel> inflex: The Grado's op-amp was 1/10th the price of the op-amp in the CMoy. LOL
[05:07:36] <Valen> trying to understand what the "ground amp" is about
http://www.tangentsoft.net/audio/pimeta2/
[05:15:22] * Valen wonders how to do a DC offset thing without input caps
[05:15:52] <Valen> I'm feeling some kind of AVR + RC filter feeding back into some kind of bias with a time coefficent in the seconds
[05:16:06] <Valen> actually just an inverting amp and the RC filter would do it
[05:19:05] <soul-d> yah! found glossy paper works for iron on only need to pre etch these old boards for a bit i guess
[05:19:07] <soul-d> http://imgur.com/2Pnox
[05:20:01] <mrfrenzy> well it looks promising, let's see the end result ;)
[05:20:05] <soul-d> tried yesterday with other paper but din't work to well
[05:25:48] <soul-d> sure need to cut board pre-etch and try again
[05:25:55] <soul-d> but firt lunch :D
[05:44:17] <mapee> hi
[05:46:37] <Steffanx> Lo
[05:59:47] <inflex> Valen: might do up a PCB for the opamp :)
[06:17:27] <inflex> http://gilmore2.chem.northwestern.edu/images2/cmoy2_6.gif <=- is it just me, or is that the DUMBEST way of doing a split supply ever
[06:18:35] <Steffanx> I would say it isn't you inflex
[06:19:40] <inflex> if it was me, I'd put GND right in between those two batteries...
[06:21:08] <inflex> holy crap, a fantastic example of a little knowledge is a dangerous thing, I see that schem is also appearing in many other CMoy opamp schems too
[06:23:15] <inflex> yep, as expected, they've done that to avoid unevenly discharging the batteries
[06:23:51] <inflex> so they then go through a buttload of "fixes" to skirt around the problem that doesn't have to exist
[06:32:43] <soul-d> grrr lol
[06:39:06] <Tom_itx> how much is a buttload?
[06:41:38] <inflex> as far as having this -
http://tangentsoft.net/elec/bitmaps/vgrounds/sijosae.png
[06:42:21] <inflex> or this -
http://tangentsoft.net/elec/bitmaps/vgrounds/gnd-channel.png
[07:13:20] <inflex> Valen: done my little opamp board
[07:13:28] <inflex> Valen: now to do the charging system and what ever else
[07:37:30] <inflex> Valen: trying to decide if I should do isolated segments on the board and then link them with cables
[07:37:43] <inflex> or run tracks all over the place
[07:37:53] <inflex> I think the 'cable' option makes it seem more "audiophile"
[08:42:43] <soul-d> here result of one
http://i.imgur.com/Y3vwM.jpg right one failed to much so put some clear tape on wich i usaly also use for large ground planes
[08:44:15] <inflex> yay toner
[08:44:24] <inflex> copper could have been more polished
[08:47:08] <soul-d> yah noticed that to i preetched but i cleaned it with sponge first wich made scratches again so tried with just rubbing
[08:49:39] <soul-d> but first board since long time so don't think a bad restart :) etchant still works also 2 years old
[08:49:57] <soul-d> http://i.imgur.com/Rl1hy.jpg cleaned up still some pitting but not bad for toner transfer i guess
[09:02:43] <Valen> inflex: not going to do a super fancy one?
[09:02:51] <inflex> why bother?
[09:03:16] <Valen> there is that aspect of it i spose lol
[09:03:43] <inflex> I mean, damn, look at the Grado one!
[09:04:06] <Valen> lol
[09:04:16] <Valen> too....much....wrongness....
[09:04:32] <Valen> i'm trying to work out what the ground channel is sposed to do?
[09:05:30] <inflex> huh? what where?
[09:05:40] <Valen> http://tangentsoft.net/elec/bitmaps/vgrounds/gnd-channel.png
[09:06:18] <inflex> oooh, basically provide a stiffer 'ground' to compensate for them not using a direct ground-tap between the two batteries
[09:06:40] <Valen> ....right
[09:06:43] <inflex> because they're concerned about one battery going flat/dropping sooner than the other, they prefer to serial-connect the cells and _then_ split the total voltage evenly
[09:07:10] * Valen votes for a 1Mhz boost converter in either direction
[09:07:25] <inflex> hahah
[09:07:38] <Valen> run it off a single lipoly then
[09:07:44] <inflex> would have been easiest to just put a pair of MCP1702 type regulators at ~6V and -6V
[09:07:50] <Valen> and at 1Mhz filtering is going to be nothing
[09:08:29] <inflex> ah but we're trying to appeal to the stupid side... the fear
[09:08:40] <Valen> omg not a switching power supply!
[09:08:49] <Valen> stuff a 317 after it then ;->
[09:08:52] <inflex> people find out you've got a switchmode converter in there they'll suddely start hearing buzzing/high-pitch-harmonics and stuff
[09:09:19] <Valen> call it a dual hybrid multi channel dampened power supply and put 2 caps on it
[09:09:31] <inflex> I'm going to mature my components in a wood cabinet for no less than 3 months...
[09:09:49] <Valen> I'm using only the freshest components in mine
[09:10:03] <Valen> capacitors degrade when they age yaknow
[09:10:12] <inflex> no, they "form"
[09:10:40] <inflex> you have to form them, and the opamp correctly by listening to the appropriate music for several hours before use
[09:11:07] <Valen> i like the $300,000 vinyl player
[09:11:11] * karlp chuckles
[09:11:14] <inflex> oh? URL?
[09:11:54] <Valen> http://most-expensive.net/turntable-in-world
[09:12:00] <Valen> Goldmund Reference II – $300,000
[09:12:10] <inflex> Valen: I was going to mention equidistant electron track lengths with 2nd-level smooth b-spine track curves to minimise electron bounceback
[09:12:27] <Valen> well you do gotta watch out for that obviously
[09:12:51] <inflex> I mean, look at all those shitty designs using 45' angle turns in them, that's just begging for lots of jarring noise
[12:50:21] <pingec> any idea on what the "S" at the end mean in LM2576S
[12:53:12] <Steffanx> Package type probably?
[12:53:27] <Steffanx> 5-Lead TO-263 (S)
[12:53:36] <Steffanx> 5-Lead TO-263 in Tape & Reel (SX)
[12:54:04] <pingec> thanks
[12:54:05] <Steffanx> Bent, Staggered 5-Lead TO-220 (T)
[12:54:05] <Steffanx> etc
[12:54:14] <Steffanx> You can find that in the datasheet
[12:54:15] <Steffanx> http://www.national.com/ds/LM/LM2576.pdf
[12:54:22] <Steffanx> page 23
[12:55:01] <pingec> thanks A LOT
[12:55:30] <Steffanx> You're welcome
[13:08:22] <pingec> I have this SIM300 kit ready to beconnected via RS232 and I get it to echo back everything I type but I cannot get it to recognize any AT commands
[13:08:34] <pingec> When I send it "AT" it does not reply anything
[13:08:44] <pingec> Is there any initialization I might have to do?
[13:24:26] <_Shurik_> pingec: try sending couple of /n
[13:24:46] <_Shurik_> My ExtremeDeals GPRS modem usually silent before I send it couple of returns
[13:33:52] <pingec> i just type /n ?
[13:43:11] <lowsider> anyone alive?
[13:48:56] <_Shurik_> pingec: I've played with the modem via terminal
[13:50:40] <_Shurik_> and to "wake" it up, I had to send return. It could be either \n or \r\n
[13:51:15] <_Shurik_> those are 0x0A and 0x0D codes
[13:51:50] <Steffanx> Which gprs modem do you have _Shurik_ ?
[13:54:04] <_Shurik_> Steffanx:
http://www.dealextreme.com/p/usb-tri-band-gprs-modem-cell-phone-radio-gsm-900-1800-1900mhz-12057
[13:54:17] <Steffanx> Oh, usb
[13:54:18] <_Shurik_> And this is hacking page for it
http://www.brodyradford.com/2011/01/hacking-the-dealextreme-gprs-modem/
[13:54:28] <_Shurik_> you convert it to TTL serial
[17:36:22] <rukuartic> Any advise on writing code for the tiny12? I don't even know where to start...
[17:36:41] <rukuartic> Specifically, using the WinAVR toolchain <_<
[17:40:07] <Tom_itx> include avr.io for the chip
[17:40:15] <Tom_itx> start writing code
[17:40:27] <Tom_itx> define the chip in the makefile
[17:40:57] <specing> xD
[17:41:12] <Tom_itx> iotn13.h is the header
[17:41:17] <Tom_itx> but you don't use that directly
[17:41:23] <Tom_itx> err 12
[17:41:24] <Tom_itx> sry
[17:41:47] <rukuartic> In the main.s file? oo;
[17:41:53] <Tom_itx> .c
[17:42:16] <rukuartic> You can't write c for the tiny12
[17:42:34] <rukuartic> Well... ...you *can* using someone's hacked together stackless C compiler
[17:42:37] <Tom_itx> there is a header file for it
[17:42:44] <Tom_itx> gcc
[17:42:47] <Tom_itx> winavr
[17:43:15] <Tom_itx> the ones i _know_ you cant are the tiny 4,5,9 and 10
[17:43:19] <Tom_itx> 6pin chips
[17:43:30] <rukuartic> #ifndef __ASSEMBLER__
[17:43:30] <rukuartic> # warning "MCU not supported by the C compiler"
[17:43:30] <rukuartic> #endif
[17:43:32] <Tom_itx> i've never use the 12
[17:43:52] <Tom_itx> then start writing asm
[17:44:00] <rukuartic> Ah! Back to my question.
[17:44:16] <rukuartic> What is the avr-as... syntax?
[17:44:28] <rukuartic> The huge assembly tutorial on the net is apparently written for AVR Studio's
[17:44:36] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/avr/tiny10/
[17:44:39] <rukuartic> And I'm using binutils from avr-libc
[17:44:45] <Tom_itx> those are for the t10
[17:45:27] <rukuartic> hmm
[17:45:31] <rukuartic> That's a great starting point
[17:45:56] <Tom_itx> you may need the .def file from studio
[17:45:57] <Tom_itx> in fact i'm sure you will
[17:46:27] <rukuartic> Just found the .def file for the tn12
[17:46:32] <rukuartic> avr-as doesn't seem to like it though
[17:46:56] <Tom_itx> iirc, i used studio for asm
[17:46:58] <Tom_itx> i forget
[17:47:06] <rukuartic> Ever used avr-as?
[17:47:14] <Tom_itx> not that i recall
[17:47:35] <Tom_itx> change the names to protect the innocent and it may work
[17:47:35] <rukuartic> I've googled for about an hour and I couldn't find anything... The only thing I've got right now is the -S flag for avr-gcc
[17:48:02] <Tom_itx> there are asm programmers here
[17:48:06] <Tom_itx> i don't consider myself as one
[17:49:08] <rukuartic> Hurr, neither do I :<
[17:49:32] <Tom_itx> you should be able to figure it out from those sample files
[17:49:39] <Tom_itx> some i wrote, some i stole
[17:50:03] <rukuartic> Well... it looks as if I should be able to include io.h somehow
[17:50:22] <Tom_itx> in the makefile, define the chip
[17:51:18] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/avr/attiny45/PWM/makefile
[17:51:22] <Tom_itx> see if that one helps
[17:58:27] <rukuartic> ...can you include a .h file in a .s file? O_o;
[18:12:34] <rukuartic> ;~;
[18:14:57] <rukuartic> Someone wrote avr-gcc... sooomeone knows how to write assembly for this!
[19:05:00] <rukuartic> Aaahaha. ;~;
[19:05:12] <rukuartic> Tom_L: Well. I got an ihex file. Now I just need to get a tiny12.
[19:05:19] <rukuartic> Seeing as I just fried both of mine X___X
[19:05:35] <rukuartic> I put it in the breadboard upside down, power got put on backward.
[19:05:43] <rukuartic> FUDGE. Ah well, I'll just put in the other.
[19:05:47] <rukuartic> Did the exact same thing.
[19:06:24] <Tom_itx> oops
[19:09:25] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/temp/tiny/tinyTPI1.jpg
[19:09:29] <Tom_itx> there's my tiny10
[19:54:55] <DanFrederiksen> what's the longest any of you have had an atmega in continuous operation? do they ever crash?
[19:57:19] <Casper> ask that to an xbox 1 gamer
[19:57:35] <rukuartic> I think we had it running for a pretty long while...
[19:57:52] <grummund> only the firmware crashes
[19:57:53] <rukuartic> Sent one out to the sahara desert, it was running for a few weeks...
[19:57:55] <tdo> 6 days data logging. but the hardware (cpu) cant crash
[19:58:05] <Casper> their remote use an AT90USBwhatever
[19:58:13] <rukuartic> Well... it *can* crash...
[19:58:17] <rukuartic> But that'd be more like a silicon issue.
[19:58:26] <tdo> only the sw can be buggy or the power supply instable, ...
[19:58:34] <rukuartic> Yea there you go.
[19:58:55] <DanFrederiksen> I'm thinking cosmic rays setting the wrong bit
[19:59:29] <Kevin`> well you are using the watchdog timer to guard against some of that, right?
[19:59:41] <DanFrederiksen> that's why I am asking
[19:59:49] <Casper> if you code well, it shouln't crash unless it have power issues
[19:59:51] <Kevin`> and internal or external brownout
[20:00:09] <tdo> hmm - never happened but read that in planes there are more cosmic rays and it could happen (but then also to ipods, laptops, anything)
[20:00:10] <grummund> that's always a great excuse... "your firmware is buggy", "no it must've been a cosmic ray".
[20:00:23] <DanFrederiksen> I'm not concerned with power failure. or bugs
[20:00:37] <DanFrederiksen> if it runs forever, how often will it crash
[20:00:53] <Casper> the hardware is quite robust, never heard of then crashing that wasn't damaged somehow
[20:01:08] <tdo> exactly, never heard of crashing
[20:01:19] <DanFrederiksen> I was hoping someone had it in continuous operation for years
[20:01:55] <Tom_itx> ask atmel's test crew
[20:02:03] <Kevin`> there's plenty of devices like that, not sure where any statistics would be
[20:02:11] <tdo> btw: if you fear cosmic rays, keep in mind that atmel, ti, microchip, all are equally affected
[20:02:43] <DanFrederiksen> that's not the point. I wonder how many watchdogs I need to keep it in line
[20:02:44] <Tom_itx> proper enclosure may help with that
[20:02:53] <tdo> (or 'would be' because the stats must be extremely low)
[20:03:10] <DanFrederiksen> I'm afraid cosmis rays are not concerned with shielding
[20:03:18] <DanFrederiksen> cosmic
[20:03:26] <Casper> oh yeah DanFrederiksen, I used to have a board which had one... many years ago
[20:03:38] <DanFrederiksen> one what?
[20:03:41] <Casper> don't ask me why it had one, used it for a few years, was a good comp...
[20:03:49] <Casper> an atmel microcontroller
[20:03:56] <Casper> I think it was a 4 bits one however
[20:03:57] <Kevin`> DanFrederiksen: if something like that is important, have a second chip that you can use to correct flash bit-flips from (insert powerful radiation)
[20:04:28] <rukuartic> DanFrederiksen: what are you trying to do?
[20:04:36] <DanFrederiksen> motor controller for electric car
[20:04:56] <grummund> if you don't let the smoke out then hardware lasts forever
[20:04:57] <DanFrederiksen> it would be nice to remain in control of the car :)
[20:05:14] <tdo> 1) see how long your stuff can be dead from the app requirements, 2) set the watchdog to this time minus whatever time is needed to restart and reinitialize your stuff
[20:06:31] <tdo> ah for the car control you could have 2 or 3 chips and some majority voting in hardware
[20:07:20] <Casper> DanFrederiksen: this is why you want to have some outside protection if needed
[20:07:44] <Casper> like for motor pwm, you might want to have a protection against 100% duty
[20:07:47] <DanFrederiksen> I'd prefer data first before adding more complexity that itself can be failure points
[20:07:49] <Kevin`> the normal think to do with a car is to have it break without any impetus but stop safely
[20:08:17] <DanFrederiksen> Kevin`, yeah graceful failure
[20:08:18] <Casper> 100% duty would mean that it's most probaly faulty, so have an hardware safety like that
[20:08:33] <Casper> but they DO use some avr for automotive use
[20:08:47] <Casper> use automotive version of them and it shall be fine
[20:08:48] <DanFrederiksen> yeah. but maybe not critical
[20:09:14] <Casper> you'ld be more worried about code mistake and hardware issues than faulty avr
[20:09:32] <DanFrederiksen> nah my code is quite simple
[20:10:23] <Kevin`> atmel has a document on this sort of design btw
[20:10:30] <DanFrederiksen> ok
[20:10:54] <Kevin`> I only ended up looking at it because it was near the top of some list :)
[20:11:02] <DanFrederiksen> :)
[20:12:23] <tdo> mechanical brakes stronger than engine, that's the current solution on cars...
[20:15:12] <Kevin`> electric systems of this sort are scary powerful
[20:15:44] <tdo> yeah - don't know the specs of his thing :)
[20:18:23] <Kevin`> even the smaller stuff, instantly switching an efficient electric motor from off to full power (and with no load, overspeeding shortly thereafter) is rather interesting :)
[20:22:32] <DanFrederiksen> explosive even :)
[20:23:05] <DanFrederiksen> some will go to 300 rotations per second before disintegrating
[20:24:13] <Casper> I've seen an ebike that the guy lost control of it...
[20:24:19] <Casper> 0-160MPG in 2 seconds
[20:24:42] <Casper> or was it 160kmh? anyway
[20:24:47] <Casper> it was insanelly fast
[20:25:36] <Casper> the guy wanted to go a burnout, but the wheel somehow sticked and the bike went like a bullet in a now you see now you don't kind of way... or more like now you see now you're almost dead
[20:25:51] <DanFrederiksen> ah the killacycle. about one second to 100km/h
[20:26:09] <Casper> I think it was it
[20:26:14] <DanFrederiksen> yes he messed up :)
[20:26:42] <DanFrederiksen> but he almost got it under control. otherwise could well have died
[20:30:02] <Casper> but that is also a design flaw in a way
[20:30:22] <Casper> but a flaw that they created and wanted
[20:30:26] <Casper> too much acceleration
[20:30:33] <DanFrederiksen> I suppose it could have a 'please don't kill me mode'
[20:31:03] <DanFrederiksen> or a burnout mode where power it cut if there is actual acceleration
[20:31:47] <Casper> yeah
[20:32:20] <DanFrederiksen> is*
[20:32:22] <Casper> that is what is worrysome with electric vehicle
[20:32:38] <DanFrederiksen> the super powerful kind yes
[20:32:40] <Casper> the acceleration is too high, and the top speed is too
[20:33:03] <Casper> and there is 2 main fault mode: open or short
[20:33:17] <Casper> so you have one chance on 2 that if the controller fry that it will go full speed
[20:34:07] <DanFrederiksen> typically the chips can't survive being full on so the explode open. but in theory if the car is at speed already they can survive being full on
[20:34:16] <DanFrederiksen> they*
[20:34:22] <Kevin`> vehicles like that are probably using a synchronous motor, so actually, short = burning and no accelleration
[20:34:32] <Casper> well
[20:34:45] <DanFrederiksen> yeah AC is hard to fail with full throttle in that way
[20:34:49] <Casper> if the controller is made for full throttle acceleration, then it will survive a short
[20:34:54] <Kevin`> unless you have a semi-software bug where the controller is working "properly" but driving the motor at 100%
[20:35:01] <DanFrederiksen> Casper, no it doesn't work that way
[20:35:11] <Casper> true they use AC...
[20:35:13] <Casper> but ok
[20:35:14] <DanFrederiksen> all electric car drives have current limiting
[20:35:23] <DanFrederiksen> without it they die instantly
[20:35:38] <Casper> driver is fine, but controller is faulty and send a 100% throttle signal
[20:35:48] <DanFrederiksen> yeah
[20:37:23] <Casper> you know
[20:37:33] <Casper> imagine.. you set the watchdog timer to 250ms
[20:37:51] <Casper> and there is a bug that cause the controller to send full throttle for that 250ms...
[20:38:05] <Casper> with the killacycle you go from 0 to like 60kmh :D
[20:39:28] <Casper> O.O
[20:39:37] <Casper> just watched a video...
[20:39:41] <Kevin`> link?
[20:39:45] <Casper> ya
[20:39:50] <Casper> http://www.youtube.com/watch?v=zcBE8HEAyi4
[20:39:56] <Casper> check it up :D
[20:41:29] <Kevin`> huh? what's that?
[20:42:00] <Casper> hand starting a truck
[20:42:05] <Kevin`> ah, starting it
[20:42:54] <Casper> didn't knew it was doable that way with just that "little" energy
[20:49:03] <Casper> oh
[20:49:07] <Casper> interresting
[20:49:20] <Casper> my new isp send me a receip when they receive the payment! :D
[21:22:26] <rue_mohr> the town just taxed my mom for 8 acres of a hourbor! this means we must own it! and were SELLING! were starting at $16M
[21:22:47] <rue_mohr> its got docks and breakwaters and everything
[21:23:05] <Tom_itx> :)
[21:24:51] <rue_mohr> I called the mayor and offered him first dibbs!
[21:24:59] <rue_mohr> (had to leave a message)
[21:26:03] <Tom_itx> you think he'll get back to you?
[21:26:16] <rue_mohr> he knows me
[21:26:18] <Tom_itx> i uploaded all my arm stuff to the server
[21:26:28] <rue_mohr> cool
[21:27:14] <Tom_itx> i installed IAR finally and got a good compile and upload
[21:27:54] <Tom_itx> it was pretty straighforward since that's what they intended you do
[21:28:36] <Casper> yesterday I looked at the ball valve, and I noticed a small thing: maybe I could replace the nut by a coupler, which would allow me to install a motor on the coupler...
[21:28:50] <Casper> and today I just remembered that I have a geared motor, but of unknown forc
[21:28:51] <Casper> e
[21:30:15] <rue_mohr> for which?
[21:31:13] <Casper> remember that I wanted a linear actuator?
[21:31:38] <Casper> I think I can finally install a geared motor with no real problem
[21:31:55] <Casper> I think it's a geared stepper that I have... hmm
[21:34:41] <inflex> lo folks
[21:35:00] <bsdfox> [16:08] <inflex> hmm... time to look for a serial-USB converter using an AVR
[21:35:16] <inflex> bsdfox: found it's cheaper to just buy modules from China directly
[21:35:19] <bsdfox> I think the arduino is using a mega8u16 instead of FTDI now
[21:35:36] <bsdfox> also microchip has a good one based on pic18f14k50
[21:38:24] <lowsider> Instead of doing rs232 using something like an ftdi chip .. can't you just use the V-USB libs?
[21:39:19] <lowsider> i'm actually just looking into a way of debugging my avr chips as well.
[21:39:30] <lowsider> having to deal with teh usb / serial cable stuff makes me sad
[21:39:52] <inflex> lowsider: v-usb can't do much better than 38400
[21:40:13] <inflex> Anyhow, as mentioned, I can just get complete modules from Alibaba for less than the cost of the chip.
[21:40:24] <lowsider> inflex: oh really? that's good to know! thanks!
[21:40:30] <lowsider> does that LUFA stuff work any better?
[21:40:38] <inflex> complete with USB connector, cable, shroud/encasing and rx/tx/5/3.3/gnd header
[21:40:55] <inflex> not sure, no tried LUFA in that capacity - though I'm sure abcminiuser could tell
[21:41:09] <Casper> inflex: li-ion that was at 2V 3 days ago and is now at 0.95V, is it bad or normal self discharge? and yes I know 2V is insanelly low :D
[21:41:21] <inflex> Casper: it's dead
[21:41:28] <inflex> those thing should hold their charge for 12 months or more
[21:41:29] <Casper> inflex: should have stayed at 2V isn't it?
[21:42:04] <Casper> even at such low level?
[21:42:13] <inflex> yep
[21:43:22] <Casper> that mean I have 2 more cells to try to fry? :D
[21:44:00] <Casper> I wonder if those 3 open cells will go up in flame at 60VDC :)
[21:44:32] <rue_mohr> Casper, that much self-discharge is bad
[21:45:08] <rue_mohr> a li-ion charged to 4.20V should sit there for months and still be at 4.1 ATLEAST
[21:45:34] <Casper> rue_mohr: they got charged to only 2V
[21:45:37] <rue_mohr> Casper, remember that lithium fires are self-oxidizing, you might need halon to put them out
[21:46:01] <inflex> or a big bucket of sand :)
[21:46:10] <Casper> rue_mohr: those I want to make go whoosh will be outside, where the pool used to be
[21:46:25] <Casper> I have a 27' sand circle!
[21:47:10] <rue_mohr> why the fff did you blow up the POOL!?!
[21:47:23] <Casper> sadly, nope
[21:47:37] <rue_mohr> did you sue the person who blew it up?
[21:47:47] <Casper> sadly, it didn't blew up
[21:48:05] <rue_mohr> HOW ELSE DOES A POOL BECOME A CRATOR?
[21:48:08] <Casper> my uncle sell the backyard to my cousin so she can build an house
[21:48:10] <rue_mohr> meteor
[21:48:17] <Casper> so we dissassembled it
[21:48:29] <rue_mohr> thats an insurance claim, act of god my ass
[21:48:32] <rue_mohr> oh
[21:48:36] <rue_mohr> hmm
[21:48:57] <rue_mohr> she will sell the house, go broke and ask for another peice
[21:49:29] <rue_mohr> or just sell it and move to somewhere else
[21:50:00] <Casper> if it wasn't of the house that was so close, I'ld have sliced it open and watched the 70000L of water flow out all at once MOUAHAHAHAHAHAAHA :D
[21:50:48] <rue_mohr> I'm gonna go to bed, have a shower, and eat supper... not in that order
[21:50:57] <Casper> I hopes not
[21:51:15] <Casper> but that's about what my stinky uncle does
[21:52:20] <Casper> imagine 70000L of water rushing out :D
[22:12:15] <h4x0r`> .
[22:13:15] <Valen> DanFrederiksen: your going to have more issues with EMI causing latchups than any other cause in that environment
[22:13:35] <Valen> also unusual code bugs
[22:13:59] <Valen> the silicon is going to be one of the more reliable components in the system
[22:14:27] <Valen> I'd be more worried about a fet latching on (if its a DC motor) and connecting power to the motor permanently
[22:14:59] <Valen> bldc less of an issue with any sort of sane control scheme. as the whole system has to work for the wheel to go around
[22:15:55] <Valen> use a smart WDT that compares output power/RPM with commanded input and an i'm alive input from the main controller and your going to cover most of the feasible faults
[22:23:55] <DanFrederiksen> Valen, wasn't transistor latching solved in 1985?
[22:24:15] <Valen> not when you blow the fet up
[22:24:37] <DanFrederiksen> then latching isn't really the problem :)
[22:24:44] <Valen> and EMI causing latchup/weird behavior is still an issue
[22:25:03] <DanFrederiksen> you can design for EMI. mot cosmic rays
[22:25:03] <DanFrederiksen> not
[22:26:12] <Valen> as i said a smart WDT will cover you, but cosmic rays are far less likely an occurrence than pretty much anything else
[22:26:16] <grummund> jeez two hours on and you guys still worrrying about cosmic rays? :P
[22:26:27] <Valen> especially given the average feature size in an AVR
[22:28:12] <DanFrederiksen> grummund, the issue doesn't go away just because you want it to
[22:28:37] <Valen> it does go away when you actually look into it
[22:29:54] <grummund> DanFrederiksen: the issue usually becomes irrelevant once the hardware is designed to account for software errors already
[22:30:49] <Valen> i'll bet $100 that a software error will be at fault 10x more than cosmic rays
[22:31:21] <grummund> Basically assume that anything with even a single transistor can fail in *any* mode, and design for that eventuality.
[23:22:25] <inflex> ah, but then you have to compensate for the failure of the failure compensation systems
[23:29:41] <rue_bed> by designing the device to extert an outwardly force in all directions which contains a reprocusive blast such to alart all users that there was such a failure
[23:30:00] <rue_bed> sorry *shock wave*
[23:32:03] <inflex> I think burning to ash is a good failsafe mode... stops people trying to reuse it.
