Back
[00:00:02] <CapnKernel> I know the company that itead uses.
[00:00:20] <CapnKernel> If someone wants the budget offering, I'll put the order through there.
[00:00:47] <ziph> The problem might be that those orders affect your reputation.
[00:01:37] <CapnKernel> I know what you mean.
[00:01:49] <CapnKernel> I think, tell people what the deal is, and let them decide.
[00:02:03] <ziph> Although at the moment they're mostly screwing up on delivery
[00:02:40] <CapnKernel> I think I can price match the other guys, and I'll be offering free shipping, which they don't.
[00:02:48] <ziph> FedEx?
[00:03:03] <CapnKernel> Free shipping is HK post. Around 7 days to USA
[00:03:25] <ziph> HK Post to Australia at least is awful though now.
[00:03:34] <CapnKernel> Yeah, it used to be better.
[00:03:53] <ziph> But if you can competently put addresses on then it'll be faster than iteed. ;)
[00:03:54] <CapnKernel> I mailed a 3G modem from my home in Shenzhen to my home in Melbourne last month, and it took a little more than three weeks. Sucked.
[00:04:19] <CapnKernel> I have several things in mind for how I can do it better than itead
[00:04:36] <CapnKernel> ziph: Meanwhile, what other websites do you know of where you can do online ordering?
[00:04:57] <ziph> PCBCart and Advanced are the only ones I know that do online ordering.
[00:05:07] <CapnKernel> jd_ce: thanks
[00:05:07] <ziph> GP now has a web page but I haven't used it yet.
[00:05:17] <CapnKernel> precision circuits?
[00:05:19] <ziph> The rest are e-mail.
[00:05:34] <ziph> There's a bunch of American ones too.
[00:06:06] <CapnKernel> doesn't matter where they are, because I want to look at the UI
[00:06:31] <CapnKernel> For example, I will also look at Olimex
[00:07:39] <ziph> http://www.expresspcb.com/
[00:07:53] <ziph> Olimex probably requires you to fax your gerbers to them. ;)
[00:10:14] <ziph> I was thinking of writing some software to handle PCB orders too.
[00:10:23] <ziph> I have most of a very robust gerber parser done.
[00:13:33] <CapnKernel> Curious:
http://www.ladyada.net/library/pcb/costcalc.html
[00:13:40] <CapnKernel> ziph: cool!
[01:52:15] <DanFrederiksen> is you set pwm to 0, do you get the pwn overflow interrupt anyway? or only for values above 0
[01:52:25] <DanFrederiksen> pwm
[01:53:52] <Casper> no
[01:53:59] <Casper> it's a special case, as said in the datasheet
[02:49:32] <chron0> heyho everybody
[02:50:45] <chron0> not exactly if I'm right here but I'm trying to build avrsim and ran into a little bit of troube (/run_avr.c:107: undefined reference to `__iob') - so I wanted to ask if anyone here ever built and used avrsim
[02:52:19] <Casper> sure some do... but simulators are often a waste of time as they ain't perfect and lack flexibility... so it's hard or impossible to make any real program to run in them
[02:53:36] <chron0> Casper: hmm thats what I thought, in this particular case I just wanted to set up a little irq driven serial system that talks to a python script/daemon via USART0
[02:54:15] <chron0> the idea was, that I could test things whereever I am, without having to take the breadboard and avrispv2 with me
[02:56:48] <chron0> however, since my avr C skills are n00bish at best I still have to figure out a lot, kinda frustrating for such a simple task, but since I want to learn something while doing it, it takes an awful lot of time :)
[02:58:19] <grummund> chron0: you could probably acheive much the same with one of the usb-avr boards, like teensy or the one on Tom_itx's site.
[02:58:55] <grummund> Small and doesn't require a separate programmer, just a usb cable.
[03:00:50] <chron0> grummund: I see, that might be an option - it just looks way to freaky for most people, if you sit on an urban commuter train and deploy a laptop, breadboard, the avrisp, the ftdi usb/serial adapter, wire it all up, at least I dont unpack a power supply, since I've modded the avrisp
[03:01:06] <chron0> and start hacking on it :)
[03:01:21] <chron0> terrorist! :)
[03:02:32] <grummund> a usb board can be used standalone, no power supply, bo programmer.
[03:02:39] <grummund> *no programmer
[03:02:40] <chron0> jepp
[03:03:35] <grummund> the at90usbkey even has a couple of LEDs and a joystick button on board.
[03:03:39] <chron0> but then I#d have to have an usb bootloader right? although it is nice and convenient most of my projects are just programmed via isp
[03:04:07] <grummund> they all come with bootloader installed
[03:04:31] <chron0> + ftdi?
[03:05:21] <grummund> no ftdi, the avr connects direct to usb.
[03:07:07] <chron0> ah, pity, that wont help, the majority of the avr projects I roll up here are connected either directly vi 3v3 rx/tx to another linux armv9 machine to communicate with each other (at least that is the plan) and I also want to go for rs-485 or CAN...
[03:09:47] <grummund> well tbh i'd be looking at boxing up whatever you need to make it look sane and portable, rather than mess with simulators.
[03:10:05] <chron0> and its specifically these communication interactions that trouble me the most, as I obviously haven't understood it completely
[03:10:17] <chron0> i see
[03:11:15] <chron0> pity though, would have been nice to have a "soft-avr" available
[03:11:15] <grummund> btw, this is my programmer -
http://i52.tinypic.com/2u4n5i0.jpg
[03:11:25] <chron0> nice :)
[03:11:42] <lovre> hi all
[03:12:05] <grummund> so you can do the same for ISP, and steal power from the usb port too perhaps.
[03:13:51] <lovre> how do i make a generic debouncing function for avr? All examples i have seen use a custom, allready known pin, like debouncing on PINB, but how do i make a generic function for this, that would take the port and the bit number as arguments? I dont know how to pass PINB as argument to a function, what data type is it anyways, and if i pass it as char, wouldnt that be by value, so its not really checking real value... should i use pointers
[03:13:54] <chron0> thats what I did already, my isp
http://www.atmel.com/dyn/resources/prod_images/AVRISPmkII.jpg - not in blue anymore - I added a switch that either puts 5v from usb to the board or 3v3 via ld vreg, depending on the setup, maybe I should just put the ftdi in there as well
[03:14:38] <grummund> lovre:
http://www.nongnu.org/avr-libc/user-manual/FAQ.html#faq_port_pass
[03:14:42] <chron0> lovre:
http://www.mikrocontroller.net/topic/226230
[03:15:34] <lovre> grummund, chron0: just what i was looking for, thank you very much :)
[03:19:04] <chron0> lovre: this one might be interesting too:
http://www.mikrocontroller.net/articles/Entprellung#Debounce-Makro_von_Peter_Dannegger
[03:20:24] <lovre> chron0: seems more complicated than needed, will try a simple approach first... thanks!! :)
[03:20:56] <chron0> lovre: sure, i love kiss
[03:21:03] <lovre> :)
[03:21:32] <chron0> lovre: just in case you needed some more examples and apporaches, sorry for the german page, but I'm very used to it, and the code should be readable anyways
[03:21:58] <lovre> chron0: yes, the code is usable
[05:36:45] <OndraSter_> huh, why is there mega128a and mega2560 on the AVR Dragon? :D
[06:13:57] <amee2k> okay, here is a conceptual question... i've got a number of devices connected to a 24V main power rail, and a control bus that has 5V bus power. my initial design puts the MCU in the device on the bus powered side, so even if main power fails it wouldn't affect access to the configuration and maintenance tasks like bus discovery
[06:15:38] <amee2k> now turns out that i could considerably simplify things by powering the MCU off the main supply rail (with a regulator, obviously). this means devices that lose power will suddenly disappear from the bus and come back up with a fresh power-up (i.e. any state is lost)
[06:16:37] <amee2k> for some reason i don't like the implications of the latter option, but i can't put my finger on why. any ideas on which advantages or disadvantages either situation could have?
[06:28:34] <karlp> is there any actual use in being able to discover a device that doesn't do anything because power's not applied?
[06:28:57] <karlp> can the mcu reliably diagnose that the problem is just that power is missing from the rest of the system?
[06:29:34] <karlp> also, is there anything in the state that can' be just loaded back from eeprom?
[06:30:05] <karlp> if it's just config, shouldn't be a problem, it's more of a concern if you want to keep a running number (like kilowatthours or something)
[06:31:48] <amee2k> it has an on-die temperature sensor that can be read while power is missing, for example
[06:32:36] <amee2k> it can't tell the main rail's voltage without main power, but it can tell that external circuitry powered from the main rail is not responding
[06:33:16] <amee2k> there is an SPI isolator between the power circuit and the bus transciever
[06:34:09] <amee2k> for some reason i keep thinking that being able to discover unpowered devices is an important feature, but i couldn't give a good reason as to why >_<
[06:34:49] <Steffanx> "because I say so" is a good reason amee2k :P
[06:35:05] <Steffanx> My parents use(d) it all the time :P
[06:35:57] <amee2k> losing volatile state isn't a problem as in it would be dangerous. but it would throw the actual state off-sync with the idea that the bus master(s) possibly have of it
[06:36:16] <amee2k> wouldn't* be dangerous, even!
[06:38:32] <amee2k> Steffanx: well, yeah... i could pull rank by being the only one on the project but that isn't very satisfactory imo
[06:39:17] <amee2k> i'm trying to weight the two options but can't find any good scenarios that push the scale way in favour of one of the two options
[06:42:05] <OndraSter> yo dudes, if I have file beginning with this (the main file, set as entrypoint)
http://pastebin.com/zwJuWBwB
[06:42:06] <amee2k> so is there any possible advantage is being able to discover devices that are non-functional due to loss of power??
[06:42:25] <OndraSter> how come that all the things, except the first JMP BOOTSTRAP go to beginning of the flash
[06:42:30] <OndraSter> rather AFTER the JMP BOOTSTRAP
[06:42:36] <OndraSter> after offset LARGEBOOTLOADER?
[06:42:42] <OndraSter> LARGEBOOTSTART*
[06:43:38] <rue_bed> [p;''''''''';ppppppppppppppppppppppppppppppppppppp/zuiiiiiiiiiiiiiiiiiiii
[06:43:45] <OndraSter> caT?)
[06:43:48] <rue_bed> ']
[06:43:50] <OndraSter> cat?
[06:43:52] <amee2k> dog?
[06:43:53] <amee2k> >_>
[06:44:05] <amee2k> too much <insert your favourite drug here>?
[06:49:40] * amee2k idly maps mapee
[06:49:41] <amee2k> >_>
[06:55:44] <amee2k> the only thing i can think of that would tip the balance in favour of MCU powered from main bus is that the board would retain state during a bus power failure and continue operating
[06:56:29] <amee2k> something tells me i want two MCUs, one for the bus interface and one for device control. with one powered of each rail >_<
[07:28:45] <amee2k> or am i trying to find the Right Thing(tm) to do where none exists again?
[07:43:43] <karlp> I think the later
[07:44:12] <karlp> just make sure you can know whether it's valid or invalid, from both sides, and what the likely cause is
[07:50:51] <amee2k> which later one? >_>
[08:04:30] <amee2k> karlp: i don't think i follow entirely... excepting hardware failure the only loss of state i can think of is a power failure
[08:04:59] <amee2k> state can be queried by the bus master for verification, but how would i explicitly check for validity?
[08:08:22] <karlp> I mean, if you can tell that the difference between "no power" and "device is broken"
[08:08:40] <karlp> even that is just a "nice"
[08:09:43] <amee2k> with the MCU on the bus powered side scenario, the MCU could tell the other side is not responding. that would indicate it is non-functional for reasons that likely require operator invention
[08:10:01] <amee2k> with power failure being the likely cause for many situations
[08:10:26] <karlp> soudns good enough
[10:01:46] <OndraSter> how much can I trust the "manual auto routing" in eagle to give me usable lines? There won't be fast switching, but it makes really weird traces
[10:02:03] <OndraSter> http://clip2net.com/s/1tKHi
[10:02:34] <amee2k> last time i tried eagle was like 2 years ago, and i didn't find it very helpful
[10:02:44] <OndraSter> I just don't want to do all the shit manually
[10:02:46] <amee2k> either i was doing something wrong, or their autorouter sucks on 2-layer boards
[10:02:57] <OndraSter> I did all the LED matrices manually
[10:02:59] <OndraSter> took few hours
[10:05:51] <amee2k> i'd say that it looks odd is kinda normal... i mean, it is like you gave it to someone else for routing
[10:06:20] <OndraSter> hmm
[10:06:22] <amee2k> so any preconceptions you may or may have about how the result could look like will be disappointed
[10:06:30] <amee2k> may or may not*
[10:07:14] <amee2k> but yeah, i found that it looks quite odd. i ended up rerouting almost everything manually anyway
[10:08:01] <amee2k> which i think is to be expected ... the autorouter can't possibly know what kind of requirements specific nets or even individual track segments have
[10:08:31] <OndraSter> ye
[10:09:46] <amee2k> for example, it ran the power tracks on IC6 there to the IC pads first, then on to the bypass cap
[10:09:56] <OndraSter> yah
[10:09:59] <OndraSter> happens
[10:10:15] <OndraSter> I'm gonna autoroute all the heavy stuff and then check it how to simplify it probably
[10:10:30] <amee2k> i tend to run the traces through the bypass cap pads and continue to the IC power pads... gets slightly better decoupling between load and rails
[10:10:33] <OndraSter> I want the board to be ready till the end of this week
[10:10:36] <OndraSter> ye
[10:11:28] <OndraSter> something like this will look my input
[10:11:28] <OndraSter> http://clip2net.com/s/1tKOz
[10:11:29] <OndraSter> hopefuly
[10:11:41] <OndraSter> the FTDI needs a bit of redesigning too
[10:11:50] <mrsun> hmm, what languages are available for avr? was thinking of vhdl it looks quite neat =)
[10:12:01] <OndraSter> vhdl is for logic stuff
[10:12:05] <amee2k> most people use C or assembly
[10:12:08] <OndraSter> ye
[10:12:12] <OndraSter> and basic is available too
[10:12:14] <Steffanx> basic :D
[10:12:15] <amee2k> i don't think you can use VHDL for AVRs
[10:12:16] <Steffanx> forth
[10:12:18] <mrsun> ye i use C normaly was thinking there was alot of nice stuff in vhdl =)
[10:12:40] <mrsun> there are compilers to compile vhdl to computer binaries but =)
[10:12:49] <mrsun> was for that i was thinking if it was available for avr =)
[10:12:54] <amee2k> iirc vhdl is intended for programmable logic devices like PLCs, CPLDs, or FPGAs
[10:13:16] <amee2k> not for ones that work like a microcomputer
[10:14:09] <amee2k> a general purpose PC has enough processing power so it doesn't really care how efficient the VHDL-to-binary compiler really is
[10:14:38] <mrsun> true =)
[10:15:02] <amee2k> but on an MCU you're working on like an 80s style computer... almost any kind of emulation layer or interpreter is going to come back and bite you
[10:15:05] <OndraSter> I have found many times crashed processes eating one whole core
[10:15:07] <OndraSter> or even two cores
[10:15:10] <OndraSter> for days
[10:15:14] <OndraSter> before I noticed it in task manager lol
[10:15:18] <amee2k> lol
[10:15:38] <amee2k> i've got a cpu load graph in my taskbar :)
[10:15:42] <OndraSter> sandybridge is the king
[10:15:42] <OndraSter> haha
[10:16:06] <mrsun> amee2k, but with a compiler i was thinking mostly it would translate vhdl into native avr code =)
[10:17:41] <amee2k> it is probably possible in theory, but getting it to work efficiently will be a tough one
[10:17:51] <mrsun> fpgas are so freakin big and the capsules are so unusable for the ordenary people :P
[10:18:05] <OndraSter> big?
[10:18:12] <OndraSter> 100 pin FPGA -> about 86 IOs
[10:18:22] <OndraSter> on small 100pin TQFP
[10:18:35] <mrsun> its the small i mean :P
[10:18:37] <amee2k> because the platforms are pretty fundamentally different... like trying to make powerpc code run on an x86 computer, just far far worse
[10:18:37] <mrsun> alot of legs
[10:20:10] <amee2k> AVRs really work like a tiny computer inside... a CPU package with a little RAM and flash memory for the program, which is executed instruction by instruction in sequence
[10:20:53] <amee2k> so you'll get the best results when sticking to a programming language that models that
[10:21:20] <mrsun> true =)
[10:21:41] <mrsun> hmm, so next question, is there any adcs with parallel output and a "ready" pin ? :P
[10:21:54] <mrsun> so when it has sampled it pulses ready and you can read the data :P
[10:23:23] <OndraSter> most likely
[10:23:44] <mrsun> i suck at searching for components :P
[10:23:45] <OndraSter> I have got here some Analog Devices AD 12bit (I know, only :P) ADCs with parallel output
[10:24:38] <OndraSter> 10bit*
[10:24:39] <amee2k> farnell and digikey have reasonably useful parametric search engines
[10:25:32] <eatyourguitar> mouser does too but they dont meta tag the items correctly thus the whole thing is useless
[10:26:33] <OndraSter> okay, AD7939 holds BUSY pin low when you start the convert (through CONVST pin)
[10:26:35] <eatyourguitar> like if you are searching IC's and you want DIP8 only it will hide half the DIP8 results cause they are missing the DIP8 metatag. making it unreliable so I have to search keyword dip8
[10:26:38] <OndraSter> and once it goes high the data are ready
[10:26:40] <mrsun> damn the 10 bits were very cheap compared to 12 bit :P
[10:27:29] <amee2k> how about AD1674?
[10:28:50] <amee2k> look at the "STS" pin on it
[10:31:11] <OndraSter> <OndraSter> yo dudes, if I have file beginning with this (the main file, set as entrypoint)
http://pastebin.com/zwJuWBwB<OndraSter> how come that all the things, except the first JMP BOOTSTRAP go to beginning of the flash
[10:31:11] <OndraSter> <OndraSter> rather AFTER the JMP BOOTSTRAP
[10:31:11] <OndraSter> <OndraSter> after offset LARGEBOOTSTART?
[10:31:36] <OndraSter> does the code segment counter reset at the beginning of each include?
[10:35:07] <mrsun> hmm, need to check why my avr development board gave up if something has burned ... want to start working with avr again =)
[11:10:05] <karlp> eatyourguitar: digikey has the same problem, you just migth not have noticed it :)
[11:10:22] <karlp> farnell is even worse at times, it's completely hidden things from search that I could get to by part number
[11:11:00] <amee2k> yeah, lots of things are untagged on farnell as it seems
[11:11:12] <amee2k> or have specs missing in the database that are on the datasheet
[11:11:19] <amee2k> imo the worst is Conrad
[11:11:23] <eatyourguitar> I never use digikey. I don't like to split up my orders between mouser and digikey. sometimes I see stuff I like but I never ordered from them
[11:11:36] <amee2k> they sort numeric columns as text ffs
[11:11:57] <amee2k> and you can filter by brand, rohs compliance and price range
[11:12:03] <amee2k> thats it.
[11:13:01] <eatyourguitar> yeah the digikey website is visually not so appealling. they just dont do anything better. maybe they are the same in every way. who knows
[11:13:38] <amee2k> i want to use digikey for this project for the first time
[11:13:57] <amee2k> their website is a usability train wreck but it works for the most part
[11:14:02] <karlp> I just refuse to use digikey because they refuse to update their list of export restrictions by country
[11:14:22] <karlp> until they'll ship me 802.15.4 parts, I'll shop my entire order from someone who will
[11:14:46] <amee2k> i didn't use them before because of the obscene shipping (18EUR vs. 6EUR for farnell) and i never got below the free shipping limit
[11:15:04] <karlp> farnell just has obscene pricing for me,
[11:15:23] <amee2k> i pay 50EUR + 6 shipping at farnell
[11:15:37] <amee2k> i would pay 45 + 18EUR shipping from digikey
[11:17:02] <karlp> 2 pounds extra on a 5 pound part for me makes it kinda absurd...
[11:17:14] <karlp> lucky you on parts they aren't screwing you on :)
[11:17:29] <amee2k> thats why i only buy these :P
[11:17:33] <karlp> hehe
[11:17:50] <karlp> at work we recently split an order, ended up saving money even after paying two lots of shipping.
[11:18:57] <amee2k> on big ones i'd believe that right away
[11:19:09] <amee2k> but i don't buy enough crap to make that worth it
[11:19:48] <amee2k> i'd probably consider splitting too if i bought 1k units of everything
[11:20:38] <eatyourguitar> in soviet russia digikey ships itself
[11:21:01] <amee2k> in soviet, digikey ships russians
[11:21:03] <amee2k> err, wait.
[11:21:36] <eatyourguitar> you need computer chip yes. make for glorious benefit the user
[11:22:12] <eatyourguitar> lol
[11:44:09] <syrou> hi all. i have a problem trying to program an 8515 via isp
[11:45:06] <syrou> i haven't got the schematics of the circuit, and it has a push button in order to toggle the device on or off. I power it on and my avrisp mkii lights green
[11:45:12] <Tom_itx> is spi enabled?
[11:45:53] <syrou> but then I click on "read signature" and it powers off (red light) and says "target appears to have lost power"
[11:46:41] <syrou> if I keep the push button pressed (cycle powering on and off) i get the signature, locks, fuses... but when reading the flash contents it gives me a sequence of bytes
[11:46:59] <amee2k> hmm... wondering if i should buy a book about technical drawing and a table book for metal working
[11:47:15] <syrou> Tom_itx: i suppose that if i'm able to read those things, spi is enabled, but i may be wrong
[11:47:37] <amee2k> 978-3808517253 and 978-3589241941
[11:47:50] <Tom_itx> what are those numbers?
[11:47:57] <amee2k> isbn numbers?
[11:48:25] <Tom_itx> get a 'Machinery's Handbook'
[11:48:58] <Tom_itx> lotsa good stuff in it
[11:49:15] <Tom_itx> they're probably up to edition 28 29 or so
[11:49:18] <Tom_itx> by no
[11:49:19] <Tom_itx> w
[11:49:41] <amee2k> amazon says 2012 edition is 29th
[11:50:29] <amee2k> if i'm going to draw something where correctness is important, chances are really slim that i'll be sending it to the ass end of the world
[11:50:48] <amee2k> so i'd want some book that reflects european / german drawing practice
[11:51:09] <amee2k> and terminology as well, i suppose
[11:51:32] <Tom_itx> i think it's becoming more universal anymore
[11:51:37] <Tom_itx> with all the iso xxx stuff
[11:55:12] <amee2k> but then, for 13EUR i can't go wrong much i guess
[11:55:25] <Tom_itx> ?
[11:55:34] <Tom_itx> for the one i suggested?
[11:55:36] <amee2k> machinery's handbook
[11:55:39] <amee2k> yeah
[11:55:41] <Tom_itx> that sounds cheap
[11:55:45] <amee2k> yeah
[11:55:57] <Tom_itx> don't get stuck with the pocket ver
[11:56:00] <amee2k> 2008 ed was 130 lol
[11:56:04] <Tom_itx> you want the whole thing
[11:56:23] <Tom_itx> that price sounds about right
[11:56:45] <amee2k> http://www.amazon.com/Machinerys-Handbook-29th-Guide-Oberg/dp/0831129034/ref=sr_1_1?ie=UTF8&qid=1326390451&sr=8-1
[11:57:24] <amee2k> err, wait
[11:57:28] <amee2k> thats not the book
[11:57:35] <amee2k> its a book about the book wtf??
[11:57:56] <amee2k> ...
[11:58:07] <amee2k> 2912 ed. of the real shit is 180EUR
[11:58:17] <amee2k> 2012* even
[11:58:18] <amee2k> >_<
[11:58:24] <Tom_itx> http://www.amazon.com/Machinerys-Handbook-Toolbox-Erik-Oberg/dp/0831128003/ref=sr_1_1?ie=UTF8&qid=1326390488&sr=8-1
[11:58:30] <Tom_itx> that's what you want
[11:58:39] <Tom_itx> unless you want the same ver but 29
[11:59:04] <amee2k> the two i'm thinking about together are cheaper than yours :P
[11:59:20] <Tom_itx> it's got alot of good information in it
[12:00:13] <syrou> Tom_itx: do you have any idea of what could be happening?
[12:00:42] <Tom_itx> not unless you have some device attached to it that interferes with isp programming
[12:00:46] <syrou> it's got a 4Mb spi flash which i've desoldered so it doesn't interfere
[12:01:14] <Tom_itx> make sure it's got power on ALL the power pins
[12:01:16] <syrou> there might be some pull-up resistors and capacitors, but no more
[12:01:31] <Tom_itx> there should be a pullup on reset
[12:01:40] <syrou> ive got 3v on reset
[12:02:03] <syrou> which other points should i measure?
[12:03:05] <syrou> i've also tried to bypass the power button by directly putting 3v on the mcu vcc pin
[12:03:19] <syrou> and then it gives me a yellow blinking light
[12:03:19] <Tom_itx> are you running it at 3v?
[12:03:28] <syrou> yep, its an 8515l
[12:03:31] <Tom_itx> i think that's a 5v part isn't it?
[12:03:41] <syrou> it's a tqfp part
[12:05:59] <RikusW> the L means max 8MHz
[12:06:07] <RikusW> it can use 5V too
[12:06:37] <RikusW> m128L can afaik
[12:06:47] <syrou> but it may break other components nope?
[12:07:15] <RikusW> • Operating Voltages
[12:07:16] <RikusW> – 2.7 - 5.5V for ATmega8515L
[12:07:17] <RikusW> – 4.5 - 5.5V for ATmega8515
[12:07:23] <Tom_itx> if the circuit is designed around 3v it could
[12:07:23] <OndraSter> ye but the other parts on the board
[12:07:35] <RikusW> • Speed Grades
[12:07:36] <RikusW> – 0 - 8 MHz for ATmega8515L
[12:07:36] <RikusW> – 0 - 16 MHz for ATmega8515
[12:07:56] <RikusW> check those part's datasheets...
[12:08:27] <syrou> maybe if i unsolder the 8515...
[12:08:44] <RikusW> why unsolder it ?
[12:09:10] <syrou> because it got some components i don't know what they do
[12:09:20] <syrou> so the may interfere with the programmer
[12:09:34] <syrou> although the board has an spi connector exposed
[12:09:41] <RikusW> your programmer don't have a level translator ?
[12:09:53] <Tom_itx> it'll be alot harder to program being tqfp
[12:09:58] <RikusW> what programmer do you use ?
[12:10:08] <syrou> i think it does. it's an avrisp mkii in 010c version
[12:10:29] <RikusW> the official atmel one ?
[12:10:32] <syrou> yep
[12:10:45] <RikusW> it does have it
[12:10:54] <syrou> and i think the spien fuse is enabled
[12:10:56] <RikusW> so it should work at 3v
[12:11:08] <syrou> since i get the chip signature if i keep the pushbutton pressed
[12:11:46] <RikusW> you need to have reset connected too
[12:12:03] <syrou> the circuit is powered with a 3v cr2302 button battery
[12:12:06] <RikusW> then no pushbutton required
[12:12:12] <syrou> but i'm using 2xaa batteries
[12:12:19] <Tom_itx> maybe the battery is low
[12:12:29] <syrou> i have the reset line connected too
[12:12:49] <syrou> but the pushbutton is not for reset, it's for powering on the device
[12:13:21] <RikusW> ah
[12:13:22] <syrou> the batteries are fresh, and i measure 3v at the reset line
[12:13:26] <Tom_itx> why would you have a momentary switch on power?
[12:13:40] <amee2k> i'd use an adjustable supply for testing, really
[12:13:54] <amee2k> to rule out power issues
[12:13:55] <syrou> you press the button for a second and it powers on. repeat the action and powers off
[12:14:40] <syrou> amee2k: do you think the voltage of those 2xaa batteries is too low
[12:15:03] <amee2k> if they're good, it should work fine
[12:15:13] <syrou> i've swapped them instead of the cr2302 so they are able to give more current if needed
[12:15:16] <RikusW> from the ds it seems it needs only 3mA at 3V to program
[12:15:18] <amee2k> but asessing the state of a battery from the outside is kinda awkward
[12:15:59] <syrou> Vtarget according to the avrisp is 3.1v
[12:16:08] <amee2k> either way make sure you have good bypassing... the output impedance of most batteries is rather shitty compared to what a power supply would give you
[12:16:11] <RikusW> does it go to sleep mode or is the power disconnected with a transistor ?
[12:17:28] <syrou> RikusW: is there any way to know that with a polimeter? i'm afraid i'm not very good at electronics
[12:17:37] <syrou> there are some smd transistors
[12:18:11] <RikusW> press the button to turn it off and measure the vcc pin on the 8515
[12:18:17] <amee2k> is there a continuous trace from the battery terminals to the IC pins?
[12:18:33] <amee2k> that would be a give-away that the shut-down is in software only
[12:18:54] <amee2k> make sure to check both positive and negative sides, it might kill power on the low side since everything is floating
[12:20:28] <syrou> 0.1v when powered off
[12:20:31] <syrou> at vcc pin
[12:20:59] <amee2k> that sounds pretty off to me >_>
[12:22:21] <RikusW> add a jumper in parallel to the switch for programming ?
[12:22:28] <syrou> same thing on the gnd
[12:23:19] <RikusW> or from battery to vcc
[12:23:25] <syrou> RikusW: thats the weird thing. if i keep the pushbutton pressed while reading, i get the signature, fuses and locks
[12:23:42] <syrou> but not the flash
[12:23:56] <RikusW> is the lockbits set ?
[12:24:00] <amee2k> is self-programming enabled? o.O
[12:24:16] <OndraSter> wtf, TQFP has normal 0.5mm pitch
[12:24:16] <RikusW> if lock is not 0xFF it is set
[12:24:18] <amee2k> err, what RikusW said
[12:24:23] <OndraSter> but that is not normal .1"/x
[12:24:26] <OndraSter> omg what a mess
[12:24:50] <RikusW> OndraSter: m128 got a 0.8mm pitch
[12:24:52] <amee2k> SMT is a mess
[12:24:55] <OndraSter> oh
[12:24:56] <RikusW> avr32 got 0.5
[12:24:57] <OndraSter> 0.8mm
[12:24:57] <OndraSter> still
[12:24:59] <OndraSter> what a mess!
[12:25:08] <syrou> the fuses are this way:
[12:25:15] <syrou> high: 0xdc
[12:25:19] <syrou> low: ce
[12:25:24] <syrou> low: 0xce
[12:25:26] <OndraSter> I had to do 0.003125 inch to position the traces to the middle of the pins as much as possible
[12:25:38] <amee2k> switch to metric grid?
[12:25:45] <syrou> the lock bits are at 0xcc
[12:26:09] <RikusW> syrou: your chip is locked
[12:26:21] <RikusW> you need a chip erase to clear the lockbits
[12:26:22] <amee2k> welcome to china :P
[12:26:46] <RikusW> syrou: what sw are you using to program ? avrdude / as4 / as5 ?
[12:26:47] <OndraSter> amee2k, but when I switch to metric grid, the other parts will have problems :P
[12:26:48] <syrou> RikusW: how do you know?
[12:26:54] <OndraSter> since they are 0.1" or 0.5" :P
[12:26:56] <syrou> as4
[12:26:57] <OndraSter> 0.05"
[12:27:19] <RikusW> syrou: about the erase ?
[12:27:22] <RikusW> its in the ds
[12:27:35] <amee2k> OndraSter: switch to metric grid and start the trace at the desired position. then while you've got the trace hanging on your cursor, switch back to imperial and draw it to a grid point nearby
[12:27:47] <OndraSter> oh
[12:27:48] <OndraSter> good idea
[12:27:49] <OndraSter> thanks
[12:28:05] <amee2k> if your CAD software restricts traces to 45 degree angles you should get a reasonably well looking trace this way
[12:28:07] <syrou> but i want to read the flash contents, not to erase it
[12:28:22] <RikusW> to reverse engineer it ? ;)
[12:28:25] <amee2k> the whole purpose of lock bits is to not let you do that
[12:29:19] <syrou> but as4 has a tick on the spien fuse
[12:29:26] <amee2k> otherwise, draw a short trace with the metric grid that leads straight away from the pad, then switch to imperial and figure out how to make it "snap" to the existing track
[12:29:31] <RikusW> it seems there is a bootloader on that chip too ?
[12:29:39] <syrou> it seems to
[12:29:56] <RikusW> syrou: click on the lock tab and read the descriptions
[12:30:23] <RikusW> spien only enables the interface
[12:31:06] <syrou> the restricting one is LB, BLB0 or BLB1?
[12:31:16] <RikusW> LB
[12:31:29] <RikusW> only chip erase will clear it to FF
[12:31:48] <syrou> ok
[12:31:52] <RikusW> no other way
[12:32:07] <syrou> i thought that if spien is enabled then you've got access to it
[12:32:18] <RikusW> unless you hack the silicon itself ;)
[12:32:29] <RikusW> with some very expensive tools
[12:32:33] <syrou> why would you enable spien if you don't allow programming nor verification?
[12:33:04] <RikusW> to allow the ISP interface to work
[12:33:18] <RikusW> if spien is off you won't even be able to read the signature
[12:33:34] <RikusW> you'll need HVPP to program the chip again
[12:33:41] <syrou> but with spien off the mcu could talk with the flash chip?
[12:33:59] <RikusW> the avrisp mkii ?
[12:34:15] <syrou> no, the 8515 with an spi flash chip
[12:34:26] <RikusW> doesn't matter
[12:34:30] <syrou> ok
[12:34:48] <RikusW> but you really don't want to turn off spien, it will brick your avr....
[12:34:51] <syrou> and those lock bits also apply with HVPP no?
[12:34:55] <OndraSter> actually Eagle can snap onto the middle of the pin
[12:35:00] <RikusW> yes
[12:35:07] <syrou> damn :(
[12:35:36] <RikusW> if it weren't for the lockbits /me would have have the avrdragons fw by now :(
[12:35:42] <syrou> i understand, that's the equivalent of the cp in the pic world
[12:35:46] <RikusW> *had
[12:35:54] <RikusW> yes
[12:36:01] <OndraSter> hehe you want Dragon's fw?
[12:36:06] <OndraSter> the bootloader of Dragon?
[12:36:09] <RikusW> of course
[12:36:13] <OndraSter> hehe
[12:36:14] <RikusW> yes
[12:36:15] <syrou> and the bootloader?
[12:36:17] <OndraSter> wanna hax it?
[12:36:18] <sabesto> about to say, why would you? to improve it?
[12:36:28] <syrou> can i do anything with it?
[12:36:28] <RikusW> want the aes key ;)
[12:36:51] <RikusW> syrou: with what ?
[12:37:03] <RikusW> OndraSter: I already have the stk500 and jtag mki fw :)
[12:37:06] <syrou> with the 256 bytes bootloader
[12:37:21] <OndraSter> you know the fw is in AS's directory
[12:37:35] <RikusW> you can load your app onto the 8515 using that
[12:37:40] <RikusW> probably with the serial port
[12:37:49] <RikusW> OndraSter: its encrypted...
[12:37:59] <OndraSter> oh
[12:38:19] <RikusW> but avrprog decrypts the jtag mki and stk500 fw on the pc itself :)
[12:38:22] <syrou> it has some kind of primitive mini usb port
[12:38:26] <jacekowski> bruteforce encryption keys
[12:38:35] <RikusW> thats how I got hold of it
[12:38:37] <jacekowski> it's only going to take few million years
[12:38:40] <syrou> it may work with avrusb?
[12:39:13] <RikusW> jacekowski: by then we will be long gone...
[12:39:28] <RikusW> syrou: did you get any docs with it ?
[12:39:36] <syrou> nope
[12:39:43] <RikusW> it should tell you what protocol the bootloader use
[12:39:48] <syrou> lets say it's a skimme XD
[12:40:00] <syrou> but im on the good side
[12:40:01] <RikusW> where did get it ?
[12:40:16] <syrou> well... work related
[12:40:36] <syrou> ive got the flash contents by directly reading the chip
[12:40:42] <RikusW> does it enumerate as a comport when plugged in ?
[12:40:51] <syrou> but i hoped i could get further with the device itself
[12:41:01] <syrou> havent yet found a compatible cable
[12:41:05] <RikusW> and its all 0xFF.... because of the lockbits...
[12:41:19] <syrou> do you remember those cheapo digital cameras? thats the connector it has
[12:41:36] <syrou> it's nicer
[12:41:37] <RikusW> mini B ?
[12:41:39] <syrou> for example
[12:41:40] <syrou> :1000000000000101020203030404050506060707B8
[12:41:56] <syrou> nope
[12:42:13] <syrou> it's kinda rectangular, similar in size to miniusb
[12:42:58] <syrou> its got a ds1302 rtc chip with its own backup battery
[12:42:59] <RikusW> get a cable and plug it in, it will probably enumerate as a comport....
[12:43:16] <RikusW> is there perhaps a FTDI ic too ?
[12:43:19] <RikusW> on the board
[12:43:20] <syrou> so there may be some sort of avrusb firm in it
[12:43:29] <syrou> nope, no ftdi nor similar chip
[12:43:51] <syrou> well it may directly expose the ttl signals
[12:44:05] <syrou> and the use an external ttl <> rs232 converter
[12:45:14] <RikusW> if the usb part does work, try using avrdude to connect using different protocols
[12:45:21] <RikusW> if comes with winavr
[12:45:43] <syrou> or maybe an hyperterminal
[12:45:45] <syrou> nope?
[12:46:08] <RikusW> avrdude might be more helpful
[12:46:19] <RikusW> stk500v2 won't work at all in HT
[12:47:13] <syrou> and about reenabling the lock bits by erasing the device, i can only do that via HVPP?
[12:47:23] <RikusW> and ISP
[12:47:33] <RikusW> just click on chip erase
[12:47:44] <RikusW> but you'll lose the bootloader
[12:47:48] <syrou> well i don't want to :)
[12:47:55] <syrou> since i'll loose everything
[12:48:15] <RikusW> then try avrdude + usb cable first
[12:48:50] <syrou> i have to find that cable
[12:49:34] <RikusW> I have a weird usb cable here, its plug is about half the size of mini B
[12:49:40] <RikusW> no idea what its for
[12:49:56] <RikusW> rectangular
[12:52:14] <syrou> this is the connector
[12:52:16] <syrou> http://img.hisupplier.com/var/userImages/old/asiadns/asiadns$829142138.jpg
[13:05:29] <RikusW> might be tha cable I have
[13:42:25] <OndraSter> uh
[13:42:29] <OndraSter> FTDI FT232RL driver is broken
[13:42:31] <OndraSter> it gets stuck often
[13:42:35] <OndraSter> won't unload on device disconnect
[13:42:49] <OndraSter> and plugging in the device again doesn't create virtual serial port anymore
[13:43:13] <OndraSter> reboot then
[13:44:53] <mrsun> windows... ? :P
[13:50:41] <Steffanx> You can crash ubunu/opensuse too with a ft232rl ..
[13:51:17] <karlp> well done!
[13:51:40] <Steffanx> Yes yes
[13:53:25] <Steffanx> OS X too btw
[13:53:41] <Steffanx> Or at least … break the usb drivers + 100% cpu load
[14:04:39] <mrsun> hmm, digital capacitor, nice ... pitty it had so low range :/
[14:04:44] <mrsun> 7 - 14.5pF
[14:05:39] <mrsun> oh wasnt to hard to build oneself either :P
[14:28:17] * specing is happy with his pl2303
[14:35:46] <syrou> thanks for your help!
[14:35:48] <syrou> see ya
[14:39:02] <Bushman> ave
[17:02:54] <myradlife> Hello, I've been having trouble for the last 3 days interfacing USART to my PC hyperterminal on the AT90USBKEY2 which uses the at90usb1287 controller. Does anyone have experience with setting this up? I've written the USART driver for other AVRs before, and I've never had this kind of trouble.
[17:05:41] <myradlife> ... I'm simply trying to print a few characters from the microcontroller to the terminal. I don't get the expected characters output. Also, I have to change the BAUD setting on the terminal from very different than I'm intending to set it per the at90usb1287 datasheet.
[17:18:32] <Roklobsta_> do any of you guys use the xemga much?
[17:18:36] <Roklobsta_> xmega
[17:23:10] * inflex hasn't - but is curious to try them in the future
[17:37:01] <BrentBXR_> more avrs
[17:37:08] <BrentBXR_> i have 2 many
[17:46:04] <Roklobsta_> i am looking for a cheap xmega devkit but the only ones i can see are 128A1's as per the Xplain board.
[18:22:56] <Kevin`> is there something obvious that would cause extremely low drive strength on an atmega328p? if I put a 10k resistor on a pin, it will mostly hold it either high or low
[18:27:19] <CMiYC> you didn't set it as an output
[18:29:19] <Kevin`> heh, sometimes all it takes is someone to prod you into trying to prove them wrong.
http://pastebin.com/8DTRWLKL
[18:29:27] <Kevin`> I suppose thanks are in order
[18:29:32] <Kevin`> this isn't a good start
[18:30:13] <Tom_itx> mama said there'd be days like this
[18:30:38] <CMiYC> they is your magic i/o pins
[20:31:54] <Kevin`> yeah, totally
[20:33:50] <vectory> os, i still need to setup interupts, figure out how to connect and drive the led matrix and figure out pwm to drive a piezo buzzer
[20:34:01] <vectory> *so
[20:34:41] <vectory> only read 30% of the datasheet, so i shouldnt ask too much questions, but ...
[20:35:05] <Kevin`> funny, sounds similar to my project
[20:35:44] <vectory> ha
[20:36:45] <Casper> vectory: just skim throught the section that you do not need yet
[20:36:53] <Casper> it will be more effective than reading all
[20:36:55] <vectory> thats what i do
[20:36:59] <Casper> good
[20:37:03] <Casper> not too discouraged?
[20:37:12] <vectory> parallel to a book on atmega16
[20:37:26] <vectory> discouraged, why? i got you :D
[20:37:33] <vectory> if all else fails
[20:38:37] <vectory> a problem for me is the interpretation of the typical characteristic charts
[20:38:41] <Casper> some get discouraged by the summary datasheet... the 30 pages one...
[20:38:43] <vectory> since there are no annotations
[20:39:05] <Casper> so imagine the 500+ pages doc that they get told to get...
[20:39:26] <vectory> well, i didnt say im too enthusiastic.
[20:39:47] <vectory> a problem for me is the interpretation of the typical characteristic charts
[20:40:10] <Casper> like speed vs voltage?
[20:41:22] <vectory> well, yes. i would think i got that figured out, but a few charts later there is pin source current vs output voltage. how do i know either?
[20:42:31] <Casper> at 5VCC, at x current the voltage on the output pin will be y volts
[20:43:17] <vectory> but current is on the y axis, so i thought current depends on voltage not other way around
[20:43:49] <Casper> you can graph voltage vs current or current vs voltage
[20:43:52] <Casper> it is not important
[20:44:30] <vectory> so current depends only on the load
[20:44:36] <vectory> and the voltage?
[20:45:11] <Casper> the graph show that when the VCC is 5V, if you take x current, the output will be y
[20:45:13] <Casper> but
[20:45:19] <Casper> you can read it another way
[20:45:32] <Casper> if you want y volts minimum, how much current must you take max?
[20:45:59] <Casper> it's the same graph, just you can read it differently
[20:46:11] <Casper> another example, supply voltage vs clock speed
[20:46:56] <Casper> you can read it as: "I have x volts, what is the highest clock I can use?" or "I need x speed, what is the minimal voltage required?"
[20:48:42] <vectory> well, i need 2 or 3 V on the output voltage, but thats different. i thought there should be a graph relating Vcc to max. Voh
[20:49:22] <Casper> it's really to tell you the effect of the output "resistance"
[20:49:33] <vectory> aha
[20:49:45] <vectory> tuts i read didnt cover that
[20:49:50] <Casper> remember, everything have resistance
[20:49:54] <Casper> this is basic electronics
[20:50:08] <Casper> you "never" have full supply voltage on an output
[20:50:13] <Casper> you always have a bit less
[20:50:46] <Casper> how much less? it depend on the technology used (bjt vs fet for example) and the "quality" of the parts used
[20:51:14] <Kevin`> is there a 24bit data type normally?
[20:51:29] <Casper> avr is fet based, so the output driver have a resistance, so the output drop under load
[20:51:32] <CMiYC> when is "normal"
[20:51:32] <Casper> Kevin`: nope
[20:51:50] <Casper> Kevin`: never hear of an architecture with native 24 bits
[20:52:02] <Kevin`> native doesn't matter
[20:52:10] <vectory> neither in gcc
[20:52:37] <Casper> it's been asked often, they refuse to make it, claiming that you should just use a 32 bits
[21:04:38] <vectory> i seem to be running into a wall with this, maybe i should just experiment with a dmm
[21:04:45] <vectory> fear to break sth, tho
[21:06:20] <vectory> oh and im running out of timers :D
[21:08:37] <vectory> i read that pwm can controle voltage levels, so i need one timer to pwm a signal according to desired voltage and another timer to keep track of the duty cycle of the signal in terms of kHz
[21:09:42] <vectory> or more likely i could neglect the voltage level and just pwm at the desired frequency
[21:10:58] <Kevin`> led array working :D
http://kwzs.be/~kevin/P1120001.AVI
[21:11:19] <Kevin`> speaking of which, what's a magic program I can use to do anything -> nice h.264
[21:11:52] <vectory> any frontend to ffmpeg perhaps
[21:14:00] <Kevin`> http://pastebin.com/XLXuYsuj
[21:28:43] <eatyourguitar> you might want to just upload that to youtube so its compressed and streaming. 18mb is kinda a lot for a random click. its probably raw video uncompressed
[21:29:21] <Kevin`> as if I have a youtube account :)
[21:34:06] <Kevin`> probably should have one
[21:37:14] <vectory> its not like 18 mb ment anything nowadays
[21:37:16] <eatyourguitar> I shoot low quality videos on my phone and upload directly to youtube. then I just grab the link on the computer. easy
[21:37:38] <Sgt_Lemming> fucking telemarketers >_<
[21:37:44] <Kevin`> phone? what's that
[21:37:52] <eatyourguitar> yeah but I have to wait 30 seconds to watch it and I need to go and delete it later
[21:38:35] <eatyourguitar> telemarketers are not so bad anymore
[21:38:39] <eatyourguitar> with the new laws
[21:38:41] <vectory> you are making all your videos owned by youtube that way
[21:39:01] <eatyourguitar> no exactly
[21:39:23] <eatyourguitar> you can choose your license or label it with your own copyright
[21:40:22] <eatyourguitar> in america you already have a copyright. if you dont say you have a copyright in the description of the video or somewhere, then you tube will claim ownership.
[21:42:11] <eatyourguitar> anyone that creates content has automatic copyright since 1989
[21:42:47] <vectory> yeah, but uploading could contract you to stand back from your rights
[21:43:30] <eatyourguitar> right but if it is already copyrighted or you dictate public license things are different.
[21:44:07] <eatyourguitar> like if warner brothers posts something that has a registered copyright then it still belongs to them
[21:45:22] <eatyourguitar> basically youtube thinks they own everything but they dont
[21:45:50] <eatyourguitar> you dont see them taking major movie companies to court
[21:46:04] <eatyourguitar> cause they would loose
[22:30:51] <CMiYC> youtube made the mistake of thinking you'd read the part of your contract with them that says "if you don't claim ownership we will"
[22:31:06] <CMiYC> or maybe it is the end users that made the mistake... tough to tell. nobody likes accepting blame
[22:57:17] <vectory> Kevin`: which avr is that for?
[23:07:20] <Kevin`> http://pastebin.com/YgH9i6gX - lovely fast math :D. suggestions welcome (although it runs fast enough apparently for what it needs to do)
[23:08:05] <Kevin`> vectory: atmega328p. way overkill, but I had a few, and it has a bunch of io pins to replace the decade counter I couldn't find
[23:10:12] <Kevin`> or is a decade counter what i'm thinking of? the one that cycles through a single pin being active
[23:10:27] <Kevin`> ring counter?
[23:10:31] <Kevin`> whatever, couldn't find them
[23:15:22] <vectory> a frequency devider?
[23:18:44] <Kevin`> http://en.wikipedia.org/wiki/Ring_counter
[23:19:02] <Kevin`> to connect to the display rows
[23:20:10] <vectory> thats a ring shift register to me
[23:20:21] <Kevin`> right, it is
[23:20:30] <vectory> didnt find any?
[23:22:23] <Kevin`> hmm, the only pins i'm not using are the programming interface. I guess I can have the speaker make noise while programming :). or find more parts
[23:23:23] <vectory> that adds capacitive resistance, doesnt it?
[23:23:34] <Kevin`> yes
[23:28:19] <Kevin`> my thingy, with the display running and counting, is apparently using 0.25ma, or 0.011w. should last a while on 3xAA batteries
[23:28:27] <ziph> What's capacitive resistance? :)
[23:28:45] <vectory> hm, i thought there was
[23:28:46] <Kevin`> s/resistance/impedence
[23:29:09] <ziph> "Reactance" works too.
[23:29:44] <Kevin`> I thought it was capacitive diractence
[23:30:22] <vectory> i probably ment impedance, though i thought it was the same as resistance
[23:30:46] <Kevin`> 1800/(0.250/100.1)/24/365
[23:30:46] <Kevin`> 82.2739758147
[23:30:46] <vectory> 0.25mA, how did you get it that low? i dont see use of sleep in your code
[23:30:48] <Kevin`> that can't be right
[23:30:53] <ziph> Resistance is just the real component of impedance, reactance is the imaginary component of impedance.
[23:31:06] <ziph> So the impedance is the resistance plus the reactance.
[23:31:28] <Kevin`> vectory:
http://kwzs.be/~kevin/clock.txt
[23:32:08] <ziph> And you'll find "capacitive reactance" more commonly said than "capacitive impedance".
[23:32:44] <vectory> ok
[23:34:19] <vectory> 0.25 mA, including the leds? hell, thats nice
[23:34:28] <Kevin`> 0.25v
[23:34:39] <Kevin`> with a 100r resistor on the power supply
[23:34:48] <vectory> oO
[23:35:28] <vectory> thought pico power was > 1.8V
[23:35:41] <Kevin`> 0.25v across the resistor
[23:35:48] <Kevin`> the power supply is around 4.5v
[23:46:09] <vectory> Kevin`: how dimm would the leds be, if you had the interupt run at 128hz?
[23:47:00] <Kevin`> vectory: the reason it's running fast is so the leds don't appear to flash
[23:47:37] <vectory> just wondering how low you could go
[23:47:41] <Kevin`> anything more than about 2ms clock for switching the segments was flashy
[23:48:22] <Kevin`> I actually bumped it up to 4khz instead of 1khz so I could lower the duty cycle to 1/32 instead of 1/8
[23:49:00] <Kevin`> (each led is only on at most 1/8th of the time)
[23:49:08] <vectory> since i use 32khz crystal, i cant get that high a frequency and still have time enough cycles x)
[23:49:42] <Kevin`> on lots of chips you can use BOTH a 32.768khz oscillator and a high speed oscillator
[23:49:48] <Kevin`> that's what I plan to do
[23:50:04] <Kevin`> run the 32khz one all the time, and turn the cpu oscillator off completely when it's not doing something
[23:50:45] <vectory> hm, yes, forgot that
[23:51:28] <Kevin`> 63hz/500hz would work for mine, anyway
[23:51:31] <Kevin`> experiment
[23:51:50] <vectory> not even nesecarry tho, since 0,25 mA will have you running for over a year all ready
[23:52:09] <Kevin`> I need to do it because the cpu's oscillator isn't accurate enough for keeping time
[23:52:24] <Kevin`> and this is a timing device
[23:53:17] <vectory> ah yes
[23:54:34] <Kevin`> if you don't need accurate time or speed you could run from the watchdog oscillator
[23:56:28] <Kevin`> what are you making, anyway?
[23:56:45] <vectory> a clock, with alarm
[23:56:50] <Kevin`> h4eh :D
[23:56:53] <Kevin`> -4
[23:57:16] <Kevin`> i'm making a clock + egg timer (countdown)
[23:57:34] <Kevin`> might include a time-of-day alarm too though, just so you can steal ALL of my code :)
[23:58:05] <vectory> haha, ive got the code figured out, except for the technical parts
[23:58:32] <vectory> since i dont know enough yet, rather abstract ideas of pull ups etc -.-'
[23:58:47] <Kevin`> hmm, I need buttons. I really need to find a shift register I guess
[23:59:09] <vectory> but yea, im trying to read your code and it reminds me :)
[23:59:36] <vectory> you could comment a bit more ;D
