Back
[03:18:59] <_abc__> Hello. I saw a weird problem. Atmega with timer2 set to async external clock for 32khz xtal did repeated watchdog resets while code was waiting on ASSR sync bits to clear. Is it normal for ASSR sync bits to not clear if there is no external clock on TOSC and async mode is selected?
[04:56:14] <megal0maniac> Quiet in these parts today
[04:56:51] <GuShH> megal0maniac: must be the time
[04:57:31] <GuShH> every other channel I'm in hasn't been active since I joined a few minutes ago
[04:58:24] <megal0maniac> It's the middle of the day for most Europeans
[04:58:44] <GuShH> That's when you are the most drunk though
[04:58:46] * GuShH runs
[04:59:14] <GuShH> 7:46AM, had to wake up at 7... snoozed for 20 minutes
[06:34:16] <tzanger> 2h ahead of me... that would put you on the east coast IIRC
[06:34:30] <tzanger> one of those states that sticks way out, probably NY
[06:34:49] <tzanger> this is weird.
[06:35:17] <tzanger> I have an LM35 temp sensor hooked up with two 1N914 diodes between its common and the real system common
[06:35:34] <dekroning> hi all
[06:35:41] <tzanger> there's a ~18k pulldown on the temp sensor's output down to the real common
[06:35:59] <tzanger> this is right out of the datasheet for measuring down to -55C (I only need about -20C)
[06:36:24] <tzanger> the output of the temp sensor goes to an ADC channel. The junction of the diode/LM35 common goes to another ADC channel
[06:36:38] <tzanger> that allows me to sample the output of the sensor + the diode drop at that time and subtract them
[06:37:38] <tzanger> the system reads really well when the two ADC channels are AD0-3 (i.e. not the ADC channels mux'd with the JTAG pins) - this is an AT90CAN
[06:38:27] <tzanger> I disabled JTAG and moved the diode drop input to AD6 which is shared with TDO since that was the one I figured would be "most loose" since the AT90CAN is floating it and the JTAG dongle is weakly pulling it up
[06:38:52] <tzanger> of course it reads like shit with the JTAG dongle plugged in, but it reads only slightly less shit with nothing connected
[06:39:20] <tzanger> so I finally broke down and installed a rail-to-rail (both in and out) opamp (TLV358) to really drive those analog values
[06:39:26] <tzanger> that cleaned it right up, which was expected
[06:39:50] <tzanger> what was not expected, though, was that my temperature readings are now exactly 4C low across temp, and I am not sure why
[06:40:23] <tzanger> the raw, unfiltered ADC values are reading a little off from what I'd expect (measuring with a voltmeter)
[06:41:29] <tzanger> the diode drop channel reads 960mV when the meter says 869mV, and the temp channel reads 1120mV when the meter reads 1086mV
[06:42:44] <tzanger> 1120-960 = 160mV = 16.0C, which is what my display shows, but the actual voltages are 217mV = 21.7C which is what the temp actually is
[06:43:16] <tzanger> I can't find a source for the high readings, especially now that the ADC channels are being fed with a couple of voltage followers
[06:44:17] <tzanger> maybe I should stick a few kohms of resistance between the op-amp output and the adc input to give it a slightly higher source impedance, although these are DC values, I don't expect that it's an impedance mismatch. besides the datasheet says the channels are designed for ~10k impedance loads or lower
[06:45:24] <tzanger> the only other thing I can think of is that the op-amp return is system GND and not AGND, but there isn't more than a few microvolts of difference between those two nets
[06:48:25] <tzanger> I'm giving the ADC 100ms between the time I select a MUX channel and the time I tell it to sample, which should be eons to the ADC
[06:48:48] <tzanger> er no, 10ms
[06:48:53] <tzanger> still should be plenty of time
[09:51:40] <[z_z]> So I'm working on a usb device. I use vendor class, and want two interfaces. interface 1 with bulk 1 in and bulk 1 out... I'm using the ASF and it has udd_ep_run() that takes an endpoint as argument. how do i tell it what interface?
[09:56:14] <OndraSter_> I don't think you can have two interfaces enabled at the same time on USB device
[09:56:20] <OndraSter_> and you have to request for the change
[09:56:24] <OndraSter_> no wait
[09:56:26] <OndraSter_> that was configuration
[09:56:27] <OndraSter_> ignore me
[10:02:26] <[z_z]> yes. i have only one config. but two interfaces... I can solve the problem by having iface 1 ep1 ep2 iface2 ep3 ep4 ... but thats err... not fun.
[10:05:21] <RikusW> OndraSter_: you do you change config anyways ? does windows allow you to ?
[10:05:27] <RikusW> how and where ?
[10:05:56] * RikusW will probably try multiple interfaces on the coco sometime ;)
[10:06:02] <RikusW> m32u2 got only 5 eps....
[10:06:53] <OndraSter_> [z_z], no idea
[10:06:58] <OndraSter_> RikusW, no
[10:07:02] <OndraSter_> I do not
[10:07:43] <OndraSter_> I thought that CDC uses two itnerfaces
[10:07:45] <OndraSter_> interfaces*
[10:08:27] <OndraSter_> I am tired
[10:08:30] <OndraSter_> and I should be learning math
[10:08:37] <OndraSter_> and I know nothing
[11:06:56] <tzanger> hm, AN6 seems to be stuck at 384 counts. That is oddly suspicious
[12:29:26] <tzanger> interesting.
[12:30:19] <tzanger> *NOT* disabling JTAG makes the AN6 input "move" from 0x180
[12:57:12] <[z_z]> I got myself confused... is it possible for a usb device to have two interfaces. where each interface has a bulk endpoint 1 IN and a bulk endpoint 1 OUT?
[13:00:46] <OndraSter_> [z_z], I am not sure it is possible - they cannot share the same address
[13:02:01] <[z_z]> same address across interfaces or within the same interface like: can't have both 1 IN and OUT?
[13:03:11] <OndraSter_> any endpoint addresses on one configuration
[13:06:35] <[z_z]> ah. somehow i find that divergence from the config->interface->setting->endpoint hierarchy counter-intuitive.
[13:07:03] <OndraSter_> USB is counter intuitive on its own
[13:07:35] <RikusW> usb is overcomplicated....
[13:07:54] * RikusW wants superspeed rs232 ;)
[13:08:08] <OndraSter_> hehe
[13:08:13] <OndraSter_> 6Gbit RS232!
[13:08:48] * RikusW wonders if sata can be abused for that purpose ?
[13:09:18] <OndraSter_> SATA is not as simple either
[13:09:26] <OndraSter_> it is not open spec
[13:09:29] <OndraSter_> well, "free" spe
[13:09:29] <OndraSter_> c
[13:10:04] <[z_z]> well. USB is everywhere...
[13:10:46] <OndraSter_> which is sad
[13:10:50] <OndraSter_> what an awful thing
[13:10:56] <GuShH> firewire :<
[13:11:01] <OndraSter_> firewire is dead
[13:11:16] <OndraSter_> I never had anything that would use fw
[13:11:21] <GuShH> marketing goes: firewire has negative connotations, it denotes fire on your wire. people won't like that, people won't buy that.
[13:11:31] <OndraSter_> it was nice thing
[13:11:35] <GuShH> and they brainstormed for years until it went bust
[13:11:35] <OndraSter_> fast and high power
[13:11:54] <GuShH> OndraSter_: there's a parody Conan did about usb vs firewire a while ago
[13:11:58] <GuShH> funny shit.
[13:12:00] <OndraSter_> eh
[13:12:12] <RikusW> and 2 different voltages on the line, 12V and 48V iirc, so it can literally cause fire :-P
[13:12:15] <GuShH> youtube must know about it!
[13:12:36] <OndraSter_> youtube is fu.ked up with the new layout too
[13:12:36] <GuShH> it was a nice standard that never got standard
[13:12:46] <RikusW> I wonder if the FW specs is any simpler than USB ?
[13:12:49] <OndraSter_> it takes me 2 CLICKS and few seconds of loading to get to new videos from subscriptions
[13:13:11] <OndraSter_> because youtube now opens in "suggestions" and on subscriptions it shows everything, not just uploaded videos
[13:13:15] <OndraSter_> if it remembered at least..
[13:13:21] <OndraSter_> also it is awful off-centered
[13:13:24] <GuShH> hm
[13:16:25] <RikusW> seems new version tend to get worse these days ?
[13:16:35] <RikusW> *versions
[13:17:24] <OndraSter_> yes
[13:17:26] <OndraSter_> also
[13:17:32] <OndraSter_> I am watching 720p video -- loads fast
[13:17:37] <OndraSter_> switch to 1080p -- keeps buffering
[13:18:01] <RikusW> what do you expect ?
[13:18:10] <RikusW> try that on GPRS...
[13:18:12] <OndraSter_> to load at least fast enough?
[13:21:37] <GuShH> OndraSter_:
http://scoop.intel.com/ajay_bhatt_on_the_conan_obrien_show/ hopefully links still work
[13:21:39] <RikusW> GPRS won't even give 1fps :-P
[13:21:55] <GuShH> ffs it doesnt, not even intel can keep links up
[13:22:29] <GuShH> http://www.engadget.com/2009/10/10/conan-obrien-talks-to-the-co-creator-of-usb-on-the-tonight-show/
[13:22:31] <GuShH> also broken...
[13:22:40] <GuShH> big sites, and they can't keep their links up
[16:04:09] <OndraSter_> nope, university math is not for me
[16:04:19] <OndraSter_> I want to work in electronics dept
[16:04:33] <OndraSter_> embedded systems
[16:15:07] <Tom_itx> pfft
[16:15:21] <Tom_itx> you need math for anything
[16:18:21] <OndraSter_> well sure
[16:18:29] <OndraSter_> but you don't need the *theory* behind every single thing
[16:18:32] <OndraSter_> you need to know how to use it
[16:18:33] <OndraSter_> and when to use it
[16:18:41] <OndraSter_> not who thought about it
[16:19:50] * specing grabs a sheet of paper to prove OndraSter_ wrong
[16:22:26] <OndraSter_> :D
[16:23:09] <OndraSter_> I mean, why should I care that some d'Frenchie thought that if q < 1 the limit is converging? Of course it is converging, because it is reaching a fixed number! I don't need to know *who* was the author of that
[16:24:46] <OndraSter_> there are 70 sentences and definitions which I am supposed to remember - including their names.
[16:30:01] <yunta> OndraSter_: personally I regret a bit that I didn't learn math better at uni
[16:30:07] <yunta> it's very useful at times
[16:30:24] <specing> Im studying math :D
[16:31:25] <yunta> simple things like integration, derivation, corelations, matrices (eigenvectors etc)
[16:32:31] <yunta> these are really useful
[16:32:43] <yunta> but I suck completely at them
[16:32:55] <specing> yunta: we really delve into the details
[16:33:34] <specing> like what is the derivative of f(x) = log_a (x)? and why is it 1/ (xln(a))?
[16:33:46] <yunta> sounds cool :)
[16:33:47] <specing> Why, why, why, why
[16:35:22] <diametric> yes.. yes... i know some of these words.
[16:35:26] <yunta> well, I always wondered how people found all those derivations
[16:35:34] <yunta> some of them are obvious
[16:35:38] <yunta> but some really aren't
[16:44:55] <OndraSter_> haha
[16:44:59] <OndraSter_> derivations? pff, middle school stuff
[16:45:01] <OndraSter_> also
[16:45:01] <OndraSter_> http://www.youtube.com/watch?v=YUhb0XII93I
[16:45:15] <Steffanx> All of them are 'obvious' when you see how they did it yunta :P
[16:47:10] <Steffanx> yunta time to follow the courses of MIT + do the homework? :P
[16:47:19] <Steffanx> or some coursera crap?
[16:49:14] <yunta> no time :)
[16:49:46] <Steffanx> you have time to talk here :)
[16:50:08] <yunta> haha, true
[16:50:19] <yunta> I value that more than math knowledge :)
[16:50:37] <Steffanx> So does OndraSter_
[16:54:05] <yunta> sure, I'm just trying to say it's not worthless to learn
[16:55:03] <specing> lim t->oo (your_math_skills) = 0
[16:55:09] <specing> ;P
[16:58:04] <Steffanx> What's next? Differential equations?
[16:59:11] <specing> well
[17:04:24] <specing> Steffanx: One could argue that the tangent on the graph of your math skills is negative :)
[17:04:47] <specing> s/is negative/is a negative function/
[17:04:56] <Steffanx> Let's check with wolfram alpha
[17:05:35] <Steffanx> "Using closest Wolfram|Alpha interpretation: tangent" too bad
[17:05:41] <Steffanx> no skills at all
[17:12:03] <specing> We had about 50 hours of analysis
[17:12:24] <specing> every hour I filled an a4 sheet on both sides
[17:12:38] <specing> that leaves me with like 100 pages of material to go through :(
[17:12:45] <Steffanx> wonderful :)