#avr Logs

Dec 24 2017

01:06 AM rue_shop3: how fast did I say that adc converts?
01:09 AM rue_shop3: 10Khz?
01:09 AM rue_shop3: if I'm trying to detect a whistle...
01:44 AM rue_shop3: hmm, can I implement a window freq filter on an avr in the order of 2Khz?
01:47 AM rue_shop3: hmm specific fft?
01:48 AM rue_shop3: no, that wouldn't work for what I want
01:53 AM Casper: I bet hardware would be easier
01:54 AM rue_shop3: right, and then I can just use a digital counter to measure the freq
01:57 AM Casper: and in some case you could just use a standard io pin that detect when there is something
01:57 AM Casper: if your filter is tight enought
01:57 AM rue_shop3: not hard to get 3db/decade
01:57 AM rue_shop3: hard not to actaully
01:58 AM Casper: and easy to get even tighter
01:59 AM rue_shop3: oh
01:59 AM rue_shop3: I have
01:59 AM rue_shop3: somewhere
01:59 AM rue_shop3: from a motolola demo board, a tunable filter
01:59 AM rue_shop3: they used it to make a freq sweep meter
02:00 AM rue_shop3: hmmmm
02:01 AM rue_shop3: 52c, but I turned it down
04:20 AM Emil: rue_shop3: damn
04:21 AM Emil: rue_shop3: you dont know anything about avrs :D
04:22 AM Emil: anycasThe ADC clock works from 50k-1M
04:22 AM Emil: woth 125k-200k the optimal range
04:23 AM Emil: if you know the frequency beforehand then the best way is imho hardware tight lp and then sample at 2.5x the frequency of interest
04:24 AM Emil: and then just calculte dft for that band
04:24 AM thardin: or tight bandpass if it is above 100 kHz or so
04:25 AM thardin: thanks to the magic of aliasing
04:26 AM Emil: and the adc takes 13.5 clocks of freerunning samples and 13 clocks if manually started
04:26 AM Emil: thardin: eh?
04:27 AM Emil: thardin: you mean we are interested in some vand above sampling frequency?
04:27 AM Emil: yeah sure, bandpass that band and then sample and it'll appear on a lower band
04:32 AM thardin: yeah. saw someone do that recently
04:33 AM thardin: got my codec2 patchset sent in: http://ffmpeg.org/pipermail/ffmpeg-devel/2017-December/222862.html
04:33 AM thardin: no replies yet tho
04:38 AM Emil: thardin: the issue
04:38 AM Emil: thougj
04:39 AM Emil: is that the input bandwidth needs to be higher
04:39 AM thardin: sure
04:39 AM Emil: hmm
04:39 AM thardin: I also recently learned that taylo-style demodulators are popular in laser research
04:39 AM Emil: iirc the datasheets mentioned something about it
04:43 AM thardin: if nothing else the input capacitance will cause some attenuation
05:16 AM nuxil: mornings!
05:17 AM nuxil: http://www.informationisbeautiful.net/visualizations/million-lines-of-code/ lol look at google compare to the rest :)
05:19 AM nuxil: me and a friend was talking about how many million lines of code MS/google has, so i googled it up :p
05:41 AM amigoj_festivus is now known as amigojapan_bnc
06:17 AM nuxil: in my manual it says that the SBIS takes clocks 1/2/3. how does this work? if bit is set in I/O reg, it skips and do 1 cycle and if its not ?
06:22 AM nuxil: Cycles:
06:22 AM nuxil: 1 if condition is false (no skip)
06:22 AM nuxil: 2 if condition is true (skip is executed) and the instruction skipped is 1 word
06:23 AM nuxil: 3 if condition is true (skip is executed) and the instruction skipped is 2 words
06:23 AM nuxil: 2 and 3 confuses me
06:41 AM nuxil: so if bit is not set. it skips instructions. but how does it work?. if bit is not set does it check the Hbyte 1st and see if its set and if not, it checks the Lbyte of in the 32I/O reg ??
06:46 AM nuxil: i mean if the bit is set.
08:52 AM polprog: Merry christmas and a happy new year everybody!
10:04 AM nuxil: https://www.youtube.com/watch?v=UIKGV2cTgqA&t=5s LOL
10:10 AM Thrashbarg: heh sounds like he was reading a CS book
10:10 AM nuxil: base 8 is just like base 10 really. if you miss 2 fingers :p
10:11 AM nuxil: best line in tha video
10:11 AM Thrashbarg: hehe
10:11 AM Thrashbarg: at least base-16 hadn't caught on in the 60's
10:12 AM Thrashbarg: heard his periodic table song?
10:12 AM nuxil: no havent
10:13 AM Thrashbarg: https://www.youtube.com/watch?v=zGM-wSKFBpo
10:14 AM nuxil: hehe
10:15 AM nuxil: this one is better tho : https://www.youtube.com/watch?v=VgVQKCcfwnU
10:15 AM Thrashbarg: heh
10:16 AM Thrashbarg: Lehrer had a warped sense of humor... listen to "Poisoning Pigeons in the Park"
10:17 AM Thrashbarg: not merely twisted but actually sprained
10:18 AM Casper: https://youtu.be/nWZMfPP34g8?t=849 <== ouchie
10:19 AM Thrashbarg: ...why?
10:19 AM nuxil: why. what dosent people do to be on TV/Ger some pr,.
10:19 AM nuxil: *get
11:57 PM day__ is now known as day