yeah that filter was cool
ha, yeah using analog feedback loops sounds cool
Posted by renesis at 07:44 | permalink | 0 comments
lordpil's blog
right but i kind of want to figure out the dsp stuff
that is neat, but dsp compressor is def on list of stuff to figure out
just because analog compression all sorts of messy
yes everyone seems to want to have a stab at that one
optical is actually prob best setup
Posted by renesis at 07:39 | permalink | 0 comments
damn the russians do not have it
they do have a bunch of books on biomedical synthesis
sounds like nico books
Posted by renesis at 06:50 | permalink | 0 comments
russians have all the books
yeah google books are usually long preview
but libgen has all the shit
Posted by renesis at 06:44 | permalink | 0 comments
so like, you could do 32b phase accumulator, and have 32b tuning words to select freq
and a 256 sample table, and you just use the high byte of the phase accumulator to index
like, itll repeat and skip samples to make shit work
and its crazy accurate freq output
and its literally just add to accumulator, mask/shift whatever for your index
damn that page looks like a lot, neat
ooh
well i can ask the russians =\
Posted by renesis at 06:25 | permalink | 0 comments
i have an older f4 discovery to play with
also have an f0, bet i could get that working pretty well
but right if i can do it 48khz on the f0, can instantly do it 192khz on the stm32
right going to try and avoid that tho
i just get it working first
the dds stuff is just adding to a uint32_t and letting it wrap
well, you have a phase assumulator, more bits is more freq resolution
and then you pick off however many bits off the top to index a wavetable
Posted by renesis at 06:20 | permalink | 0 comments
haha, doidic at line 6 is like, OH YEAH THE EQUATIONS ALL USE FRACTIONS OF Fs FOR FREQUENCY!!!
he said it took him forever to figure that out, shit made a lot more sense after he told me that
right
sweet
i got dma stuff working on stm32 so should work pretty cool
want to see how many channels of dds i can get going at once, and then have a couple sections of filter
Posted by renesis at 06:15 | permalink | 0 comments
i kind of want to see if i can get some sort of cap touch velocity/aftertouch thing working
but yeah if i can make digital filters happen ill be way happy
i prob just start fucking around with the dac on the f4 discovery board
are you using a library or coding your own stuff?
im looking for old dsp book had really good filter explanations
i think i found it, computer music something something, its like 80s shit
yeah the dsp guide thing
this thing was about music and was just kind of goofy and old
like academic + synth dork
oh cool
Posted by renesis at 06:10 | permalink | 0 comments
jero32: if you can think of it, its probably been done
digital synths been around since giant dip logic gates
you just need mem and an adder, really
synth done
one of the places i worked was founder by olberheim geeks
*founded
and a few engineers there worked on the alesis digital controlled analog synths
jupiter series i think?
oh, adromeda i guess, jupiter is roland thing
cool
Posted by renesis at 06:05 | permalink | 0 comments
kostix: anyway suggestions for synth dacs?
yeah, just the ic
i want to play with dds -> ADSR stuff
http://www.ti.com/product/pcm1754/description
jero32: yeah im actually considering 24b r2r dacs
resistors and a decent opamp buffer
right i dont want to deal with that
fuck nyquist i want 96k minimum
also just fun coding against a fast clock
Posted by renesis at 06:00 | permalink | 0 comments
it didnt die!
i thought of way to do it without vented bolt and tiny wires, bearings would be loaded up more, prob okay tho
Posted by renesis at 05:52 | permalink | 0 comments
so like, fast PID for rotation speed with disco gyro feedback, less fast PID for pendulum
then integrate speed for trimming target angle to keep it from rotating like crazy
anyway i kind of want to fuck with opencv at some point
jero32: did you get to end with standoff nunchucks
Posted by renesis at 05:47 | permalink | 0 comments
i posted little imgur vids of it like month ago
seems to
not like i worked out the physics
but i think whats happening is the bearing is off axis loading until friction mostly locks the beam
else i think it maybe has to keep accelerating, but i dunno maybe the inward acceleration is enough force
it stabilizes at around 75%, i have plot pwm and angle vs time of behavior when it spools up without intervention, and when its just kind of shuffling back and forth
shrug, doesnt need it
and i wouldnt do anything crazy
some sort of pwm integral to angle trim to keep it at a consistent position
what would be really neat is using the gyro on the discovery board to smooth speed
because geartrain is all work out
Posted by renesis at 05:42 | permalink | 0 comments
jero32: neat
did not fail to fail
thats because you didnt make sexytime curved traces
jero32: did you watch spinny deathbot video?
GO WATCH
i did
15:21:18 <@renesis> http://www.dailymotion.com/video/x6bsqzr
the video or the bot?
and yes
Posted by renesis at 05:37 | permalink | 0 comments
like, you need to use diodes with lower Vf or doesnt matter
if its same or close, prob lot of energy goes through internal diodes anyway
right
pretty sure, ive gona over rails on adc pins and shit still works
*gone
i checked all this stuff before, its in the manual
Posted by renesis at 05:27 | permalink | 0 comments
for the last stage?
and right diodes to ground and rail are pretty typical
http://electro-music.com/forum/phpbb-files/thumbs/t_that_xlr1203_1206_132.png
thats a bit overkill
but the 100R input followed by hanf nano caps to each other, then caps to ground
i dont do it that way, usually i have caps on each path to ground and then caps tied
but same general ide
you want to mix so some common mode rejection, and try and dump the energy
you can do same type of thin, diodes to rails
stm32 already has them, pretty sure
nope
if theyre schottky in the stm32 you maybe kinds gotta live with it or buy expensive schottky
but i think theyre normal
right
Posted by renesis at 05:22 | permalink | 0 comments
from esd?
big voltage divider will usually do it
resistors were usually the most effective, compared to ferrites and caps
you need an amp to offset anyway, no?
so you want to protect those inputs, they should protect the ADC from anything external
Posted by renesis at 05:16 | permalink | 0 comments
http://www.dailymotion.com/video/x6bsqzr
bad robot^
now with nunchuck action
Posted by renesis at 05:09 | permalink | 0 comments
Top | Add to Technorati Favorites
© 2007 lordpil. XHTML 1.0! CSS! Site design by GNAA Blog Engine by pbx | MULTI2 | ian hanschen | lolwat