sculptor: it tries to like backspin stab peoples
is not a very social robot
Posted by renesis at 20:24 | permalink | 0 comments
lordpil's blog
also with base off, pendulum acts like giant speed control lever
Posted by renesis at 19:26 | permalink | 0 comments
zigthats so many moneys
er
also zigler nichols methods seems to work
Posted by renesis at 19:11 | permalink | 0 comments
can run pid thing at 1024 hz instead of 256 hz
gearbox is pretty fucked
Posted by renesis at 18:38 | permalink | 0 comments
guys i have reduced log packet from 25 to 4 bytes \o/
Posted by renesis at 12:41 | permalink | 0 comments
k so my pwm freq and sampling rate seems to have converged
and im like, oh i can put the pid control into the pwm timer isr
but then im going to end up waiting like almost entire period before pwm update
so its like, shit got twice as fast but somehow not at all
but maybe i can chain the current timer off the pwm timer
and then like, sync the pwm update
Posted by renesis at 11:42 | permalink | 0 comments
the volume and power keys were stuck onto non standard gpio or some shit
so you couldnt boot into recovery
someone blogged about it, the normal button positions are unpopulated and theres no buttons in the case there
anyway i gave it back
but that shit runs windows
tho 4gb ram 64gb random flash, almost useless
Posted by renesis at 11:09 | permalink | 0 comments
k i try and cut packet in half then i can double sample rate without upping uart anymore
that looks kind of cool
i almost bought one from a engineer guy who goes to china a lot
Posted by renesis at 11:04 | permalink | 0 comments
that emitter cap seems assymetric as fuck
on base upswing, that cap gonna cause signal to pull harder on the base
increasing base current
but on down swing it maybe reverse biases the Vbe drop
which would kill base current completely
Posted by renesis at 10:57 | permalink | 0 comments
it doesnt work without dc offset
no for ac signals
*not for
Posted by renesis at 10:13 | permalink | 0 comments
jero32: depends on circuit config
jero32: what like class-a?
shrug
eh
feedback comes from the ratio of resistors
its a negative feedback loop
i dont think you understand
which is negative feedback
sure but mathematically its negative feedback
and you can get pretty accurate gain with it
its negative feedback and the explanation is being snarky
Posted by renesis at 10:06 | permalink | 0 comments
i wonder if i can put the uart much faster
upping sample rate def seemed to smooth shit out
or i have to cut packets out of the logging data
right now i just send all vars same as status request, but for plots i only need a couple
Posted by renesis at 07:31 | permalink | 0 comments
says the guy who cant get pours to work in eagle
prof says i can reset integral at TDC
eggsalad: boop.
k
you can change that
pretty sure its in DRC
Posted by renesis at 06:36 | permalink | 0 comments
i think if i reset Ki at top dead center itll actually be usable
should ask prof if thats cool
and i think the only way to keep it from rotating would be to do some sort of integrator that trims the target angle
Posted by renesis at 05:28 | permalink | 0 comments
rab: weird i bet that changes dominant freq w/ duty cycle
Posted by renesis at 04:29 | permalink | 0 comments
rgb scope lamp sounds neat
i need lamp
so my project is mostly done but my motor gearbox is also
Posted by renesis at 04:13 | permalink | 0 comments
w00t, spinny death now serves gnuplot pngs along w/ text data after logging session
Posted by renesis at 03:43 | 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