inorite
Posted by renesis at 15:26 | permalink | 0 comments
lordpil's blog
yeah i gotta go back to school =\
this is my third now
its likw i have a career or some shit
Posted by renesis at 15:11 | permalink | 0 comments
so you can visualize ringing in a cable as a turbine with a chamber between it and a door (switch)
turbine is spooled up, then door closes, the pressure (voltage) builds up in the chamber, causing it to expand
eventually the inductor slows and the chamber pressure (capacitive charge) is released back through the turbine, spinning it the opposite wave, creating reverse pressure and flow
if the impedance is high on the other side of the turbine, it will pressurize and eventually push back and flip the turbine around again
friction in the flexible chamber walls and in the turbine bearings are your resistive damping component that eventually stop the ringing
yeah but its the same
likely the math is exactly the same
same shit
so youre saying i should learn that shit?
Posted by renesis at 15:06 | permalink | 0 comments
yeah well figure that out and you got some of the most complex electronic problems understood
haha, its funny when shit like fluid flow comes up, i will explain shit at people will look at me and ill be like I KNOW ALL ABOUT ELECTRONICS, TRUST ME
because resistance is resistance, math is the same
yeah
inductors are turbines
transformers are turbines linked with chains or sprockets
caps are membranes, or pistons
Posted by renesis at 15:01 | permalink | 0 comments
precisely lame!
yeah no one got that until im like GUYS IT TAKES TIME FOR SIGNAL TO GET THROUGH OPAMPS BECAUSE REACTIVE COMPONENTS AND PARASITICE
SO IF THE TIME IS HALF A WAVELENGTH THEN THE WAVE FLIPS AND THE NEG INPUT BECOMES PSOTITIVE AND YOU GET FUCKED
and everyone is like OOOOOOOOOOH
so you just dont do that
poles and zeros is a lot of math about hey dont be phase shifting your feedback until youre fucked
wtf everything is reactive
early opamps were rolled off way early on purpose, so they basically had no output by the frequencies that would cause oscillation
youre doing poles and zeroes for on opamp class, no?
oh just for controls in general
Posted by renesis at 14:56 | permalink | 0 comments
so you figure out how mayne watts a candle generates
*many
then you find out how much the power dissipates with distance
its a spherical wavefront so its likely for every doubling of distance you get a fourth the power
marcecko: try and rouch math it to see what you need within an order of magnitude
zeeshan: yeah when you learn more of it, you realize a lot of it is just variations of the same equations
everything can be defined by equations
some oscillators are defined by how they break equations
but if you cant model it mathematically its not very useful
neat
lame
Posted by renesis at 14:51 | permalink | 0 comments
dunno i imagine its like acoustic wavefronts, so falls at 6dB when you double distance
so you figure it out from that
maybe
hmm maybe no
because temp is related to frequency?
yeah weird at the amplitude of the specific temperature
mind slightly shattered
so yeah i have literally no reference for this shit
A thermopile detector sensitivity defined as the ratio
of thethermopile detector output signal generated in
response to a unit input radiant power.
Responsivity[Rv, V/W]
so it looks like its voltage output per watt of IR energy recieved
Posted by renesis at 14:46 | permalink | 0 comments
marcecko: i dont think a single turn will be hard to route a cable to
so now its a normal controls project
you can go steppers and do the ram into a mechanical limit for a full rotation at startup to calibrate init position
or use a servo, step or DC motor based
quick way to prototype might be one of those fluke IR probes
i think they might have RS232 ports or something similar
yeah
ha sorry im used to just having that shit around
at work to try shit like this wed prob use a MDF platform glues to a stepper shaft with the fluke IR probe duct taped to the platform
we have lots, speaker lab
Posted by renesis at 14:41 | permalink | 0 comments
1.5 seconds
3 seconds for 360 degree sweep
just have it do a turn, then scan again in the other direction
so you dont need special couplers
hehe, like unwinding speakers on turntables so you dont know them off the platform doing off axis measurements
heh everyones so paranoid about that
im maybe the only one hasnt done it yet, heh
*knock them off
shit i test will have power, ethernet, toslink
=\
and ill have to run at least L, R, and L+R
so yeah you can throw shit up there and wrap the cable three time itll just pull the speaker off, and platform is on a pipe like almost 6ft in the air
er, you cant throw shit up there
i rabbed
tho i guess it makes sense both ways
Posted by renesis at 14:36 | permalink | 0 comments
depends on what youre following
if its something that moves slow, a single IR sensor and a motor to spin it might be enough
its a static object?
get some sort of narrow angle single sensor, and put it on a motor to sweep it back and fort
sample every couple degrees
use a serve motor with a fine encoder, or a stepper
http://www.ge-mcs.com/en/temperature/infrared-ir-sensors/ztp-115.html
that sensor has a time constant spec
20mS, very slow
like 50Hz lowpass
so you want to move to a new position and take a sample, as fast as every 20ms
so lets say 5 degrees ever 50ms, a 180 degree sweep
Posted by renesis at 14:31 | permalink | 0 comments
tpa81 is an 8 pixel linear array
yeah this shit isnt cheap
they prob make 1000 and throw away 300
most of electronic component manufacturing is like this
make good bins, fail bins, and 'it depends on how desperate for sales we are' bins
high precision and sensor shit prob have the lowest yields
everything only works per spec on the client side, on the manufacturing side, lots of shit sucks, but you just throw it away and try for better yields next time
Posted by renesis at 14:26 | permalink | 0 comments
http://imgur.com/f0udC18
heh
marcecko: its more a mechanical/optical engineering project
and signal processing
the electronics part of IR sensors isnt so difficult
Posted by renesis at 14:19 | 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