feedback resistor values for log pots

Posts about MIDIsense. open source MIDI sensor platform

Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.
Locked
evilxsystems
 
Posts: 406
Joined: Thu Jul 07, 2005 10:23 pm

feedback resistor values for log pots

Post by evilxsystems »

I'm following this here:
http://www.ladyada.net/make/midisense/sensormodel.html

So I have some 50k log pots I'm trying to use with the midisense....I think I understand the principle behind choosing a feedback resistor value, but I can't figure out one that works with a 50k log pot. If I understand what's going on correctly, the voltage can vary from .5 to 5V. If I choose a feedback resistor that's say 10k...that means at the lowest resistance of my pot (which is zero but for the sake of argument let's say) is 500 that would give me a voltage of 10v and at the highest resistance of 50k I get .1V. OK so that means when the pot is at 10k the voltage has already dropped to .5 volts so the top 40% or so of the pot's travel is not registered since the voltage has bottomed out. You can adjust for this by increasing the resistance of the feedback resistor but then you have the exact same problem at the bottom end of the pot.....

I'm really BANNED confused...it seems like a linear pot would make the most sense...but limor clearly said to use a log pot, so somebody please tell me what the hell I'm doing wrong here.

-thanks
jason

evilxsystems
 
Posts: 406
Joined: Thu Jul 07, 2005 10:23 pm

Post by evilxsystems »

aghhhhhhh this sucks! help! did you guys who used pots use linear pots?

adafruit
 
Posts: 12151
Joined: Thu Apr 06, 2006 4:21 pm

Post by adafruit »

http://www.ladyada.net/make/midisense/sensormodel.html

the first case, 'minimum' resistance is 0 so it will pin (any feedback resistor will work)

the second case, 'maximum' is 50K, Vout = .5V * Rf / 50,000 and you want Vout to be as small as possible but not smaller than 0.5, so make Rf = 50K and you're done

short version: feed back resistor == max resistance of pot

dont forget there's two ways to hook up the pot, try both.

that help? :)

evilxsystems
 
Posts: 406
Joined: Thu Jul 07, 2005 10:23 pm

Post by evilxsystems »

yeah that makes sense but the problem is the response of the pot is not linear and doesn't utilize the full range of the pot. Basically the midi output of the pot's motion is only changed when the voltage is between 5 and .5, correct? If I turn the pot up (clockwise) the voltage goes to infinity as R goes to zero, but the midi output maxes out once the voltage goes over 5V which happens when the 50K pot reaches 5K ohms, basically all of the pot's range between 5k and 0k is useless "dead space" on the pot...should I just put a 5k resitor inline with the pot?

I guess I'm confused, it seems like a linear pot works better....I hooked up a linear pot and it gave me a much nicer response...it seems from the math that the best pot would be a 5k->50k ohm pot coupled with a 50k ohm resistor...that would give me a linear response from .5 to 5 volts. Why am I wrong?

also the midisense is constantly sending out signals for midi channel 2 cc 33 which I don't have specified for anything...any ideas what that could be?

Help me Limor Fried, you're my only hope....

-jason

evilxsystems
 
Posts: 406
Joined: Thu Jul 07, 2005 10:23 pm

Post by evilxsystems »

ok I got some thing that works but maybe you can explain why...i put a 10k ohm resistor inline with my 50k ohm log pot. Then I wired the pot backwards, so the calibration slider goes up(right) as I turn the pot down(counterclockwise) this setup gives me a nice clean linear looking response in ableton live....albeit backwards...but that's easy to fix in live

adafruit
 
Posts: 12151
Joined: Thu Apr 06, 2006 4:21 pm

Post by adafruit »

the reason a logarithmic is better is that Vo = Vout = .5V * 50K / R where R is the potentiometer resistance. therefore Vo is inversely proportional to R ( V = 1 / R * const) if you graph 1/R you get a logarithmic curve. But if R is not linear on the scale, but inverse log...you'll cancel out the logs and get a linear.

soundsl ike its working now, though?

adafruit
 
Posts: 12151
Joined: Thu Apr 06, 2006 4:21 pm

Post by adafruit »

evilxsystems wrote:yeah that makes sense but the problem is the response of the pot is not linear and doesn't utilize the full range of the pot. Basically the midi output of the pot's motion is only changed when the voltage is between 5 and .5, correct? If I turn the pot up (clockwise) the voltage goes to infinity as R goes to zero, but the midi output maxes out once the voltage goes over 5V which happens when the 50K pot reaches 5K ohms, basically all of the pot's range between 5k and 0k is useless "dead space" on the pot...should I just put a 5k resitor inline with the pot?

I guess I'm confused, it seems like a linear pot works better....I hooked up a linear pot and it gave me a much nicer response...it seems from the math that the best pot would be a 5k->50k ohm pot coupled with a 50k ohm resistor...that would give me a linear response from .5 to 5 volts. Why am I wrong

-jason
i cant understand your comments because you're trying 10491843 things at once and asking why each one may or may not work.
you can only ask one question at a time :P

Locked
Please be positive and constructive with your questions and comments.

Return to “MIDIsense”