Freq to voltage using only caps, resistors, diodes?

Thread Starter

SoCo

Joined Mar 13, 2009
3
I need to build a circuit that will take a square wave signal (between 1KHz and 10 KHz) and output a voltage that is linearly related (or as much so as possible) to the input frequency. I am only allowed to use capacitors, resistors, and diodes. Can anyone give me any tips on this circuit? The circuit also should be able to handle any type of wave and any frequency between 100Hz and 250KHz without, you know, blowing up or whatever. Thanks to anyone who can help!
 

thatoneguy

Joined Feb 19, 2009
6,359
Hint: Capacitive Reactance decreases with frequency

short RC Time Constant for charge/drain cycles, average over a short period with another RCR filter.

Center input waveform on ground. The voltage will continue to climb when frequency is constant if negative peaks aren't seen.

I hope this gives you a push in the right direction. It may be entirely useless if I skimmed your post too fast. The square Wave part may add a glitch...

Others may have input as well. There are usually many ways to skin a cat, err, measure a frequency.
 
Last edited:

Thread Starter

SoCo

Joined Mar 13, 2009
3
Thatoneguy- your suggestion sounds a bit like the idea we have been playing with, but ill look more into what your saying. thanks for the tips
 

Thread Starter

SoCo

Joined Mar 13, 2009
3
The idea that we have so far is by using a rc circuit with a discharge time constant larger than the period of the lowest freq the capacitor cannot fully discharge. Also, the negative peaks of the wave are blocked using a diode. The time alloted for the discharging changes based on the frequency (the cap is discharging during the blocked negative peak but halted and recharged when the wave becomes positive) and thus the average voltage changes due to the freq. Is this on the right track?
I feel like this will never create a linear correlation between the freq and voltage because of the nature of the voltage drop across the cap over its discharge time (V=V0*e^-t/RC) but my project partner thinks that if the discharging time contstant is long enoguh the portion allowed will aproach rough linearity. Which one of us is right?
 

thatoneguy

Joined Feb 19, 2009
6,359
Since you are only allowed the use of caps, resistors, and diodes, there are only a few configurations, all having the drawback of of limited accuracy. The most linear response would occur with a period below 2/3 of the Time Constant.
 
Top