I want this to generate white noise. I understand that I should omit R4 through R9 and C4 through C9, but what about C2? Looking at the schematic, it is apparent that the output is a square wave or train of pulses that should be low-pass filtered or integrated into an analog output waveform. R10 has the value 220 kilohms, and it appears that C2 is necessary to form an RC network with R10. Is that correct? If so, what should be the value of C2?
C2: 100pF (see page 9 of manual).
But the RC time constant of 220 kilohms and 100 pF is 22 microseconds, and the -3 db cutoff frequency is 7958 Hz. I want white noise spanning the full audio spectrum to 20 kHz. I’m going to change the value of R10 to 100k and use 51 pF for C2. That will make the cutoff frequency 31.2 kHz, so there should be virtually no attenuation at 20 kHz. I did see the recommendation on page 9, but it appeared to be a typographical error.
There will be no loading on the RC network, as I replaced T1 with an op-amp voltage follower, to provide enough current to drive a small speaker. I changed the output capacitor C12 to 220 µF, but I plan to eliminate the output capacitor by bridging the speaker from one output of the LM324 op amp to the output of another voltage follower op amp on the IC whose “+” input is at the same DC potential but signal bypassed to ground.