diyAudio Chief Moderator
Join Date: Oct 2002
Originally Posted by nnx
Thank you for your advice!
I have made the modification to the Riaa generator as you described, and it'a quite surprising to see the frequency response in the HF area. The green curve shows +20db around 1 Mhz and that is quite a lot. I wonder if adding a zobel network as suggested will make any difference (blue curve)?
I also made a noise simulation with a different input capacitor, C1=500nF. It shows less noise, but only at the very low frequencies around 20Hz, and has almost no effect above 50Hz. The Riaa curve doesn't change much, only minor adjustments may be needed.
The strange noise described above seems to be a transistor related phenomenon called "popcorn noise", and the cure is finding and replacing the transistor.
Changing the first transistor may be an option, I have 2N5209, 2SC945 and BC550 in my collection. Reducing R6 to 4k7 will double the gain and reduce the noise a little. It will also alter the riaa equalization, so some adjustment will be needed there too.
On the other hand, if we reduce R4~R6 and R24 with a factor 100, and increase C2 and C2 100 times, then the noise will be reduced around 4 times. Maybe that is the way to go?
Anyway, the preamp is now inside the main amp case, so any modifications will be made next time it is opened for service and modifications. In the meantime, maybe it would be fun to start on another Riaa preamp project.....?
You don't need the E2 dependent source either in this simulation. Relaxing the supersonic response should make a difference indeed. As for achieving less noise, since it integrates from the circuit's full bandwidth to an RMS value its good to control its quantity even in places below or above the audio band. Strange noises can also stem from some instability in a not yet finalized circuit.