I'm trying to pick some better capacitor values for the part of a circuit shown below (dont ask me what it's for, you'll just laugh). Before this final amplifying stage there's a bunch of transistors summing signals from some DACs and a low pass filter (22nF to ground). The signals are 0-5v with 2.5v offset.
When I calculate the cutoff frequencies ot the bias filter and the output AC coupling (assuming I'll connect the the outputs to the line level input of a Hi-Fi amplifier) I get: Output = ~0.03Hz, Bias = ~7Hz.
I've read somewhere that the bias cutoff should be 1/10 of the amp cutoff to minimize the noise from the bias circuit. Is that right?
The design on the picture is doing the opposite, and I'd like to correct that if it's possible (or even necessary?). I'd like to use a pair of 10uF MKP4s for the output AC coupling. 100uF seems a bit overkill. But can I use a smaller value? How "high" can you make the cutoff frequency before it starts getting audible?
If I go with 10uF caps on the output, I'll have to use 22000uF caps on the bias instead of the 100uF to get 1/10 of the output cutoff (unless I'm doing all of these calculations wrong... there's a good chance of that). I guess I can do that by paralleling 10x 2200uF. But will it have some bad side effects (22000uF sounds huge to me)? Is it even worth the trouble?
When I calculate the cutoff frequencies ot the bias filter and the output AC coupling (assuming I'll connect the the outputs to the line level input of a Hi-Fi amplifier) I get: Output = ~0.03Hz, Bias = ~7Hz.
I've read somewhere that the bias cutoff should be 1/10 of the amp cutoff to minimize the noise from the bias circuit. Is that right?
The design on the picture is doing the opposite, and I'd like to correct that if it's possible (or even necessary?). I'd like to use a pair of 10uF MKP4s for the output AC coupling. 100uF seems a bit overkill. But can I use a smaller value? How "high" can you make the cutoff frequency before it starts getting audible?
If I go with 10uF caps on the output, I'll have to use 22000uF caps on the bias instead of the 100uF to get 1/10 of the output cutoff (unless I'm doing all of these calculations wrong... there's a good chance of that). I guess I can do that by paralleling 10x 2200uF. But will it have some bad side effects (22000uF sounds huge to me)? Is it even worth the trouble?
Attachments
It's not my design, but thanks anyway
I've been told that they need to be as small as possible (within reason) to keep the feedback resistors out of the equation. And the whole circuit only has a single 5v rail to work with, so the current draw is quite low.
That would be the optimal solution. But at the moment the I'm limited by the 5v supply.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.