Why not a line output? And could the squeaking noise be feedback?
Microphone "emphasis circuit"
- Thread starter pasarski
- Start date
Help Support GroupDIY Audio Forum:
Why not a line output? And could the squeaking noise be feedback?
pasarski
Well-known member
- Joined
- Jan 29, 2010
- Messages
- 482
Scratched my head all day trying to measure the frequency response of my build. I couldn't see much of anything change on the screen when connecting and disconnecting the emphasis circuit. After hours of sweat and tears something about calibrating U87 creeped out of my subconsciousness and I realised it doesn't seem to do anything, because I'm injecting the test signal straight to the gate, inside the feedback loop 
For measuring the frequency response it seems I have to use similar setup than the U87 calibration input. What a delightful learning experience!
For measuring the frequency response it seems I have to use similar setup than the U87 calibration input. What a delightful learning experience!
Last edited:
pasarski
Well-known member
- Joined
- Jan 29, 2010
- Messages
- 482
I gave up measuring the circuit. Tried to inject the signal across a ~500 ohm resistor à la Neumann messeingan, but it didn't work at all. Very weak high passed output. Maybe it's some parasitic capacitance thing or something, I really don't know. In simulation it works ok.
Also, I don't know if the whole thing makes much sense tbh. Literally every component in the circuit affects the frequency response, including the bias resistor, which makes it really cumbersome to reproduce the same curve with different fets while maintaining a perfect bias. Also the preamps input impedance affects a whole lot. Not a circuit with someone with OCD for sure
Also, I don't know if the whole thing makes much sense tbh. Literally every component in the circuit affects the frequency response, including the bias resistor, which makes it really cumbersome to reproduce the same curve with different fets while maintaining a perfect bias. Also the preamps input impedance affects a whole lot. Not a circuit with someone with OCD for sure
pasarski
Well-known member
- Joined
- Jan 29, 2010
- Messages
- 482
Okay, I din't give up after all. Turned out that I've just fried the fet. So here's the circuit as it is and the simulated response:
and here's the measurement (not calibrated):
The very bottom end is not totally accurate because my dummy cap is ~110pF instead of 75pF (I used a dummy cap because I'm not measuring in a quiet place, don't know if that makes sense or should I measure with the capsule connected regardless the noise?). The difference above 8Khz is probably caused by some parasitic capacitance? (Notice how I want to attribute at least something to the parasitic capacitance).
It's not a perfect curve for what I was aiming for. The rise should start later, some where around 8k me thinks. But it sounds good to me. And again, I don't think this capsule (wonderful vintage accurate Arienne Audio's cardioid K47) needs this. I was just trying to match it with the other mic (which is wonderful Martin Kantola's NU-47V mk1 modified at some point with a Thiersch blue line M7).
No I just have to bias the new fet and this goes to a friend for testing
and here's the measurement (not calibrated):
The very bottom end is not totally accurate because my dummy cap is ~110pF instead of 75pF (I used a dummy cap because I'm not measuring in a quiet place, don't know if that makes sense or should I measure with the capsule connected regardless the noise?). The difference above 8Khz is probably caused by some parasitic capacitance? (Notice how I want to attribute at least something to the parasitic capacitance).
It's not a perfect curve for what I was aiming for. The rise should start later, some where around 8k me thinks. But it sounds good to me. And again, I don't think this capsule (wonderful vintage accurate Arienne Audio's cardioid K47) needs this. I was just trying to match it with the other mic (which is wonderful Martin Kantola's NU-47V mk1 modified at some point with a Thiersch blue line M7).
No I just have to bias the new fet and this goes to a friend for testing
