chrissugar
Well-known member
I made some measurements and listening tests and will report this night my results. Unfortunately now I have to go.
chrissugar
chrissugar
Zobel network for OEP mic transformer
- Thread starter chrissugar
- Start date
Help Support GroupDIY Audio Forum:
chrissugar
Well-known member
Hi all
OK, so here are my not yet final results after some days of measurement and listening tests.
I made a setup with a variable load from zero to 100Kohm and a Zobel network with variable R and variable C.
I built the test circuit in a way that permited me to listen/measure by switching to the next four configurations: no load no Zobel, only load, only Zobel, and zobel plus load.
Everything was done with switches and pots.
Also I had three diferent setups: for dynamic microphone, for condenser microphone, and for music and sounds played from harddisk and attenuated to microphone level on a 150ohm resistor that simulated the impedance of a mic.
First I tried to see how the transformer behave without load and Zobel. On the scope there was a large overshoot.
Next I tried what Joe Malone sugested, to use a 27K resistor on the secondary. Lowered gradualy the value from 100k to 10k. The waveform started to look much better but still a little overshoot and the signal was a bit attenuated. Joe was right,at aproximately 27k the compromise between waveform and loss was good.
The third test was to find an optimal Zobel network that produced a good looking square wave. I tried to find a small value capacitor that will tame the transformer but couldn't find anything less then 800p that works. The result is the same as before 800p and 21.7k resistor.
Next step was to ad to the Zobel the load. The waveform already looked good and the load resistor helped to smooth a bit more the square wave. It seemed to me that to go lower than 50K didn't produce significant improvement and the level droped too much.
Listening test.
The most interesting thing is that all three setups produced very close results in the listening tests.
With no load and no Zobel the transformer seemed to sound more open than compared with the loaded situation, but after a short listening it was clear that it was something not natural with the sound of the cymbals and the air in the voice. It sounded ringy and harsh.
The next test was the variable loading and no Zobel. I started with the 100k pot at maximum value and lowered it to 10k continously many times with closed eyes concentrating on the sound of the test material that was so familiar to me. At the value indicated by Joe Malone (27k) the sound was a bit dark, in the Neve direction, the harmonic structure of the sound was well preserved, but for my taste was a bit too dark and had the feeling that somewhere around 50K it was a good balance between darkness and corectness of the harmonic structure.
Next test was with only the Zobel. Listened to the material with closed eyes trying many times to rotate the pot to the point when the sound seemed ok. The same thing, at one extreme it was ringy, at the other dark. Before doing the audio test I made a mark on the pot with a marker to have a reference where is the position where the square was the best in the objectiv test (square on scope). Interesting thing is that with closed eyes I managed to repeat many time this test and after finding the point where the sound seemed to me ok I verified the relation between what I measured and what I listened. At every test the mark made with the marker pen was very close to the position I had in the listening test.
The last test was to add to the Zobel the variable load. After many tests I found that the load should be above 50k or the sound will become too dark.
At this point I have the Zobel formed by C=800p and R=22k plus for loading a 100k pot. This is not a final result but i think I'm very close to the right values. Also it looks like there is a corelation between the measurements and the listening tests.
chrissugar
OK, so here are my not yet final results after some days of measurement and listening tests.
I made a setup with a variable load from zero to 100Kohm and a Zobel network with variable R and variable C.
I built the test circuit in a way that permited me to listen/measure by switching to the next four configurations: no load no Zobel, only load, only Zobel, and zobel plus load.
Everything was done with switches and pots.
Also I had three diferent setups: for dynamic microphone, for condenser microphone, and for music and sounds played from harddisk and attenuated to microphone level on a 150ohm resistor that simulated the impedance of a mic.
First I tried to see how the transformer behave without load and Zobel. On the scope there was a large overshoot.
Next I tried what Joe Malone sugested, to use a 27K resistor on the secondary. Lowered gradualy the value from 100k to 10k. The waveform started to look much better but still a little overshoot and the signal was a bit attenuated. Joe was right,at aproximately 27k the compromise between waveform and loss was good.
The third test was to find an optimal Zobel network that produced a good looking square wave. I tried to find a small value capacitor that will tame the transformer but couldn't find anything less then 800p that works. The result is the same as before 800p and 21.7k resistor.
Next step was to ad to the Zobel the load. The waveform already looked good and the load resistor helped to smooth a bit more the square wave. It seemed to me that to go lower than 50K didn't produce significant improvement and the level droped too much.
Listening test.
The most interesting thing is that all three setups produced very close results in the listening tests.
With no load and no Zobel the transformer seemed to sound more open than compared with the loaded situation, but after a short listening it was clear that it was something not natural with the sound of the cymbals and the air in the voice. It sounded ringy and harsh.
The next test was the variable loading and no Zobel. I started with the 100k pot at maximum value and lowered it to 10k continously many times with closed eyes concentrating on the sound of the test material that was so familiar to me. At the value indicated by Joe Malone (27k) the sound was a bit dark, in the Neve direction, the harmonic structure of the sound was well preserved, but for my taste was a bit too dark and had the feeling that somewhere around 50K it was a good balance between darkness and corectness of the harmonic structure.
Next test was with only the Zobel. Listened to the material with closed eyes trying many times to rotate the pot to the point when the sound seemed ok. The same thing, at one extreme it was ringy, at the other dark. Before doing the audio test I made a mark on the pot with a marker to have a reference where is the position where the square was the best in the objectiv test (square on scope). Interesting thing is that with closed eyes I managed to repeat many time this test and after finding the point where the sound seemed to me ok I verified the relation between what I measured and what I listened. At every test the mark made with the marker pen was very close to the position I had in the listening test.
The last test was to add to the Zobel the variable load. After many tests I found that the load should be above 50k or the sound will become too dark.
At this point I have the Zobel formed by C=800p and R=22k plus for loading a 100k pot. This is not a final result but i think I'm very close to the right values. Also it looks like there is a corelation between the measurements and the listening tests.
chrissugar
dukasound
Well-known member
Hi chriss
Thanks for tests.
If I understand zobel can be 22K and 800p and load can be 68K.
Best
Duka
Thanks for tests.
If I understand zobel can be 22K and 800p and load can be 68K.
Best
Duka
chrissugar
Well-known member
Yes Duka, so far this values for zobel (800p + 22k) gave me good results at measurements and audition. I'm still not decided about the load. I think a value between 50k and 100k is the right one. I will play for some more days and will decide what to do.
Will report back about my decission.
chrissugar
Will report back about my decission.
chrissugar
dukasound
Well-known member
Thank you Chris :thumb:
rafafredd
Well-known member
That´s very nice info! Thanks for this!!!
