Zobel network for OEP mic transformer

Today I was playing with the OEP A3E [1+1:6 : 6.45+6.45] and tried to optimize a zobel network for best square waveform.
I came up with the next values: 21Kohm in series with 800p [330p ll 470P]
Source impedance 150ohm, test signal 1Khz square wave.
Can anyone confirm?

chrissugar

See;
http://www.groupdiy.com/index.php?topic=2050&highlight=ringing+transformer

http://www.groupdiy.com/index.php?topic=2049&highlight=ringing+transformer

I had a nice note from Jensen and it was posted here
but I cannot seem to find it. I will look.

Adrian, I obtained the results by doing the procedure from the Jensen document.
The document is in this thread: http://www.groupdiy.com/index.php?topic=1221
I needed a confirmation because I supose many people experiment with the OEP transformers.

chrissugar

Chris;
Yes that is a good doc but the one I had emailed to me is better.
It tells how you approach the zobel network.

I have never seen that anywhere.
I have it at home.
I will get it.

It is a PDF document? I downloaded it from a site some month ago.

chrissugar

Please, post a link for this, if you ever find it again.

Interested on this doc also! :thumb:

Rafa and Fabio, look at what Mark Burnley posted here:
http://www.groupdiy.com/index.php?topic=1221

chrissugar

Hey Guys,

JLM posted this in answer to another question...

The A262A3E is the only one that is really usable. The A262A2E rings badly and has to be zobeled to death to get it sound ok. But the A262A3E can just be loaded by a 27k when the secondaries are used in series or 6k when used in parallel. It has a dark tone to it which I really like and would call it neve-ish sounding.

Click to expand...

Justin.

>> I was playing with the OEP A3E... and tried to optimize a zobel network... I came up with... 21Kohm in series with 800p (330p||470p) Source impedance 150ohm, test signal 1Khz square wave.

> the A262A3E can just be loaded by a 27k

I'd say that is confirmation. 27K, 21K, not a big difference. Your 800pFd equals the 21K at 9.5KHz, so below 5KHz your version is unloaded and a part-dB less loss than Joe's version. Above 20KHz your version damps a little more than Joe's. But really the differences are small enough that either will work, perhaps "the same". (Joe's version does save a dime capacitor.) If not quite the same, it is down to your ears to decide which suits your situation better. Scope is good for getting a starting point, but the eye can't judge what you will hear.

If you really want to play: wire some small (this needs to be tight and compact) rotary or DIP switches with 15K, 20K, 25K, 30K, 50K, infinite and 2,000p, 1,000p, 500p, 220p, and do many sessions over many days and use your ears.

Thank you very much PRR for the confirmation.
That is what I was thinking, to conect a 50k pot in series with a rotary switch with capacitors and make some listening tests.
I will do it the next few days and report back what I discovered.

chrissugar

Hi
I will use OEP in my API. I saw a couple different information about zobel network.
First was 27K only, than 27K with 220p, and 21K in series with 800p :?: :?:
Also I saw Jensen's explanations about load resistor across the secondary of the transformer.
They said:
This value is typically 1500 Ohms x the turns ratio squared.
For OEP it is 1500 x 6,45 x 6,45= about 62K (61,9K and 68K standard value)
Any comments and answers?
Thanks
Duka

Any comments and answers?

Click to expand...

chrissugar was talking about the zobel network, not the load resistor. I'm not sure whether he uses a load resistor, but I would. The load resistor is in parallel with the zobel network. Just look at some Jensen datasheets.

Not sure either about the exact config he uses.

The "Jensen equation" sets the input impedance at medium freqs (say around 1 kHz) to 1.5k, whereas the zobel network loads at higher frequencies to reduce peaking in the frequency response.

It is likely that the different zobel networks result from not using the same transformer. OEPs are probably not as well matched as Jensens. And even with Jensens, there are differences in the zobel network from unit to unit. To be sure, you would need to check every single transformer...

Samuel

[quote author="Samuel Groner"]

Any comments and answers?

Click to expand...

chrissugar was talking about the zobel network, not the load resistor. I'm not sure whether he uses a load resistor, but I would. The load resistor is in parallel with the zobel network. Just look at some Jensen datasheets.

Not sure either about the exact config he uses.

The "Jensen equation" sets the input impedance at medium freqs (say around 1 kHz) to 1.5k, whereas the zobel network loads at higher frequencies to reduce peaking in the frequency response.

It is likely that the different zobel networks result from not using the same transformer. OEPs are probably not as well matched as Jensens. And even with Jensens, there are differences in the zobel network from unit to unit. To be sure, you would need to check every single transformer...

Samuel[/quote]
Thanks Samuel :thumb:
I was talking about both zobel network and load resistor.
Also, I need informations about it because I will use OEP in my API.
Duka

BTW, how do people usually connect the windings? In series or parallel? I guess for a reasonable max. input level we need them in series?

Samuel

[quote author="Samuel Groner"]BTW, how do people usually connect the windings? In series or parallel? I guess for a reasonable max. input level we need them in series?

Samuel[/quote]
1+1:6,45+6,45 is both windings in series

Jakob use both windings in series for the G9. I don´t think the inductance would be high enought for good low frequency response with the windings in paralell...

I´ve used a 47k load resistor parallel to the secondary winding (gives 47k/(6.45*6.45) = 1.1 k input impedance). The zobel network which is parallel to the load resistor consist of a 5k6 resistor in series with a 220pF capacitor. This gives a good frequency response with a 1 dB peak at 15 kHz. This network also tames the behaviour when nothing is connected to the input, without the load and zobel network my "discrete american" inspired preamp oscillated at high gains.

Klas

[quote author="klahal"]I´ve used a 47k load resistor parallel to the secondary winding (gives 47k/(6.45*6.45) = 1.1 k input impedance). The zobel network which is parallel to the load resistor consist of a 5k6 resistor in series with a 220pF capacitor. This gives a good frequency response with a 1 dB peak at 15 kHz. This network also tames the behaviour when nothing is connected to the input, without the load and zobel network my "discrete american" inspired preamp oscillated at high gains.

Klas[/quote]
And now, I am confused :sad:
Duka

[quote author="dukasound"][quote author="klahal"]I´ve used a 47k load resistor parallel to the secondary winding (gives 47k/(6.45*6.45) = 1.1 k input impedance). The zobel network which is parallel to the load resistor consist of a 5k6 resistor in series with a 220pF capacitor. This gives a good frequency response with a 1 dB peak at 15 kHz. This network also tames the behaviour when nothing is connected to the input, without the load and zobel network my "discrete american" inspired preamp oscillated at high gains.

Klas[/quote]
And now, I am confused :sad:
Duka[/quote]
We stopped talking about very important thing :roll: