Zobel networks

[Rob Flinn]

At the other place there was a thread which explained how to select the components for a Zobel network. Does anyone have this info on their machine? Rafa you always seem to have copied everything like the rest of us should have !

 

[rafafredd]

http://www.groupdiy.com/index.php?topic=1221&start=0&postdays=0&postorder=asc&highlight=

 

[rafafredd]

http://www.siteswithstyle.com/VoltSecond/Damping_ringing_XFMRS/Damping_ringing_in_xfmrs.html

 

[rafafredd]

double

 

[rafafredd]

some info from PRR, on the old site:

Test-bench the transformer with the proposed source and load impedances. (For hi-Z windings like mike to grid, you really must build the amp; there is no way to estimate parasitic capacitances accurately enough.) Sweep a sine up to 100KC or 200KC. FOr a few 600:600 transformers, response will be flat far beyond the audio band, and your job is done.

More likely the response will droop or peak. If it droops, try a higher load impedance (if possible). Peaking is tough and sometimes can NOT be cured. If it peaks inside the audio band, it is never going to be "uncolored". If the peak frequency is above 20KC you can probably make audio response pretty good. Try a lower load impedance. This will generally flatten the peak, but in some cases by the time you have the peak tamed the midrange loss is too high or the input impedance is too low. Pick a trial resistor that seems to tame the peak, then compute a series capacitor with reactance equal to the resistor at some frequency between the top of the audio band and the peak frequency. Run the response again, checking midband loss and impedance too. Try different values of R and C until it looks best.

> with solid state input stages, I would have to add a capacitor paralel with the resistor on the sondary of the mic input transformer.

Almost never. A cap on a transformer usually makes things worse.

However the transformer winding already has significant capacitance, and will ring. In some cases you can do "better" with a series combination of R and C.

> Any tips to find the apropriate pF value for this?

Equal to the transformer nominal impedance at 20KHz. That is almost never the right answer, but puts you near the right drawer of your cap assortment box.

 

[Rob Flinn]

Thanks very much Rafa

I think you should be the official librarian for the Lab !!

 

[zmix]

Here is an online Zobel calculator...

http://www.selectproducts.com/calculator_zobel.html

 

[rafafredd]

Interesting calculator. Very useful, if it actually works. It seems that the resistor value is always Z + Z/4...

I don´t think it would work, as diferent transformers will have diferent peak points.