Posted by Bill Brown (A) on April 24, 2015 at 12:20:30

I am thinking of building a PP amp without global NFB. I have long wondered about the LL1620/CFB transformer and set out to see how it would potentially work.

If I assume a fixed bias output triode with a gain of 6 (just looking at one-half of the pair), needing 35VRMS to be driven to Vg=0, then:

-applying +1V to the grid causes the anode to decrease by 6 volts,
-The CFB winding is not connected to B+ so does not see its DC voltage
-the 6 volts at the anode is coupled via the transformer, 25% to the CFB winding, so +1.5V appears at the positive phase of the CFB winding and is applied to the cathode, decreasing Va-k, such that-
-1V to the grid produces 4.5V of change, so the gain with CFB is 4.5. So..

NFB amount = 20logA/A' = 20log6/4.5 = 2.5db of NFB(?), and
drive voltage required to reach clipping is now 35Vrmsx1.33= 46.7Vrms(?)

This amount of NFB seems modest relative to what I have seen described with CFB and makes me question my thinking, but I have pondered it, and can't see my mistake if there is one.

Thank you!

Bill Brown