AB2 Anyone else done similiar

[richwalters]

Anti grid blocker ?
Anybody tried this little diode trick on a typ MI 25W AB1 amp output stage ? (rounded in red). In fact any spare tube section botched as a diode can be used to the same effect.
The 350V B+ on this amp is low for the output power so forcing the power tubes into harder conduction (subtle class B2) up’s the output power and then sounds like a 50W amp.
There is an equivalent silicon diode variant around (which sounds harsher) but since I’m using ECL82 pentodes as triode drivers
( o/p Z = 2K); the triode sections were redundant so I config’em as soft diodes and strapped them across the o/p stage G1 inp drive signal and neg bias line. Anyone else done similiar ?
Some care is required. The power supply has to cope with the fierce anode current hike on high signal which can be easily doubled when compared to standard AB stage. A cathode res value
of 47K is used to limit to 150mA per tube on high signal levels. Lower resistor values increase current and power dissipation.
The results are quite revealing. The square wave is remarkably good to 10Khz.
Aikenamps covers an article of grid blocking in detail. I fell into the trap of checking the amp at 1Khz at overload which is too low a frequency to reveal the effects of grid blocking. This is far more revealing when overloading around the 4-8Khz range. The effect of grid blocking is the influence of the anode o/p tube voltage and effects of o/p transformer leakage inductance partically rectifing the g1 drive signals below cutoff that forces o/p tubes into crossover cut off with instant loss of output power. So with circuit time constants the amp “burps” on overload.
Anyone else with another perhaps better description is welcome to add it.
Alot of MI amps get around this by using lower coupling caps but I find this not always reliable and often makes MI amps sound even tighter. The 6550 tubes prefer a very low grid bias resistor value so I used 39K + odds in the trimmer.
As this amp stands, for ordinary music listening it behaves as class A.
Not mentioned here and can be dealt later, is by venturing into class AB2, subtle changes of the 1st stage phase compensation components is required (not shown) . The 20dB global nfb remains.

reading; Radioton h’d bk. 4th ed p.590 “grid driving conditions”.

richj

[gingertube]

Rich,
I guess the answer is NO.
I certainly haven't but thanks for posting. New info is always of interest.
As you say variations have been done with SS diodes/zeners, most recently there was an Audio Express article which was titled something outrageous like "Push Pull Amp Design Breakthrough". The "design breakthrough" was g1 clamping diodes (SS).
Cheers,
Ian

[Allen Wright]

Ian,
Agreed the article's title was a bit over the top, but it actually works, completely stopping the crossover distortion in a test amp, with no negative side effects I could see or hear.

Regards, Allen