[Lord Valve]

I really hate to give you poor ******s a clue, but...please tell me where
R60 and R61 are *physically* located in your particular unit.

LV



Patrick Turner wrote:

So I can only guess that in class B during cut off for one 6V6, the circuit
changes to that of an HF oscillator.

The solutions tried were 15r & 0.27uF across Vo, 680pF & 8k2 across anode to
anode, placing small C from Vo to pot wiper for phase correction of NFB.
None of these worked. Then I remembered the right solution, TWO R&C zobels
across each 1/2 primary.

** Take a look at the schems on this page for Deluxe Reverb:

http://www.webphix.com/schematic%20h...om/fender.html

Many of them have 1200pF caps from grids to ground on the 6V6s.

I believe this is for the same problem that happens with some combinations
of OT and speakers.

Using caps across the OT primary is dodgy, high voltage ( ie 2 kV ) ceramics
will survive - 600V film caps will not.

275VAC caps " class X " caps should be OK too.

.... Phil

The Deluxe Reverb is a re-issue. In the amp here there are no 1n2 caps between 6V6 g1 and 0V. None on the schematic I have.

I strongly agree caps across 1/2 primaries are a dodgy thing to do.
***BUT***, I have 4k7 5W resistors in series with 2n2 which ARE ceramic and 2kV rated, and muso can now blast away all day and not have tubes making ultra high F oscillations.

Having 1n2 between 6V6 grids and 0V means that output from 12AT7 anodes is effectively shunted and if AT7 anode Rout = 12k, Ra//RLa//Rg, approx, then expect HF pole at 11kHz. But above this F, there is increasing open loop phase
shift making oscillations still likely with lowish levels of GNFB.
Other methods used include a small value cap used between 6V6 anodes and g1, but this also causes ultimate extra phase shift, and only suits old radios where GNFB is frowned upon with a single 6V6 left to muddle away vainly to make horrid sound once above a whisper level.

Using the pair of Zobel networks with 4k7&2n2 gives a pole at 15kHz and above this F the load seen by 6V6 becomes a pure resistance, 4k7, with less phase shift than otherwise. 6V6 tetrode gain is thus curtailed in region where the speaker offers virtually no loading due to its inductive property at HF. Using a Zobel at OPT sec does not stop the problem because the OPT has high LL and when loaded with secondary R load the 6V6 still decouples from sec because of the effectively in series leakage inductance.

Square wave testing reveals a lot about amplifiers. I like to use a 5kHz sq.wv, and have output set to say 1V, low level at first and where OP tube class A operation means the open loop gain is highest. Once some GNFB is added, the waves at OP tube anodes can have far higher overshoots on square waves than at the secondary output. Using higher levels in class B working may bring bursts of HF oscillations on peaks.

High levels of LF, say 100Hz sine waves can cause the same HF crap,
and if very low F at say 10Hz are tried, and OPT saturates then chances are this will also cause bursts of HF oscillations. If the OP tubes overheat and anodes turn red - not at all an unknown phenomena in a guitar amp - then tubes can break into HF oscillations during their self destruction mode of operation, so ALWAYS there must be some way of stopping very high oscillations. And in all amps, not just guitar amps.
The Watt ratings for R in Zobels don't have to be equal to power produced by tubes because their presence STOPS the HF ever becoming present and there is so little average audio signal power above 5kHz.
In old books about old amps, sometimes adjustable spark gaps were recommended, with two pointed screws facing each other and each taken to each PP anode. Two much Vac = Vac arcing, ( NOT DC arcing ) and thus excessive voltage at anodes is controlled somewhat, but methinks the Zobel is better. Some guitar amps have used say 3 x 1N4007 in series from each anode to 0V. Diode anodes are to tube anodes, and if either anode tries to make a huge negative going voltage higher than the B+ value, then diodes conduct, thus shunting the stored energy in leakage inductance and protecting insulation from Va maybe 3 times B+ if signal is turned up with no loading.
But this "excessive Va" function is distinct from the HF oscillations in class B when there is GNFB and the is no loading above say 20kHz.

Patrick Turner.