Fortunately with
pentodes, it is easy to drive the output to zero (negative modulation).
The real problem occurs when the suppressor is driven very far into the
positive voltage region and it starts drawing current. Typically the
suppressor grid can only dissipate a few watts. Since current is drawn
during the positive peaks, an audio driver must present a low impedance.

73, Barry WA4VZQ

Ah, yes, "(negative modulation)"!

You seem to be "hep" on old stuff, do you remember "negative peak
clipping"?
I thought I was in Heaven in about 1958 with a single 2E26 final on 15
Meters.
In class C, with a pair of 6L6 modulators, push-pull, class AB1, and used a
VR
tube across the modulation transformer secondary clipping the negative
peaks, while
allowing the positive peaks to go "over" 100%.

Technical gurus of the day poo-poo'd the scheme. It looked a little
rough on the
scope, but unsolicited signal reports said it "packed a lot of modulation".
I am thinking it must have been a little like more modern amateur
"speech
processing"

I also seem to remember those modified 1646's getting almost 20%
suppressor
grid modulation after tinkering around with power supply voltages.

Old Chief Lynn, W7LTQ