[Amps] designing and testing RF Bipolar Power bias circuit

[Dan Sawyer]

All, I am converting a TH430 based Class C amplifier to Class AB. The 
basic circuit is a traditional single stage layout with input and output 
binocular transformers. Things are proceeding, however the bias seems 
unduly noisy. I measure over two volts of peak to peak ripple at the 
bias source.

The bias is a current source, from a Tek 577, to the common side of the 
input transformer. When power is applied the bias sags. This may not be 
the 'best' circuit, however it is convenient because it is possible to 
easily make measurable adjustments.

1. any thoughts on the source of the noise?
2. would a standard voltage divider and diode circuit be better?
3. would it be less noisy.
4. would a capacitor across bias input help?

Sorry for the low level questions, but your analysis would be helpful. 
Thanks Dan