[HBR] Taming 6EH7 IF strip?

[Ian Wilson]

I recently made a start on a 455kHz IF strip / product detector / BFO. The
lineup is:
  - mechanical filter on the input, matched to 50 ohm output of the RF
section
  - 2x 6EH7
  - 2x Miller IFTs (the 'interstage' type)

Since I haven't added the product detector or BFO yet, I'll stop there.

Power supply to this section is 200V. Each IF stage power tap is decoupled.
Each 6EH7 has a 25k screen resistor, a cathode resistor of 22 ohm unbypassed
plus 100 ohm bypassed. The screen and B+ side of the IFTs are bypassed.
This has been built 'ugly style' on a sheet of one-sided PCB. As a result,
grounds
are very short and the layout is much better than my normal mess.

If I disconnect the IF stages then they are stable with nothing connected
to the
input. They have about 300x voltage gain measured at the output side of the
IFT.
This doesn't seem excessive: 10uV input will give about 0.9v at the product
detector.

The layout is: tube - IFT - tube - IFT. This section of circuitry is about
7" long.
The IFTs are connected correctly (according to the legend on the IF cans).

When I cascade the two stages, the combination oscillates *very*
enthusiastically
at around 455kHz. Increasing the (bypassed part of the) cathode resistor
from 100
to 200 ohms made no noticeable difference.

At the moment there are no tube shields.

Things I can think of trying:
  - reduce the stage gain by increasing the screen resistor, etc
  - kill the Q of the IFTs by adding parallel resistors
  - add some screening.

Anything obvious I've overlooked? What else would be worth trying? I don't
think
I can reduce the gain too far (but perhaps I have miscalculated the
required gain).

73, ian K3IMW