Patrick What sort of bypass caps does the screen connection have. How long are leads? I would try to put a good VHF bypass in there, if possible, something with very short leads to chassis from the tube pins. One cap that comes in mind if the American Technical Ceramics 100E style leaded chip capacitor. It has wide strips for leads, very low inductance. 390 pf value tends to have zero impedance at 200 MHz (self resonance) and higher values look like an excellent RF short at correspondingly lower frequencies. You can check the ATC website and see a chart of impedance versus frequency for these caps I believe.
In addition to the ATC cap, have some good lower frequency bypass like two x 5000 pF or so in short leads ceramic disk or doorknob style RF ceramic RF caps (5 kV). A series grid 'stopper' resistance isn't a bad idea, but you shouldn't need that high a value (much over 50 ohms) if it is working. I would be suspicious of the anode circuit having a parasitic resonance that is getting back to the grid and regenerating. I cannot remember from the photos if it had parasitic suppressor on the plate lead? I never shipped you those Raytheon RA1000 power chokes, do you still need one or two to replace the one with broken studs? 73 and good luck with the 4x1K John K5PRO New Mexico > The status on that beast is that there is a parasitic oscillation in the > final's grid circuit, between the 4-1000 control and screen grids, at > about 140MHz. Plate tuning, current, and neuralization setting do not > matter or vary much at all with this issue and the grid current and > screen current interact alot. I found that a 50 ohm resistor in series > with the grid helped. Maybe I need to increase that value or I don't > know.. The screen grid is heavily bypassed directly at the G2 terminals > on the socket, and an RF choke is optionally in series with the screen > to the chassis connector to no effect. This oscillation happens no > matter if the grid drive connector is shorted, open, or terminated, so > this tells me it is in the control grid circuit somewhere. Nothing's > been changed from the original design there, so it is further mysterious. > >
