[Amps] A Meeting Ground

[Will Matney]

Bill Fuqua wrote:

> This has always been an issue. Particularly when high-Q circuits are 
> involved.
> One time while helping the accelerator tech try to find out why our 20 
> MHz, 10KW bunching amplifier signal was getting into the sensitive 
> instrumentation I noticed that in haste they only installed every 
> other screw on a panel shielding a very high-Q tuned circuit from the 
> outside world. This was a 3 turn ( really looked more like twist) 
> water cooled coil with the ends of the tuned circuit going to 
> defection plates used to deflect 5 MegaElectron Volt protons at a 20 
> MHz rate.
>
> Well, I took an insulated screw driver and put it between the shield 
> (where a screw was supposed to be) and the chassis. And there was an 
> RF arc. The there were screws on either side only about 6 inches away 
> and we were producing an arc half way between them on a 1/8 inch thick 
> aluminum plate which they assumed was grounded.  Needless to say, they 
> reinstalled the screws and the problem went away.
>
> I have noticed some homebrew amplifiers I have purchased have the 
> plate tuning capacitor grounded to the front panel. This is bad. Since 
> all the fundamental and all the harmonic currents from the plate of 
> the tube flow thru this ground path and can produce a good deal of 
> radiated TVI off the front panel.


On smaller amps I used to like using the "APL" series of air variables 
by Oren Elliott. The reason being is I could order them with extra 
spiders (wipers) that you could connect straight to ground on the 
chassis pan. These I had put on the aft end of the capacitor between the 
last plate and the ceramic end. Actually you can have this added to the 
larger air variables they make also. They come standard with one on the 
front if no back ceramic end is used. But this wiper has a solder 
terminal which can still be connected to the chassis pan and before the 
faceplate. Since currents take the closest path to ground, it should go 
through this wiper lead to the pan instead of the threaded shaft through 
the face plate.


>
> Also, it is bad when the plate capacitor return path to cathode on the 
> chassis is common to the input circuit. Even though the circuits are 
> on opposite side of the chassis there can be a good deal of coupling.
>
> Just some thoughts.
>
> 73
> Bill wa4lav



Best & 73's

Will Matney