[Collins] 30S-1

[Dr. Gerald N. Johnson]

What Collins called a screen supply in most any rig greater than 1 KW 
was a misnomer. It really was a cathode supply and had to carry the 
total of the plate and screen currents because the plate supply was 
grounded, it didn't float. Whether 30S-1 at 1KW or 821A-1 at 250 KW.

That being said, I found while in the Army that a MARS 30S-1 had lousy 
punch on high speed CW, that was cured by adding an old 350 mfd 300 volt 
electrolytic to the screen supply filter. I based that on Ray Swanson at 
Cedar Rapids having the same problem with the 208U-10 with Swedish 
silkscreens where they wanted them to run quality CW. The symptom showed 
up on the scope, that at key closure, the screen voltage takes a yoop, 
dropping enough to cut the output more than in half for about the length 
of a dot at 22 wpm. Then on dashes it rises to normal power after that 
delay. In the 208U-10 (10 KW output linear) it took a few thousand 
microfarads, maybe as much as a few 10,000, but 350 mf did the 30S-1 
just fine. This poor voltage regulation has to have contributed to less 
than perfect transient intermod distortion on SSB as well. I don't know 
that 100 or 200 may do as well, I had a 350 and it worked. I wasn't 
being paid a proper engineer's salary at the time to allow me to go to 
town and acquire a variety of capacitors to see which was enough.

Before you go wild with the variable "screen" supply, try my capacitor mod.

Remember too, that the tank circuits are designed to handle 1 KW input 
power, not 2 KW output power and to supply a load impedance to the PA 
tube for that 1 KW input rating. Cranking up the plate current requires 
a lower impedance plate load, which means lower inductance and greater 
input tuning C in the PI output network which is more than a little 
inconvenient to achieve. That then means more circulating current (where 
the circulating current is at least Q times the input or output current) 
to heat the coils and bandswitch contacts. Its on the order of a compact 
linear so the coils may not have a lot of extra dissipation capability 
for half again (2.2 times the heat) or twice (four times the heat) the 
current.

73, Jerry, K0CQ

On 9/20/2010 12:28 PM, Alan Chandler wrote:
> The 30S-1 is one of the lowest IMD amplifiers ever produced for amateur use.
> It uses a grounded screen grid and operates the cathode at about -200 volts
> as the screen supply.  The RF drive is supplied to the cathode so the RF
> adds about 100 volts to the cathode which results in a total screen supply
> of about 300 volts when fully driven.  The Eimac 4CX1000A maximum screen
> voltage is 400 so adding 125 volts to the screen supply will exceed the maximum rating under RF drive.  Typical AB1 operation for the 4CX1000A is with 325 volts on the screen which is about where the 30S-1 actually operates.  Increasing the screen voltage does increase the gain of the tube, requiring less drive power but it reduces the available plate voltage swing before the screen current rises to the level that exceeds the screen dissipation (35ma at 325v is just under 12 watts,
> the maximum screen dissipation.)  and the tube is in saturation. The net result can be more output power but with a substantial increase in IMD.  You will typically get about 1 to 2 dB increase in power (0.2 to .04 S units) and about 10 dB higher IMD (2 S units).  Be kind to operators on adjacent frequencies and leave the screen supply where it was designed to be operated.
>      I have just about finished setting up a variable screen supply and will
> make power and IMD measurements with a 4CX1000A and 4CX1500B in a 30S-1 later in the fall.  I started the project in the Spring but good weather forced yard work to take precedence.  When I complete this, the results will be sent to The Signal.
>
> 73, Alan - K6RFK
> ______________________________________________________________
>