If I’m interpreting this correctly I think the amp is faulting on high grid current. If you load the amp up in low voltage then switch to high voltage the amp is not properly loaded for the higher gain you’ll get from higher voltage. Without advancing the load the grid current can soar and trip the grid out. I’m guessing this could be the problem. Using low voltage should cease now that the old power rule changed. No reason to run any amp on low voltage anymore. The manuals were written to satisfy the FCC so the loading instructions are no longer valid under the new rules. No amp with dual voltages are ever perfectly loaded when switched from low to High plate voltage. Based on the old power rule it would be illegal to load the amp up correctly in the high voltage mode. Therefore the manual instructions had to be such. Today you can now key down and load an amp like that and be legal. Look at any newer amp, they don’t have those dumb voltage switches.
I would not want to run a SB-220 at 1200 watts out on CW anytime soon. Not with it's paltry 500 va xfmr. I'd be a bit leery with my L4B as well....and both have inadequate cooling. Notice on all those older amps with a high or low B+ position, that the higher B+ position is always 1.414 X higher. The idea here is when tuned to 1 kw dc input on the lower B+..... then switch to the higher B+... then INCREASE the drive, such that the plate current increases by 41%, on paper, the plate load Z did not change at all, and the amp should be tuned up correctly. ( I would just pulse tune the amp on the higher B+ position, using a pep wattmeter, and be done with it).
Some AM broadcast stations used the same concept to reduce PO after dark. If you reduced the B+ in half, and also reduced the plate current in half, the PO would then be 1/4. Same deal, plate load Z remains the same. By shifting the B+ up/down..and also the plate current (in the same % proportions), several different PO options are available, and all will have the same plate load Z. I would still re-tune the amp for each new B+ / plate current/ drive required combo.
In some ham applications, max PO is not required, or not allowed. We are only allowed 750 watts out on CW / FM/ data modes here in Canada. If only the higher B+ was available, and drive reduced, and amp retuned for lower PO, then the plate load Z would DOUBLE. The tank loaded Q would also DOUBLE. Tank coil taps are all in the wrong place. uh on each band would have to be doubled. On my hb amps, I have a 13 kv model 88 switch to change HV taps on the secondary of the big 253 lb dahl xfmr. The multi kw output PI tank circuit is independent of the manually tuned, tuned input (NOT ganged for that reason). To run lower PO efficiently, I can either use reduced B+, drive......OR use the higher B+ position, use less drive and plate current, then increase the uh on each band. IE: use the 75m tap when operating on 40m etc.
AFAIK, on some of the older ETO-alpha amps, the hi-low B+ function was done via a tap on the secondary of the plate xfmr....via an (optional) $80.00 spdt vac relay. Most folks would order the amp without the optional vac relay..and leave it hardwired for the higher tap.
For other applications like testing an amp, gettering a tube, lower B+ makes sense. Heath, Drake etc, used taps on the primary to reduce the B+. That's not an optimum method to reduce B+, as the magnetizing current will increase a bunch. It works though, since you are just increasing the turns ratio of the plate xfmr...by stuffing 240 vac into the inner taps on the pri. In the dahl commercial catalog, you can see the same thing, called 'PO cut back taps' for AM broadcast station use, to reduce PO after dark.
