4cx3000A high level mod

Tom,

I have modulated the amp with the Yamaha and the result was excellent from fidelity and quality sides.
The ham society here in Greece found the sound more elegant and complete than the pll exciter with absolute linear amplification could do. Myself couldn't believe how smooth and sweet the voice came out from the monitor. I made the blind test between console's phones monitor and the phones monitor of the Lowe HF125 with 7khz & 10khz filters and 3 of 3 thought that the radio sound was the console sound. The chain was experimental without any processing just a mixer and a condenser mikrophone.
The thing is that with this amp - at 3200w bridged- I couldn't get more than about 75% modulation to check the sidebands for artifacts or for splatters. For the 3 nights of testing nobody complained about splatters but all said that I need 20%-30% more volume to modulate the huge carrier. The big amp will be ready next week and I'll test it further. I could also upgrade the transformation to 41:1 and get more audio from the mod transformer but I want stay closed to the right calculations so the truth is that I need 5000w af at least.
In here there are very powerfull linears with 4x15s that modulate 5-10kw at 130% without negative peak limiters and annoy at least 15-20khz left right. All that story is a local fest out of mf band that in Greece stops at 1602khz, then we start at about 1620-1720khz with nvis horizontal antennas for covering with 70-90 degrees take off and extreme power 300-400 miles max. It is however an elegant ham-pirate experimental community trying to mix rf power with modulation quality in a local race (Totally oposed to CBers, except power). The mod transf is not that simple as an ac hv transf but it is more closed to a pushpull output transf with 4 windings of primary between 5 windings of secondary. W2PFY, I ordered a 7kw pushpool output transformer from a company that used to make same transformers for valve B class amplifiers used for musicians in open area concerts. They did first time so heavy (360lb) but they had the know how from 1-2kw ones. They also made the 3 reactors (30H,10H and 2H).
I used to have big problem with the kick back of the reactor but Steve gave me the golden solution and now all are in right order. My only worry is how to protect the solid amp in case of smthing goes faulty...let's say if the driver dies or shut down during transmitting, the reactor will spark and the trancient will jump as a small in time but huge in energy spark on the sensitive solid state output...A capacitive coupling could help but I am not sure, does anybody have an idea if a series cap would help and what are the minus or plus for that?

Stefano