Mains Voltage is running at 244VAC
The Toroids are rated at 240V (2x 120) out to 2 x 22V totalling 384VA,
the eight outputs are all 17VAC within 0.1V under load. The voltage switch is 115v/230V.
In the Power smoothing Chassis image, the ones running hotter are marked 0 and 1, respectively MJE15030G MJE15031G :-
Very little head room to add to heatsink, only a 1/4"; 6mm height increase.
Max permissible heatsink size increase to:- H= 36mm x W 40MM x 18MM.
On arrival to the Preamp chassis it's clearly engineered to output two dual rails of 24VDC however with the heat imbalance, within either 'rail' there is a 1 - 1.2v difference and between rails there is a 0.6 - 0.7 V difference, not much but i would expect this to be spot on 24VDC given the over engineering.
Krell suggests/says:-
" Minor power supply variations are tolerated because of the servo circuits in each section or stage. Since the circuits you describe are using discrete transistors (as opposed to regulator IC's like 7818 or 7918) if you wish, you can replace the zener diodes with "matched" devices to obtain closer voltage tolerances."
and
"There are a number of our products that were designed long ago, with some transistors operating at levels sufficient to cause high temperatures. If the original transistors in the device are operating too hot, and space permits, add a small heat sink to them."
I think we are talking about the 1N4750's in the above image.
Will update when the pieces come together, preferably from Krells archive...
On another note:-
To quote from another post from 2006...
Also those obsolete VN0210N5 and VPO210N5 devices only operate at about 3 mils so good subs are Zetex ZVP3310 and ZVN3310 and they are readily available and dirt cheap"
Not having experience with vertical DMOS power FET's,
has anyone else experience in such an application and can advise please?