Been thinking about "solving" the heat issue on the power board.
So we have to "get rid" of at least 18Volts of energy.....that's a lot really !
The solution IMHO should have some important aspects:
- must work without problems
- solve the heat issue
- be cheap
- use as few (extra) components as can be
- the less mods on the board the better
The first thing I thought about was using a voltage regulator instead of the zener diode.
This would be possible, here's an example schematic:

HOWEVER, the problem with this is that whichever way you turn it, the whole thing _still_ needs to "get rid of" (offical word: dissipate) 18Volts....so the heat is going to exist anyway because the LM317 is a lineair voltage regulator....
The one thing I don't know is wether a set-up like this will heat up as crazy in unloaded condition...if not, this would be a good idea.
Then I googled a bit more, and found out that there are beefier versions of the LM317, like LM150 and LM350, which are basically the same but can provide currents up to 3 Amps.
The Q*bet manual claims that the 30VDC section is designed to deliver 1.5A so that would be more than sufficient.
The best thing is that the LM150/350 come in a TO-3 housing so......could I put that on the spot of the 2N3055 ??
Well yes I could, but I'd have to mod the board/connections because the pins don't match (Vout is the case on the LM, but that's the Collector on the 2N3055, essentially I'd have to swap the case connections and one pin. I already thouhg of a clever and easy way to do that but.....read on)
OK, so if would really be pretty easy to do this mod, you'd have to remove the zener and put a pot there instead and bridge the input resitor. A second resistor would have to be put in between Vout and the reference voltage pin....could be done.
However, again , these are still linear regulators, and the amount of heat produced would still be the same ! The big advantage though, is that the heat will be pumped right into the heat-sink.....
I don't know how hot the heat-sink normally becomes, but seeing that the black is totally gone on my heatsink I guess it does get pretty hot.
The next idea I played around with was finding a cheap "buck" DC-DC switching converter and "hacking that in". It wouldn't look pretty, but it would GREATLY reduce the heat because switchers of course have efficieny of up to 95%....
So I searched and searched, but there don't seem to be any of these devices that AND can handle 2 Amps, AND can handle 50V input AND put out +30V......
Then I thought....why can't we take away the SOURCE of the problem, being the rather high 48VDC "source" voltage after the rectifier ?
Of course, I cannot change anything about the 35VAC coming out of the transformer (unless I'd use an extra transformer, but that's not the way I want to go)
Since the AC-DC conversion is done by the usual (and efficient) diode-bridge set-up (=full wave rectification) couldn't I go back to something basic and use only 1 diode do we'd get a half-wave rectifier.
Usually you don't want half-wave rectifiers because they waste half the energy, but that's about what we're after here

The problem with this idea however is that the ripple will be much higher with half-wave rectification and usually audio amps don't like that (or you can hear it).....
In my basic idea the voltage _would_ be (about) half of the 48V, so a nice 24V. This is definitly enough for the LM379 sound amp (minimum power voltage is 10V) and the knocker might be a little bit weaker but not much....
Now I wonder if the two filter caps of 2200 (in parallel, so 4400uF) would be enough filter for half-wave rectification.
I guess there's only one way to find out.....try it and give my oscilloscope some good use

In the mean time, if wiser electronic buffs then me have any opinions about this, please chime in (after you stopped lauging about my stupid ideas

)