[Amps] HV Diodes

[Manfred Mornhinweg]

Paul,

> Regarding Alpha, the 77Sx (dual 8877) uses a quad set of Semtech SCH5000 
> modules rated at only 5KV/0.5A.  The current rating is increased to 1A with
> forced-air cooling.  I?m sure these have failed owners at one time or another
> but they seem to have held up well for most 77/70 amp owners.

Their rating of 0.5A in a bridge circuit is just as good as 1A diodes used in
doublers, which was done in many older amps. It's pretty much at the limit, and
should last if nothing out of the ordinary happens, but they have little margin
to tolerate any overload. Things like a power microcut during TX, that allows
the filter caps to discharge but doesn't allow the diodes to cool off before
power returns, might kill them. A flashover in a tube might also kill them, if
the transformer and powerline are stiff enough to provide much current in that
situation.

> I?m all for ?super-sizing? components, but not until a circuit?s minimum safe
> limits are calculated.    Over-sizing for the sake of over-sizing seems like
> a lazy way to impress friends!

My take at safety margins is to first calculate as precisely as possible what
stresses a part will have to take, in all imaginable situations that it has to
survive. Then see what parts fulfill all requirements, with smaller and larger
safety margin on top of the actual requirements.  And then see what the cost and
other possible penalties of the stronger parts are. For example, when choosing
power transistors for an amplifier, you can't just use one 3 times larger than
needed, because both the cost and the capacitances will kill the project! In
other cases, like rectifier diodes, the cost penalty might be tiny, and the
other disadvantages, such as more bulk, weight, and higher capacitances, might
be complete non-issues. In that case it's fine to super-size parts, within
reason. There is a point where it simply makes no sense to go to an even larger
part, because the failure rate from processing defects will mask that from
overload failures. That's the point where one should stop, even if the larger
part's cost still isn't significant.

With your amp, 1N5408 diodes indeed would be a good choice, if the originals
ever fail, and you have room to fit them. If room is tight, I would just put in
1N4007's. Having forced-air cooling, 1N4007's are more than plenty big enough,
while 1N5408's still cost nearly nothing. And anyway Carl has enough of them to
give some to anyone who needs them! ;-) I too have a lot of them. I must have
around 1000 1N4007 and maybe 400 1N5408, a gift from CE2HS when he closed his
radio parts store at age 92. That was about 15 years ago.

Steve,

your account of the SCH-5000 is really interesting! So these were very high 
quality diodes for special high-rel applications. Using circular dies must be a 
pretty rare thing, I guess, because of the additinal etching and the lower 
amount of parts per wafer.

How did those high cost parts find their way into ham equipment? I mean, a power 
rectifier in an amp's power supply really doesn't need to have such ultra low 
reverse current specs! As long as the better diodes in a string can handle the 
leakage of the worst ones in avalanche mode, without failing, that should be 
good enough!

Could it be that at some point these quality diode modules became available as 
surplus at low cost, and then the amplifier manufacturer grabbed them?

Just guessing again. Darn. Carl told me to stop guessing! ;-)

Manfred



========================
Visit my hobby homepage!
http://ludens.cl
========================