[Patrick Turner]

On Oct 15, 6:25*pm, flipper wrote:
On Fri, 14 Oct 2011 22:56:03 -0700 (PDT), Patrick Turner





wrote:
On Oct 15, 3:47 pm, flipper wrote:
On Fri, 14 Oct 2011 19:42:32 -0700 (PDT), Patrick Turner

wrote:

snip

In fact the 75V
zeners had a zener voltage of around 82V when new. Their working
condition was with about 20mA in each, and 82V, and that's 1.64Watts,
and I'd cramped them between plate of aluminium and the aluminium
chassis with lots of heat paste to keep 'em cool as cucumbers.
Didn't stop them frying it seems.

You didn't mention what the part number was but, FYI, if they're the
typical plastic case 5 Watt zener then heatsinking the case doesn't
accomplish much, if anything. The rating depends on dissipation
through the leads soldered to the 'heatsink', which is typically a
'large' PC board trace area but can be something more robust.

The leads 'to the heatsink' should also be as short as possible.

The "75V" zeners are of unknown asian origin, maybe 8mm long x 5mm dia
hard plastic bodies.

Sounds like the 1N53xx series. Those are about 8.5mm by 3.5mm.

So when they are cramped between Al plates with
generous heat paste, they still may get warm internally.

Not just 'warm' inside, 'hot'. The body is not a thermal interface for
them.

1.64W isn't
much though. But at just after amp turn on, and before OP tubes
conduct, the current is 50mAdc, which is 4.1Watts, which one might
think was OK

I don't know. They don't give a free air transient thermal response
but there's hardly any thermal mass there to heat up, meaning I'd
imagine they get damn hot damn fast.

until the screens conducted some current within 15
seconds and Ea dropped to reduce the total Ig2 supply to a steady 42mA
with about 1/2 that in Zeners and 1/2 in KT90 screens. So probably the
repeating heating cycles on the zeners affected their zener voltage.
I've known zeners to be a bit dodgy in this aspect, but as long as one
doesn't exceed 1/3 the Pd rating for continuous use they are OK at
least for the short term. I could have used say 10 x 33V x 5W Zeners,
which may have coped better.
Sometimes I've wrapped a U shaped piece of copper around zeners with a
screw to hold them to a chassis and with silicone or heat paste, and
such allows one to hold a finger to the zener indefinately, ie, T
45C, which should be OK.

That's what I'm trying to tell you. The 'body' surface being a
'comfortable' 45C doesn't mean anything because they are not
heatsinked through the body.

You can't really calculate anything without knowing the 'heatsink' to
ambient thermal resistance because everything is based on lead
temperature, which one derives from the power dissipation and
'heatsink' thermal coefficient to free air, and lead length, which
gives the internal diode to heatsink thermal coefficient.

I.E. The 5 Watt rating is with 3/8 inch leads at 75C (heatsink temp).

There is no 'free air' or '25C' rating. However, one could calculate a
heatsink rating at 50C ambient (not an uncommon rating, although maybe
too low for a tube amp), which would allows a 25C rise on the heatsink
while dissipating 5 Watt. That would mean a 5C/W heatsink soldered to
the leads no longer than 3.8"

Heatsink effectiveness would depend on convection or forced air
cooling. For example, assuming copper traces were 'perfect', the
convection thermal resistance of free air is around 166C/W per square
inch (conservatively high) so to get to 5C/W you'd need 33 square
inches, which isn't chicken feed. On the other hand, it's likely (with
solid state) one end is soldered to a ground plane that acts as a
heatsink.

You can pretty much bet that anyone using the 'full' 5 Watt rating is
employing forced air cooling.

Ironically enough, all that 'heat' in tube amps can be an advantage as
I'm sure you generally put convection cooling holes around your power
tubes. That means the internal air isn't 'still' but is moving due to
the convection currents created by the power tubes. However, if those
aluminum plates blocked the convective airflow they could have caused
more heat retention than what little they could squeeze through a
plastic body that is essentially a thermal blanket around the zener.

Anyway, the point is it isn't trivial.

Indeed, nothing is trivial. All things must be carefully considered
when building tube amps, and sure, I have lots of holes around each
KT90 in the chassis and a bottom chassis cover with 70% open
perforated sheet steel, so air flow is pretty good.
The new arrangement without a zener string should work better, IMHO.
I've nearly completed the much tidied up 8585 PSU schematic, may
upload it later today.

Patrick Turner.