Phil Allison wrote...
Winfield Hill
(snip massive overkill solution)
** You haven't got a fucking clue what the app is have you - Win ?
Never stops you from make a whole string a unjustified assumptions.
Ever heard of "Occam's Razor" ??
Although I'm offended by Phil's tone, which isn't conversational,
and is certainly impolite, I'll address his remarks. First, we
still don't know Ross Tucker (NS7F) rjtucke's actual application,
although he's come back with a short response. Let's assume it's
a tube amplifier of some sort, audio or RF. We also don't know if
his 520V is regulated or not, but if it was we'd assume he'd seek
a way to simply change its regulation voltage.
One might assert that actual supply voltages aren't very important
for tube amplifiers, but unregulated power supplies suffer from
ac-line variations, voltage droop under load, and ripple losses.
Arguably the voltage range resulting from the sum of these effects
compromises the best-possible performance of a tube amplifier, and
indeed one often finds a regulator employed in tube circuits.
Furthermore, if one takes an already-wide unregulated voltage range
and applies it to a lower voltage, such as 350 vs 520 volts, using
a fixed voltage drop such as with zeners, the variation compared
to the new lower voltage is nearly 50% higher. If one didn't have
a regulated supply before, he'd certainly want one now! That's my
story, and I'm sticking to it.
Phil suggested six 5-watt 28V zener diodes in series to drop 170V.
The 5-watt dissipation rating comes from assuming the junction is
operated at 200C (which is too high, BTW) and the zener leads are
maintained at 75C or lower. But the latter would be very difficult
to do with six closely-spaced diodes dissipating 14 watts, so even
though each zener is dissipating only 2.3W, it's likely the zener
junction temp will still approach 200C. So Ross would have to add
9 volts for 5% zener tolerance to his already-wide voltage spread,
plus up to 32 volts for the zener tempco, according to the 1n5352
datasheet. That's another 41V or 12%. It doesn't look good to me.
As a last comment, one should consider what will happen to these
series zeners under the condition of a short-circuit fault. Ask,
how much current will the 520V supply deliver into a 170V load?
While I'm not enthusiastic about fixed-voltage-dropping schemes,
if I were still to choose this, I'd make an active zener using a
power transistor and resistor to dissipate the heat, and thereby
avoid dissipation in the voltage-reference components, e.g.,
.. 20-watt adjustable-voltage active zener
..
.. n-channel MOSFET
.. 520V in 1.5k, 100W IRF620, 10W sink Vin-170V
.. ----+-------/\/\--------- d s --/\/\-+---- 100mA max
.. | g 10 |
.. | | 1n4744A |
.. | +----|<|---+ 75k 2W
.. | +---/\/\---+--/\/\--,
.. | | 1.5k |
.. | 22k gnd
.. | 1n4757A 51V 1n4753A |
.. '-|<|--|<|--|<|-+--|<|--+
.. | TL431 |
.. ,--+--|<|--+--,
.. | | |
.. '-/\/\-+-/\/\-'
.. 500k \__| 36k
The 1.5k resistor is a large one (to run cool normally, and to
handle a short-circuit), say three 500-ohm 50-watts in series.
If the voltage and current capabilities of the 520V supply are
well understood, this rating can be reduced. The resistor only
dissipates 10W under normal conditions at a full 100mA load.
The 22k resistor in the MOSFET's gate limits the zener diode
current to protect them during a short-circuit fault.
There's a 75k minimum load resistor in case there's no load.
A TL431 adjustable zener IC with 500k pot allows fine adjustment
of the voltage-dropping value, from about 155 to 190 volts. If
this feature isn't wanted, replace the 1n4753 36-volt protective
zener with a 1n4744A 15-volt zener and eliminate the '431. (One
can change the voltage-dropping range by selecting other zeners.
IRF620 mosfets are limited to 200V, but one could use an IRF720.)
A power resistor with a power transistor is better suited to
dissipating high power than modern zener diodes, and allows for
a few other features as well. But it's interesting to realize
that the circuit above is only a few steps away from becoming a
full-capability voltage regulator, rather than a mere voltage
dropper. When one considers that the most of hard work goes
into the power-handling components, heat-sinks, etc, this is a
compelling thought.
