Using my UNI requires certain considerations. Firstly, it is seldom possible to overload this
Universal Power Transformer prototype with more than some 10% drawn power. The only situation
is when all secondary windings are NOT used and(/or) ALL primary windings are in use. In this case
it's actually possible to overload one or two of S1-S8 with some 100% higher current than nominal.
Please be aware of the generated heat (which never should exceed 80C) though. For a more specific
Mains and/or OPT analysis, see The UNI Loss Model and/or Using my UNI as an OPT respectively.

Figure 1 speaks for itself but it should be noted that output voltage may be tuned from some
500V to 450V with the use of R as noted. It is recommended that C never may exceed some 66uF while
being driven directly from GZ34 (or whatever). This is due to the fact of high capacitive
current peaks which may be estimated by calculating w*C*Up (=10.3Ap) vs Up/(Zdrive + 2rp + R) and
whichever comes first (4.6Ap@Up/Zdtot in this R=0 case). Even this 4.6Ap will however probably reduce
the MTBF for the rectifyer tube.

The figure 2 resistive and capacitive values are calculated with the just mentioned considerations
(220uF/3=73uF). It is furthermore notable that you more or less have to parallel the capacitors
with leakage/bias-resistors. The reason is different dielectric losses or ageing which otherwise
may destroy one or two capacitors due to over voltage. I've used the lowest possible 1W value
(47k/1W) for this purpose. Please note the quite long time-constant of 22 seconds.

The 0.1uF/1.6kV capacitor is used to eliminate HF current peaks which are more prominent in
solid state rectifyers (due to faster as well as lower internal diode resistance).

The reason for 2*3 diodes in the FW HV end, is that 1N4007 can't withstand more than 1kV reverse
polarity whereas the maximum reverse voltage actually is 2.6kV (1.3kVdc--1.3kVac) here.

Figure 3 shows my version of an Universal High Power Supply. You may of couse configure the UNI
in some optional other manner (and for instance use the S7-8 windings also) but this is more than
satisfactory for me. It is notable that you may draw as much as 2*4 Amps from the series
connected secondary windings (S1-6) while they only can withstand 4 amps. The reson is simple
because only half-wave rectification is used. The drawback is however the need for twice as much
capacitance. Except for the 5V/8A output you may however lower the (C2-C3) capacitance and expect
a lower maximum stabilized output voltage or current (30V/4A with 10mF should also suffice AND is cheaper!).

Figure 4 displays my sense of "humour". While it actually seams like a configuration like this
is possible, you should be aware of the difficulties. One major problem is the 7.8V battery
(which however may be implemented with 6V plus a number of 1.5V rechargeble batteries in parallel).
Another related fact is the fact of uncontrolled battery voltage drop but this may be controlled
by some sort of feedback.

Anyway, something like this configuration IS feasible and the fact of low leakage inductance might
even enable some 500-1kHz DC to DC conversion (with the consequently lower capacitance values).
Please let me know if anyone of you customers out there make something like this work.
I'm especially intertested in DC to AC conversion which, by the way, above uses one of my dedicated
Williamson OPT's as a pure inductance for series resonance elimination of the third (as well as higher)
order harmonics. You may however (due to L(U)) have to tune the capacitance a bit for optimum
higher order harmonic rejection.

There is however also the possibility to use the 5V windings instead but this still
requires an additional obsolete battery of some 3.7V (as well as lower output power).

I'm grateful for any tip what so ever. Thanks!