USING MY UNI AS AN OPT
Using my current UNI as an OPT requires special care and consideration. Firstly, it must be emphasized
that this UNI prototype (KE87060-A) is wound on standard "hard" iron (see OPT HF & LF Performance Test)
which in short means that the UNI is extremely sensiteve for unbalance DC currents. This
conseqvently means that the current UNI prototype may ONLY be used in well-balanced pure Class A
amplifiers (as well as almost pure resistive speakers...)
My current UNI is in other words a prototype only and should be considered as such. My next version,
if profitable, will be wound on the same type of iron as my dedicated superb Williamson OPT
(KE57613-C+) which in turn will yield a remarkable universal, good and cheap tube transformer!
All drawbacks considered it may actually already work quite well as an OPT (as well as Mains Transformer)
if connected as in fig.2 above. After recent separate testresults I'm proud to announce that my current
UNI actually works as an OPT IF the standard primary windings are omitted (or perhaps integrated like in fig.3)
and the secondary S7&S8 are used instead.
The S7&S8-leakage inductance is only 50mH (i.e 40kHz@KT66/10ka-a) and the inductance is 135H@5V/50Hz.
The critical L(dI) is however only 37H or Lp/4@+/-7mA (compare with 33H or Lp/3@+/-20mA!
for my dedicated OPT...). A bonus is however the fact of 700V@40Hz (read 260V@15Hz) which is
somewhat better than the L(U) tolerance for my dedicated OPT (230V@15Hz).
While the low dI tolerance IS a big problem, there actually still excists some 40 Henrys at reasonably
high offset currents (<7mA) which makes it practically possible to use the UNI as an "Class A" OPT.
Please observe the need for serious (static as well as dynamic) tube matching as well as an
(automatic?) integrated bias offset feature, like Williamson actually integrated in his
original design but I has, due to excessive cost and almost too good an OPT, omitted.
In regard to fresh testresults, I really don't know what or how I obviously have failed to
measure and plot the UNI L(dI). The good surprise is however that the UNI obviously is much better
than the original testresults showed(!) I really don't know why but in view of my nowadays much
simpler test method, it is possible that the "damned" Murphy has been interfering again...
Instead of using my customized high current OP, I've now used the much simpler and more
direct and certain "battery-method" as described in OPT LF & HF Performance Test.
Anyway, it's not that easy to do these tests with only a car-battery charger as an AC-source and
a 9V battery as an DC-source and it actually seams like it almost only is a scaling flaw cause the
actual newly re-verified high dc unbalance sensitivity has almost the same proportions
(Lp/3@+/-7mA and Ls/4@+/-7mA for the primary and HV secondary windings respectively)
It should finally be noted that you have to be somewhat "religious" while using and testing the HV secondary
winding leakage inductance. Depending on actual grounding technics I've verified everything within
50-83mH leakage inductance as well as more or less small stepresponse oscillations. The best way to
configure this UNI as an OPT may well be to omit unused winding's grounding totally (I had the best
leakage inductance as well as insignificantly small HF oscillations with ungrounded primaries
while series grounding the rest of the unused windings). Generally I however recommend using
the fig.1 grounding technic which has been proven alright even though ungrounding the primary
windings slightly lessens the 100kHz ringing amplitude (<5%).
Watch out and good luck!