Ian Iveson wrote:

Bret Ludwig wrote

Partridge figured out that tolerance of up to 10% DC imbalance was a
necessity for the highest level of OPT performance and he was right.

10% of what?

Thanks, Ian

In Tremain's Audio Encyclopedia there is a graph for IMD vs DC imbalance
between primary halves.
So where you have 45mA in one tube, and 55mA in the other, then you have a
20% imbalance,
since 10mA more Ia flows in one of the tubes compared to the other.

At low levels its should not cause too much trouble and if you work out
the dc magnetisation
in Tesla for an average OPT then as long as the dc field isn't more than
1T then
there is still some ability for ac magnetisation. The lower the effective
µ, the greater the
tolerance for dc imbalalance which is the equivalent of the net imbalance
applied from a to a so
maybe 10mA dc a-a in the case above.
I don't exactly know what 10mA dc will do with a toroidal OPT with max µ =
40,000,
amd say Afe = 50 mm x 30mm, ML = 350mm, and Np = 2,000 turns.

But a core of E&I won't magnetize so easily; placing a fine air gap in a
pair of C cores or E&I lams
in a PP tranny can still give plenty of primary inductance, but very much
improve the tolerance
for dc offset currents.

Patrick Turner.