[Amps] AL-800H transformer

[Manfred Mornhinweg]

Carl,

> With much lower core loss and less than half the primary power and half 
> the secondary power consumed the regulation should be much better than 
> the 15% Ameritron rates it at in the 811H.

No. Unless you managed to change the laws of physics.

The voltage sag in a well made transformer (which I hope this is, but I
don't know) comes mostly from the resistance of the windings, and only
to a much lesser degree from the leakage inductances. The resistance of 
the windings is the same regardless of operation voltage, and the 
leakage inductance is also pretty constant, except when the core 
saturates, which is not the case here.

So, the _absolute_ voltage sag in this transformer depends on the 
_current_ it is delivering, and not on the power. It will suffer 
essentially the same absolute voltage drop at a given load current, 
regardless of the voltage level at which it is operating. So, the 
_relative_ (or percentual) voltage drop at a given load current is twice 
as bad when the transformer is running at half the voltage.

> In intermittent duty at half voltage you should be able to pull a bit 
> more current at that 15% regulation but temperature rise should be 
> measured first.

It's true that from the thermal point of view you can pull _slightly_ 
more current when running at half the voltage, as long as the voltage 
sag is acceptable for the application. Since the core loss will be much 
lower at half the voltage, you can afford some more loss in the 
windings. But only some more, not much more, because transformer 
windings, specially in larger transformers, are usually limited by the 
thermal conductivity from the winding center to its surfaces. Working at 
half voltage, the core will stay much cooler, so the windings can 
dissipate some heat through the core, but you will soon run against the 
limit given by thermal conductivity between winding layers.

But there is more to it: The core loss of a given transformer depends on 
applied voltage and frequency, so it stays essentially constant 
regardless of load current, except for the minor effect of voltage drop 
in the primary winding slightly _reducing_ flux density and thus core 
loss, when operating at higher current. Instead the loss in the windings 
depends on the square of the load current.
Transformers designed to stay idle for much of the time, then deliver 
full current for short times only, are optimally designed by accepting a 
peak loss in the windings that is much higher than the constant core 
loss. A typical HV transformer for a linear amp might have a core loss 
of 20 watts or so, almost constant as long as the amplifier is on, while 
the loss in the windings will vary from almost zero while the amp is 
idling (only a tiny load due to the bleeders), to as much as 200 or 300 
watts during peak load! So, during relatively high duty cycle operation 
the transformer will get hot, while during low duty cycle use it will 
stay pretty lukewarm, and during long time idling it will be only 
slightly cooler than that. The alternative would be designing a 
transformer for higher core loss, which allows somewhat lower copper 
loss. That would be done with a transformer intended to run at full 
power whenever it is on, but definitely not for a ham linear amplifier!

So, with a transformer designed for ham amps, saving most of the core 
loss by running at half voltage will allow only a _slightly_ higher 
power loss in the windings, and due to the quadratic dependance of 
wiring loss on current, that will allow only a _very slightly_ higher 
output current, before overheating the windings and making the 
insulation fail.

And of course, higher load current makes the voltage sag problem larger too.

The bottom line: When using a given transformer at half its design 
voltage, it can be used only very slightly above half its power rating, 
and the percentual voltage sag will be twice as bad. On the positive 
side of things, its loss during idle time will be extremely low. And the 
exact amounts of these effects depend largely on the tradeoff between 
core loss and wire loss chosen by the designer of the transformer.

Manfred

========================
Visit my hobby homepage!
http://ludens.cl
========================