On Dec 4, 1:22*am, "N_Cook" wrote:
Robert Macy wrote in message

...
On Dec 3, 3:29 am, "N_Cook" wrote:

In a closed steel case, no air cooling ( fan air is ducted only through
the
power amp heatsink ) and no passive/convective vents.
Torroidal structure, no inherent saturation current with open secondaries
and otherwise well behaved. What sort of plateau temperature would you
expect if it was delivering continuous 600 watt ?

I would expect most transformers to temperature rise on the order of
30-35 C in free space, better designed transformers maybe 10-15C

But, you've said you've closed off the air space the torroid is in?
Not good. *Heat then must transfer from torroid to air, air to box,
box to outside air. *When you add up all the thermal resistances, that
torroid will run VERY hot.

Rule of thumb: 1 degree rise for every 1 watt with an area of 100
sq.in. located in free air.

Assume 95% efficient transformer *30W!

Two layers, we're talking over 60 degree rise from outside air to
inside torroid. *Just an estimate. *Try using some finite element
analysis software [free femm 4.2] to analyze the static thermal rise.
I've found it was fairly accurate in estimating rise of temperature of
fat/flesh for a SAR analysis.

******

That was worse case scenario. It is an existing amp and wanted to add a
thermal fuse external to the transformer as I did not like this almost
thermally enclosed siting. I added a 110 degree C , 10 amp thermal fuse to
the outer surface of the tx. Should never be run at 600 watt and there is
some forced air ventillation as the amp heatsink is not totally airtight and
slight holes around the casing, but these transformers were made without
internal thermal fuse.

Is that a power transformer, or the audio transformer?

If power, tends to run hotter, because no one cares much

If audio, it must operate very well over a broad band of spectrum,
tends to run COOL, because it can't run hot unless the core is
operating around its full range BH curve, which means distortion! So,
I think most designers, back way down the operating point and that
transformer should be running fairly cool. Did you ever measure its
temperature?

A way to measure INTERNAL temperatures in a transformer without having
to dig into them:
Run the thing until temp is stable. Shut off and immediately start
measuring winding resistance as a function of time. Plot the curve
and recreate the missing section of plot between t0 to t1, where t0 is
when you shut off and t1 is your first measurement. The resistance at
t0 tells you what the winding temperature was. Yes, you need to know
what the winding resistance was at room temperature to get a
calibration point.