Re: 547 scope HV transformer problem. One practical solution.

Chuck Harris

Just to add some other flies for the ointment:

When I first started my 547 HV repair project, my scope,
which had sat idle for several years before I got it,
would only run for 15 minutes before the HV crashed.

After I experimented for several days, the run time kept
increasing, until after a week, it ran continuously.

I thought it had cured itself, and I was done!

Well, then I got busy with something else, and the scope
sat idle for a couple of months, and the process had reset
to the original 15 minutes before crash.

The issue is heat induced epoxy loss, and seems to be
related to moisture absorption by the epoxy. It seems
also to be related to a chemical change in the epoxy.

There is a temporary improvement when the internal
heating of the transformer winding dries out the winding...
but it is only temporary.

The oscillator stops because the ferrite core is operating
at a bad point in the ferrite's curves. Its frequency is
just below the frequency where core losses start to go
exponential. And, to add insult to injury, when the core
heats, it approaches the Curie temperature (about 150F),
which is where the ferrite ceases to be a magnetic material.
This causes the frequency to rise... which you guessed it,
makes the losses rise even more.

Back in the old days, when the EHT ran at 60KHz, and beeswax
was used to impregnate the winding, the HV section was right
in front of the scope's fan... out in the open. It stayed
cool and breezy.

Tektronix wrecked all that when they decided to make an
extended temperature range scope (647) and switched over to
an epoxy varnish potting material. They also decided to
unify the HV section to fit in a plastic box... to keep
it clean, and allow reuse on all of their future scopes.

The epoxy varnish, even when new, was much more lossy than
the original beeswax, so the engineers lowered the frequency,
and added a provision for heatsinking the transformer core.

They also started to think about using solid state rectifiers
to replace the old power hungry 5642 tubes.

You can achieve an improvement by cooling the transformer,
through most any means: fan, heat sink, solid state rectifiers,
lowering the operating frequency, lowering the drive voltage...

Lowering the drive voltage will make it hard to see fast, low
repetition rate waveforms... kind of like hobbling a race horse...

Lowering the EHT's operating frequency can be done by parallel
addition of a capacitor to the 1000pf resonance capacitor

Ultimately, I found that rewinding the transformer was the
only sure long term solution.

We probably wouldn't have ever noticed this problem if Tek
had chosen an MOPA (Master Oscillator Power Amplifier) design
to drive the transformer, instead of using a modified Hartley
oscillator. It only takes a slight lowering of the feedback
efficiency to reduce the oscillator's gain below unity, and
make the oscillator stop.

-Chuck Harris

Ernesto wrote:

Hi Morris,

Yes, the deflection factors, which I called the "gain of the CRT" is an inverse function of the high voltage, and I have taken this into account. It requires simple readjustments of the gain of the vertical and horizontal amplifiers. When all this is over I will care about finding good standards to calibrate the scope with. In the meantime I have made some more observations:

After being relieved with the good results of lowering the HV to 1500 V I decided to try out a "thermal enhancement".
After removing the plastic cover of the HV section, I placed over it an old PC supply whose little fan did blow mostly over the transformer. And here is what I got:

- With HV = 1500 V, the supply current raised slightly in the first minutes, then it stabilized. One and a half hours later it was still the same, and the supply was still running fine.
- With HV raised to 1655 V, the same happened and I ended the test after one hour.
- With HV returned to the nominal 1850 V, exactly the same happened. The HV supply DID NOT FAIL.

So the COOLING OF THE TRANSFORMER will solve the problem of my 547, which originally failed after running for about 10 minutes.

The easiest implementation of the cooling can be to drill holes in the cover right above the HV transformer and place on top a little fan (with a filter), blowing on the transformer with the plastic cover off.
But I don't like the aesthetics of the fan sticking out on top of the instrument, so I will do the cooling internally. A change from electrical engineering to industrial engineering, haha.



Join to automatically receive all group messages.