Tektronix 547

Update: 8/28/18

The method described below did NOT hold up over time sadly, so what is written below is purely for interest.

Update: 1/28/16

I dried and vacuum sealed the HV transformers in beeswax at the end of August. Since then we have have 4 months of high humidity before the onset of winter. I have just run the two scopes for over 48 hours continuously with no HV issues at all. Previously they would run maybe 4 hours before the HV collapsed. To recap, this issue is well know in Tek circles and came about with the change of sealant from beeswax to epoxy and I surmised that it was due to the tendency of epoxy to absorb moisture 0ver time. Having baked one transformer out previously, it then worked but as expected, failed again after a few months in my damp basement. So, I decided to bake the transformers out under vacuum in molten beeswax, then gently restore the atmosphere and let them cool. It looks as though this solution does work.

This post is part of my effort to complete and clean up loose ends as I disperse my collection.

I have two of these superb instruments, the final bow of the valve era for scopes and what a bow! Sadly though, that bow was assisted by a mostly transistor Y system. In mitigation of this heinous offence, the magical X system was mainly valved and could have been implemented using valves alone. I am not going to provide a full-blown discussion of the 547 scope at this juncture, rather to focus (pun intended) on a problem specific to the 546/547 models, failure of the HV system due to the substitution of epoxy for the traditional beeswax HV transformer seal.

The problem is typically that the scope will work when turned on and then, after some time, the HV collapses. If using it, this is first evidenced by the traces lengthening as the CRT sensitivity increases with the falling HV, though the traces actually brighten, at least initially. I first experienced it working through the lengthy calibration procedure. I discussed the issue with Stan Griffiths several years ago, he told me that  it was thought to be dielectric absorption due to the epoxy. Hmm I thought, I wonder? So I baked a transformer out and it worked properly for a while, however it soon failed again; I am not the only one to have done/discovered this. My thoughts are: That the issue is not an intrinsic epoxy dielectric issue, rather power lost to heating of any moisture present in the epoxy; epoxy does absorb moisture (by diffusion) so what if I could dry it out and then seal the result? Knowing that beeswax is basically moisture impermeable, I decided to bake a transformer out in molten beeswax, evacuate it and slowly allow the pressure to return to ambient thereby sealing the newly dried transformer. I described my thoughts to Bruce Baur and Stan Griffiths by email back in May of 2014 but it has taken a while to get a “rountoit”.

Accordingly I bought a vacuum pot and pump for the quite reasonable price of $150. The pot is aluminium with a thick lexan top. The instructions provided with it insist that one MUST NOT continue to pump once the gauge has dropped to 29 in Hg otherwise there is a danger that the pot will implode! With respect, this is bunk. In round numbers, one atmosphere is 15 lbf/Sq in corresponding to 30 in Hg. Pumping down to 29 in Hg results in an external pressure of approximately 14.5 lbf/Sq in (or “squinch” as one of my excellent teachers used to say). Pumping down the final inch results in the effective external pressure rising to 15lbf/Sq in, that is to say roughly an increase of 3%. If this increase will cause the pot to collapse then it should not be on sale for this purpose. In fact for external pressure, vessels are customarily designed with a safety factor of 3*. That means it should be impossible to collapse the pot when pumping down from atmospheric pressure.

(*For external pressure, the design safety factor on yield has to be greater than that used for internal pressure because the failure mode due to compression results in failure by buckling which unlike direct tensile failure, is very difficult to predict accurately. Geometric factors such as dents are critical.)

I used USP beeswax that I bought from Amazon, for a fair price if there is any such thing anymore.

Here is the HV transformer and rectifier assembly partly removed:

Transformer removed showing heater power loops:

First, I put two transformers in a can with beeswax powder in the oven for 24 hours at 170 °F. (The melt temp is around 140 °F and the lowest setting for my oven is 170 °F) There was a nice smell of honey! To ensure that the wax remained molten as I evacuated the pot and returned the atmosphere, I needed to safely heat the vacuum pot. I heated a 5lb weight in a small oven at 220 °F that I then placed into the pot just before placing the transformers with their wax. The vacuum pot was placed into a large saucepan with water that I gently heated until I could just see wisps of steam escaping up the sides. (The weight also served to keep the vacuum pot firmly down in the saucepan.) I then transferred the transformers into the pot and evacuated it. I did go to far though because the wax started to boil below about 25 in Hg of vacuum. At that point, I turned the heat under the pan arrangement off and allowed the atmosphere to return by leakage, having turned the pump off. This took about 30 min and at the end, the wax remained molten as I intended. I then removed the top and hung the transformers over the tin to drip (having soldered on copper hook wires before starting the process).

If I do this again (I have a third transformer) the only change I would make to the procedure is to hold the transformer under vacuum carefully at 20 to 25 in Hg for some number of hours (with the wax molten of course) before letting it leak back to atmospheric pressure.

Ready to evacuate:

Evacuated:

Dripping:

By the way, my mother always looked in the garage if anything was missing from the kitchen!

I ran the first unit for 10 hours straight and it did well. In particular, the oscillator screen grid voltage was a few tenths lower than when I turned it on (not powered it up, what a pretentious BS way of saying it). The HV had stayed dead stable;  however stability of the SG voltage is the best indicator, it will rise if the oscillator load increases, topping out at 152V or so. Note the real meter (on the 3000V scale) on the HV test point!

The oscillator tube is run quite hard so you should check the transconductance. The few I have are around 25% down. The other thing is the B+ feed to the oscillator is via a 270Ω carbon resistor. If run hot, these develop internal cracks and that can be a reason for failure. The resistor in my unit measured 434Ω cold, most likely higher when hot so I replaced it with a wirewound resistor.

The question now is whether this fix will stand the test of time. I keep this equipment in an unconditioned basement and there are 10 weeks or so of high humidity left so if it works in the winter, that will be a solid sign that all is well.

8/30/15 Update: I have three of these transformer units, all three were failing in the way described and all three now hold up over a 12 hour period. I will test again at the onset of winter once the units have been exposed to what is left of the summer humidity.