I have now ended the resuscitation from 40 years hibernation of my 547 scope. I could not be more pleased with it.
It has now inside an added little centrifugal blower fan that cools the HV supply. This has solved the classical HV transformer problem, that caused it to fail after 10 minutes.
This internal cooling was not easy, since the HV supply on the top left back corner is so removed and inaccessible to any air flow.
I removed the big metallic electrolytic cap. C802 so the hole left in its place is an opening to the metal enclosure, and replaced this cap with a small modern equivalent.
I cut a large opening into the front plastic box in front of the transformer to leave it exposed to incoming air.
I placed the blower in the space behind the vertical amplifier, and made an air conduit with plastic sheet, paper tubing, etc. around the CRT neck, then through the hole left by the capacitor, to the opening of plastic box in front of the transformer, a messy fix that I may improve some day when I am motivated. The plastic cover of the HV transformer, diodes, is now left off. Messy, but it works, and I am the only one aware of it.
In gratitude for working well again, I eased the strain on the supply by lowering the HV from 1850 to 1750 volt, a 5% change I don't notice on the screen, and I readjusted the gains of the horizontal and vertical amplifiers.
To further ease the work of the 6AU5 oscillator tube I lowered the supply voltage to the unit from the 325 unreg by 110 volt with a string of zener diodes.
Finally, to let the scope run cooler I changed the selector of line supply to the input transformer taps from MED to HI, since my home power is about 120 VAC.
If the HV transformer starts failing again after some decades in the future, I might not be around anymore to notice. In the meantime, the 547 is in my office on a cart, together with two TM504 supply boxes with miscellaneous plugins. I won't be able to resist powering up the instruments every day just to see the nice traces on the oscilloscope screen.