So, your Tek 7904 stopped working.

This is really quite a nice scope - when it works. Bear in mind, there's plenty to have gone wrong in the last 37 to 49 years since yours was built, though. Some parts are completely unobtainable and others are common as dirt. Pray yours are on the latter list.

Here is a list of reasons mine has failed, and repairs I or others have made. I have attempted to keep them in chrological order but that's not guaranteed as some of this info is over four years old at time of writing, so I'm working from my known-fuzzy memory.

When I got this scope it worked great, aside from some minor dirty contact issues in many of the switches that went away after they were all exercised and some were sprayed with contact cleaner. I allowed it to sit for several years unused at one point and when I turned it back on, rather exciting noises and an EXTREMELY concerning amount of quite noxious smoke was emitted. I feared the worst, but dug in because what did I have to lose at this point?

I quickly determined that the two tantalum-electrolytic capacitors filtering the unregulated 17/-17 volt DC supply rails from the second PCB in the switchmode power supply were to blame. How did I determine this? Well, one had been replaced already by someone else, poorly, and the other had been replaced by someone else, poorly, and then had gone nuclear again. The great thing about tantalum electrolytics is that they're super easy to diagnose when they fail - they dead short and either smoke the weakest link in the circuit that charges them, or they dead short and go all Chernobyl when re-energized. A blast mark like that is pretty hard to miss. The not so great thing about them is, well, they leave a hell of a blast mark, molten metal spatters all over your board, etc etc. And sometimes the part they smoke when they short is an unobtainium one, not the junk cap. Anyways, these caps are common as dirt. I replaced them with modern aluminum electrolytics because we've made large strides in reducing ESR and ESL on those since this scope was produced. I also recapped the rest of the board's aluminum and tantalum electrolytic population at the same time because I didn't want to have to clean up any more blast marks. I don't recall the sizes or part numbers I used but will write them down next time I open it up.

Anyways. As soon as I had completed the recap, I turned it back on and a few more things smoked. Notably, my 7A26 dual 200MHz vertical amp did its best to catch fire (but failed) and after removing it, the unit seemed to work fine for several power cycles. Upon cooling down and powering back up however it failed to power on at all. Further investigation determined that I was far from the first person to work on the switch mode power supply (which I already suspected since I'd just replaced two already-replaced capacitors, but hadn't been sure of.) Among other things, the main high frequency (by 1970s tech standards... it's only in the mid tens of kHz range, rather pedestrian for an SMPS today) transformer had been desoldered and resoldered by a butcher, and it appeared the main bridge rectifier had been replaced on the AC input board. I discovered that the main AC fuse had blown and started tracing out what could possibly have caused that. Eventually I found that the already-replaced main bridge rectifier indeed had a shorted diode in it, replaced that, and all was well again. The bridge rectifier is an NLA (no longer available) custom Tektronix part number, but the ComChip BR1006SG-G I installed is better specced than the original, rather widely available at a whopping $1.02, and fit perfectly.

One thing to note - this power supply is ABSOLUTELY MISERABLE to work on. It's packed too tight, and it was never designed to be disassembled again after being put together. Removing the low voltage precision regulator board on top requires unbolting all the transistors from the heatsink and getting covered in heatsink grease. Removing the rectifier/filter board on the back looks like it should be easy since it's got a connector along one edge, but when you dig deeper you realize the custom transformer in the flyback converter stage is soldered to both it and the AC input/rectifier/chopper board, from the back. There's really no good way to get in there which means you'll find a lot of crappy repairs done from the top surface of both boards without accessing the back. If you find yourself contemplating this, I strongly recommend removing the side panels from the supply along with unscrewing both main filter capacitors and removing them and the piece of stamped aluminum they mount to. You won't get full access to the back of the first two boards this way, but you'll get close enough to do the job most of the time.

If I had had to get too much further into this thing - or anything truly NLA like the flyback transformer or SMPS controller IC had failed - I probably would have looked at redesigning it from the ground up. Most of the voltages it supplies are well within the realm of DIY hobbyist construction methods and parts availability aside from the HV DC supply to the CRT anode, which shouldn't even be too bad for a determined amateur.