[glyphmon]

Thank you everyone for chiming in.

First, to MrFixr55's suggestion of measuring the resistance from pin 5 on the 6A8 to chassis ground with the tube removed, it tests at 34K ohms.  When I replaced this resistor, I couldn't get a 32K, so I went with 33K instead, so this feels like close to where we should be - certainly within 20%, though higher than I would expect from the path to ground.  And don't be sorry for falling asleep while thinking about my problem.  I needed to sleep two nights to get the ambition back to look in the darn thing again.

Second, because it was fairly non-invasive, I snipped the wire running from the junction of R17 and R29, which itself runs to F7 on the band switch, and tested for continuity as best I could (20M ohms setting) between the oscillator section side of that wire, and every terminal on the terminal strip, the pins on the 6Q7 and IF 6K7, and at the junctions of the electrolytics up adjacent to the field coil.  The good news, I suppose, is that I couldn't find anything other than infinite resistance, so if there's short, there isn't a strong one.

Would it be problematic to power the radio on briefly to see if stronger current is successful in finding a path?  Without the oscillator section powered, it simply just shouldn't work, at all, right?  And if not, then we know, whatever is coming in is coming in through that snipped wire, and it can't be more than what the voltage is at that junction, and therefore, is just a wonky reading related to the meter used to do the measurements?  I haven't tried powering it on without the 6A8 installed yet, but imagine the result would not be too different.

Does it make sense that if the radio is simply terribly out of alignment now, that that could be introducing the weird readings with this meter, even though it had relatively nominal readings for the oscillator before?

I know there has been discussion about how to simulate older equipment, i.e., 1000 ohms per volt meter, with modern equipment.  Would that be worth trying to see if that has any effect?  I do have a Heathkit IM-11 VTVM, but not sure if that would be any better at this particular measurement.  Goodness knows the probe certainly won't fit where I need to stick it to reach the pins.

I do have a signal generator, but no signal tracer as yet.  I think though, with the above diagnostics, I can assuage my concerns about the potential for a short.

Based on the outcome of these next steps, I guess the next thing I should do is try an alignment and see what that gets me.  Maybe everything falls back into place.

When this is all said and done, I'd be curious to hear people's opinions on the design of at least this chassis, and, I suspect, any Philco utilizing unit construction vs designs from other manufacturers.  For example, the RCA 6K3 that I worked on had the same essential band ranges, a three-band switch, a two-gang tuning condenser, but when you look at the schematics, and indeed, inside the chassis as a whole, it is far simpler and less-cramped than this 37-620.  I'm sure some will argue that by separating all the different coils and transformers out, and having the extra gang, and every other difference, that maybe it will have resulted in a higher quality product.  But in your experience, did it really?  Was it noticeable?  Or did it simply make a charlie-foxtrot out of something that could have been much simpler?

Best wishes,
Roy