It was tuning near in frequency, so it probably wouldn't have been a whole section in or out where it shouldn't be. Finally, I just happened to be looking around with a bright light and....

Using a grid-dip oscillator which you dial in a frequency and the current will dip when the circuit under test is resonant. The tank circuit should be able to resonate around 3.8 MHz for 80 meters.

I took it apart and started poking around. There were several disk capacitors that are switched into the tank circuit depending on the frequency. I checked the switches and they seemed ok. Then I realized it would tune ok on 40-10 meters. It was just the lower frequencies with issues.

Working on a Kenwood TS820S that I had repaired several months ago. It wouldn't load up on 80m suddenly.