[Chas]

Has the area of the noise been identified as to being before or after the volume control. Determining that helps...

If the suspect area is the RF/Osc/IF and to eliminate the AVC, clamp the AVC at a given negative voltage with a battery or make a simple AVC voltage "generator" from a 9 volt battery and a linear potentiometer, 50-100k, across it. Attach positive to Ground (of which appears to be common with AC line) and the variable negative from wiper to AVC . Operate radio and set for comfortable volume level as well as AVC (which will control volume too). If radio continues to go up/down in volume it is the AF side, suspect leakage in AF transformer to frame or transformer. (That would exhibit a warm-up failure characteristic). If the radio is stable in volume disregarding aether fading, do test in daytime to be sure, then problem is in "possibly" the AVC, new defective bypass caps or defective resistors or leakage in the AVC circuits to ground. Note available AVC current is very small in order of pico amps so any meter except old school EICO VTVM which has a 25 meg input impedance cannot measure the AVC voltage correctly. That said, any leakage will effect volume.

To further confirm AVC circuits, turn off radio and unplug. Measure AVC for resistance to ground using a 20k ohms/volt meter AND lift R5 from the volume control. BTW a defective volume control can cause the problem, usually because it has been doused with incompatible cleaner. The resistance reading should be infinity if all is well. Any leakage is a problem and can be found by lifting branches of the AVC...

To make a socket cleaner use any dud tube with all the required pins, take very coarse sandpaper to pins like 80 grit, thoroughly roughen the pins all around, all pins. Insert this tube "tool" in/out of the socket repeatedly, some 25 times. The pins are a match for the contacting socket surfaces so any (soft) oxides will be abraded safely away without removing excessive base metal.

I would choose a metal tube, as a glass tube may loosen in the base or break in the hands.

Last blurbs:

Any connections via a rivet to the chassis are suspect! Use a large soldering bolt (much copper) and burnish connection at rivet (wire brush) and any terminal device involved, use flux and warm chassis with heat gun if need be and solder. Note: Absolute NO-NO! Don't do this if you are timid! Use a tiny dab of plumbers liquid acid flux with a cotton swab if there is cadmium or zinc corrosion present, rosin flux will not cut that oxide. The flux will puff-off when heat is correct and solder will flow like water into the joint. An alternate is to drill out the rivet and use hardware to make a sound connection. Real zinc chloride/hydrochloric acid liquid Plumbers flux is hard to find in some states. Do NOT use any grease/oil based version of this flux as that cannot be removed nor burns away and WILL corrode...

There is my experience and other ARF members of re-soldering chassis mechanical/electrical connections. These defective joints do NOT respond to DC measurement with an ohmmeter, so, are often ignored as a potential problem. Reminder: It will take a lot of heat to properly solder a chassis connection such that it does not look like "chicken poop", Spot warming the chassis to 120 F helps, a lot...

Do not ignore mechanical connections like the frame of the tuner to chassis...

GL

Chas