[morzh]

Brenda

1. Did the experiment, simply fully assembled the radio in the cabinet with the grill cloth and all.
The distortion persists.

2. While still out of the cabinet, some songs sound better than others.
For instance, "Hello Marie-Lou" by Rick Nelson sounded quite OK.
"Pretty Woman" - so-so.
Some of the Credence C. R. sound OK.

3. Now, I also soldered that 0.25uF cap across the choke in the filter and the hum got reduced significantly, at least on the scope 400Hz signal when applied to the 37 preamp tube's grid via a cap, does not look 60Hz- modulated so much anymore. Again, this did not help distortion much.

What I need is an audio gen, so I could check the distortion across the range and not just at 400Hz or 1000Hz.

4. The undistorted sine wave comes out to about 5V peak which makes it about 1.7Watt out, which is close to what the datasheet lists as the UPO for the 45 tube (1.6W) at 250V plate Voltage.

The last thought is, Philco 20 (2x71A), and 90 (2x45) use the same 2766 transformer for push-pull which has 100:1 ratio and it wants to see app. 7kOhm of resistance (6700 Ohm from the table) which is what 0.7 Ohm time 10,000 will get you;
The tubes at the max voltage want to see 4.8kOhms each which is 9.6kOhm for CLASS A (not sure but probably this is what we have).

Mine is 60:1 across the push-pull (30:1 each half) which makes the expected impedance for CLASS A = 3600*6 Ohm*1.25=27kOhm. The output expects ideally 7.8kOhm (each 45 expects 3.9kOhm load), and with the original speaker voice coil and transformer they would get the same 7kOhm which is close.

SO I am quite a bit away from optimum.

I might try a different transformer.
Just for the heck, to see if this is my problem. I should probably keep a spare 125 Hammond pushpull at home.