I did a few more experiments with my BC-221 and think that I may have possibly made the BC-221 better than it was at least before I started messing with it. That is, of course after it had sat around for about 80 years. 80 years is a bit hard to comprehend, but they are that old! By making a few mechanical adjustments I have brought this particular BC-221 back into top notch calibration.

I found that very slight movements of the VFO shield plate on the bottom of the chassis can be used to adjust the inherent inductance of the coil and bring it back close to the original settings in your CAL book. There are four screws that hold the aluminum plate to the unit. I put a pair of 0.030" spacers on the front two screws located under the chassis. It moved the crystal checkpoints closer to the original points on the scale. I then played with the spacing on the back side where the screws go thru the top of the chassis and secure the shield plate with two standoffs on the opposite side underneath. This is a BC-221N made by Philco.  Some other units may not have the same shield. All of these added spacers help to move the aluminum shield a bit farther away from the ceramic coil. This affects the residual inductance of the oscillator coil. I tried 0.020 spacers on the rear and saw the 4.000 MHz checkpoint go above the original value by about 4 or 5 divisions. The original error was  +12.4 divisions.  So I increased the spacers to 0.040 and the crystal checkpoint at 4 MHz was within about -1.6 divisions of where it was back in 1942. That is awfully close.

My method of testing involved calibrating the BC221 at 2.000 MHz on the high range and then noting each frequency marker without touching the corrector knob. My thinking was that the crystal markers would be close to the "book" value if the coil was exactly the same as it was in 1942. What I saw was a gradually increasing error as I moved up from 2.000 MHz. At 4.000 MHz I was 12.4 divisions off what was listed in the book.  I then looked at fixing that error.

Here is a listing of a few iterations across the 2 to 4 MHz range  I compared crystal cal points from 1942, and then 2024 without modifications. Then I show two examples of varying the coil shield plate by .020" and .040". I also show the cumulative error from the original 1942 book settings. The dial markings ate peculiar to only my BC-221, but all of them will show the same trend I am sure.