How many times has an Accutron fork hummed?

Take an early Accutron, where the one-transistor oscillator beats at 360Hz., or beats 360x per second, x 60 seconds x 60 minutes x 24 hours, so it beats 31,104,000 times per day? 933,120,000 times per month, or 11,197,440,000 per year, or 559,872,000,000 (that’s 559 billion) in 50 years? Many accutrons are working still being as old as 64 years. How many electrical devices still work after so much use? So, if it is off by 5 minutes per month, that’s 108,000 beats (too many or too few) per month. That’s roughly what, an error of one in one thousand, or 0.001% which is well beyond the 10% tolerance of normal electronic components, beyond the 5% tolerance of very good components, beyond the 1% tolerance of the best 1960’s components and pushing the tolerance of laser-trimmed 1980’s components.

But it should be possible to trim the resistor and capacitor in an Accutron “somewhat” in order to get a 0.001% value change, by adding an additional component of the required small value.

Perhaps someone would know which component (the resistor or the capacitor?) would need to be trimmed in order to change the resonant frequency of the oscillating transistor, faster or slower? I belive that in a coil rewind instance, adding more turns of the wire would also do the trick.