[R-390] Collins R 389 PTO

[2002tii]

Roy wrote:

>The R-389 PTO is a different animal.  According to info I have the 
>thing has a frequency range of  .470  to .980  Mc. and operates over 
>51 turns (probably just 50 turns in use.)  That is 470 KC to 980 KC 
>and FIFTY ONE turns.

Correct.  Back in the day, a rather unfortunate choice for civilian 
use.  R-389s are not exactly Tempest qualified, and enough LO leaks 
out that they are very definitely "bad neighbor radios" if other 
folks on your block are trying to listen to AM.  Mine would create a 
strong heterodyne two doors away when my neighbor tuned in his 
political talk shows and blasted them out the window so he could hear 
them in the yard.

>I don't know what resolution you'd need.

If you want to copy SSB with reasonable voice quality, 10 Hz steps 
are pretty much mandatory and 1 Hz is better (with the 389's 
continuously-variable BFO, I suppose you could get away with 100 or 
even 1,000 if you had to, but I'd prefer more precision to get the 
audio centered in the IF filter).  If you don't care about SSB, 100 
Hz is adequate.

If I were to design a replacement, I think I'd be inclined initially 
(that is, until such time as terrifying problems suggested there must 
be a better way) to investigate an [almost] all-digital VFO.  I'd 
load the coefficients for a 6- or 8-bit sine wave approximation into 
firmware, then just clock it out into a D/A converter to 1 Hz 
precision by dividing down a logic-level TCXO clock.  All of the 
quantization noise in a 6- or 8-bit sine wave approximation would be 
far above the fundamental, so following the D/A with a pretty simple 
1.5 MHz low-pass filter would clean it right up.  You'd still need a 
10,000 count/revolution encoder (only 1,000 if you were willing to 
accept 10 Hz resolution, and you can buy those off the shelf -- but 
I'd hold out for 1 Hz tuning, just because).

Best regards,

Don