[Premium-Rx] Racal genius

Charles P. Steinmetz charles_steinmetz at lavabit.com
Wed Aug 15 16:48:00 EDT 2012


Tisha wrote:

>Yes, it is a sin but whenever I see a parted out 6790 on eBay I almost
>cannot help myself and if the bandpass filter is the part for sale I will
>spring $30 for it. I have added that filter to many other receivers (even a
>venerable Hammarlund SP-600) to reduce out of band interference. It is a
>great design.

In the 6790 itself, the IF filter switching circuitry leaks pretty 
horribly -- a consequence of feeding all of the filter inputs in 
parallel and switching only their outputs with diodes.  If you have 
the 16 kHz filter installed (most of the ones I've measured are more 
like 20 kHz at 6 dB), it spoils the ultimate out-of-band attenuation 
of the narrower filters from their cutoffs up to 25 kHz or more.  The 
simple solution is to remove the "16 kHz" filter.  Presumably the 
narrower filters all leak too, when you are using an even narrower 
filter, but I have never been bothered by it.  However, my widest 
filter is 6 kHz.  If you have an 8 or 10 or 12 kHz filter installed, 
that might also exhibit the problem.

It has long been known that if you are going to switch only one end 
of your IF filters, it should be the input end, for this very 
reason.  For a radio of this stature, it is (IMO) an unpardonable sin 
not to switch both ends, and even worse to switch only the wrong 
end.  All you would need to modify a 6790 to switch both ends is a 
handful of cheap components.  If I were doing this, I would change 
the existing diodes to something more suitable than the installed 
1N916s, and use them for the input side, as well.

I never bothered to modify the filter switching because removing the 
"16 kHz" filter works for me, and I really don't much like the 
performance of the radio even after this change (to say nothing of 
the hateful ergonomics).  The only reason I've kept mine is because 
it is one of the ultra-rare 2174B models (factory equipped for 
LF/VLF/ELF, using 5 or 6 modules that are unique to this model, as 
well as the normal complement of modules from the 2174/2174A).  Only 
200 2174Bs were made, according to Gary W., and only a few seem to 
have escaped from the original owner (the USAF).

Best regards,

Charles










More information about the Premium-Rx mailing list