[HBR] HBR2K -- Chapter 13 -- The VFO Revisited and Futher Work

[hbr@mailman.qth.net]

Jim comments on my report:

> >  Replacing the 6DZ4 with a 12AT7 with the second 
> >  triode serving as a cathode follower cleaned up the slight roughness 
> >  in SSB signals.   
> 
> Glad to hear it!

The roughness was *very* slight.   Since then I've found that the 
impedance of those filters is actually 500 ohms; I had guessed 1000 
from the FT-101 circuit.   Getting the termination resistors right 
improved quality slightly.

> In situation where the actual VFO freqs must match a predetermined scale or 
> dial, or be linear, the design can be complicated. I use a different method 
> for homebrew projects - the old "make the scale/dial match the VFO". The 
> trick I use is to  get the VFO working just the way I want it (drift and freq 
> coverage), then make up a hand drawn scale out of cardboard using the LM freq 
> meter. 

> Then I use CAD software to make up a nice dial scale, print it out, and use 
> it for the actual dial. 

I have done that on some other projects but I am working here with 
that wonderful FT-101 calibrated-every-1-kc dial -- big payoff for being 
able to track that.   With three adjustments (coil, series and parallel 
trimmer caps) you can get three points exact, then bending plates 
brings it spot on.   The trick is in doing and describing it so that 
someone who has not done it before can duplicate the results.
  
> >  Next is another alignment and then doing the two-tone dynamic 
> >  range measurement.  Also need to check the skirt selectivity -- may 
> >  need a shield across the filters, or ???  
> 
> Good idea, but also check decoupling caps everywhere. Do not be afraid to add 
> more bypasses, add chokes, rewire with shielded wire, etc. Sometimes it takes 
> a combination of capacitors of different values to get a low Z bypass at all 
> important freqs. Do not overlook the heater circuit as a feedback or bypass 
> path. 

Amen, brother!    I'm pretty  encouraged, though, because when 
there are serious problems in these departments there is usually 
enough regeneration to put bumps in the filter passband.   I passed 
the filament circuit to the front end area through a choke and 
feedthrough caps.   I used the same boxed-in area as in the FT-101 
for the front end.   Used lots of bypass caps and lots of decoupling.   

However it takes much less stray coupling to cause degraded skirt 
selectivity than to distort the in-passband response or show up as 
regeneration.   If the output side of a filter just 'sees' the input, that 
can do it.  The filters are built with ground pins between the input and 
output pins and I tried to do the wiring the same way ... but testing is 
the only way to know.   

So far it's looking like a pretty easy radio to build, for 14 tubes.   
There have been none of the kind of problems you're pointing at -- the 
problems have been of the "D'oh!" variety, or pretty simple just-get-
the-darn-design-right issues.   Meaning it should be possible to just 
follow the schematic, chassis layout, and notes and get good 
results.   

> What is really neat about this project is the concept of recycling an old 
> unwanted rig that is not worth restoring into something completely new and 
> unique. 

Yep.   Of three parts rigs, two are clearly ex-CB'er and substantial 
parts of the PA or driver bandswitch area were fried.   Two came with 
no transformer.   And it's pretty cool to be able to say "The receiver 
here is homebrew, OM." 

Walt Hutchens
KJ4KV