[HBR] HBR2K -- Chapter 14 -- Large Signal Performance, Part 4

[hbr@mailman.qth.net]

Re my use of a cathode follower to drive the crystal filter, Jim said:

> I've thought of doing that sort of thing to match an xtal filter but being
> the lazy sort I've used LC matching instead. Some sets have used a
> slug-tuned coil and capacitive voltage divider arrangement (see ARRL
> handbook), others a tuned-toroid-with-link setup.

One man's lazy is another's 'too hard for me.'   I tried that first, and 
tied myself in knots, trying to get a match.   With a cathode follower, 
everything can be calculated -- the source impedance is 
1/transconductance in parallel with the cathode resistor so even for a 
triode at a funny plate voltage you're in the ballpark to start.

> This just doesn't sound right. A 6ES8 Pullen making all that noise? 

I agree -- it doesn't sound right.  I'm betting that most of the noise is 
from a combination of the VFO and the unwise inclusion of 4" of 
chassis in the injection circuit.   Even microvolts of noise in the 
chassis would cause a problem and my experience is that it can 
easily be much more than that.   I think the problem is just a design 
error.   

I had a difficult layout problem in that area because the basic 
structure (where's the VFO relative to the front panel, for example) 
was designed around the use of very small devices that generate no 
heat.   The FT-101 mixers are directly adjacent to the VFO, in a 1-
1/2" alley between it and the frame of the front end coil slug tuning 
mechanism.   A tube socket barely fits and controlling VFO heating 
is a problem.   I stuck the 1st mixer there because I felt there was no 
choice -- it needed to be near its grid coil and the 1st oscillator.  The 
2nd mixer was placed a couple of inches behind the VFO because I 
thought I could get away with it.  But I might have foreseen the need 
to deal with the ground loop that my casual approach to mixer 
injection caused.

A possible alternative to grounding the mixer via the VFO is to link 
couple it.   We'll see... A lot of the noise is definitely coming from 
elsewhere and I'm pretty sure the problem can be solved.

>>  (Can't use an untuned buffer at 9 mcs because the stray
>> capacitances are too large to allow a high enough plate >  
>> impedance for the necessary gain unless circuit is tuned.   

> Cathode follower, just outside the VFO, maybe? 

I need some voltage gain.   Using a Pullen circuit in which the mixer 
tube grid is swung zero to cutoff with a tube that's remote cutoff calls 
for pretty much voltage and I'm not pushing the oscillator that hard.   
But the other half of a 12AT7 with a low-Q buffer plate tank works fine 
-- I just need a voltage gain of 3.   The tank (and everything else 
possible) needs to be in the VFO assembly to keep spurs to a 
minimum.

> And the mechanical design could use a little work - build at least
> the front end on a removable plate so the thing could be used as a
> test bed for 7360 vs. 6ES8 Pullen vs. 6JH8 or what have you. 

That needs to be a fairly large removable plate.   A triode mixer vs. a 
beam tube is comparing apples to flat paving stones -- that's why 
there are no good converstions to a beam tube mixer.   Local 
oscillator injection is higher for the beam tube and should be push-
pull; the thing has so much gain that any RF stage should have a 
gain of no more than unity and may well hurt the noise figure.   Beam 
tube output should be push-pull as well but with so much gain you 
really want single conversion if possible and you want to feed the 
mixer right into the 'knothole.'   Basically you're looking at a totally 
different design up to the last IF (may not be a 1st IF in the beam 
version!) and with a whole different set of issues and trade-offs.

> And I don't have the test gear to put numbers on how it actually 
> performs.

Well ... don't be too sure it's impossible.   The URM-25's are actually 
okay, at least within the range up to 'excellent' receiver performance; 
they run $100 and down.   (And they have the basic design needed 
to go all the way -- lots of buffering, plenty of internal isolation, no 
AGC loop.   I have to study them a little more when I get the receiver 
finished.)   Stability leaves a bit to be desired but you just let 'em 
warm up for a couple of hours.  They're a little more trouble to use 
than a first rate lab instrument but in ham work, t'was ever thus.  Add 
a home made hybrid combiner and a home made step attenuator 
(directions for both are in the *Handbook* and you're there.   

> Trouble is, back when I had the time and space I didn't have the 
> parts, and now that I have the parts I don't have the time or space. 
> Anybody interested in undertaking such a project?

I'll probably do it when this one is done, probably as a general 
coverage receiver.   But that's way too far in the future to make any 
plans now!  

Walt Hutchens
KJ4KV