[HBR] Another Project -- 1MHBT -- Part 1

[waltah at earthlink.net]

Over the last month I've returned to the 'One month HBR' project to 
clean up loose ends.   This was a project along the lines of the Ted 
Crosby HBR's, using four 85 kcs command IFTs for the 2nd IF, two 
1415 kcs IFTs retuned to 1665 kcs for the first, with a single IF 
tube at each frequency, and rewound command set plug in coils 
through the front panel.  

The thing works surprisingly well.  The dynamic range is not terrific 
but it's in line with the better receivers of the 60's, stability is 
perfectly adequate -- it's tough to do better than 'adequate' with 
plug in coils.  It sounds great.  The calibrator is kind of a mess -- I 
stupidly put a 100kcs calibrator in the same dual tube as the 
85kcs BFO, so there are beat notes everywhere if both are on at 
the same time. But that's not an operational problem:  For rough 
calibration I turn off the BFO and just peak on the calibrator; for 
precision I set to the 100 kcs mark on the dial and trim to zero 
beat.  

One problem I'm going to work on is blocking by very strong 
signals maybe 10 kcs away.  This is a frequent problem with 
double conversion receivers where the selectivity is at the 2nd IF 
because the 1st IF isn't sharp enough to greatly attenuate these 
signals and the RF, 1st mixer, and 1st IF boost them to the point 
where the 2nd mixer bias is upset, reducing its sensitivity.   
However I can reduce the gain of the first three stages 
(compensating in the 85 kcs IF) and possibly change the 2nd 
mixer circuit slightly to minimize the effect.  

Most double conversion designs of this type use no gain at the 1st 
IF (1st mixer directly feeds the 2nd mixer) but I wanted to keep 
stability issues to the minimum by distributing the gain across all 
frequencies.   Definitely a trade-off, probably one I got wrong.   

The plug in coil scheme works fine.  I've had coil sets in and out 
maybe a few hundred times and there's no sign of significant wear --
 all they've done is smooth up considerably.  I do have them well 
lubed.

The pictures of this radio were lost in the termination of one of our 
web sites; I'll get new ones up with the next post on this subject. 

The set works well enough to deserve a matching transmitter and 
I'm starting to think about that.  I'm considering generating the SSB 
signal at 85 kcs, using those IFTs to do the filtering.   Then 
conversion to a tunable IF (500 kcs tuning range), probably just 
above 2 mcs, and a crystal controlled conversion to the channel 
frequency.  

There are obvious issues: (1) whether adequate suppression of the 
opposite sideband can be obtained with a reasonable number of 85 
kcs tuned circuits, and, (2) rejection of the 85 kcs image in the 2nd 
(~2 mcs) IF.  I can check the calculations on these points with 
measurements on the receiver.

(Images in a receiver are only an issue if there's a signal at the 
spurious frequency.  Images in a transmitter are spurious output 
signals.)

Other ways to go would be following the Yaesu scheme of 
generating the SSB signal at 3180 kcs or the common (early) ham 
system of using a 455kcs mechanical filter.   Those approaches 
are of course 'better' but if that is the main concern, then there are 
several fine recent generation ham transceivers one can consider.

I'm looking at using push-pull 22JF6's for the final; they should be 
good for ~50-75 watts PEP out in class AB1 -- and I have a bunch 
of them.   (Series string filament design.)   Push-pull should 
significantly improve distortion at the price of more complex tuning. 
Bandswitching would be a real mess so I'm considering separate 
plug-in PA assemblies for each band and building an antenna tuner 
into the bottom of the case -- the space occupied by the coil 
drawer in the receiver.

The standard way to go in this power and radio size ballpark would 
be parallel tubes -- 6146's or larger sweep tubes -- to get 100-200 
watts PEP out.   However, there are *no* tubes in this range that 
were designed for linear SSB service and it's tough to get distortion 
even close to 30 db down; many such radios are mid-20's when 
correctly adjusted and as a few minutes listening will show, much 
worse when otherwise. 

Parallel tubes must be perfectly matched to achieve reasonable 
distortion levels, otherwise one tube is overloaded and generating 
3/4 of the power with distortion products 10 db down.   Perfect 
matching is darn hard to do with current prices and availability of 
these tubes.  Push pull eases that problem too -- the tubes operate 
semi-separately rather than in tandem.

A minor consideration is that push-pull operation reduces 
circulating currents in the chassis.   With a reasonable conversion 
scheme those won't cause instability but they can spread the 
spectrum of the VFO causing distortion.

(The much larger modern tubes used these days in linear amps 
can keep distortion products close to 40 db down.   (8877 to ~40 
db, 3CX500A7 to high 30's in cathode drive circuits, ~5 db worse in 
grid drive)    I don't think it would have been hard to make good low 
power tubes, but by the time these issues began to be 
appreciated, tubes were fading fast on the low-power transmitter 
scene; there would have been no money in selling tubes with 
annual production in the few thousand units range against 6146's 
and TV sweep tubes.   

(The only tube I'm aware of that was even adapted for this class of 
service is the 8236 which is a 6DQ5 with a carbon plate and hard 
glass envelope rated at 50 watts plate dissipation.   AFAIK it was 
used only in the Hallicrafters SR-500.   The 6DQ5 is a pretty good 
SSB linear amp as sweep tubes go (distortion products down high 
20's db) but looking at Bill Orr's numbers for a sample of common  
sweep tubes (they're in some of the Radio Handbooks), the feeling 
I come away with is that they are better used at power levels well 
below the limits set by the plate dissipation.)

6550's are the closest approach to low power linear SSB tubes but 
darned if I'm going to compete with the Hi-fi guys to buy them.  And 
they are BIG and single ended.   

Push pull will improve distortion but the more complex tuning 
arrangements (even without bandswitching) will take up some 
space, meaning smaller tubes must be used.   Cutting the power 
level in half seems an acceptable trade-off for a set that's not going 
to be a pileup buster anyhow.   

A possible alternative is plug-in tuning units with a fixed (single) 
pair of tubes.   That makes the mechanics more complicated 
because high voltage and high impedence connections must be 
plugged rather than the reverse but it's worth thinking about.

Walt Hutchens
KJ4KV