[Hammarlund] Old Hammarlunds

[Kenneth G. Gordon]

On 12 Apr 2011 at 21:27, Richard Knoppow wrote:

>      When I got the RX it has been modified in accordance 
> with an article in CQ magazine of some years earlier.

Was that possibly the one entitled something like, "Souping the Super Pro" ? 
I never saw that article, never having access to "CQ" magazine in my 
younger days.

> The 
> mod used 6SL7s as common cathode pairs and both RF stages 
> and the mixer stage were so modified.

Gee...sounds like a lot of what I would call unnecessary work.

> The gain was far down 
> from what it should have been and I suspect the loading on 
> the tuned circuits was increased, not sure of that. I 
> restored it to original and then made adaptor sockets for 
> the RF and mixer stages so that I could experiment. I 
> modified the LO to an ECO design copied from a General Radio 
> frequency meter. Very stable but I think now the output was 
> much lower than the original.

That makes sense. I think you're correct there.

> Of course, being in my teens I 
> had no instruments. I tried several variations of RF stages 
> and mixers including cascode types for the first RF with 
> manual gain on it to prevent overload. I wound up with 
> 6BA6's for the RF and a 6BE6 for the mixer, very 
> conventional but I found that the dynamic range and overload 
> resistance was better than the more exotic circuits.

OK. Sounds good.

> I 
> eventually restored it to original including the oscillator. 
> I then discovered how much the oscillator varied with B+ 
> changed caused by the RF gain control and AVC. I had used a 
> VR tube for the ECO and went back to it. The arrangement of 
> the B+ to the tube must be changed a bit to maintain its 
> output so that the conversion gain of the mixer stays up.

Yes, but that is fairly easy to deal with.

>     While the 6BA6 tubes undoubtedly have lower noise than 
> the original 6K7's I think there is a problem with loading 
> due to the length of leads etc.

I have heard of that problem in other receivers: specifically, a mod to the 
RME-69. The author had to use a parasitic choke in the grid (as I remember 
it) among other things.

> The ECO definitely affected 
> the dial calibration and probably RF tracking, which is 
> normally very good on these RX.

Hmm....that is somewhat strange: I suppose something like Miller effect had 
something to do with it. That could be compensated for though, it seems to 
me. I really like ECOs...

> So, I think the only mod I 
> would recommend is adding the VR tube to the LO. Ideally, 
> the screens of the mixer should also be regulated  but that 
> gets more complicated.

I'll have something to say about that below. I have a couple of ideas that may 
have merit.

>     It should not be difficult to add a product detector

Actually, it is very easily done. I did a partial, but effective, job on my original 
BC-779. The result was dramatic, to say the least! As I have mentioned on 
the Hammarlund list, I removed the 6N7 noise limiter and built a triode 
product detector, copied from the Heathkit SB-101, in its place.

Thinking back on it a few years later, I realized that there must have been 
something wrong with the BFO coupling to the detector, since when I tried to 
use the receiver for CW when I first got it, the BFO injection was so low that I 
had to run the RF gain WAAAAAY down in order to not overload the detector 
stage, and the receiver was extremely noisy with the BFO on.

In any case, I wanted to use a product detector, so at the time, that gave me 
a good excuse.

After I installed the product detector and then fired up the receiver, I was 
totally blown away by 1) how utterly quiet the receiver was, and 2) how utterly 
and amazingly sensitive it was! It quickly became my very most favorite 
receiver. I used it constantly for CW and RTTY with an ST-6 and model 15 
and 19 TTY gear.

The increase in apparent sensitivity, and reduction in receiver noise was 
absolutely amazing to me. I would have never thought that a better detector 
would have made such a huge difference. 

BTW, the RTTY transmitter was a Heathkit VF-1 I had rebuilt and had added 
a "shift-pot" FSK circuit to. That drove a modified and upgraded DX-35, 
which then drove a rebuilt and modified BC-610 to a kw input.

I used the combo to handle traffic for both NTS and MARS for a number of 
years. I installed a 304TL in the BC-610.

> and 
> slow AVC which would make the RX a very good SSB receiver.

Yes. And modifying the AGC for slower release shouldn't be difficult either. 
The SP-200 AVC circuit is one of the best and most effective I have seen. 
Perhaps carefully examining the "Hatch" AGC circuit for the R-390A from ER 
Magazine #208 would give us some excellent ideas. That AGC circuit is the 
very best of all those I have been studying for the past several years. The 
funny thing is, it looks very much like the one used in the SP-200.

> I 
> used it for SSB using manual gain control and it was quite 
> satisfactory.
>     One feature of the Super-Pro is that the BFO is 
> completely isolated from the AVC allowing rather high 
> injection levels without desensitizing the RX.

Yes, which is another factor that makes adding a product detector such an 
easy and effective modification.

>     I suspect that if it had not been for the war time 
> contracts Hammarlund would have refined the design. As it is 
> the SP-400 has some decided improvements but I think by that 
> time the SP-600 was already in the works so not much effort 
> went into updating the older products.

That also makes a lot of sense.

>     I think Collins rather surprized everyone else with 
> their 75A and 51J receivers, which were very innovative at 
> the time and essentially obsoleted conventional design. I 
> don't think National, Hammarlund, or Hallicrafters ever 
> really recovered from that and competition to Collins 
> eventually came from smaller companies like Drake and Racal.

Agreed.
 
Now, here are my thoughts on the front-ends of the SP-200 receivers: first of 
all, those have two RF amp stages. The gains are adjusted so that high "Q" 
is maintained, and signal levels into the mixer are, or seem to be, about 
optimum.

Although I really thoroughly dislike noisy mixer stages, pentagrids being one 
of my pet hates, after my experiences with the added product detector in my 
BC-779, I really cannot see any advantage to "adjusting" the mixer to a much 
lower-noise version. Or even to changing the RF amp tubes to quieter ones, 
for that matter.

I think the gain from the 2 RF stages adequately forces the mixer noise to 
become a non-issue...at least that sure seemed to be the case with my BC-
779 after the PD mod.

I think more experimentation can be indulged in here.

However, I am also a fan of the so-called "Pullen" mixer. This mixer (which, 
by the way requires FOUR, not TWO triodes) exhibits unusually low noise-
output levels, high conversion gain, coupled with low required RF and HFO 
input levels. It has been described as an honest to goodness "linear" mixer, if 
such a beast really exists...which I doubt.

So, I propose to experiment with this mixer type. I'll use a pair of submini 
dual-triodes, the 7963, which have a transconductance of 13000 umhos 
each, and build them onto an octal-socket based adapter, somewhat like you 
did with your experiments.

Lastly, as far as the product detector mod is concerned, instead of the 6N7, I 
should have used a 6BL7, which would have enabled me to keep the noise 
limiter, yet add the product detector. The 6BL7 has two completely 
independent triodes in the envelope, whereas the 6N7 is designed to be a 
push-pull class B audio amp tube and has one common cathode for the two 
triodes.

Well, I had better shut up and go to bed now. Thanks for all the very very 
informative discussion, Richard. I really appreciate it.

vy 73 for now,

Ken Gordon W7EKB