[HBR] G2DAF receiver

[hbr@mailman.qth.net]

In a message dated 10/3/03 9:31:30 PM Eastern Daylight Time, 
waltah@earthlink.net writes:


> It appears to me that the early G2DAF receivers are entirely 
> conventional double-conversion-with-tunable-1st-IF designs.  

I have a 4th edition RSGB Handbook. The rx described uses an ECC84 cascode RF 
stage (!) and a 6BE6 first mixer (ugh). Oddly enough, the second mixer is a 
6BA6!

The filter is made from FT-241s and there are two of them - a  four-pole 
after the second mixer and a two-pole between the  first and second IF stages. 

Later in the chapter, there's a section on using a 6BZ6 RF amp and ECC85 
Pullen first mixer. A different heterodyne oscillator is also shown. Also shown is 
a suggested replacement of the homebrew filter with the Kokusai mechanical 
filter, and an ECC85 Pullen second mixer. (They don't identify the mixer circuit 
by name, but it's the Pullen all right, even down to the lowering of the gm 
by lowering the plate - er anode - voltage through a divider).

 They 
> 
> should beat the W6TC designs on stability and since they use a 
> mechanical filter, perhaps on skirt selectivity; otherwise they're 
> unremarkable.   


I think they are VERY remarkable! Unlike the typical American practice of 
using a bandpass circuit for the tunable IF, a tracked tunable high Q filter is 
used, so the second mixer doesn't have the whole band pounding on its grid. 
('DAF had to do that because the second IF is so low that with a bandpass filter, 
secondary images would have been a problem). Even the original design uses 
bandswitching, selectable sideband from an xtal BFO, built in Q multiplier, 
Eddystone dial and 20 tubes. 


Image and spurious rejection are improved because the IF is at least 5 MHz.

Back in the adverts, an outfit called "E.J. Philpott's Metalworks, Ltd." 
offers to make up the metal work for you. 

> 
> What gives the Mk II version of these receivers such remarkable 
> specs is the push-pull design of the front end.   By cancelling out 
> second harmonic distortion, this makes the strongest third order 
> products (which are a second order product combining with a 
> fundamental) very much weaker.   


Agreed. Plus the use of high performance triodes allows low noise.

> 
> (For example, consider a weak desired signal at 3840 and very 
> strong unwanted signals at 3800 and 3820.   Distortion in the RF 
> stage produces the second harmonic of 3820 = 7640 and also gives 
> the difference between this and 3800:  7640-3800=3840 which 
> interferes with your desired signal.   "Third order intermodulation 
> distortion dynamic range" (a mouthful if ever ...) is "How strong do 
> the two unwanted signals have to be to produce a detectable 
> interfering signal?"   
> 
> (But push-pull amplifiers tend to cancel out second harmonic 
> distortion products -- that's why hi fi amps are push-pull designs.   
> Less second harmonic means less to combine with the other signal's 
> fundamental and a correspondingly higher 3rd order IMD dynamic 
> range.)
> 

Bingo. 

Also why I use a pushpull 12BH7 audio output stage.


> One would not have to build the entire G2DAF receiver or even 
> include the bandswitching feature in order to get that performance.   
> Plug in coils for a push-pull front end would be a little more trouble 
> than for the conventional design, because they need to be center 
> tapped and tapping down for bandspread (as in the HBR-series) 
> requires two tapes/coil.   However the narrow bands tuned by any 
> ham-band-only receiver mean that balance will be determined by the 
> shunt caps.   You would need a five section tuning cap of the regular 
> sort, using four sections in two pairs to tune the push-pull RF stage 
> and the remaining section to tune the 1st oscillator.   
> 

The parts are the real killer. Five section tuning cap? 



> Circuitwise there's nothing much to either the RF stage or the mixer.  


Note that the triode RF stage is not neutralised but the first mixer is!

> 
> In other words, the critical idea can be grafted on to the regular HBR 
> design.   It's just a different way to do the RF stage and mixer, not a 
> wholly different receiver design.   
> 
> I speculate that you could make it even easier:  Omit the RF stage 
> entirely -- in the G2DAF design it's just a controllable attenuator 
> anyhow.   Drive the grids of the mixer in push-pull from the antenna.   
> Inject the VFO on the cathode and take IF output push-pull from the 
> plates.   The mixer's where most all the distortion comes from 
> anyhow.   Three gang tuning cap that way; use a toroid for the 
> antenna coil and keep the antenna coupling very loose.   


The RF stage serves an important purpose: it makes up the losses in the RF 
coils, which can be considerable even with high Q coils. The stage runs at unity 
gain but the tube sure doesn't! And you need two tuned circuits to get decent 
image rejection. That's even more important with the lower IF of the HBRs. 

The Mk2 also includes filter switching. Although the original intent was to 
use separate upper and lower sideband filters, the feature can be used for 
selecting SSB and CW filters. 

The lack of attention to the CW op in Handbook and many homebrew receivers 
has been a sore point with me for a long time. The most modern Handbook tube rx 
with CW selectivity is the DCS-500, circa 1961 (and it was a QST article 
first). It used width coils and selected coupling caps to obtain CW, SSB and AM 
bandwidths. 

After that, the Handbook rx's mostly assumed we all wanted SSB selectivity. 
Only the "Miser's Dream" used a real multipole CW filter, and it wasn't even 
given a complete description in QST. The "Junior Miser's Dream" and the "2X4" 
receivers could be configured to do 2 pole CW filtering - sort of. The W2LYH 23 
tube masterpiece used separate IF strips for CW and 'phone, but it never made 
the Handbook. 

The G2DAF article has me rethinking the Southgate Type 8 to something like 
this:

Pushpull RF lowgain stage (6ES8?) 

Pushpull first mixer (12AT7 is pretty close to ECC85) fed by VFO to give 
fixed second IF. (5.1-5.4 MHz VFO, 8.9 MHz IF on 80 and 20, 12.4 MHz IF on 40, 
selected by small dpdt sealed-can relays).

Pushpull second mixer (another 12AT7) with triode xtal oscillator to get to 
1.4 MHz second IF freq.

Front end is tuned by a pair of dual capacitors ganged. VFO is unit 
construction. First IF doesn't have to tune at all. 

I wonder how much I could get for my 7360s? ;-)

73 de Jim, N2EY


--- StripMime Report -- processed MIME parts ---
multipart/alternative
  text/plain (text body -- kept)
  text/html
The reason this message is shown is because the post was in HTML
or had an attachment.  Attachments are not allowed.  To learn how
to post in Plain-Text go to: http://www.expita.com/nomime.html  ---