[HBR] HBR2K -- Chapter 14 -- Large Signal Performance, Part 1

[hbr@mailman.qth.net]

In a message dated 2/23/03 8:47:48 PM Eastern Standard Time, 
waltah@earthlink.net writes:

> Jim quotes me
>  
>  > >  The 1st mixer gain is low partly because of the circuit but also 
>  > >  because of the low injection voltage from the 1st oscillator.   And
>  > >  *that's* low because with the small crystals of the FT-101 my simple
>  > >  Pierce oscillator wouldn't deliver more without drift.   
>   
>   and adds:
>  
>  > Two words: Buffer stage.
>  
>  Maybe.   But it's not a power issue -- 1/2 12AT7 will deliver plenty of 
>  power.   The first thing I'm trying is a super-Q oscillator coil and after 
>  that, a transformer.   Only if there isn't another way will I get to a 
>  buffer stage.   

OK! My point was that I prefer to have the crystals run easy and build up the 
signal with a buffer. Tubes are easier to find than xtals.
>  
>  The overall goal is a bit different than some receiver designs you 
>  mention.   I'm interested in an outstanding all purpose HF ham 
>  receiver -- the one designers might have produced in 1970 had there 
>  been an effort to do it -- rather than a specialized machine with a lot 
>  of knobs. 

Trouble is, any design involves compromises. The FT-101 conversion schme 
means three stages before the selective knothole of the filter. Two of those 
stages are mixers. 
>  
>  > The Type 7 is single conversion, using a 6EH7 RF amp and 7360 mixer.
>  > After the mixer is an 8 pole 500 Hz xtal filter which I think was
>  > originally meant for a RACAL receiver. 
>  
>  Single conversion is definitely the way to go but it trades internally 
>  generated spurious signals for the extra mixers taken out of the 
>  signal path.  

All depends on the conversion scheme used. The Type 7 scheme is not optimum - 
it evolved around the available filters.

>  And I felt that at my level of experience there was a 
>  huge advantage to staying with an existing conversion scheme and 
>  bunch of recycle-able hardware.  You can only answer so many 
>  questions wrong at the same time and make any progress.   Call 
>  that Walt's (rusty) razor.

Works for me!
>  
>  The 6EH7-7360 sounds like a good combination, although I wonder 
>  at the need for that much gain in front of a 7360.  

The stage normally runs at much less than max gain. Its purposes are:

- to make up the losses in the front end tuned circuits
- to isolate the antenna from the mixer
- to provide a gain control opportunity

>  And beam tubes 
>  have so many difficult design challenges, from the DC voltages to the 
>  very high LO drive requirements (and should be push-pull, at that); all 
>  that means a lot of heat which complicates the stability challenge.

I used th single-ended version shown in the ARRL Handbooks. Drive is only a 
few volts. Maybe a Pullen would do the job better. We'll find out in the 
Southgate Type 8.
>  
>  All of which is another way of saying that one day I'd like to try a 
>  beam tube receiver.   Just not there yet!
> 
Some time back I acquired a pair of 2215 kHz filters as used in the "Miser's 
Dream" receiver. Have been thinking of building a version of same for some 
time, perhaps with a double-tned input circuit rather than an RF Q 
multiplier, and another IF stage for more gain and control. Perhaps a 
premixer type LO chain. 
 
>  > The basing of the 7044 is such that a single shield across the
>  > socket will ground both grids and shield input from output. Most dual
>  > triode pinouts won't allow that. The 2C51 and 5687 have similar
>  > bases. The RF amp has only enough gain to overcome the loss of the 4
>  > tuned circuits - it's not there to boost the sensitivity much. 
>  
>  The trade off is peak noise performance for dynamic range, is it not?  

Not really. The conventional reason for rx RF amps was to make up for noisy 
mixers like the 6L7. But a 7360 is so quiet that the only real reason for an 
RF amp is to make up for tuned circuit losses (the W1KLK circuit has four 
tuned circuits before the mixer!) and perhaps control the gain. 

>  I'm betting that a cascode 6ES8 comes close on the noise (6.5 db in 
>  a TV tuner) and kicks anatomy dynamically.  1% crossmod at 0.5 
>  *volts* of signal as you near cutoff -- and not bad at higher gains.   
>  And since it's impossible for a conventional mixer to equal the large-
>  signal performance of an amplifier, a key function of an RF stage is 
>  to have a place to do AGC ahead of the mixer.   

All true. Also why high performance designs got away from conventional mixers!
>  
>  I'm thinking in the direction of a 12AU7 1st mixer.  That would pretty 
>  well deal with the large signal issues but would require a whole 
>  bunch of LO drive so we'll see.   Added advantage -- you can put 
>  some AGC on that tube.

Interesting!  

>  Jeff said:
>  
>  > ...Yaesu Mark V, at least mine. It hears  SWB signals on 40m in the cw
>  > band  that do not tune and are wide and sound muddy, They are not
>  > there on the Tentec Omni lV+. 
>  
>  While that could be a design flaw, there are also radio and other 
>  hardware failures that could produce the problem.  I'm not 
>  experienced enough with modern gear to have an opinion.   Have you 
>  compared notes with other owners of the same rig?
>  
>  But it definitely could be a design flaw.  An interesting comment that 
>  I heard recently: "The basic RF performance of modern ham gear is 
>  going down steadily, even as computerization gives us more bells 
>  and whistles." 
>  
>  In the 60's the ham marketplace accepted SSB sets that drifted a 
>  few kcs the first hour and a good fraction of that, forever.   Signal 
>  quality and reception was variable, at best.   By the end of the 
>  decade Yaesu gave us better performing gear but saved money on 
>  quality control.   (Anyone out there running an FR-100B/FL-100B 
>  combo?)   The radios with strong all-round performance cost much 
>  more and never sold comparable numbers -- Collins is the leading 
>  example.   Computerization vs. sound basic RF performance is 
>  another such issue.  

Then along comes a rig like the Elecraft K2. Which goes all the way back to 
the single conversion ideal of 1957's "What's Wrong With Our Present 
Receivers". 40 years and back to the fundamentals...
>  
>  Another quote, possibly more controversial:  "The FT-1000D is 
>  today's Collins and has a similar future."
>  
Maybe. One of the things that made S-line popular was that they could be 
worked on by a knowledgeabel ham.

>  Let's face it:  the average ham set today is a transceiver used an 
>  hour a day for talking to buddies in a 1000 mile radius.    Every set in 
>  the market, from an eBay FT-101E upwards, will do that job -- at 
>  least when it isn't out for a few weeks for repair.   

Yep. The top end market is driven by the contesters and the DXers, not the 
ragchewers. 
>  
>  The question of what could have been done in 1970 remains very 
>  interesting.   Thanks for the discussion and suggestions!

Thanks for the great ideas!

Here are some more:

Back in the mid-90s or so, there was a series in "ER" about the R-390A, 
including some front-end mods. Perhaps the circuits and tube lineups used in 
that rx could be adapted to the '101.

The Squires Sanders SS-1R was probably the closest thing to what you seek. 
It's my understanding that the SS-1R had excellent dynamic range and low 
noise. The manual is on BAMA, btw. The main reason it did not achieve 
popularity was the price and the lack of a matching transmitter. Perhaps some 
of its circuitry could be adapted.

The RSGB Handbooks of the mid-60s have some good ideas, too. More info on the 
Pullen mixer, although not identified as such. 

73 es GL de Jim, N2EY