[hpsdr] An high dynamic QSD front-end

[Giancarlo Moda]

Hi Bob,

--- Robert McGwier <rwmcgwier at gmail.com> wrote:

> 
> Hello Gian:
> 
> Thank you for your many contributions in the mixer
> area, to the N2PK VNA 
> group (both of mine use your circuit boards), and
> more.  

I do appreciate your comments. Thanks

 Here is what I 
> know from my own experience,  measurements done by
> Phil Covington,  Phil 
> Harman, and others.  From several QSD's I have
> measured and have seen 
> measured the rise in noise figure from (say) 14 to
> 54 MHz.   

You are right on these. With Phil I had some e-mail
exchanges as he was looking for any experiences I had
with I-Q mixer and to find a possible solution for an
I-Q mixer that would not increase its NF going up in
frequency. From these exchanges Phil did a good
exercise and the SoftRock40 V7 with the I7SWX - VK6APH
I-Q double balanced mixer was born, using the FST3253
in place of the FST3125, operating on the 28MHz band
and with the simple RC LO quadrature generator at
receiver frequency.
It looks like it did work as many kits were sold by
our good frien Tony, KB9YIG.
At this point it would be interesting to have comments
from Phil if this solution had important or no
benifits, particularly on Noise, versus the classic
I-Q QSD (SR V6 and previous) using FST3253.

I see that 
> Marco has done the same with his very nice work.  
> In each case,  the 
> results are similar, if not identical to those I
> have measured or others 
> have measured.   In those cases where I have seen
> these kinds of mixers 
> done without the FST switches, but with different
> parts,  there is <<no 
> such rise in noise figure>>. 

It is interesting to know which kind of components are
used with no FST switched mixer, maybe Gilbert cells
and the like? Maybe these different solutions may
present IMD problems. I believe we have no perfect
component for mixer or QSD. We may have some benefits
with new switches particularly the ones that are fast
and well balanced with Ton-Toff, and break before
make. 
It is important we do present discussions ... there is
always something to learn sharing even the wrong
ideas.

 Marco's work has one
> thing different from 
> all of the other ISD implementations I have seen. 
> It uses the FST 
> switches you discovered for us.  ;-).   Marco's
> circuit has the rise in 
> NF which is NOT present in those ISD's I have seen
> measured which do NOT 
> have the FST's.  I am certain Marco is not sloppy
> and his work is 
> excellent and I am taking his graphs as the absolute
> truth.  The logical 
> implication from this is immediate.  The FST
> switches are the cause of 
> the rise.   I would like to minimize the amount of
> gain needed to make 
> the receiver usable at 10 and 6 meters.

Marco has knowledge and also intelligent experiences
with measurements and design. 
The idea of using the FST3125 versus the FST3253 is
very simple. With the FST3253 you have no possibility
of adjusting the phase on each of the 4 switches as
these are driven by internal decoding.
The FST3125 solution may or may not obviate quadrature
and phasing problems and so give or give benefits.
Simulation is an important techique but we have to
come down to earth and see if the circuit really does
what it has been simulated ... we are using components
that are not perfect and we too are in the same range
...hi

I do feel, although I am one of his friends, that
Marco has done real measurement and analysis that I
have not seen reported by anyone fiddling with SDR.

In the amateur field we do certainly appreciate the
best design and the best solution but we have also to
find KISS projects in the advanced telecommunication
field that will help to keep alive the "homebrewing"
part of the hobby and be reproduceable by as many hams
as possible and with the least of test equipment.
An all digital solution may help on this but looking
at the wonderfull HPSDR project it is not so much KISS
and I do not know how many Hams will be able to
implement it. For this we will need the availability
of expensive kits ... the SDR Roll Royce versus the
little Fiat Topolino of the SoftRock40 ... hi

> 
> I acknowledge your work and that of Marco.  I am
> giving my opinions on 
> the best way to proceed, fully aware of all of this
> work.  I read this 
> group every day and am grateful for all of the
> wonderful contributions 
> and sharing.

Opinions are very important ... are the basis of
knowledge, learning and friendship.

Thanks for your phrase:
 > > But it is time to leave them (FST switches)
behind FOREVER.> >

I am sure many members of this and other SDR groups
have started thinking more...

BTW ... I support Lyle's heretical thought ending
with:
"We need something in front of the QSD to suppress LO
radiation, and the "front end" filters and what not
are going to be required as well.  An H-mode mixer
will preserve the dynamic range characteristics we
want."

I experienced this whem I applied H-Mode Mixers and
mods to an Yaesu FT1000MP ... I had to retrofit the
1st mixer replacement ... but very strong performances
obtained with the 2nd mixer mod (H-Mode)...

> 73's
> Bob
> 

73 to you too Bob

Gian
I7SWX



 
____________________________________________________________________________________
Cheap talk?
Check out Yahoo! Messenger's low PC-to-Phone call rates.
http://voice.yahoo.com

 1168960056.0