[hpsdr] An high dynamic QSD front-end

[Robert McGwier]

Congratulations.  Very nice analysis,  very straightforward,  not 
requiring the machinery of Laplace transforms to understand the reasons 
which the "intuition impoverished mathematician" had to do to gain 
intuition.   If you study Leif Åsbrink's WSE hardware you will see the 
great lengths he went to do his "transient fix".   He was working very 
hard to supply the current from a suboptimal current source to overcome 
the transients problems with the four phase  capacitor 
pseudo-integrator.  He had a heuristic understanding of the need some 
time ago but did not go fully virtual ground, current mode with the 
necessary integrator to solve this in the utterly simple ISD Ahti has 
shown us.

Thank you both for a serious elevation of this discussion.  It really is 
time to build hardware.  WARNING:  We definitely want to steer away from 
the FST switches.  FROM MEASUREMENT,  coupled with DDS issues,  they are 
the source of the increasing noise figure.  Marco's plots, as well as 
Phil C's and my measurements of various QSD's using them show this 
inescapable fact.  I think various pieces of analysis have incorrectly 
attributed the measurements but the switches are a major contributor.  
They solved an early problem and we can thank Gerald for finding this 
part and his very gracious publication in QEX and with the SDR-1000.   
But it is time to leave them behind FOREVER.

73's
Bob
N4HY




Ahti Aintila wrote:
> ***** High Performance Software Defined Radio Discussion List *****
>
> Hi Marco and all,
>
> After some studies and simulations I have a pleasure to congratulate
> Marco for the excellent QSD mixer configuration. However, as Marco
> refers to my ISD experiments with OPA1632 and also mentions 90 degree
> sampling time better than the 180 degree sampling, I want to add some
> comment.
>
> Marco is completely right (as well as Dan Tayloe) in the case of
> "passive" integration, that is used in the original "Tayloe detector"
> and many other configurations after that.
>
> In the case of my "active" integrator version (ISD), where the
> integrating capacitor is in the feedback loop of an operational
> amplifier, the opposite is true. That is because the current to be
> integrated flows against to the virtual ground potential (and and also
> the current flow time is longer), hence the current is not influenced
> by the accumulated voltage level of the capacitor. In the Tayloe mixer
> the current is related to the difference between the signal voltage
> and the accumulated capacitor voltage.
>
> I made a couple of simple simulations using my ancient Pspice  Student
> version. Please, note that the simulation results may be compared only
> passive against passive and active against active like lemons to
> lemons and oranges to oranges. I did the simulation using only one
> channel (I or Q, but not both at the same time) The clock frequency
> used was 10 MHz and the signal 10.01 MHz. Two amplifiers were used for
> ISD to get the differential outputs, that is the reason for two
> sine-wave signals. The QSD was run without amplifiers, so only one
> output trace. See the pictures:
> <http://kotisivu.dnainternet.net/ahti/sdr-1000/QSD_ISD_180_90.zip> .
>
> By the way, my earlier real measurements in last August gave same kind
> of results, but I did not record them because the 180 degree active
> circuit looked as an obvious choice for me.
>
> 73 and Prosperous Integration for the New Year,
> Ahti OH2RZ
>
>   


-- 
AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL,
TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair
"If you board the wrong train, it is no use running along the
corridor in the other direction. " - Dietrich Bonhoeffer


 1168814226.0