[time-nuts] Sub Pico Second Phase logger

[Joseph M Gwinn]

Bruce,


time-nuts-bounces at febo.com wrote on 12/16/2008 10:21:55 PM:

> Joe
> 
> Joseph M Gwinn wrote:
> > Bruce,
> >
[snip]
> > 
Bruce wrote:
> >> 
> >> A CB stage probably isn't optimum for the mixer preamp so that lower
> >> value caps can be used provided that they effectively short the
> >> amplifier input resistor Johnson noise at the frequencies ofinterest.
> >> 
> >
> > But is a CB stage adequate?  Elimination of hum pickup is worth a lot.
> >
> >
> > 
> [Bruce] Text should have been:
> 
> A CB stage probably isn't optimum for the mixer preamp so that 
> a low noise preamp with a higher input impedance can be 
> employed allowing lower
> value coupling caps to be used provided that they effectively 
> short the amplifier input resistor Johnson noise at the 
> frequencies of interest.

OK.  The form of input amplifier is one of the tradeoffs one must make.



> >> It may be simpler just to select a mixer for which the IF ground can 
be
> >> isolated from the RF and LO grounds.
> >> However a preamp with a transformer output may be useful if one uses 
a
> >> mixer where all the grounds are connected together by the package.
> >> 
> >
> > It has to be far easier to select the right mixer than to deal with a 
1 Hz 
> > transformer.  And cheaper.
> >
> > 
> 
> I've been advocating this for some time, however one can then no longer
> just buy an off the shelf mixer complete with SMA connectors, one has to
> design and assemble a suitable PCB.

Actually, for quantity one, I don't bother with PCBs.  I use 1/16 inch 
thick glass-epoxy Vectorboard with a 0.1" hole pitch, and thread bare wire 
through the holes.  Given that these are effectively prototype boards, the 
ease of wholesale change is very useful.  For transmission lines, I would 
just run miniature coax or homebrew twisted pair from point to point on 
the board, although it has not been neecessary yet.  If microphonics is an 
issue, the board can be conformal coated to glue the wires in place. 

The traditional coat, wax, allows easy alteration and repair.  In the 
1960s, a friend was building things using RTL (Resistor Transistor Logic) 
ICs and solderable magnet wire between the leads.  What a rats' nest that 
was.   The problem was how to package this so it wouldn't fail when 
bumped.  The solution was to build the circuit on a sheet of vectorboard 
in a 12" by 12" rectangular baking pan, and then fill the pan with hot 
wax.  Whenever something had to be changed, melt the wax, pour it off, 
make the change, pour the wax back into the pan, allow to cool.

Modern ICs connected with point-to-point wires and potted in wax - it's an 
odd mix of technology ages.  Like implementing stone-age tools with modern 
ceramics.

If I were building receivers, I suppose I would be forced to use 
surface-mount comoponents and PCBs.


> Obtaining suitable mixers for 5MHz and 10MHz input frequencies or even 
100MHz is easy.
> However for the higher microwave frequencies most mixers come complete
> with connectors attached and share a common ground.

True.  However, I don't think we will be going from 1 GHz to 1 Hz in a 
single step, and the last mixer can have separate grounds.


> The noise problem with audio balanced drive chips can easily be overcome
> with a discrete implementation.
> That is discrete resistors and IC opamps.

Yes, even on vectorboard.  I do it all the time.

Eventually, the supply of through-hole components will dry up, but it 
hasn't happened yet, and some components from the 1970s are still 
available.  Even if the original manufacturer is long dead.


> > 
> >> I have read similar papers from that era on radar signal processing.
> >> They either used a Hilbert transform or a pair of digital filters 
whose
> >> outputs were in phase quadrature.
> >> The quadrature accuracy for a given bandwidth depends on the number
> >> of taps.
> >> The beat frequencies (in a dual mixer system) won't match exactly and
> >> some correction for the resultant phase shift errors will need 
> >> to be made.
> >> This may be less of a problem when the 2 beat frequency signals are
> >> identical in frequency and just differ in phase.
> >> 
> >
> > So long as we know the exact frequency, even if it isn't the exact 
desired 
> > frequency, all may be well.
> >
> > Joe
> >
> >
> > 
> I'm reminded of some phase recovery algorithms used in phase shift
> interferometry that largely negate the effect of small fixed 
> phase errors.

Yes.  A detailed math analysis of the test setup will be needed.

 
> If we can devise a suitable test setup then one could just log the
> samples to a file for whatever sound card one has and make the data
> available to others for analysis.

Yes.


> This allows a wide variety of sound cards to be evaluated without one
> person having to test them all.

And evaluation of the same test data by multiple people using different 
tools also allows us to distinguish test artifacts from processing 
artifacts.


Joe