[time-nuts] Early Christmas - DMTD up and running

[Bob kb8tq]

Hi

When doing DMTD “noise floor” with a common oscillator, you need to use a “line 
stretcher” between the inputs to the DMTD. In the case of a 5 MHz system, you
would need about 30 to 50’ of coax to get the job done. There is no need for infinitely
fine steps, a set of 3 cables at 20 / 10 / 5 feet (or anything near those numbers)
would do fine. This is very much a “use what you have” sort of setup ….

Bob

> On Nov 30, 2020, at 11:34 AM, Skip Withrow <skip.withrow at gmail.com> wrote:
> 
> Hello Time-Nuts,
> Some time back Bert Kehren supplied me with a DMTD board that he and
> Jeurig (in Germany) did.  Bert was kind enough to include the
> transformers and mixers with the board that he sent.  This board is
> quite similar to the Bill Wriley (Hamilton Technical Services) system,
> and is just the dual mixer portion.  I have finally managed to acquire
> all the rest of the parts, a 4.999990MHz oscillator, power supplies,
> and packaged everything is a 2U chassis (first attached picture).
> 
> The system drives my HP 5370B (second attachment), which is obviously
> overkill.  A TICC or modest counter with TI capability would be
> sufficient. First tests are quite encouraging.  Right now I only have
> 5MHz capability, and will add a doubler to the other port of the
> offset oscillator to give me 10MHz ability as well.  Even though
> running a 5MHz input with a 10Hz offset gives you a 500,000 increase
> in resolution, noise in the system will degrade performance.
> However, it does look like my noise floor has been lowered by over two
> orders of magnitude.  The 5370B can get you to about 1x10E-10 at 1sec,
> with the DMTD it looks like it is about 7x10E-13 at 1 sec.  I now also
> get data between .1 and 1 second with the DMTD.  The third attachment
> is a timeLab plot of several different noise floor runs (and a couple
> of other random tests).
> 
> This will bring the true performance of many systems out of the noise
> for lower taus, and is necessary for high-performance oscillators.
> I'm looking forward to having some fun.
> 
> Need to learn a lot more though.
> 1. I have a three-way comparison of three oscillators done on a
> borrowed TSC-5110.  Need to make sure that the DMTD gives me similar
> results.
> 2. Corby Dawson uses a very good crystal oscillator for short taus and
> a maser for longer taus as the reference.  Need to figure out my best
> reference for specific measurements.
> 3. I do have an HP 8660B synthesized generator.  I need to investigate
> whether it may be a good offset oscillator candidate.  This would
> allow different sample rates and measurement of oscillators at odd
> frequencies (which I do have).
> 
> Thanks for the bandwidth to show-and-tell.
> 
> Regards,
> Skip Withrow
> <dmtd-top-3.jpg><dmtd-system.jpg><dmtd-plot.jpg>_______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.