[time-nuts] Re: Using a nanoVNA as DMTD, simulation only

[Erik Kaashoek]

Bob,
The IF frequencies for downmixing to DC are inside the dsp part so they are
guaranteed to be correct as demonstrated in the simulation.
The risk is in generating the two LO' s for the two mixers as these need to
be coherent otherwise the noise from the LO's  won't cancel out.
Thanks for pointing this out
At least I have the math now correct to have some freedom in choosing an IF
frequency.
Erik





On Fri, Sep 23, 2022, 20:34 Bob kb8tq <kb8tq at n1k.org> wrote:

> Hi
>
> With different IF frequencies ( at least if I understand things right ….),
> I
> think you risk the frequency difference between the two IF’s getting into
> the result. Depending on how high the IF’s are that may be a big deal or
> maybe not so much of a deal …..
>
> Bob
>
> On Sep 23, 2022, at 12:11 PM, Erik Kaashoek <erik at kaashoek.com> wrote:
>
> Bob,
> I extended the simulation to fully include both channels.
> Using some careful math it was possible to use two different IF
> frequencies for the downconversion to DC. This implies that any leakage
> after the two mixers can be reduced by using different IF frequencies
> because he phase calculations for each measurement sample acts as a one bin
> FFT effectively filtering out the leakage from the other mixer output
> The errors due to the 16 bit quantization and mismatch between IF
> frequency and downconverted input signals can mostly be removed by
> choosing IF frequencies that are very close to the frequencies of the
> downconverted input signals, which, due to the new calculation method, is
> now possible. It's like creating a PLL where you tune the IF for best
> accuracy, e.g. identical to the input frequency after downmixing
> The only leakage that is still relevant is the leakage from one channel
> into the other channel before the mixers and, probably, leakage between the
> two LO frequencies (to be investigated)
> I've also tested the impact of having harmonics in the mixer output and as
> these are all blocked by the one bucket FFT there is no visible impact.
> It's weird that a frequency measuring device becomes more like a dual
> heterodyne receiver with the final IF at zero Hz acting like an
> analog/digital PLL where you can measure the phase errors very accurately
>
> ADEV with 0.1 Hz input frequency difference after the mizers, -80dB
> leakage and with equal IF frequencies
> <image.png>
>
> Same with different IF frequencies
> <image.png>
> Whether this method will enable to measure ADEV in any useful way (e.g.
> are the ADEV pictures above a good representation of the actual
> performance)  is still to be investigated.
>
> Erik.
>
>
>
> Op wo 21 sep. 2022 om 00:20 schreef Bob kb8tq <kb8tq at n1k.org>:
>
>> Hi
>>
>> With signals in the < 0.1 ppb offset range, you should see effects at the
>> -80 db isolation level. They should show up as ripples in what otherwise
>> should be a straight line ( ADEV drops vs tau in a straight line ….. ).
>>
>> Bob
>>
>>
>