[time-nuts] My DMTD (actually: TMTD)

[Tobias Pluess]

Hallo all

I've got a question concerning the DMTD methodology.
I want to compare different oscillators to each other in a 3-cornered hat
fashion such that I can work out (hopefully at least approximately) their
stability.
For this, I thought about the attached schematic.
I plan to connect my three oscillators under test to three mixers and use a
common transfer oscillator for all three. In my opinion, a time-tagger
instead of a TIC would allow me to compare any two of the three
oscillators: 1 vs. 2, 1 vs. 3 and 2 vs. 3.
With this, I could determine the Allan variances of the different
combinations and then solve the equation system for the Allan variances of
the individual devices.
Would this scheme work?
I would like to use a STM32 microcontroller as a time tagger. When it is
clocked by 10MHz only, it would provide 100ns time tag resolution, but
apparently, this should be enough:
If I compare 5 MHz oscillators and heterodyne them to 5 Hz, the resolution
should be somewhere at 2e-13 (for tau=200ms). And I think at tau=1sec it
would be five times higher, i.e. 4e-14. Is that right? In this case, I
wouldn't even need a TDC7200 as an interpolator. On the other hand, I could
use the STM32's internal PLL to generate a 50MHz clock for the timers, and
therefore increase the time tag resolution to 20ns.
I think the attached schematic should definitely work, but I am unsure
whether this is a "smart" idea. Maybe there are better methods to compare 3
oscillators directly.
Any hints for this?

many thanks.
Tobias
HB9FSX





[image: dmtd.png]
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dmtd.png
Type: image/png
Size: 26665 bytes
Desc: not available
URL: <http://febo.com/pipermail/time-nuts_lists.febo.com/attachments/20200818/0378fffa/attachment.png>