[time-nuts] Fwd: cheap frequency extension for timepod

[John Miles]

> FWIW, I have done PN measurements at 144 MHz with the TimePod and a
> Frankenstein's laboratory lashup of 80 MHz ULN, HP doubler to 160 MHz,
> HP mixer, and several amp and filter blocks.  As far as I can tell, it
> worked pretty well; at least, I could discern the PN difference among
> several sources, some of them quite good.
> 
> Not sure why you are doing two oscillators/mixers, though.  I used a 5
> or 10 MHz reference directly into the TimePod, and the DUT fed through
> the mixer.  No need to increase the frequency of the reference to match
> the DUT.  But if you do want to mix both the REF and DUT channels,
> wouldn't it be better to use a single source to drive both mixers?

He's driving the Ch0 and Ch2 input channels directly with a topology like this (attached).  This allows the TimePod to average the LO noise out along with the rest of the instrument noise.  The reference is provided as usual from another independent source. 

The FPGA in the TimePod had room for only two DDS cores, one for the Ch1/Ch3 reference channels and the other for the Ch0/Ch2 input channels.  So while you can use two independent LOs for downconversion, they do need to be kept close together in frequency.  Otherwise the cross-spectrum averaging process will be forced to work with baseband signals at significantly different frequencies relative to the Fourier bin size, which yields inconsistent results.

The 3120A that Bob mentions is identical to the TimePod under the hood.  Their actual 'big brother' is the 53100A at https://www.microsemi.com/product-directory/phase-noise-and-allan-deviation-testers/5565-53100a .  The 3120A is a nifty piece of gear (if I do say so myself :) but the 53100A provides a lot more flexibility at a similar price.  In this case, a 53100A measurement can leverage independent DDS cores in all four channels.  This feature makes it a better fit for downconversion applications, although it's certainly possible to get good results with the older units as well.
 
-- john

-------------- next part --------------
A non-text attachment was scrubbed...
Name: VHF_test_jig.jpg
Type: image/jpeg
Size: 265224 bytes
Desc: not available
URL: <http://febo.com/pipermail/time-nuts_lists.febo.com/attachments/20201017/34aa3ba8/attachment.jpg>