MIMO vs OctoClock

[cochenob]

I have a setup similar to the attached picture.  The path length between the Tx and Rx1 is fixed, while the path between the Tx and Rx2 varies.  Using a MIMO cable syncs the two Rx's, and I can measure a relative phase delay between the two Rx's corresponding to the varying path length between the Tx and Rx2.  I say relative phase delay because I am not accounting for the constant phase delay that will always exist between the two N210/WBX radios.  Ok...so far, so good!

Making things a bit more complicated...I'm transmitting at a slight IF from the carrier (50kHz).  This would result in a time varying linear phase at each Rx.  Still, since I'm comparing the *relative* phase between each Rx, I'd still end up with a constant phase corresponding to the path delay.  Ok!

Making things complicated even more...if I were to toggle my IF from +50kHz to -50kHz (think FSK), I might expect to see some change in the relative phase since the IF's are different...but...if my IF is small relative to the carrier (and it is...my f_c=70MHz), then this difference is likely to be in the noise.  

Here's the problem....when I use the MIMO cable, everything I said above is true.  I measure a constant relative phase, and toggling the Tx IF doesn't measurably change the relative phase.

However...If I synchronize the Rx using the Ref in/PPS from an OctoClock...I observe the phase going haywire when I toggle the IF.  Given that the IFs are small relative to the carrier, and that I'm not actually changing anything about the Tx or Rx configuration (other than how the Rx is sync'd), this behavior seemed odd.

My only guess at this point is that there are some differences in the actual ways in which the two methods go about the synchronization.  This is never really articulated anywhere though (for example, the only thing I've seen about the MIMO cable is that it passes "timing messages" between the two radios...what does this even mean?).

Any help here?

---

Brandon

[cochenob]

Some additional info....

- I've checked the Ref and PPS levels from the OctoClock.  They seem good. 

- The Ref and PPS cable lengths are similar.

- The Ettus boards have some folks that observed odd behavior with their OctoClock which was remedied by updating the OctoClock driver.  How might I do this exactly?

[ashley_gee]

Hi Brandon,

The firmware and update instructions for the octoclock can be found here via the Ettus support site: http://files.ettus.com/manual/page_octoclock.html

If other users were successful after updating, it's worth giving that a shot. 

Also, in case you haven't seen them, I wanted to point of the Ettus support mailing lists, found here under technical Support: https://www.ettus.com/support

These are the Ettus version of forums, so you may be able to crowd-source some info there as well.

Hopefully you'll get some additional insight here from the NI community - however, since this is pretty specific to your particular setup, this might be better addressed in an actual service request with Ettus. You can create a service request by emailing support@ettus.com or going to the support site above.

[vvkakade]

This ettus link talks a little bit about how MIMO cable based sync works:

https://files.ettus.com/manual/page_usrp2.html#usrp2_mimocable