WARP Project Forums - Wireless Open-Access Research Platform

You are not logged in.

#1 2019-Apr-11 06:51:53

vutran
Member
Registered: 2017-Jul-01
Posts: 46

Carrier Frequency Drift

Hi,

When synchronising the transceivers on WARP v3, I notice the frequency drift. After sometime (1 or 2 hours), the phase shift between 2 transceivers is considerably increased. It's worse if I measure the phase shift of 2 transceivers on 2 different boards, even I use CM-PLL to connect 2 boards.

My questions are:

(1) What should be the sources of local oscillators in 4 transceivers in 1 WARP board? Can I assume that the phase shift between any 2 of them are constant if I don't change the configuration of the transceiver?

(2) Can I assume that the transceivers on the second WARP board have constant phase shift  compared with the first board (e.g. RFA on 1st board)? If not, is there any way to achieve that constant phase shift?

Thank you very much.

Best,
Vu

Offline

 

#2 2019-Apr-12 09:02:09

murphpo
Administrator
From: Mango Communications
Registered: 2006-Jul-03
Posts: 5065

Re: Carrier Frequency Drift

1) Yes, this should be a safe assumption. Any change to the PLL config on any radio will shift the relative phases (radio reset, changing center frequency, power cycle, etc)
2) Yes, with the caveat that the nodes must share a clean reference clock source. Any noise on the external clock signal could manifest as an extra/missing clock edge at the PLL input, which would cause relative phases to jump.

Offline

 

#3 2019-Apr-15 04:16:21

vutran
Member
Registered: 2017-Jul-01
Posts: 46

Re: Carrier Frequency Drift

I'm using a CM-PLL cable to synchronize the 2 boards. Is it considered a "clean" clock source (from first board to second board via CM-PLL)?

Thank you,
Vu

Offline

 

#4 2019-Apr-15 09:26:06

murphpo
Administrator
From: Mango Communications
Registered: 2006-Jul-03
Posts: 5065

Re: Carrier Frequency Drift

The CM-PLL outputs a clean clock to its external header; care must be taken when routing this clock to other nodes (shorter cables are better, high-quality coax is important, isolation of the clock cable from other emitters, etc).

Offline

 

Board footer