You are not logged in.
Hi
we have use up to 9 WARP boards to build a Massive MIMO transmit beamforming system with up to 32 transmit antennas. Now we want to upgrade our system to a distributed MIMO system. That means we are not able to connect the debug headers together in order to trigger all boards together. We plan to build a external distributed synchronization system and use that system to trigger WARP board simultaneously. Our question is what will be the required accuracy of synchronization in order to build a beamforming system.
The main idea for beamforming is tuning the phase to transmitted signals so that they will be add constructively at the receiver side. If we consider transmitting two identical signals with carrier frequency of 2.4GHz to a single antenna. At the receiver side, theoretically, the received signal will be null if the phase difference of two transmitted signals is Pi, which means one of the signal is delay by 1 / 2.4e9 / 2=0.21ns. Then in that case, is it meaning that we need to trigger all boards with in around 0.1ns or even smaller period?
Is the above analysis correct? What will be the time difference among boards if we use a single trigger source? I think the difference may come from the jitter of oscillators in different boards?
Thanks
Offline
We've never build a distributed MIMO system like that, so we can't provide much help other than pointing you to some other resources. You might want to check out Vlad Balan's excellent PhD thesis where he used WARP to build a distributed MIMO system. In that work, he describes many of the challenges synchronizing distributed nodes and it might serve as a good starting point.
Offline