Antenna cable or synchronized clock cable


I have a ground robot that’s about 1m in size and am developing a PCB that can synchronize four DW1000 modules on said robot.

Since the robot is smaller than your average wired synchronized UWB network, I was trying to decide if:
1. I should put all the DW1000’s on one PCB and run ~0.5m antenna cables to the UWB antennas, or
2. I need to put out a wired synchronized clock to four individual DW1000 modules?

I’m hoping for the first option, since keeping the clock on a PCB is easier in my mind than converting it to LVDS and taking it off board, then needing to clean up the clock signal. Then again, I would imagine that with option 1 I might need to add an LNA to each of the DW1000’s because of receive signal loss in the cable.

My precision requirements are quite strict, so I would like to note lose too much of the DW1000’s capabilities. Accuracy in antenna delay etc. is not as much of an issue, ie. I have software that performs continuous self calibration.

Thanks in advance for your help!

1 Like

A 0.5m run of good quality cable shouldn’t be an issue. The losses you get would at least in part be offset by the ability to get the antenna further away from any metal and in a better environment generally. A clean antenna pattern can do wonders for system accuracy.

Two issues I found when using cabled antennas:

  1. The difference between a loose connector and a tight one can be 5cm of antenna delay. Sounds like this won’t be an issue but is worth being aware of.

  2. I have seen issues with signals at short range effectively jumping the cable or most of the cable. I would see a sudden step decrease in range when getting very close. The step was roughly equal to the cable propagation delay. I added more and more shielding to the DM1000 and cable junction (it was a little hacked together, we replaced a smd antenna with a socket.). The problem persisted. In the end I tracked it down to the signal leaking directly out of / on to the cable. The signal level was tiny but it was enough to count as a first path. A few layers of foil over the whole thing fixed it in the end.