April 11, 2007

Rain fade update

I posted previously about the wireless LAN bridge we installed from our Central Campus to our new leased office space, and I mentioned the problems we've had with rain fade.

Since my last post I learned that the radios nominally use the 80 GHz band, but since it's a full-duplex link, there are two diffferent frequencies in use: around 80 GHz in one direction and around 72 GHz in the other direction. Consequently, the signal strength in one direction isn't the same as the signal strength in the other direction. The graph I posted previously was for the stronger of the two.

Since my last post our vendor has repositioned the radios again and wrung out another 3 dB, which is great. Unfortunately, the link is dropping at -66 dBm, whereas the spec says it won't drop until somewhere in the -69 to -72 range. And according to Bridgewave, our radios tested to -69 in the factory before they were shipped. So we still have a problem.

Here is the signal strength graph from yesterday when in rained starting around noon and continued at varying intensities for the next 24 hours:

In all of that we had only one drop lasting around 30 seconds at 3:39 pm. You can't see the excursion below -65 because the monitoring system is on the other end of the link. (The monitor needs the link to be up to grab the signal strength from the other end.)

Now, one 30 second drop doesn't sound to bad considering it's been raining for 24 hours. But the problem is, if we had staff working here yesterday, every phone call in progress at 3:39 pm would have been dropped. Most likely every Shelby session would also have been dropped. That doesn't seem good enough to me. What do you think?


Anonymous said...

what happened to the link?

Clif Guy said...

It stayed up with the rain we had yesterday. The signal never dropped below -62.

Bridgewave sent out new radios that arrived yesterday morning, but the weather didn't allow us to install them. So we're hoping for an opportunity early next week.

Anonymous said...

so do you think you still have a bad radio? Was yesterdays ran very heavy? I'm asking all these questions because I am istalling 2 Bridgewave links this week- 1 and 3 miles...

Clif Guy said...

Rain intensity varied over the 24 hours changing over to a rain/snow mix for the second 12 hours. It didn't have any thunderstorm cells or periods of especially heavy rain. Total accumulation for the storm was 1.65 inches.

We can't be 100% certain if we have a bad radio, but the link does seem to be dropping at least 3 dBm higher than it should. I think we'll know for sure after we install the new ones and have our next storm. Right now the forecast is for three chances of showers over the next 10 days.

Ralf Baum said...

Rain attenuation is a very interesting thing. In IEEE Transactions on Antennas and Propagation there is an article called "Spatio–Temporal Rain Attenuation Model for
Application to Fade Mitigation Techniques"
that discusses the topic and builds plots of PSD distribution of rain attenuation against the signal frequency. They also prepared a small MATLAB script of the function that builds the model.
Additional article can be interesting as well.
With that your radio link, which BER it shows?
I also found an additional diversity simulations MATLAB script.

Anonymous said...

What was the outcome of installing the replacement radios?

Clif Guy said...

It increased the signal strength by about 3 dB. Unfortunately, we get intense rain events here that can attenuate the signal over the 1.7 mile length of the beam by 30 dB or more, so we still have some drop outs. We recently installed a high gain antenna on one end to increase the signal by another 6 dB. We'll have to wait for spring rains to know whether that's enough to make it stable.