disappointed in 5.2 BH20

heya folks,

so, we just a baught pair of new 5.2 Bh20 to try and improve on a mikrotik setup we were using. The mikrotik could pull up to 18 mbit, but rarely, and it was very inconsistant. I was getting a lot of “so, the network was slow last night, right?” from customers.

The situation is this:

We are trying to backhaul half of a DS3 to inject it into our network. our link is one mile. radio stats look like this RSSI 1893, db -49, jitter 1-4.

Problem is performance. We can barely pull 4Mbit. at the switch at the DS3 location we pull 20mbit + consistantly. One mile up the hile on a server we get terrible performance (I tried my laptop directly to the CMM switch too, same results). I am thinking it may have been the setup. any tips? I pretty much left the defaults. Out of frustration, I upgraded to 8.2v2 with no apparent improvements.

Any other 5.2 in the area?

Is it running in 2X mode?

-49 is pretty hot. Try reducing the power to the low power setting (-18dB). This should put about -68dB at each radio.

The BH20 should give you 14Mbps usable, 7Mbps each way assuming you are configured for 50% DL

Heya Jerry,

Yes, in 2X mode, and yes, another 5.2 backhaul on the same tower, but upper limit freq while this one is lower limit freq.

I am seeing a ton of overrun counts. Not sure what that means. Pings seem pretty consistant from backhaul. ave about 14 ms. I was expecting better, but that is obviously not the prob. I tried different freqs with no improvement. I am pretty lost here, and wondering if moto gear can’t handle many packets per second (thus the overruns). I checked the MRTG graph for the DS3, it is at 4%.

I will try lowering the db from 5 to 3 and see if any difference.

what does link test show? 9 second test, how clear is the spectrum you can see it in spectrum analyzer.

Is it bandwidth related problem or packet per second problem. I have BH20s at 80 Km, and they give 5 mbps duplex at max

here are the 9 sec results:
Stats for LUID: 2 Test Duration: 9 Pkt Length: 1522

Downlink RATE: 7270400 bps
Uplink RATE: 6707882 bps
Aggregate RATE: 13978282 bps
Pkt Xmt (Act/Exp): 5325/0
Pkt Rcv (Act/Exp): 4905/0

Downlink Efficiency: 98 Percent
Downlink Index (Act/Max): 98/100
Frag Count (Act/Exp): 129305/127800

Uplink Efficiency: 97 Percent
Uplink Index (Act/Max): 97/100
Frag Count (Act/Exp): 120646/117912


I don’t know if it is a bandwidth or PPS issue. I have plenty of bandwidth at the base NOC.

what I am wondering about is the overun errors:
Overrun Count : 7387
after being on for 6 hours.

There is nothing wrong with your BH20.

Upgrade to 8.2, that will give you more pps.

Play with the uplink, downlink ratio. Probably you don’t need full duplex.

Find a way to monitor the pps and troughoutput.

hello, thanks for the replys. I mentioned in my first post I had already upgraded to 8.2v2. honestly, ping times went up after that. But in any event, I just can’t seem to get any real bandwidth. Can you tell me if these variables matter?

location of master/slave
recommended setting for "Prioritize TCP ACK :"
recommended setting for "Transmit Frame Spreading :"
recommended setting for "Schedule Whitening "

Transmit frame spreading should be disable if possible.
Schedule Whitening - only if you use DFS.
Prioritize TCP ACK - Enable.

This was ported a while back. Not sure who posted it I just printed it. So here it is again

RxOverruns:

1) RxOverrun occurs when a network interface cannot get a packet because of overfilling. Look at the traffic running from switch port to radio. Also look at the NI BUF STAT page of radio

2) Also try switching from auto to fixed speed. Some ethernet ports do not auto-negotiate well.

3) Whenever I’ve seen this it is usually indicating a broadcast storm on a non-routed network

So as mentioned above, the device or network connected to the ethernet port on the radio is going ballistic spewing broadcast replies ot just plain garbage from a bad device.