HI,

I have a Mant 2 12S with four client LHG 2 clients. All on 6.49.13. This install is 80km from the nearest town so any interference is my own. The Mant is configured for NV2. All four clients connect with TX/RX Signal of around -70/-70. There is very little vegetation to get in the way.

With a bandwidth test between the Mant and a client, I can send UDP at over 15MB/s but only receive at 200KB/s.

I can't see any interference at that location. I limit my 2.4Ghz channels to 2412, 2437 & 2462.

The Mant 12 wireless is set to

Band: 2Ghz only N
Band With: 20Mhz
Freq : 2412
Antenna Gain: 16 ( usually 12 but it think it works better at 16 )

The NV2 is set to ..

tdma period size = auto
cell radius 20km
mode: dynamic download
download ration: 50

The rest of the config is default. No firewall, simple bridge.

There is another Mant 12 on the same pole that is using NV2, but is on same windmill that is on freq 2437.

I have used the groove access point there to scan the frquencies being used and I can't see any interference.

Does anyone have any clues as to why my download speeds from the sector clients is so poor? Yet upload is fast?


Any help appreciated.

regards

Justin.

First of all, show screenshots of how the links work.

Antenna Gain: 16 ( usually 12 but it think it works better at 16 )

You are mistaken by putting 16 you have reduced the output power by 2.5 times. It will work better if you set the gain to 0.

I was trying to reduce the power of the mant just in case it was too loud. I thought a signal strength of -70 was ok.

I tried again with gain of 0. The downbload speed went up for a short time only. After a few seconds it reverted to 240Kbps and stayed there. Upload speed was bou 80Mbps. Thats a big difference.


Do you mean a screenshot of a map or layout of the sector and clients, or screenshots of winbox?

How do you do bandwidth test? When using the builtin test, it will consume cpu power (and hence loose performance). In that case, better to use a tool like iPerf.

In regards to the antenna gain...just use the antenna gain. This will make sure that you remain within regulations.

I am using the built in bandwidth test. The network has very little traffic and I have used the bandwidth test many times without seeing this sort of imbalance.

still some more testing to do, but....

I change one Mant to be a sync master and the other to the a sync slave, put them on the same frequency and increased the bandwidth to 40Mhz XX.

The bandwidth test from the 4 sector clients now goes over about 5Mbps.

When both Mants were on different frequencies (and 20MHz) I checked the spectral scan, using a Groove at the same location and, and there wasn't any noticable overlap.

I thought a signal strength of -70 was ok.

That's what you think. What does the standard require?
The recommended signal levels are specified in the file "802.11 Data Rates & SNR Requirements" search the Internet.
Show me the link parameters, you're interpreting them in your own way.
Show me a scan of the airwaves without NV2 and with it.

I checked the spectral scan, using a Groove

You need to scan with the device itself, not a separate one.

The LHG 2 is 18 dBi, the Mant 2 12s is 12 dBi.
Since one has to abide by EIRP regulations, the antenna gain works only to increase reception.
In other words, the signal from the Mant to the LHG is improved by 18 dB. The signal from the LHG to the Mant is improved by 12 dB. We should be seeing different signal strengths by 6 dB, yet you indicate -70/-70 which I don't see how it can be the case.
You should be getting higher throughput from the Mant to the LHG than from the LHG to the Mant.

With a bandwidth test between the Mant and a client, I can send UDP at over 15MB/s but only receive at 200KB/s.

What do you mean "send UDP" what direction is that? What direction is "receive"? From the LHG to the Mant or from the Mant to the LHG? Which frame of reference are you using?

Btw have you tried 802.11 or nstreme?

Hi,

-70/-70 was an average. They were changing in Winbox all the time.

The only time I change the antenna gain is to reduce the power of it. I assumed that under "regulatory-domain", increasing the "antenna gain" would force the transmitted power to be reduced accordingly.

I was using the built-in bandwidth test from the sector radio (Mant). Sending is from the Mant sector to the LHG client. I was using the option to send UDP packets instead of TCP. When I first tested I was able to send from UDP packets from the Mant to the LHG at over 15Mb/s (not MB as I had written earlier). Transferring data from the LHB to the Mant was about 200Kbps. It varied a lot, but that is an average.

The four clients are at distances of 1.8km, 3.2km, 5.3km and 10km. THe closest client is an SXTsq Lite 2. It's current Tx power is 23dbm and I can't find a way to reduce it. The rest are LHG 2's.

The furthest client has a txrx ccq of 90%. Below are resutls when the Mant is running 20/40MHz in NV2 ( synch master )

I have tried 802.11 but it gets worse. I haven't tried nstreme yet.

in which country are you ? what freq are you using?
also - the signal are quite bad for this distance, do you have a lot of trees in between? ccq is also bad.
for that distance it should workwell, i have 4km link over a hill with trees, on 2ghz i get about -77 and can pass about 75-80mbit (because one chain is around -82... otherwise it was passing about 100-105mbit before)

You don't want to show the full picture of what's going on. You don't seem to understand what parameters are important. The above is not enough to help.

This install is 80km from the nearest town so any interference is my own.

So it's your fault the CCQ is so low.
The only advice I can give from what you have shown is to disable the short guard interval. Instead of any of them, choose a long one.
In general I don't see why the rate is so low, apparently you don't show the necessary data, or there is a mistake in the way of measuring.
PS
Why is the signal from a 3.2 km station 10 times smaller than from a 10 km station?
Draw the station layout on the plan. the angle of direction to the station matters. It also matters which station the base point is pointing to.