Hello,

A recent NStreme2 link we have configured is demonstrating an anomaly that I have seen twice before, but only resolved after replacing the RB532. I'd rather get to the bottom of this problem as this is the third time it has appeared.

The problem is clearly revealed during a simple UPD HDX test between the radios, but is aggravated further to the point of dropping packets and severe data throughput loss during FDX tests.

It appears the wlan TX interface will report twice the bandwidth that is actually passing across the NStreme2 link. Here is a snapshot showing a BW test between the radios in the TX direction and then another test showing a BW test running in the RX direction.

[/url http://www.belwave.com/NStreme2_Problem_4-6-06.bmp

Clearly RX is not having any trouble, but the TX is.

Previous encounters with this problem were approached by replacing the radio, upgrading firmware and countless configuration changes, but only after the RB532 board was replaced while keeping all other hardware in place (same radio, pigtail, PS, Ethernet cable etc, etc.) did the problem go away.

Thoughts?

Many thanks,


Brad

RB532 and Nstreme? Forget about it. Man you can't even run 5Ghz-turbo coz processor too slow to make any difference. Use a PC instead.

Nstreme2 requires processing power, 800Hmz of CPU is recommended (in a X86 system) for full potential of Nstreme.

Nstreme2 requires processing power, 800Hmz of CPU is recommended (in a X86 system) for full potential of Nstreme.

Agreed, more CPU power is always a plus when working with NStreme2. However, this problem does not appear to be related to CPU as one direction will work fine and the other does not. Did the URL I provided work for you to see what I am describing? Why is it that the wlan TX Interface is indicating twice what the NStreme Interface is reporting? Certainly there has to be an explanation for that, right?

When we are not plagued with this problem we have consistently achieved 20-22MBps FDX on a properly configured NStreme2 link using over-clocked RB532s. Could it be that some RB532s are just not up to the task? That would explain why simply replacing the RB532 board has resolved this problem the last two times we've seen it.

Thoughts?

Many thanks,


Brad

It looks like you have more interference on one side and the is why the wireless shows almost twice that actual traffice (retransmits) -- I don't know this for sure, Uldis can check this for you.

John

First what is the distance for the Dual nstreme link? What anntennas are you using? Are both antennas very close on the mast?
What frequencies are you using for each side?
Have you tried to change the frequency that has bad transmit?

Also, you can make the support output file from the routers and send it to support@mikrotik.com so we could analyze that.

First what is the distance for the Dual nstreme link? What anntennas are you using? Are both antennas very close on the mast?
What frequencies are you using for each side?
Have you tried to change the frequency that has bad transmit?

Also, you can make the support output file from the routers and send it to support@mikrotik.com so we could analyze that.

Hello guys,

Yes, RF re-transmits was my first thought as well the first time I saw this behavior, however all freqs exhibit the same results. Distance is approx 9 miles perfect LOS with no fresnel incursion. Antennas are mounted via pipe-to-pipe standoffs one each tower leg. MT v2.9.19. RSSI readings on each end are between -56 & -58. Antennas are Gabriel 2' DP & PAC 3' DP. This is the first time to use a PAC DP antenna, but again we've seen this problem twice before and it does not appear to be radio card, pigtail, frequency or configuration related.

The first two times we saw this problem simply replacing the RB532 and keeping everything else the same fixed it. Here is another screenshot this time showing a FDX TCP test running at a set 7Mbps cap:

[/url] http://www.belwave.com/NStreme2_Problem ... 4-7-06.bmp

You can see the wlan TX interface is approx twice the actual carried bandwidth. Once the wlan TX interface begins to reach its max as we continue to bump the BW test higher and higher it begins to fall apart resulting in a nearly a complete loss of link due to dropped packets.

The test can be run from either end and still show the same TX interface with the problem.

Many thanks,

Brad

Hello again,

Just an update...

Replaced the RB532, TX & RX radios, u.fl pigtails, enclosure etc. Basically it was far easier to replace the entire "box" this time given the location.

No improvement. In fact exactly the same problem with the TX interface showing approx twice the bandwidth the NStreme interface is actually passing.

Regarding the re-transmit theory, why would there be re-transmits on an NStreme2 TX interface? How would the radio "know" to re-transmit if it isn't receiving anything?

Let's assume it is a re-transmit problem. Can a bad LMR jumper from the enclosure to the antenna cause re-transmits?

Another thought is if we've had three RB532s exhibit this problem why should we not assume the replacement RB532 is not the fourth to show the problem?

Thanks,

Brad

Hello, I have the same problem.

My config: WRAP2C+2xCM9 , dual pol. antennas, 16km clear LOS, 63/67 dBm signal.

Any news in this topic?

Jacsa