Greetings

Im going straight to the point

I have a ptp wireless link with a pair of NetMetal 5 and shielded antenna running with Solid 100%/100% CCQ 866/866Mbps Data Rate
But i cant even reach 400Mbps

My results are

NSTREME = 250Mbps TCP and UDP
NV2 = 210Mbps TCP and UDP
802.11 = 240Mbps TCP 330Mbps UDP


So my question is
Why i cant reach 540Mbps ? or at least 400Mbps


PS: I think NetMetal 5 should have at least 2 cores, and why noise-floor-threshold doesnt work

How are testing the speed? Do not use the built in bandwidth test as it will use too much CPU.

I`m running the btest from a RB3011 to another

-33 is too strong of a signal, try lowering output power and testing again.

I tried with -55, but the data rate starts to drop. my current TX power is 22dbm on both sides

But even with this signal the ccq is running at 100%
It should be able to reach it's full potential right?

I'm actually wondering if someone has achieved over 500mbps

What CPU load do you have on NetMetals and 3011 during Btest? What happens with TCP results if you change channel width to 40 MHz? Are you testing it with one TCP session or more?

What CPU load do you have on NetMetals and 3011 during Btest? What happens with TCP results if you change channel width to 40 MHz? Are you testing it with one TCP session or more?

CPU load on RB3011 reached 46% and NetMetal 69%

my results with different channel width
20Mhz = 130Mbps
40Mhz = 200Mbps
80Mhz = 250Mbps

all results with nstreme

i ran the btest with 1, 20, 50 tcp conn

Theoretical best would be 50% of the data rate. It has been my experience that real world yields 35-40% of the data rate for a PtP. It depends on the environment. I have seen rare instances of interference causing lower than expected throughput that had high and stable data rates and ccq. If you test from each direction, are your numbers lope sided, or fairly even?

2frogs wrote:

Theoretical best would be 50% of the data rate. It has been my experience that real world yields 35-40% of the data rate for a PtP. It depends on the environment. I have seen rare instances of interference causing lower than expected throughput that had high and stable data rates and ccq. If you test from each direction, are your numbers lope sided, or fairly even?

Each direction gave me the same results, the weirdest thing is with 40Mhz the results are pretty close to 80Mhz
i tried with the older wireless package cm and fp (6.30.4)

i achieved 402Mbps udp in 802.11 but tcp was way too low, like 160Mbps
with nstreme 300Mbps udp and same thing with tcp, something like 170Mbps
NV2 gave me 340Mbps UDP and 170Mbps TCP

udp was way better but tcp wasnt. so i went back to the lastest version (6.40.5)

i'll keep running some tests.

cheers

Wondering if there was a resolution to this? I am having similar issues with high air connected rates, but slow speeds, I got air rate of around 590mbit, but only 75 mbit real data. This is the same on 80Mhz or 40 Mhz, using Nstreme, on either AC or N.

Wondering why this happens. Seems as every thread about this issue, or similar, just ENDS and no postings of if it was solved, or what came of it.

Lets start asking... When was the hardware purchased. This may be a particular batch of Netmetal products that were sold that have this issue?

One big difference between interface data rate and application data rate is the shared electromagnetic medium.The 100% CCQ excludes adjacent channel interference, but any other device on the same channel(s) will make the radio WAIT till the other is silent. If the or those others are slower in interface data rate, then you will wait longer between 2 packets.

The wider the channel, the more chances for co-channel interference, and waiting to transmit. This waiting does not alter the CCQ, RX power or interface data rate indicators.

Are there any other 5 GHz (wifi) signals? NV2 is not co-channel interference to 802.11, as I understood. They don't see each other except as noise.
And what about encryption for 802.11? Only "WPA2/AES" or "none" should be used.