5km - Groove 5Hn - What is the expected throughput?

Hi,
I have this 5km test link using Groove 5Hn on both side, there are not much interference from either side.
The most i’m getting out of this link were 55mbps TCP and 88Mbps UDP (one direction)
Comparing to previous setup using 5Ghz - Turbo mode, R52H i can get similar result of 55mbps TCP and 60Mbps UDP, nstreme (for some reason using nstreme is better than nv2).

My question is, by looking at the parameter & connection what is the expected throughput (please look at the image attached)? Seems to me it’s good enough to achieve max result as advertised (125Mbps Aggregate Throughput), but i’m no where near it.

CCQ TX is almost 100% all the time while RX varies from 80% to 58%.

Any suggestion or advice?
GROOVE-5km-Test-UDP.png
GROOVE-5km-Test-TCP.png
GROOVE-5km-Test.png

Try using advanced rate-selection for data-rates and make sure you are running v5.6
Also what setup you have: you use WDS or regular routing?

Hi Uldis,
The units are running ROS v5.6.
It’s WDS-Bridge mode.

Will try advance with fixed data rate, report back soon.

Thanks!

Back with test results:

Scheme 1: PC to RB1100 (Distance 5km but reported as 7km by Groove)
PC-> GROOVE-AP -->><< — GROOVE-ST ← RB1100

TCP:
RX: 67Mbps
TX: 34Mbps

UDP:
RX: 95.8Mbps
TX: 37Mbps

Groove CPU < 20%
RB1100 CPU < 6%

Scheme 2: Ethernet Test Cable Length 40m
PC-> GROOVE-AP

TCP:
RX: 68Mbps
TX: 32Mbps

UDP:
RX: 98.8Mbps
TX: 38Mbps

Groove CPU = 100%

Scheme 3 Radio to radio
GROOVE-AP -->><< — GROOVE-ST

never get near the 2 schemes above, and CPU were 100% on both radio.


Overall results are improved after data rate set to advance, ticked all A/G and Basic A/G Rates
HT MCS all ticked as well.

Tested using BTEST multiple test instances.

So, possibly TCP limited (bottle-necked) by STP quality/length (?)

That’s lower than I’d expect from a direct ethernet connection. You using half-duplex for some reason?

The test were done one direction. Ethernet is set to 10/100mbps full duplex.
It’s definitely bottleneck-ed by the cable length (or even the quality of the RJ45). Tested again using factory made UTP at 1m long, direct ethernet, btest, I got 95-98.8Mbps both tcp and udp one direction, when trying both direction it drops to 70mbps Rx and 25mbps Tx (TCP).

Hi there,

Is it a real 29DBi Antenna?
for both side?

Regards

11km, 30dbi antenna, -53 signal, ~70mbps TCP

I have a link by 29.5 db parabolic antenna and -58dbm signal
AP-station bridge
Nstreme and Nv2 are active but I have not both side bandwidth stable!
I have in one way 30-40mb and both 15/3 Mbps but I need 30mb full dulplex in both side
please help me

it’s kbt 30dbi 5Ghz Grid on both side :slight_smile:

providing more info like distance, band, ccq, signal noise etc, would help others to help you with your needs…

What is the antenna Db to be used for 5 km range if groove is used..I want to select omni or patch of which DB to have perfect point to poin link with clear line of site?
Please help me…

hey.. reviving an old topic but what happened?? is the link working fine now?