Force 180 Downlink MCS counter MCS 1 QPSK ticking but home page says its MCS 14

Hi,

Currently running a Force 180 pointing at a ePMP 1000 120 Degree  Connectorized sector anternna with 14 SM's being serviced off it.

Signal Stats - 

RSSI - 55dBM

SNR - 36 dB

UL and DL MCS - 14/14

My questions is that on the Monitor > Performance page i am see this -

MCS 15 - 64-QAM 5/6

7014(2.4%)

MCS 14 - 64-QAM 3/4

10026(3.4%)

MCS 13 - 64-QAM 2/3

993(0.3%)

MCS 12 - 16-QAM 3/4

335(0.1%)

MCS 11 - 16-QAM 1/2

0(0%)

MCS 10 - QPSK 3/4

0(0%)

MCS 9 - QPSK 1/2

0(0%)

MCS 7 - 64-QAM 5/6

0(0%)

MCS 6 - 64-QAM 3/4

0(0%)

MCS 5 - 64-QAM 2/3

0(0%)

MCS 4 - 16-QAM 3/4

0(0%)

MCS 3 - 16-QAM 1/2

0(0%)

MCS 2 - QPSK 3/4

0(0%)

MCS 1 - QPSK 1/2

275558(93.8%)

 

The Uplink is stat is fine as its set on the Scheduler for MCS 14.

Usually i wouldnt be concerned with a weird stat its just that i am troubleshooting this service as end user is recieving a solid 20Mbps DL but a fluctuant 10Mbps - 20Mbps UL. Operating in the 5.7Ghz range on a 20Mhz channel with software upgraded to 3.2.2 on all SM's and APs. There is a bit of interference at the AP end as its noise floor sits at about -80 to -75 dBM. 

 

Any help would be appreciated.

 

MCS1 (or MCS0 if selected), is used for management frames (beacons, scheduling, etc). 

When the radios are idle (not passing ethernet data) the Wireless side sees only management frames, therefore, incrementing the MCS0/1 counter over time, but at a very slow rate. 

as there is no spread arround the bottom of the table, i think those are really just the normal management frames.

What is the uptime on this radio? if you run a wireless-link test , do the higher-mcs counters increase?

the rssi and snr are both very good. Where did you get that noise floor measurement from?

At the AP side, what is the uplink-packet-drops counter for this radio?

1 Like

Hi daball,

Please also note Multicast traffic is running on MCS1 in case Reliable Multicast option is disabled.

Broadcast is always running on MCS1.

It would be nice if you could attach Monitor->Performance and Monitor -> Wireless pages screen-shots.

Thank you.

1 Like

Wireless Link Test increases the MCS14 counter. The system uptime is about 42 days but i reset the stats which have only been up for 1 day 19 hours. 

Thank you for you explaination. 

I got the noise floor measurement from a spectrum anaylsis with another ePMP 1000 intergrated unit which was next to the AP.

Uplink Packet drops for that radio is from the AP is 0. (attached)

Hi,

See attached. 

Hi daball,

Actually I don't see any issues with your link.


Looks like data traffic is running on MCS14 as is displayed on Dashboard.

But Multicast + Broadcast + Management Traffic are running on MCS1.

So you see so many packets transmitted on MCS1.

Thank you.

2 Likes

@daball wrote:

Usually i wouldnt be concerned with a weird stat its just that i am troubleshooting this service as end user is recieving a solid 20Mbps DL but a fluctuant 10Mbps - 20Mbps UL. Operating in the 5.7Ghz range on a 20Mhz channel with software upgraded to 3.2.2 on all SM's and APs. There is a bit of interference at the AP end as its noise floor sits at about -80 to -75 dBM. 

 

Any help would be appreciated.

 

As others have mentioned the downlink stats will often show unexpected activity at MCS1.  But looking at your underlying concern (variable uplink), I'm assuming you have the customer provisioned for 20/20?  How are you arriving at the "fluctuant 10Mbps-20Mbps" determination - radio link test, network speed test, monitoring data, or what?  How many customers on the AP?  What uplink provisioning do they have?  And what's the AP's config - 75/25 TDD?

If the uplink MCS is stable at MCS14 (as it appears to be) then the noise floor at the AP isn't a factor - whatever that noise floor is will be reflected in the uplink MCS, lower MCS max for more problematic interference levels.  (now if you see the uplink MCS regularly hopping around, like MCS14 to MCS5 and back, that's a likely sign of intermittent interference)

My concern is that collectively, all the clients on the sector are sometimes consuming enough uplink timeslots that this customer cannot reach 20Mbps consistently.  E.G.: Say you have 10 clients on the AP, all but this one configured for 5Mbps uplink.  3 of those customers using their full 5Mbps uplink running at MCS14 would have far less impact on sector capacity than 3 customers using their full 5Mbps but running at only MCS5.

The sector limit on uplink capacity is often reached on a per-timeslot basis long before you reach the theoretical throughput limit, unless all clients are running consistently high modulation.

You /may/ be able to ease the situation by setting that SM's "Subscriber Module Priority" to High, giving most of their traffic (up AND downlink) higher priority that most other client's traffic on the sector.  Really not the 'right' solution, but a possible quick'n'dirty fix.

j

3 Likes

Fluctuant uplink was  tested through an iPerf test from PoE of SM to the same router that AP plugs into.

AP is set to 50/50.

13 SM's on sector. 

UPLINK MCS's - 11, 15, 13, 12, 15, ,15 ,14 ,15 ,13 ,15, 9, 14, 14.

I have seen the MCS on the SM in questions do jumps from MCS14 to MCS5 which i also thought would be because of interference.

I do find it strange that the AP is advertising a 20MHz channel and achieves up to 50Mbps DL. On a 50/50 Frame Ratio, theoretically that tells me that the AP should be maxing out DL Mbps at 38Mbps. Just a thought.

My initial thought was that the uplink time slots were being used, but that stat had been taken out a few software versions ago. It is one of our busiest sectors.

I have applied the SM priority high for now until we can upgrade capacity.

1 Like