450i Configuration Suggestions

Products PMP
1

Pulled our test 450i equipment out. I've got a shot of 50' in the office. I tried to mimic the settings we have on the PMP100 we are using currently, but am not seeing the most exciting numbers on speed tests. Any suggestions on what I could be missingin the config to tweak? I'm down to adjusting a setting one at a time to see if it makes a difference.

2

What antenna are you using on AP and SM?

What is your power output on the AP?

What do you have your SM Receive Target Level set to on the AP?

What signal is your SM hearing the AP at and vice versa?

1 Like
3

Thank you for the specific questions! I'm still hitting the learning curve with this:

  • What antenna are you using on AP and SM?

AP has the 5 GHz 60 degree sector antenna. SM has the 900 MHz 12 dBi directional yagi antenna.

  • What is your power output on the AP?

I'm assuming you mean Transmit Power? I bumped it up to +10 dBm during testing.

  • What do you have your SM Receive Target Level set to on the AP?

-52

  • What signal is your SM hearing the AP at and vice versa?

 I'm assuming this is the receive power when looking at the alignment tool on the SM? It is at -40.9.

Interestingly when I fired it up this morning to check all these stats I was able to get 52 down on link capacity test with a 20MHz channel. The speedtest on the laptop hooked to the SM though was only giving me 13 down.

1 Like
4

Suggest you turn the power WAY down on the AP (down to 0 dB or even less).  As you point out, the target level of the SM is set to -52, but it's is seeing -40... this is overdriving the signal and probably resulting in poorer than expected performance.

A couple other methods for getting "lab test" results that make sense, any RF attenuating obstacles in the path (like absorbing foam) would knock down the signal effectively.  Even intentionally misaligning the antennas can help, but may also result in weird multipath anomalies.   If you're performing this indoors, you need to be aware that signals are bouncing all over the place, and multipath reflections can casuse self-interference resulting in sub-optimal performance.

Also, a little confused by your first statement - "AP has the 5 GHz 60 degree sector antenna. SM has the 900 MHz 12 dBi directional yagi antenna."  Maybe you meant that the AP is using the 900 MHz Cambium Sector antenna (N009045D003A)?  Are you trying to test the 900 MHz 450i or the 5 GHz version?

5

In reference to that first comment we are testing the 900 MHz 450i. I know it is the 60 degree sector antenna. I got the 5 GHz from the manaual as the external antenna option for the AP, so I'm thinking a typo in the 450i User Guide, I kind of thought it was a bit weird myself. See page 2-23 of the Cambium PMP/PTP 450i Series User Guide System Release 14.0

I've turned the power down to 0 dB and the receive power is now at -51. Using iperf I was able to get 15.2 Mbps. I am implying from your comment that you want the SM signal strength to be near the target level set on the AP. It makes sense, haven't made that connection before. Something for me to file away to check on with troubleshooting.

I understand we don't have it in the best environment for testing, but on the widest channel I would still think we would be doing better then 15 Mbps? If we can get this test to be resonable we were going to take it out to do a test over a longer distance.

1 Like
6

I agree that you should be doing beter... it would help to post a screenshot of both the Link Test page (with results) as well as your settings on iPerf.  Sometimes a parameter change in iPerf is all that's needed to see what's really going on...

7

Regarding the Rx Power and Target Power Level, I erroneously connected the two above.

The Rx Power Level of the SM is the power that's being received at the SM.  This is only dependent on the Tx Power of the AP.

It is NOT dependent on the "Target Power" that you set in the SM.  The target power is using the ATPC (Automatic Transmit Power Control) mechanism of the system to adjust the SM Tx Power, so that it doesn't overwhelm the AP receiver.  This adjusts every two minutes, and in your case will try to ensure that the AP Rx Power Level is -52.

However, setting the power lower on the AP will surely help prevent an overload (and hence throughput degradation) at the SM.  A great number for the highest modulation level is in the low -50's, maybe shooting for -55 or so, -51 should work fine.

1 Like
8

Hopefully this is legible screenshot of the LInk Capacity Test.

2016-02-12 11_02_35-No Site Name - Link Capacity Test [none].png
2016-02-12 11_02_35-No Site Name - Link Capacity Test [none].png1359×746 54.7 KB

As for Iperf we did both UDP and TCP

-u -b 50M -t 60 -i 20 = average of 49.9Mbps

-w 10M -t 20 -i 4 = average of 12.1Mbps

When I do speedtest.net I get 15.39Mbps up/15.56Mbps down

1 Like
9

OH yeah... doubtful you can read that link test... here it is as an attachment.

10

An interesting observation that on the Speedtest.net we get quite a burst up to 30+, but then it drops right back down to 13 or so. Sometlhing in the QOS settings maybe?

11
@kendali wrote:

An interesting observation that on the Speedtest.net we get quite a burst up to 30+, but then it drops right back down to 13 or so. Sometlhing in the QOS settings maybe?

Can you post a screen shot of your SM's QoS page?

12

Here is the SM. I tried changing the sustanined rates, but it won't save off of the 65000. I can lower the burst allocation and I thought maybe that would then allow me to change the sustained rates, but it won't. This is one page on both the AP and SM that I did not mess with simply because I figured the default would get me something decent and I don't know what I'm doing in it.

13

Can you do another link capacity test with Signal To Noise Ratio Calculation During Test enabled?  It provides a lot of useful information about modulation and such which can really help.  The modulation data is completely independent of channel width, so you get a much better idea of how much of the potential capacity you're actually getting.

Do you have any PMP100 900 APs that you're colocating with?  Are you using a 5 ms frame width to coordinate with any PMP100 noise in the area?

Speedtest adds a lot of other factors to the test, so I'm not sure I would trust that as a proper indication of the speed of the SM-AP link.

14

Sorry for the delay in the response. Attached is a screenshot of the link capacity test with the only setting changed is the one you asked about. I would be curious as to what you are intepreting with this.

We have this system currently set up in the office, so no other radios around. We don't have any other PMP100 900 APs, we  would be colocating with 5.2 and 5.8 Ghz. I have not adjusted any settings to compensate for colocation since we aren't doing that yet.

15

Edit:  I appear to have two separate accounts, and logged into the wrong one.  Odd.

The percentage of each modulation is helpful because it will show you how efficient the links is regardless of channel size.  An absolutely perfect shot will give you 25% in each modulation, in both A and B antennae.  A good QAM-64 show will give you 33%, a good QAM-16 shot will give you 50%.  It's a little nonintuitive, but it works.

In my experience, 256 QAM is rare.  It requires a perfect shot, though to be fair, I'm in a worst case environment with a lot of hills, trees, and noise.

Your uplink looks  good to me.  It's using both antennae, and it's pushing 64 QAM.  The downlink is more worrisone.  It's sending the same data on both antennae, which in my experience is usually a response to noise or reflections.  If you're doing the test in an office, that could be a problem.

Your S/N is also high enough that it shouls be using more 256-QAM.  That makes me wonder if it might be out of sector.  The 450 SMs REALLY don't like being outside the AP sector beamwidth.  You might still get a good signal, but it will wreck your capacity.  The PMP100s were pretty forgiving that way, especially with the 900s, so you could do things like face the AP away from the SM to reduce signal strength, but doing that with a 450 causes a lot of problems.

1 Like
16

I really appreciate the breakdown of what I'm seeing. I messed with the alignment a little bit and reran the link test. I did get A and B antenna on the downlink and wanted to share.

Have a meeting in a few minutes to hopefully determine if this is just something simple I am overlooking.

17

What is your downlink % set to on the AP?

18

Ah yes, I forgot I really dinked with some settings... For the last test it was set at 50%. In a typical environment we are going to have more use on the down and less on the up, so I would imagine the default of 75% would be best.

19

Hmm.  Your uplink S/N is really good, but your downlink isn't.  That's oddin a test environment.  There's still something wonky there, but I would assume it's something to do with the test environment.  I know for a fact the equipment performs better than that where I'm located - and we're in worse case so you should do even better.

1 Like
20

Ted, what kind of speeds are you getting and what size channel? I had confirmation that the config looks good and agreement that our test environment is weird. I think we are going to try a parking lot test with direct LOS and if we get good enough numbers then maybe we'll put it on the tower.

The issue that I think we are seeing is also that we assumed we would have way better speeds with the advertising hype that we would be getting 100Mbps on a 20Mhz channel.