PMP 450i 900Mhz

We are new to the Cambium world, former Telrad customers.

Question, this morning our Frame Utilization shot up too 100% on the Uplink and around 60% on the Downlink virtually knocking everyones modulation on the sector down.  We have about 15 customers off of the AP and most are set for a simple 3m/1m package.  A few with lower modulations but nothing crazy.   

Is this interference, I don't believe it is usage related.  I downloaded the tech support file if needed. 


Welcome!  It doesn't sound like interference, rather how your QoS is setup on the devices.  Do you have an application or network appliance that is controlling the bandwidth to your customers?  Or are you limiting them at the radio level (using QoS)?

If you are using the default settings (and letting the radio manage bandwidth), it is possible that one or more devices is going well beyond the 1Mbps upload speed.  You would want to set your sustained downlink/uplink speeds to a data cap for the plan (i.e. 3000kbps / 1000kbps) and either consider using bursting (see attachment) or configure the burst downlink/uplink speeds the same as the sustained.  Otherwise, the data cap per customer is the entire bandwidth of the sector.

Here are some additional links to get started with PMP450 (in addition to the training at ):

1 Like

Formula to calcute QOS in PMP 450i AP if we want to connect 10 SM Canopy, 4Mb throughput each SM.
We use this for camera setup.

Attached is our QoS for all of our SMs.  They are uploaded via a config template we have created.

I do believe we will start allowing some Burst but have yet to make that change an implement it.  So I'm not quite sure if that is the issue.  All of our radios are Bridged and we do make the subscriber establish a PPPoE connection at their end. I did turn on Protocol Filtering on all SMs (except PPPoE) after this happened and SM Isolation on the AP thinking perhaps a subscriber plugged something into the POE of the 450i SM.

1 Like

Can you post a screen shot of all your stations on the Link Status page, so we can see what modulations things are running at... ALSO... under frame utilization stats... what's the Bandwidth Request Success % at 15 min average. What firmware revision are you running on everything? Please update everything to 16.1 if possible.

1 Like

Attached files requested.  We are running CANOPY 16.0.1 AP.  I just upgraded it a few weeks ago, so there is a newer version?

@Scott B wrote:

Attached files requested.  We are running CANOPY 16.0.1 AP.  I just upgraded it a few weeks ago, so there is a newer version?

So I looked over everything and things actually look decent on this AP. I wouldn't say that it's having massive interference issues or anything like that.

Yes there's a new version, 16.1... and it fixes a number of issues and adds some really cool features like auto contention slots (which you should enable on all your AP's), and follow me feature, which makes channel changes quick and easy as SM's will follow/jump to the new channel nearly immediatly.


Thanks, maybe I will upload hat on Monday, hate to make changes on a Friday an hour before I leave for the weekend.

So if I update the 2 APs we have online right now it will auto-update all of the SMs?

@Scott B wrote:

Thanks, maybe I will upload hat on Monday, hate to make changes on a Friday an hour before I leave for the weekend.

So if I update the 2 APs we have online right now it will auto-update all of the SMs?

If you're using CNUT's AP auto update feature... then yes.

I agree with Eric about the AP looking pretty good. The Bandwidth requests at 71% is a little lower than ideal - how many contention slots are set up? Chris posted a link to a forum post on ideal contention slot configuration. 

Like Chris mentioned your issue may just have to do with congestion whenever the AP is loaded with many active SM's all trying to compete for the AP's resources. Did you see their modulation modes drop during the busy times or was it their throughput that dropped? Typically a 900MHz AP in a 7MHz channel with typical 900MHz interference will achieve around 12-16Mbps, so if all 15 are active at once trying to get their full 3/1Mbps then they will certainly see less than their package speeds because they're all competing for a finite resource. Your actual sector throughput may be higher since your LQI stats are pretty good and the average modulations look pretty healthy. 

Also, with the burst rates set to 0 you're allowing max bursts for a duration of the burst allocation, which looks to be set to the default value of 2.5Gb. Burst traffic above the sustained data rate gets lower priority than traffic for/from other SM's that are up to their sustained data rates so I don't think this is a problem, but if you'd like to limit it you can configure that in that page. The burst bucket white paper is attached for your reference. 

1 Like

Thanks, we are going to do some testing wtih the Burst bucket, but agree we need to add a Burst rate.

I attached the radio screenshot for contention slots.

You should update to 16.1 and enable auto contention slots on all your AP's.

You should set your broadcast repeat count to zero, unless you're actually using this, and you would know it if you were as there are very specific use cases for this.

On the client side, personally, we don't like to enable burst bucket rates. We keep these all at zero. We only enable/use the MIR sustained downlink and uplink bandwidth settings. I'd also disable Broadcast/Multicast Data Rate... again, you would know if you were using it... 

After all of this... watch your AP's during peak hours when things are the busiest and watch your frame utilization and bandwidth request success % and see where you're at. Another thing that might help you monitor the situation a little more easy would be to install the "Canopy GUI Enhancer" plugin into Chrome. This allows you a quick and easy way to monitor your AP's and SM bandwidth from the Statistics -> Ethernet or Radio tabs.



My mistake on the description of traffic priorites earlier. Burst traffic exceeding MIR is given the same priority as traffic under MIR so you'll want to cap those burst rates or set the burst allocation to 0 if you dont want burst traffic potentially exacerbating things when congestion occurs. 

The FPGA handles the scheduling and the software handles the MIR. So if SW gives the packet to the FPGA scheduler this is the priority:

Legacy scheduler:

High priority CIR (up to CIR rate for all)

Low priority CIR

Broadcast CIR

High Priority

Low Priority


The FPGA scheduler doesn’t know about burst. Burst is just temporarily higher MIR.


What would this message where all SMs drop off of the AP.

MAC : 0a-00-3e-46-a6-1f Out of Range 09/26/2019 : 12:44:20 UTC : Status : 1 Flag : 812575
MAC : 0a-00-3e-46-ac-d4 Out of Range 09/26/2019 : 12:44:13 UTC : Status : 1 Flag : 1130600
MAC : 0a-00-3e-46-ac-d4 Out of Range 09/26/2019 : 12:44:01 UTC : Status : 1 Flag : 1061825
MAC : 0a-00-3e-46-ac-45 Out of Range 09/26/2019 : 12:43:28 UTC : Status : 1 Flag : 1127975
MAC : 0a-00-3e-46-aa-77 Out of Range 09/26/2019 : 12:42:41 UTC : Status : 1 Flag : 806700
MAC : 0a-00-3e-46-ac-81 Out of Range 09/26/2019 : 12:42:31 UTC : Status : 1 Flag : 859675
MAC : 0a-00-3e-46-ac-d4 Out of Range 09/26/2019 : 12:42:01 UTC : Status : 1 Flag : 789925
MAC : 0a-00-3e-46-ac-45 Out of Range 09/26/2019 : 12:41:54 UTC : Status : 1 Flag : 735875
MAC : 0a-00-3e-46-ac-d4 Out of Range 09/26/2019 : 12:41:45 UTC : Status : 1 Flag : 924875
MAC : 0a-00-3e-46-a6-1f Out of Range 09/26/2019 : 12:41:43 UTC : Status : 1 Flag : 1136050
MAC : 0a-00-3e-46-ac-45 Out of Range 09/26/2019 : 12:41:34 UTC : Status : 1 Flag : 802650
MAC : 0a-00-3e-46-ac-d4 Out of Range 09/26/2019 : 12:39:20 UTC : Status : 1 Flag : 1384350
MAC : 0a-00-3e-46-ac-45 Out of Range 09/26/2019 : 12:38:48 UTC : Status : 1 Flag : 1382400
MAC : 0a-00-3e-46-af-e7 Out of Range 09/26/2019 : 12:37:18 UTC : Status : 1 Flag : 717800
MAC : 0a-00-3e-46-ac-45 Out of Range 09/26/2019 : 12:35:53 UTC : Status : 1 Flag : 1316575
MAC : 0a-00-3e-46-aa-77 Out of Range 09/26/2019 : 12:35:45 UTC : Status : 1 Flag : 1315975
MAC : 0a-00-3e-46-ac-69 Out of Range 09/26/2019 : 12:35:32 UTC : Status : 1 Flag : 1436675
MAC : 0a-00-3e-46-a6-db Out of Range 09/26/2019 : 12:35:20 UTC : Status : 1 Flag : 1375000
MAC : 0a-00-3e-46-ab-ae Out of Range 09/26/2019 : 12:35:12 UTC : Status : 1 Flag : 1245475
MAC : 0a-00-3e-46-ab-ae Out of Range 09/26/2019 : 12:34:54 UTC : Status : 1 Flag : 1433750

Range is set to 2miles.  All SMs are within 1 Mile of the AP.

Eric, reading this I am curious where the Auto Contention slots can be enable on the AP?

Yesterday I added 2 more sectors to one of our sites and one of the Sectors is hearing a lot of noise across the 900 spectrum.  We are also seeing sporadic high frame utilization on our existing sector that was in place for testing.  

We are using a CMM5 for timing and sync.   

I am also curious on the Max Range if they all need to be set the same for timing.

I am not sure how I missed this post.  This is great.  We had the broadcast set to 2, and I know we are not using it.

@Scott B wrote:

Eric, reading this I am curious where the Auto Contention slots can be enable on the AP?

I am also curious on the Max Range if they all need to be set the same for timing.

To use auto contention slots you'll need to have 16.1 on the AP. This setting can be enabled on the AP's radio settings tab. Once you set auto, you'll want to make sure that other AP's are set to auto, and at this point, the contention slot number is no longer relevant.

Max range is one of three important settings in regards to sync... the other two being downlink data % and contention slots. Typically these three settings should match across all sync'd AP's. If you find yourself in a situation where you need to deviate from this rule, you can use the frame calculators built in the radio OR available in Cambium's software section to calculate the proper settings.

Also in regards to Max range... you'll want to set it to as low a value as possible that is within your target coverage area... this setting has a (small) impact on overall performance the larger the value you use.

1 Like

Thanks Eric, that explains a lot.  We thought if they were all the same the APs would transmit and receive at the same time which would vastly help with timing.  


Are you unable to get better speeds to end-users than 3m/1m? I want to deploy 900 PtMP for some NLOS applications (half-mile) but would want to get at least 20m to end-users.

The 450 equipment is certainly capable of speeds up to and including 100 Mbps of throughput. The real constraint with 900 MHz is the spectrum availability. If there is NOTHING in the area, there is only 26 MHz of space available (902-928 MHz). Therefore, if you have one isolated sector, you can operate in a 20 MHz channel, and push up to 125 Mbps or so of aggregate throughput.

This is not the typical use case for this product, however. In a lot of areas, you might have 10 MHz or less channel bandwidth available for use.

If you would like to get 360 degrees of coverage at a specific site, the 450 system can operate as a 3-sector (120 degrees) or 4-sector (90 degree sector) site. Doing this with limited spectrum will result in site capacity of up to 90 Mbps or so (3 sectors * 7 MHz channels = 90 Mbps), or up to 200 Mbps (4 sectors * 10 MHz channels (employing frequency re-use N=2, ABAB) = 200 Mbps).

So, results can vary quite a bit depending on what else is out there...