PMP 450 Ask the Experts May 13

Will a current 450 5ghz sm be able to talk to the new 450i in the 5.4 and 5.8 bands?


@Sean Heskett wrote:
Will a current 450 5ghz sm be able to talk to the new 450i in the 5.4 and 5.8 bands?

Yes, it absolutely will.  However, this feature won't be there at launch. Interoperability with PMP 450 will be implemented in Release 14.1, which is expected within a few months of the PMP 450i release.

When will I be able to sync PMP450 ap's and ePMP ap's?


@Kevin Simms wrote:

When will I be able to sync PMP450 ap's and ePMP ap's?


Hi, 

There is ongoing work to allow ePMP to sync with PMP 100s. This is planned for a late summer release.

For PMP 450, there are additional technical challenges we have to overcome on ePMP to sync. We have not planned work on this yet.

Thanks,

Sriram 


@adammoffett wrote:

Ok thanks.  I want to hammer on the NLOS question a little more because right at the moment it's near and dear to me.

Has the 3.65ghz PMP450 been compared to 3.65ghz WiMax products for NLOS performance, and if so, how did it compare?  Assuming we're using the (current) legal FCC limit of +40dbm EIRP, it seems the 450 should be able to match the output of a wimax BS.  Meanwhile the SM+reflector has 6-7dbi more gain that a typical WiMax CPE.  On the surface, it sounds like the 450 should be better than wimax because we've got the same output from the AP, and more receive gain at the CPE, but I've got wimax (and one LTE) vendors telling me that their product is better for NLOS.

I'm wondering if Cambium has done it's own testing on this topic.


In the scenario above, in LOS conditions PMP 450 will perform better than PMP320 because of the higher link budget. Tests done by Cambium show a better range for some NLOS PMP 320 deployments, but a consistent better throughput for PMP 450 deployments in the same conditions.

Thanks everyone! The development team will be back at 9PM tonight Central time, but feel free to post your questions here throughout the day.

My Questions are for the PMP 450 900mhz:

- How well will the AP and SM cope with interferiance (both self and exsternal)?

- Will the AP and SM have inbuilt interferance filters or filtering?

- Will there be some kind of GPS sync allowing for frequiency re-use in 4 and 6 gang configuration?

- Willl there be  MIMO and in what conguration?

- What antenna selection wll be avalible for the PMP450 900, both for SM and AP?

Lastly, please dont bring back that god alful reflector dish... The Fore 110 was a great step forward.

Thanks.

Hello,

I have three questions:

First: The new frame utiliization statistics in 13.4 build 7 is it valid/accurate to use them as a rough guide to capacity/use on an AP. IE: If the downlink frame utilization is 80% it would be a good rough estimate it's roughly 80% used if the remaining customers/new customers meet the same average modulations/specs of the existing customers.  Up till now we've been determinening AP's capacity through the planning guide and a dart board (to guess average customer modulation per data bit).

Second: The high priority vs low priority queue. If I have two SM's one configured with Sustained Downlink Data Rate of 6000 and the Low Priority Downlink CIR set to the default 0 and another SM with Sustained Downlink Data Rate of 6500 but with a High Priority channel enabled and High Priority Downlink CIR set to 500 and Low Priority Downlink CIR set to 6000 does the AP give priority to the SM with a setting in low priorty uplink/downlink over the SM with 0 in the CIR? Or is the only difference between the SM's at that point that one has a high priority 500 Kbit channel?

Last question: The high priority channel, is there a negative effect to provisioniong larger then needed high priority channels if they are not being fully used? IE: giving an SM that needs 90 Kbits for a VoIP call in DSCP 46 a high priority channel of 100 kbits vs 500 kbits. (Assuming they only use 90 kbits not that they use more then that)

Thanks for taking the time and keep up the good work. Many of changes in 13.4 seem very nice and feature improvements really help the argument for staying with Cambium equipment.

Tim

1 Like

The engineers will be joining us in a few minutes. We will respond to all open questions in the queue, but feel free to post your questions at any time.


@Timothy Alexander wrote:

Hello,

I have three questions:

First: The new frame utiliization statistics in 13.4 build 7 is it valid/accurate to use them as a rough guide to capacity/use on an AP. IE: If the downlink frame utilization is 80% it would be a good rough estimate it's roughly 80% used if the remaining customers/new customers meet the same average modulations/specs of the existing customers.  Up till now we've been determinening AP's capacity through the planning guide and a dart board (to guess average customer modulation per data bit).

Second: The high priority vs low priority queue. If I have two SM's one configured with Sustained Downlink Data Rate of 6000 and the Low Priority Downlink CIR set to the default 0 and another SM with Sustained Downlink Data Rate of 6500 but with a High Priority channel enabled and High Priority Downlink CIR set to 500 and Low Priority Downlink CIR set to 6000 does the AP give priority to the SM with a setting in low priorty uplink/downlink over the SM with 0 in the CIR? Or is the only difference between the SM's at that point that one has a high priority 500 Kbit channel?

Last question: The high priority channel, is there a negative effect to provisioniong larger then needed high priority channels if they are not being fully used? IE: giving an SM that needs 90 Kbits for a VoIP call in DSCP 46 a high priority channel of 100 kbits vs 500 kbits. (Assuming they only use 90 kbits not that they use more then that)

Thanks for taking the time and keep up the good work. Many of changes in 13.4 seem very nice and feature improvements really help the argument for staying with Cambium equipment.

Tim


Hi Tim -

I'll answer your last 2 questions:

In this configuration, yes the AP will give priority to honoring the SM with the CIR configuration up to the configured CIR.   The order of precedence in the scheduler is as follows:

1)  High Priority (HP) CIR

2) Low Priority (LP) CIR

3) Multicast/Broadcast CIR

4) High Priority Data (beyond HP CIR)

5) Low Priority Data (beyond LP CIR)

6) Multicast/Broadcast data (beyond Mcast/Bcast CIR)

The CIR is honored on demand, but if the demand isn't there, then the scheduler moves onto the other priorities.  As far as device to device, it is a round robin scheduler.  

In your scenario, SM2 would be honored with 500 of HP CIR data, then 6000 of LP CIR data, and then, if there is room in the frame, SM1 would get its high priority data scheduled, then the SM2's remaining HP data, then SM1's LP data, then SM2's LP data.   Basically something to that effect.   The important thing to note is because you gave one SM an actual CIR, that will be attempted to be honored before getting to the SM without a CIR configured...  Hope that makes sense!

For your last question:

If the CIR is not filled, meaning there isn't enough data demanded, then the frame will be filled with other data following the priority flow I covered above.  The danger in over-provisioning CIR is that if there is the data demand there, and the scheduler can't honor all of the CIR, then it is a bit unpredictable as to what the data flows will be.  

Thank you for your feedback and questions!!

1 Like

@Timothy Alexander wrote:

Hello,

I have three questions:

First: The new frame utiliization statistics in 13.4 build 7 is it valid/accurate to use them as a rough guide to capacity/use on an AP. IE: If the downlink frame utilization is 80% it would be a good rough estimate it's roughly 80% used if the remaining customers/new customers meet the same average modulations/specs of the existing customers.  Up till now we've been determinening AP's capacity through the planning guide and a dart board (to guess average customer modulation per data bit).

...

Tim


Tim,

Thanks for the questions!

  • Regarding Frame Utilization: Yes, it is absolutely valid and accurate on PMP 450. And your understanding is correct, with the modulation level of the SMs being important. If you’re peaking out at 80% then you only have 20% left, if you have the same modulation level on the added throughput. If another SM comes in at a lower modulation level, it will fill up the frame faster than SM operating at high modulation rates.
    Note we provide 3 levels of monitoring and changing in in 13.4 (Build 9). 1 minute, 5 minutes and 15 minutes.

-Charlie

1 Like

@test123 wrote:

My Questions are for the PMP 450 900mhz:

- How well will the AP and SM cope with interferiance (both self and exsternal)?

- Will the AP and SM have inbuilt interferance filters or filtering?

- Will there be some kind of GPS sync allowing for frequiency re-use in 4 and 6 gang configuration?

- Willl there be  MIMO and in what conguration?

- What antenna selection wll be avalible for the PMP450 900, both for SM and AP?

Lastly, please dont bring back that god alful reflector dish... The Fore 110 was a great step forward.

Thanks.

  • In the 900 MHz band, the planning guidelines to minimize the interference are equivalent to those in any other band. The APs have to be GPS synchronized, and the downlink and uplink frames aligned (this always happens if configuration parameters are the same). Frequency planning guidelines should also be followed like in other bands.
  • The AP and SM do not have inline filters. The AP has a super heterodyne receiver to suppress out of band interference.
  • GPS synchronization is available. Frequency reuse is possible in back-to-back sectors.
  • Both MIMO-A (added robustness) and MIMO-B (higher throughput) are available in the 900 MHz product. The radio dynamically adapts the modulation according to varying channel conditions.
  • Both the AP and the SM are connectorized (no dish). For the AP, we are planning to have a 60⁰ sector and a 90⁰ sector options.

1 Like

Thank you to everyone who participated in today's session, and a special thanks to the PMP 450 development team. Feel free to continue to post questions to this thread or recommend it to your colleagues.

Our next Ask the Experts session will be Wednesday 20 May, when we will have three WISPs from North America answer any questions you may have. Details on that session are HERE.


@CambiumMatt wrote:

@David Hannum wrote:

Questions about PMP450 900MHz

1)  What will the AP and SM total throughput be?

2)  What is the suggested upgrade plan (co-exist with PMP100 gear???)

3)  Any chance it can be backward compatible with PMP100 (450 was touted to do that when first announced)

~ And last but not least ~

4)  Is Q4 avability still possible?


1) First release of 900 MHz will support 5, 7 and 10 MHz channel bandwidths that will have the same throughput of other PMP 450 systems. If you have clean spectrum in a 10 MHz channel, you can get more than 50 Mbps.

2) We will publish a migration plan prior to launch.

3) As stated, we are not planning compatibility with PMP 100, but "co-location" via synchronization will be capable, and necessary for effective migration.

4) Absolutely, the launch is planned for Q4 of this year!


Sorry, but if #2 & #3 work as well on the 900MHz as the 3.65GHz does migrating from/colocating with PMP320 then we won't waste our time.  Our test with a PMP450 in the middle of a PMP320 network on 3.65 has been a disaster with no useful solutions.

j

If the PMP 450 900mhz does not have amazing NLOS thoughput, then I think that will be the final nail in the Cambium Networks coffin.

Since most WISP's and Business customers will move to Telrad LTE for its better NLOS and throughput capability.


@test123 wrote:

If the PMP 450 900mhz does not have amazing NLOS thoughput, then I think that will be the final nail in the Cambium Networks coffin.

Since most WISP's and Business customers will move to Telrad LTE for its better NLOS and throughput capability.


Sure it has better throughput? It does only do qam64 at the moment? And the uplink is even lower? So with LOS it is slower at much higher cost.


@test123 wrote:

If the PMP 450 900mhz does not have amazing NLOS thoughput, then I think that will be the final nail in the Cambium Networks coffin.

Since most WISP's and Business customers will move to Telrad LTE for its better NLOS and throughput capability.


Are you thinking Telrad LTE is great because you tried it and found it to be great?  Or do you think it's great because that's what they told you?  I moved from a Cambium shop to another company running Alvarion/Telrad.  I would give my left arm to have the reliability and mangaeability of Cambium back.

The current state of that product does not have better throughput than the 450....and whether NLOS is better is arguable.  And if NLOS is better, it's the only way in which that product is better.  The rest of what you've heard about its capabilities is stuff they think they're going to do in the future. 

If I go on about it, I risk ranting.

2 Likes

The issue is that the NLoS performance of the 450 is poor.  Then it doesn't matter what the throughput capabilities of the 450 product is since the installation cost becomes to prohibited to install the customer in the first place. 

I currently run PMP FSK900, PMP 320 and 450.  Any LTE platform is going to have the same or likely better NLoS capabilities then the 320.  We heavily invested into the 320 platform, replacing our FSK900 network with it.  We where successful 90% of the time in migrating FSK900 to 320.  Granted to obtain this level of success, ruffly 40% of the time a modification of the original installation was required.   Some examples would be moving the mounting location to a cleaner shot or installing a tripod to clear the majority of the tree canopy. 

There is no real comparison in NLoS between the 450 platform and the 320.  The 320 wins hands down in nearly every situation.  I suspect it is mainly due to 320's 1024fft (subcarriers) vs the 450's 512.  Along with other error correction mechanisms such as HARQ.  Also the 450 is not able to operate at lower signal CINR, S/N ratios with the same reliability of the 320 platform.  This is again likely due to the lower subcarrier count. 

We are continuing to install the 320 platform on new customers and have regulated the 450 deployment to situation dependent or ruffly 5% of new installations.  Unless the 450i platform is going to significantly improve NLoS capabilities over the current 450 platform then it is a non starter.  And 450 in a 900mhz platform is a band-aid not a solution due to limited spectrum and the noise level in that band.  And now it will have to share that band with our existing FSK system for migration is nearly impossible.  Again due to the limited spectrum availability. 

I hope my experience is the odd ball and others are seeing the opposite.  Has anyone seen results contradictory to mine?

This debate between PMP320 non-LOS and PMP450 non-LOS coverage has been going on for awhile. There was a pretty good thread regarding this where I outlined an apples to apples comparison of PMP320 vs PMP450.

From a technical perspective, assuming that you're running legal EIRP's on everything, the PMP320 should only have slightly better non-LOS performance due to the items mentioned by anthonyw... e.g. HARQ, and more sub-carriers. Aside from that EVERYTHING else is better with the PMP450... much much better throughput, and much lower latency, the user interface and manageability, the future looks far brighter as well.

On the Telrad side... something that a lot of people don't know is the fact that the  LTE build they're distributing is only a 2x2 implementation, and it's only 64QAM ATM... which again, side by side comparison, if you're running legal EIRP's... the PMP450 overall beats the Telrad platform when you factor in entry price, ease of management, and overall sector capacity (e.g 256QAM).  By the time that Telrad implements 4x4 and/or 256QAM.... I'd wager that Cambium will respond with the next generation platform that will not only build on the PMP450's success, but surpass the competition.

4 Likes

@paulchops wrote:

On the PMP 450....  we are attempting to read the value that the SMs are set to as far as 1x/2x/4x/6x/8x using OID

1.3.6.1.4.1.161.19.3.3.2.85.   It returns no such object.  The notes say "Engineering Use Only".   

We can rate the actual connected rate using: 1.3.6.1.4.1.161.19.3.2.2.20.  We use both of these variables in our custom analysis tool.

Is there a way to get the value (setting in the radio for Adaptrate) on PMP 450?  if not today, is is coming?

With the 100 series, platform we can do this without problem.

Paul McCall, PDMNet / Florida Broadband


Paul,

I will open an issue for this.  Thanks for letting us know!

Best,

Cambium Jonathan