What Chris said. Our 450Ms have been a huge game changer for us. We have a tower with 300 people on it, pretty much all on one side (180 degrees is busy and the other 180 degrees is pretty silent). In the past we’d add an AP to offload capacity and that AP was used up day one. There just wasn’t a path to catch up with demand much less start to offer faster speeds. If there is any kind of density, greater than 100+ clients on the tower, I’m not sure how you use anything other than 450m at this point.
Joe > On Jul 18, 2017, at 11:54 AM, Chris Wright <[email protected]> wrote: > > ePMP might make sense to build in a new area, but when you have 100+ clients > in a single 90 degree sector, replacing your triple-stacked PMP450 APs with a > single 450M, saving 40MHz of spectrum, offering faster speeds, AND not having > to swap 100+ radios feels mighty nice. > > Chris Wright > Network Administrator > > From: Af [mailto:[email protected] <mailto:[email protected]>] On > Behalf Of Adam Moffett > Sent: Tuesday, July 18, 2017 8:58 AM > To: [email protected] <mailto:[email protected]> > Subject: Re: [AFMUG] PMP450 ~50mbps cap > > I've had ePMP's do some weird things here and there, like refusing to reboot > or accept config changes. Nothing a power cycle didn't fix. > I've had to power cycle a PMP100 or 450 to fix a problem approximately zero > times. > > I know a neighbor who had a tower hit by lightning an ePMP and a Ubiquiti > Rocket died while the PMP100 and 430 stuff kept on chugging (he didn't have > 450 there yet). > > 450 has a few management conveniences like remote spectrum analyzer, RF > private IP, and SM proxy access via AP. > > So overall my experiences say the 450 is better than the ePMP, but I still > use a lot of ePMP for all the reasons others have stated. ePMP is good bang > for the buck. > > -Adam > > > ------ Original Message ------ > From: "Mathew Howard" <[email protected] <mailto:[email protected]>> > To: "af" <[email protected] <mailto:[email protected]>> > Sent: 7/18/2017 10:23:17 AM > Subject: Re: [AFMUG] PMP450 ~50mbps cap > >> It seems to me (as someone who hasn't actually used PMP450 to speak of, >> other than 900mhz), that PMP450 has some advantages for high density >> deployments... particularly if you're talking 450m, or even the ability to >> easily upgrade to 450m. But in a network like ours, where the average SMs >> per AP is somewhere around 15, I just can't see any way that it could >> possibly be worth going with 450 over ePMP. >> >> On Tue, Jul 18, 2017 at 8:52 AM, Bill Prince <[email protected] >> <mailto:[email protected]>> wrote: >> The PMP450M can make ~~ 12° sectors. On a busy tower, and surrounded by >> noisy neighbors, we're often seeing 30 dBm SNR. Nothing else comes close. >> >> bp >> <part15sbs{at}gmail{dot}com> >> >> On 7/17/2017 10:10 PM, Steve Jones wrote: >>> Serious question, not my usual sarcasm, how are you 450 folks justifying >>> the substantial proce difference between 450 and epmp, seriously, if you >>> dont freeze in the winter, i couldnt justify it. Granted we dont sellbover >>> 12\2 and we dont have more than 40 per ap, i just dont see the value, they >>> did too good a job on epmp >>> >>> On Jul 3, 2017 12:10 PM, "Craig Schmaderer" <[email protected] >>> <mailto:[email protected]>> wrote: >>> I have never seen more than around 55mbps on 450sm in bridge mode with any >>> firmware. Maybe I missed a 14.x that it did, but I haven't seen more than >>> 55mbps on any 15.x firmware. I am working on a bug with then on 15.1 where >>> it looks like the qos speed limiters are not enforcing speed settings. I >>> have never tried a 450i sm but I would assume those work fine, I have many >>> 450i PTP that work great. >>> >>> From: Af <[email protected] <mailto:[email protected]>> on behalf of >>> Kurt Fankhauser <[email protected] <mailto:[email protected]>> >>> Sent: Monday, July 3, 2017 8:38:49 AM >>> >>> To: [email protected] <mailto:[email protected]> >>> Subject: Re: [AFMUG] PMP450 ~50mbps cap >>> >>> The original 450 hardware will max out at around 70mbps TCP. Even if you >>> have no traffic on it that is the limitation. I don't remember which >>> firmware it was but the 14.0 something sounds right and I have personally >>> gotten the 70mbps to a SM on a AP with 3 clients. >>> >>> Now the 450i AP basically the 70mbps cap is gone and you can get whatever >>> the link tests show (as long as your testing to an 450i SM). I have two >>> customers with 450i SM on a 450i AP and have seen well over 125mbps TCP >>> easy. Now that same AP talking to older 450SM's those SM's can still only >>> get 70mbps max. >>> >>> So basically what you need to do is put the 450i AP up where you need to >>> total AP capacity of more than 50-60mbps and then only use the 450i SM on >>> the clients that need more than the 50-70mbps. >>> >>> I have not tested speeds since I upgraded firmware to 15.1 so if a bug was >>> recently introduced I know nothing of it. >>> >>> On Sat, Jul 1, 2017 at 7:51 PM, George Skorup <[email protected] >>> <mailto:[email protected]>> wrote: >>> Aaron demonstrated some throughput improvements with both NAT and bridge >>> mode back around 14.2 development. IIRC, they were able to get a little >>> over 70Mbps downlink in bridge mode on a standard 450 SM. Real TCP >>> throughput, not a link test. >>> >>> What I was seeing on some of our most heavily loaded 450 sectors is the AP >>> seemed to max out around 55Mbps downlink and 10-11k PPS. In that case I'm >>> thinking the issue was simply the PPS limit and not that the AP was limited >>> to 55Mbps. >>> >>> As far as the SM, I don't know. Perhaps a regression with 15.x. 15.0.x did >>> have some high-priority issues. >>> >>> On 7/1/2017 2:25 PM, Craig Schmaderer wrote: >>>> Ok, I have tried to get cambium to publicly admit to this ever since I >>>> started using 450 a few years ago. The current sm in no way can pass more >>>> than about 50-55 mbs of real traffic. This has never been a real issue >>>> however because I don't do those kind of speed plans. I have been waiting >>>> for their new sm design to take care of this. If anyone disagrees with me >>>> I would love to chat. (Really) >>>> >>>> From: Af <[email protected]> <mailto:[email protected]> on behalf of >>>> George Skorup <[email protected]> <mailto:[email protected]> >>>> Sent: Wednesday, June 28, 2017 8:01:22 PM >>>> To: [email protected] <mailto:[email protected]> >>>> Subject: Re: [AFMUG] PMP450 ~50mbps cap >>>> >>>> 14.2.1 has its own issues, mostly the high-priority bugs, but I thought >>>> that would achieve 60+? >>>> >>>> I guess someone could try a 450i SM on 15.1 and see if it's any better. >>>> That should rule out or confirm the CPU suspicion. >>>> >>>> On 6/28/2017 6:53 PM, Ryan Ray wrote: >>>>> That's about all I get on a 450 as well. I think it's CPU bound. >>>>> >>>>> On Wed, Jun 28, 2017 at 3:18 PM, Eric Muehleisen <[email protected] >>>>> <mailto:[email protected]>> wrote: >>>>>> See attached. This is a 20mb SM. Again, it will burst up to 52 mb >>>>>> <image001.jpg> >>>>>> >>>>>> >>>>>> On Wed, Jun 28, 2017 at 4:42 PM, Chris Wright <[email protected] >>>>>> <mailto:[email protected]>> wrote: >>>>>> What QoS settings in the SM? >>>>>> >>>>>> Chris Wright >>>>>> Network Administrator >>>>>> >>>>>> From: Af [mailto:[email protected] <mailto:[email protected]>] On >>>>>> Behalf Of Eric Muehleisen >>>>>> Sent: Wednesday, June 28, 2017 1:36 PM >>>>>> To: [email protected] <mailto:[email protected]> >>>>>> Subject: Re: [AFMUG] PMP450 ~50mbps cap >>>>>> >>>>>> Yes. Just yesterday tested a 450i AP (40mhz) with a 20mb 450 SM. >>>>>> Linktests showed 137 x 58 but most we could burst to was 52 x 45. Both >>>>>> AP and SM are on 15.1. >>>>>> >>>>>> I currently have a support ticket open for this as well as increased DFS >>>>>> issues. >>>>>> >>>>>> On Wed, Jun 28, 2017 at 3:22 PM, Colin Stanners <[email protected] >>>>>> <mailto:[email protected]>> wrote: >>>>>> We setup a customer on a very lightly loaded 5ghz PMP450 AP and SM, >>>>>> 20mhz channel, expecting them to burst up to 70mbit speeds for download; >>>>>> linktest is reliably 85-90d and there's effectively no other usage. But >>>>>> we can't seem to get over 50mbit speedtests. Has anyone else seen such >>>>>> issues?
