Are ePMP GPS issues resolved?

Where and when the Cambium Network announces that the GPS problem has been resolved.
On several occasions I get information on the comunity that it has been resolved but officially not.

I believe they squashed that in 4.5.5, but I’m not sure exactly. We are running 4.6 on pretty much everything.

Resolved Issues in 4.5.5
System Release 4.5.5 resolved issues
Resolved regression in 4.5.4 causing ePMP3000L to lose GPS sync.

Or maybe it is still an open issue they are still working on? I don’t know for sure. I’m just saying at some point, I remember them saying “OH, we found a bug in bus/memory/something” and “thats what was causing GPS issues”, and for us, we have upgraded everything to the latest firmware, and we haven’t had any issues since they fixed that bug.

There are also a bunch of people reporting that their ongoing 3000L issues cleared up when they attached external puck antennas. Again - I’m not certain that resolved 100% of them or not, but I do remember a bunch of people saying that they put pucks on their 3000L’s and they have been right as rain since then.

1 Like

All but one of my 3000L units have pucks installed except for one. They all have GPS issues - some of them multiple times per day, some of them only a few per month. There doesn’t seem to be any rhyme or reason to it at this point.

Running 4.5.6 on most of them and the latest 4.6 RC on one or two.

4.6Beta35 resolved my GPS problems without any pucks. I had two access point 3000L that would lose GPS signals every other day. Since installation of the BETA not one loss of GPS


4.6-rc35 solved the gps issue for us. It also solved the random nat lockup issues that plagued us.

1 Like

Why doesn’t someone from Cambium officially announce on the page that the GPS problems have been resolved. The problem with the GPS caused me a lot of damage, I regularly take tranquilizers. ISP business on the verge of bankruptcy.
I made a terrible business mistake by migrating from Mikrotik to Cambium.

1 Like

Hi Zica. I’m not sure if they 100% know yet or not? Maybe they do – difficult to say, since they haven’t responded to you. However, if I had to imagine, I’d say that as a programmer, I know that a ‘bug’ can be for a dozen different reasons. So, sometimes a programmer ‘thinks’ they’ve squashed the bug, but there may be multiple reasons for it. SO - I can’t speak for Cambium - but I can tell you that for us, when we moved to 4.6-RC35 at the beginning of December, we haven’t had a single GPS issue that I’m aware of on any AP’s since going to 4.6.

I feel your pain. For us, we haven’t been badly affected by GPS issues, but I also feel like things aren’t all what we were told to expect in Cambium world. So - my advice would be to do what we did. Make your needs clear and concise to Cambium, and if they choose not to accommodate those needs - move to a different brand.

I mean - if you’re facing bankruptcy - and if Cambium is unable or unwilling to provide what you need, then you have a clear path forward. You’ll unfortunately have to do what we did, and move to other brands of gear that does provide what you need.

Have you tried version 4.6 on the AP & all SM’s? There is nothing definite in the 4.6 Release Notes either – but there is this one mention of ‘GPS sync stability improvements’

SO - in my opinion, running 4.6 is your best chance for GPS on 3000 series AP’s. My advice would be to try that, and then report back if that helps or not.

Thanks for the info. I have never digested a single beta! Cambium distanced himself from the problems that such versions could cause.
I have an AP2000 I didn’t embark on buying an AP3000 because I started thinking about going back to Mikrotik.
I will try version 4.6 and no one would be happier than me to solve my problem.
Thanks again.

I suggest this too, 4.x is only required if you have a 3000/2000/1000 mixed network. for 1000/2000 gear, 3.5.6 is the best. I held my ptp links back to 3.5.6 until 4.6 because of issues between my 1000 ptps and my 3000L ptps.

as for GPS issues, ran 4.6-rc35 since it came out and havent had a single GPS based outage on the 3000Ls (no sync over power) and haven’t had force300s dropping of 1000 APs for odd reasons despite good signal and SNR either.

GPS is still not ok.
And that problem has existed since the Cambium ePMP existed.
Normally 15 satellites are heard and without any sense none.
Without touching the GPS antenna or reset after two or three days he hears a large number again, then gps sync starts working
During that time, not a single satellite is heard even with 2 to 5 dB SNR.
How, where are they? Isn’t there at least one up there in the next minute, hour or day?
I had to turn off GPS sync and continue without it.
And what is my non-GPS sync benefit compared to competing devices?
Nothing but outrageously more expensive.
When will one of the engineers from the cambium explain what the purpose of the internal and which is the external antenna for GPS sync. Two antennas for receiving GPS signals and so many problems.
And why GPS sync with GPS Ubiquiti devices has no incidents like Cambium.
I’m not a manager from Ubiquiti but I gave an example and I got angry.
I gave a huge amount of money for something that doesn’t work as presented .

I talked about this in a previous post. It monitors 18 satellites and SNR on all zeros. It’s impossible to constantly get a catastrophic failure in writing GPS firmware.
In an hour, a day or two, the SNR will go over 30 dB at least 10 and get GPS sync. Without any movement or replacement of the GPS antenna.
A reboot of the device would also restore GPS sync, but my clients are already so angry that they don’t forgive me for that.

Two hours later and sometimes it lasts up to three days, the same GPS tracks a large number of satellites with an excellent signal.

The GPS chip that Cambium uses has a ceramic resonator antenna as part of the package. So in theory it can work without an external antenna. We have tested this and as you observed it is not reliable for this application. By adding the external antenna and reading about this antenna from the manufacturer, we implemented a virticle mounting policy. This means we mount to virticle surfaces only, cable pointing down and for us only facing south.
This allows our epmp3000L radios to perform properly and the do not loose sync.
We still have packetflux timing for our 1000s and 1000-lite APs despite having the internal gps as backup.

Make sure you are not using the round gps puck, only the square ones.

Screenshots from my previous post say that there is a terrible omission in writing the firmware for GPS It is impossible to hear it detects a large number of satellites and the SNR is zero. How is it even possible to hear something that is heard on the lita and the signal does not exist. Funny, isn’t it.
All GPS packs are square and I turned them off as a potential problem, I transferred them to Ubiquiti airFiber links, they work there by
several years without loss of GPS.

Snr of 0 isnt as bad as you think. Thats why they provide the rx signal strength too.
Think of the satellite showing up at snr 0 as the same as listening to someone speak in a very noisy environment, since you can hear the other person a bit does not mean that you can hear them above the noise by an appreciable amount. But you can hear enough to know that they are there.
Also it takes 4 satellites with an snr of 6 to make the fix needed for gps timing as per the firmware. Ofcourse if you have some LTE cell towers near (relative term, think if your cell phone has signal in lte mode then its near) then you can desense the gps antennas (both internal and external) and thus not get gps timing pulses. We have one tower where we had to make a metal box open on one side for the gps antenna to prevent the new cell tower from mangling the gps signal at the antenna. The box needs to be twice the width of the antenna, twice the height and no more than about 2 inches deep. This prevents the box from hindering the gps signal. We just used 16ga steel and havent lost gps sync on that tower since.

Hello Douglas …

Please don’t misunderstand me.
I don’t need lectures in electronics and radio engineering. I am 51 years old and I have been with this subject since the time when many who are help desk in Cambium were not even born.
Additional armor adding a mesh or steel box is unnecessary if the firmware is written properly and is not.
The same ceramic antenna active and other passive components that were mostly made by the Chinese are used by everyone.
The topology is the same because it is done as the component manufacturer suggested, why change anything.
Why doesn’t the GPS crash on the Ubiquiti device.
Not because someone knew how to make firmware properly.

Best regards

Though I am not yet in my 50’s do not discount what others say, we come froma wide variety of backgrounds and may have an idea or two about the problems.

The GPS chip in question comes with or without a ceramic resonator cavity antenna. The cost difference between the two is literally a couple of pennies. Cambium chose to provide an option of not using an external antenna but include an antenna in the box (except for the e3kL). For what its worth, the sync over power has been solid without issue and the onboard GPS is a toy that is nice but really not necessary and could be dropped in favor of using only sync over power but the point is moot.
The GPS didnt crash until Cambium trued to make sure that the N based devices didnt get left behind on the upgrade path to AC making sure you didnt have to run multiple generations of APs on a tower (which by the way UBNT devices still have problems with) saving you money and costing Cambium sales revenue. You dont have to forklift your network just to get more throughput to a sector not a whole tower as you can upgrade just the sector you need to. Again costing Cambium revenues that UBNT forces you into. Yes Cambium is more expensive up front but you also get support from Cambium, just try to call UBNT and get a person to help you in a meaningful way including unlocking a shell and letting you see whats under the hood while fixing a problem.
The chip manufacturers plant location is simply a governmental issue that drove companies to place those facilities where it was economical. You want chips made locally, then pressure your authorities to grant incentives to manufacturers. Again moot point.

Compairing UBNT and Cambium technologies and calling them the same is like calling an apple an orange. They are based on the same standards and use compairable chip sets but they are not the same. Open a few up and compare them and you will understand.

Firmware stability? Really compairing UBNT to anyone regarding stability is an industry joke. I have an UBNT link that has more down time than a politician! And yes one bad egg shouldnt spoil the batch but isnt that what your claiming?
Building of firmware, well since both devices run a stripped down flavour of linux and add the features of choice it is really an implementation choice. Do I use dnsmasq or bind? Do I use isc-dhcp or udhcp? Apache or nginx? J-node or just use http calls? Also consider the base flavour and its code base, run a couple revisions behind but is really stable or run bleeding edge? There is also the mentality of one root and never upgrade until no other choice path.
There are really too many options to try and say one is the best and how they interact with each other can be both frustrating and amusing.

Before you claim I am blind to the faults of the epmp series, I am not and Fedor can confirm that I have no problem pointing them out. But I also provide my insight as to what could be done about it as well and not always in open forum. But I am not here to fight with you.

We are not here to fight but to discuss in the name of progress.
At least I want to resent anyone.
And I don’t want anyone to insult me ​​either.
My personal position and I do not want to lead anyone to make the wrong decision.
The problem with GPS has been there since ePMP.
And no additional nets and steel boxes are needed, this can be solved by neatly writing the firmware.
If nets and these steel boxes were needed, Cambium would sell it properly as its product and it would cost at least $ 100 per piece.
I begged someone to look at that GPS.
A colleague talked for days with support from Cambium regarding the GPS problem. Once he told me I’m giving up, which firmware, which antenna is south
do you have a GSM, TV or Radio transmitter on the tower is it round or square etc.
Have you tried to move the antenna.
But the firmwre team made sure the web looked a little different (very important)
And version 4.6.0 comes out when someone there has a problem with the characters%, $, #, & / and the like.
Version is coming out urgently!
Is that the support we are talking about.
Why did anyone even touch that?
When there are more important problems.
Cambium inherited from Motorola a reputation that has seduced many like me and we will see how it continues.

Best regards


Far from being fixed. This has been an ongoing issue for years now.

Everything Cambium has done in terms of fixing GPS problems is ridiculous.
Older firmware versions had synchronization for up to 68400 seconds or 1 day.
Someone raised it to 684000 seconds or 10 days and it is considered a solved problem.
The solution to the problem of dying of laughter.
Best regards.