GPS Sync - I'm tired

Please contact our support team at support@cambiumnetworks.com. They will be able to help?

What size network are you operating?

are any of the big cell providers doing work in the area?  

the GPS chips can be overwhelmed by some of the LTE gear.

if that is the case, you'll need to use ground GPS.    the CMM or the packet flux solution. 

the ground base GPS circuits are able to deal with the interference the onboard GPS chip can't.

in our area, when we co-locate with verizon we have to use ground based GPS for stable results.  sprint and ATT don't mess with our topside GPS here. 

1 Like

Chris,

Does CMM and packetflux give you the same  timing as the onboard gps?

Packet flux and cmm units can all be intermixed with the gps pucks to maintain snyc. I’d suggest on the same time source all timing from the same point. From tower to tower it doesn’t matter. By specifications they can be mixed on the same site, but I chose to avoid it. Just another peace of a puzzle if something isn’t working smooth.


@Luis wrote:

Hello Matt,

Do you have any cellular towers nearby your ePMP equipment? Some ePMP users have complaint in the past about potential interference from cell towers.

http://community.cambiumnetworks.com/t5/ePMP-2000-and-1000/Random-Loss-of-GPS-Sync/m-p/51193/highlight/true#M5413

Increasing your holdoff timer may alleviate the situation.

Also, out of curiosity, are you using the old ePMP1000 GPS antenna with the ePMP2000 or are you using the GPS antenna included with the ePMP2000 radio? The new antennas have better filtering/gain than the old one, plus they will allow the GPS chip to also use GLONASS satellites for tracking.

Regards


@Luis, do you have a partnumber for that new antenna?   google and I aren't getting along so well with this. 


@Chris_Bay wrote:

@Luis wrote:

Hello Matt,

Do you have any cellular towers nearby your ePMP equipment? Some ePMP users have complaint in the past about potential interference from cell towers.

http://community.cambiumnetworks.com/t5/ePMP-2000-and-1000/Random-Loss-of-GPS-Sync/m-p/51193/highlight/true#M5413

Increasing your holdoff timer may alleviate the situation.

Also, out of curiosity, are you using the old ePMP1000 GPS antenna with the ePMP2000 or are you using the GPS antenna included with the ePMP2000 radio? The new antennas have better filtering/gain than the old one, plus they will allow the GPS chip to also use GLONASS satellites for tracking.

Regards


@Luis, do you have a partnumber for that new antenna?   google and I aren't getting along so well with this. 


Is this information you need?


@Cheree King wrote:

@Chris_Bay wrote:

@Luis wrote:

Hello Matt,

Do you have any cellular towers nearby your ePMP equipment? Some ePMP users have complaint in the past about potential interference from cell towers.

http://community.cambiumnetworks.com/t5/ePMP-2000-and-1000/Random-Loss-of-GPS-Sync/m-p/51193/highlight/true#M5413

Increasing your holdoff timer may alleviate the situation.

Also, out of curiosity, are you using the old ePMP1000 GPS antenna with the ePMP2000 or are you using the GPS antenna included with the ePMP2000 radio? The new antennas have better filtering/gain than the old one, plus they will allow the GPS chip to also use GLONASS satellites for tracking.

Regards


@Luis, do you have a partnumber for that new antenna?   google and I aren't getting along so well with this. 


Is this information you need?


yes sir,  ATT has been upgrading their gear here which has been causing gps problems for our 1000 gear. going to switch a site out to these antennas to see if it helps.   I can't find where to buy these for spares?   plenty of the old round GPS pucks around though. 

We did(and still have) canopy with packetflux for years and I don't remember ever having problems with GPS sync... really I can't remember a single instance were GPS failed in the last 10 years on a packetflux/canopy, 450 or even Airfiber. Since we already had canopy on most of the towers we put ePMP on we just went ahead and used packetflux to power the ePMP but used the pucks for gps. It seems like once very few months an ePMP radio will lose GPS sync, sometimes even though it showed it could see a 20 sats and was tracking 18 of them it wouldn't have GPS sync.  Other times it would lose the sats and of course lose GPS sync.  So on the towers were we already had them powered with SyncInjectors I just changed them over to sync over power. 

I have 3 towers that never had Canopy on them so I don't have packetflux on those towers. All 3 towers have at least 1 ePMP AP or backhaul that I have had to disable sync on because they kept randomly losing it or it just stopped working period.  I am in the process of getting the three towers changed over to packetflux simply so we can get reliable sync on the ePMP stuff.  Luckily 5Ghz don't travel as far as 900Mhz and we aren't doing ABAB on any of the affected towers or losing GPS sync on a radio would be a much bigger problem.

All 3 sites are standpipe water tanks , the radios are on the very top and the pucks have no obstruction in any direction, the entire sky is visible. All the radios see 19'ish sats at -19 to -48 and tracking 18 of them. Two of the 3 ePMP radios that have this problem are radios and have the new style GPS puc, only one is the old style GPS puck.

While the old canopy, new 450i and airfiber give me no GPS problems at all the ePMP AP's have always had random GPS problems.

I ordered a sync pipe for this site to try it out. Can’t get the radios to stay at gig connection, that being said we are pushing cable length here, 99 meters according to the netonix switch. I had horrible luck with the packetflux gear getting hit and taking out mysites. In the last 6 years we’ve had 6 major strikes, all of which was a packetflux injected or timed site (also only makes up about 10% of my network. ) they come off as to much of a hazard to my network for me to swallow. When i called support at packetdlux over tge sync problem remaining st this site, the request was to bypass the LPU, not gonna happen. The gps has been stable for us with the epmp until ATT started changing nodes here. Everything with glonass antennas have been fine.

I also have noticed with both antennas, the orientation of the gps puck matters, a lot. Laying flat here takes 10 to 15 db snr from our signal. Upright so the coax is going towards tge ground from the sectors works the best for us.

Ive already pulled the packetflux gear to return… won’t run gear without full protection anymore. Anyone used a cmm4 and compared them to this problem?

The problem is a Chinese satellite out of its orbital position. I understand that they are already interceding for the owner to correct the problems that are causing.

best regards

Same issues here. I've contacted support, and I get the same reply that's of no help. "Change your Antenna, go test GPS with your cell phone" When I have 10 APs on the same site, and I colocate their pucks in the same orinatation and we will see random loss of GPS across the sites and it will almost never auto recover. Requires a hard reboot of the AP. 

I have noticed our sites that are aginst a hill seem to be impacted the most, but it's never ALL aps at one site. It's only ever one at a time, and they don't recover typically need a hard reboot.


@matkix wrote:

Same issues here. I've contacted support, and I get the same reply that's of no help. "Change your Antenna, go test GPS with your cell phone" When I have 10 APs on the same site, and I colocate their pucks in the same orinatation and we will see random loss of GPS across the sites and it will almost never auto recover. Requires a hard reboot of the AP. 

I have noticed our sites that are aginst a hill seem to be impacted the most, but it's never ALL aps at one site. It's only ever one at a time, and they don't recover typically need a hard reboot.


We had am issue a long time ago simular to what you described, i actually traced it down to one specific AP causing interference to the rrst of our APs. Next time you loose sync,  hard power cycle the APs that ARE getting sync, wait 5 minutes between cycles to see when your def one starts getting sync.  If you can pin it down to one ap, replace that one.  I have no idea how or why that happens, but ive found that one time.  It progressively got quicker and quicker.    The issue the rest of the thread discribes is a interference issue related to being drown out from a carrier, or atleast thats what ive gathered from them. Fighting that problem now myself. 

Funny, I use PacketFlux's gigabit sync-injector, a sync box junior and believe it or not Ubiquity's Eth-SP at my tower sites and haven't lost a single radio, switch or timing unit yet. One tower gets struck so often that it should have melted by now. Burn marks on the top-rod show repeated strikes.  

Now the important question: have you heard of R-56 protection standards? We setup all our sites to this spec that exceeds CEC and the NEC specifications. It might be overkill but I say not loosing equipment is worth the extra investment of time and a small amount of extra protection equipment. Most of the important points simply come down to single grounding point interface, dont create a loop through earth, ground everything at the correct place and remember that a lightning bolt always takes all paths to ground and that the bolt you see may not be the direction of current flow! 


@Douglas Generous wrote:

Funny, I use PacketFlux's gigabit sync-injector, a sync box junior and believe it or not Ubiquity's Eth-SP at my tower sites and haven't lost a single radio, switch or timing unit yet. One tower gets struck so often that it should have melted by now. Burn marks on the top-rod show repeated strikes.  

Now the important question: have you heard of R-56 protection standards? We setup all our sites to this spec that exceeds CEC and the NEC specifications. It might be overkill but I say not loosing equipment is worth the extra investment of time and a small amount of extra protection equipment. Most of the important points simply come down to single grounding point interface, dont create a loop through earth, ground everything at the correct place and remember that a lightning bolt always takes all paths to ground and that the bolt you see may not be the direction of current flow! 


We ground properly.   Ive shared stories here and you can see some of our grounding work.  Our issue was most definitely linked to packet flux gear.  That being said that was mostly pmp100 and 320 gear. I have not used any of thier new gear for any length of time to determine if they have or have not fixed the issues we had with them.   None of thier old stuff had grounding lugs and 100% plastic housings.   I ordered 1 unit to test, and didnt deliver sync reliably to the epmp and 2 radios kept falling to 100 meg, but like i said we are pushing wire length.   Im unwilling to remove any of our LPU for long testing or operation as i stated above. 

I do not know the older equipment, but I do know that the latest revisions do work properly and the syncpipe 2.2 and the syncbox jr  (replaced the syncpipe) work very well with no timing drops. I have epmp1000 gear at 1gbps (force 200) working just fine beside older canopy 100mbps gear through the same syncinjector. I did notice that the epmp grea is more sensitive to cable type (cat5e-300 vs cat5e-350) and termination issues. On one tower we used 300Mhz spec cat5e and had nothing but troubles getting 1gbps to work, on another tower we used 350 spec cat5e and had no troubles at 98m long cables to the syncinjector with a 1m cable to the LPU and switch.

AP-------lpu--injector--switch

I would never tell someone to remove there LPU for more than a few seconds to test if it was bad. That is just stupid and not worth the dangers of blowing an expensive switch.

Is there any way to use some in line sync source and still run POE on a switch for the ePMP 1000 APs?

Mat

At this point in time, no.
Inline sync while using a poe switch to supply power is not possible. There are devices such as the CMM (cmm micro had a 10/100 switch built in) or PacketFlux sync injector that allow adding power and sync to different switches.

I think over integrating devices in one box is a bad policy to begin with. If the switcg dies in a multi box setup, you replace the switch and its back up. If its all in one box, you spend more money to toss the replaced unit for what could have been a mucg cheaper repair.
Dont get me wrong, replacing one big box over 3 small boxes is a nice notion, just that much harder to fix.


@Chris_Bay wrote:


We ground properly.   Ive shared stories here and you can see some of our grounding work.  Our issue was most definitely linked to packet flux gear.  That being said that was mostly pmp100 and 320 gear. I have not used any of thier new gear for any length of time to determine if they have or have not fixed the issues we had with them.   None of thier old stuff had grounding lugs and 100% plastic housings.   I ordered 1 unit to test, and didnt deliver sync reliably to the epmp and 2 radios kept falling to 100 meg, but like i said we are pushing wire length.   Im unwilling to remove any of our LPU for long testing or operation as i stated above. 


For what it's worth, the newest SyncInjector units we've gotten from PacketFlux support two power supplies, jumpers onboard to decide pinouts and which PS feeds each radio, and the power connector is 4-pin: DCneg, DCpos-A, DCpos-B, and shield, which is tied to the ethernet shield in the ports.

j

I am using the packeflux sync injectors and even with long cable runs I dont lose sync. One of our towers we have the syncbox jr up on the tower with 120ft of cat5e for the sync run back to the sync injector. Still never lose sync.
For LPUs we chose to use ubiquity Eth-SP.
These are fully gigabit friendly and we did a horizontal cable run between buildings for a client that is over spec by 3m (about 10ft) and still get gig speed links. I could not do this with other LPUs as they would fail anytime we got close to max cable length.

If you dont want to go packetflux then your looking to the CMM4 option. The other CMM units do not do gigabit.

We just had another ePMP radio just suddenly lost all sats for about 3 hours then come back up. Then about 2 hours later it lost them again.  It either sees 20 and tracks 16 or it sees none at all while the other radios on the tower (A mix of ePMP and Airfiber) work fine and show no decrease in number of visible sats, tracked sats or SNR.  

Already in the process of converting all the epmp on the tower to packetflux for sync. As for packetflux actually causing lightning strikes or lightning damage...  I will say that way back in the day when we first started and used the parasitic syncpipes we noticed a big decrease in lightning damaged a few years later when we changed over to the SyncInjectors but since then I don't believe we have lost a single piece of equipment that I could in any way attribute to the Packetflux gear (or anything else in the cabinet actually).  Though in all honesty using the parasitics put powered equipment at the top of the tower and changing over to the sync boxes put the all the sync equipment at the bottom of the tower so I would expect less lightning damage as a result.