Confusing, Disappointing Peplink Pepwave Equipment Performance

Sorry, I’ve tried to be as succinct as possible, but this is kind of long.

Cellular Equipment: Peplink Pepwave MAX BR1 Pro 5G, Peplink Pepwave Maritime 40G antenna

My friend’s cabin is in the middle of a 35 sq. mi. township in Michigan. There are AT&T Wireless towers on the NW, SW, and SE corners of the township, as well as one about mid-way along the western border. The towers all have bands 2, 12, and 66. Three of them also have band 14 and one also has bands 14 and 5.

The terrain is mildly hilly and heavily-wooded (mostly coniferous trees [spruce, pine, fir]).

The cabin and pole barns are located in a relatively low elevation, relative to the northwest through northeast. There is also a rather tall-ish hill to the SSE.

He’s been using an AT&T NetGear Nigthhawk (M1?) 4G LTE-A WiFi mobile router at the location with acceptable results.

The goal was to provide wall-to-wall WiFi coverage within the metal-roofed and -sided structures, using a cellular data connection for Internet access.

Upon our Peplink vendor’s recommendation we acquired the hardware noted above. Our experiences with it have not met expectations.

Testing the Pro 5G at home, in an AT&T and T-Mobile tower-rich environment, I was seeing great download speeds (up to 135Mb/s down on AT&T, 50Mb/s down on TMO [more on this, later]) but upload speeds were inexplicably disappointing.

I’d earlier tested an Elsys Amplimax LTE modem. I regularly saw 12Mb/s up on both AT&T and TMO, and sometimes as much as 14Mb/s. The Pro 5G: Never saw more than 7MB/s up on AT&T and 8-10Mb/s up on TMO. (With the stock antennas.)

When we got the equipment up to my buddy’s property the results were very erratic and very disappointing.

With the Marine 40G antenna eight feet off the ground we were pretty consistently seeing ±90Mb/s down, but the upload was usually no better than ±1Mb/s. We actually got better results with the stock antennas, fastened to the back of the modem, with the modem sitting on a bucket, about a foot off the ground, but still never better than 2Mb/s up.

Meanwhile, the Nighthawk, sitting on an outside table about waist-high, was getting ±3Mb/s up. (±35Mb/s down.)

Another inexplicable thing: When we turned the “omnidirectional” Maritime 40G antenna on its axis we could get the upload up to 2.5Mb/s, but the download would drop to ±50Mb/s. Huh?

We went ahead with the installation, anyway, assuming that, once we got the Marine 40G attached to the western end of the pole barn, some 25 feet or more in the air, performance would improve.

Not only did performance not improve, but it got far worse. Download dropped to no better than 30Mb/s, at best. Upload went down to less than 1Mb/s - sometimes less than 500Kb/s. The next morning it was even worse, with download dropping to ±9Mb/s.

Knowing something of antennas: I have a hypothesis as to what might be happening with that Marine 40G antenna where it’s mounted, but I’d like to see what the brain trust here comes up with.

The other performance issues: I’ve no clue. E.g.: How that little Cat 12/16 Nighthawk could achieve more than double the upload speed, located a mere five feet or so from the Peplink Cat 20 modem with the Marine 40G antenna, and about six feet lower.

The T-Mobile numbers and Other Inexplicable Numbers (when I was testing at home):

On TMO the Pro 5G claimed it was connecting with 5G, yet I never got more than 50Mb/s down and usually ran around 8Mb/s up. Before that I’d seen 12Mb/s up with that Amplimax non-LTE-A modem. The Pro 5G claimed it was connecting to AT&T with LTE, yet it was getting up to 135Mb/s down. (The modem would occasionally display “LTE-A” for AT&T.)

Far more skilled radio/signal engineers in this forum may dispute this blanket statement, but IMO - for “fixed” wireless - you should utilize directional antennas.

That may not be ideal if you’re doing SIM failover and the secondary SIM carrier is in a different direction - however the fact is that you would likely experience much better performance with directional antenna.

3 Likes

Thanks for the follow-up, @erickufrin.

Skilled cellular radio/signal engineers is whom I’m hoping will chime-in.

Sadly, Peplink doesn’t make a 4xLTE + GPS directional antenna. Though maybe two IoT 20G’s would work?

He currently has only one SIM in the modem and I don’t see that changing. Where he is, AT&T has the best coverage by far.

But none of this explains why a little, inexpensive NetGear Nighthawk LTE mobile hotspot bests over $2k in Peplink Pepwave gear in upload sitting in essentially the same spot.

1 Like

You are not sharing details. When you did your various tests, what bands were you connected to? carrier aggregation? done at same time of day? signal strength? you sure you were connecting to the same tower and not a mini cell somewhere?

look at poynting antennas. maybe x polarized.

2 Likes

The Poynting xpol v3 is what I use in my long range portable setup. You could just mount two of those one on top of the other.

Panorama makes a 4x4 directional.

FWIW, the devices only utilize 2 antennas for transmit. All 4 antennas are used for receiving.

There used to be a thread which described which two ports mattered on these devices but its locked now. Sure would be great if @peplink would unlock this thread. https://forum.peplink.com/t/Can-I-connect-only-2-antennas-instead-of-4-for-a-5G-modem/618d1205e71adfac4df66245

2 Likes

Don’t know what bands to which the Nighthawk might have been connected. Doesn’t look like there’s a way to tell? IIRC, the Pro 5G was usually connecting to Band 66 (1700/2100MHz) at the time.

Carrier aggregation? Don’t know.

The A:B tests, Pepwave vs. Nighthawk, were done at the same time: Late Friday afternoon, early evening.

Signal strength: IIRC, the Pro 5G was showing an RSSI of ±-72dBm. The Nighthawk simply displayed “Good”

A mini-cell tower out in the middle of, literally, nowhere? Not very likely. The next closest habitation is at least a couple miles away, as the crow flies.

Regarding the signal strength: Again: At the time of the A:B tests on late Friday, when the antenna was mounted on a ±8 ft. pole sitting on the ground, the RSSI on Band 66 was ±-72 dBm. After it was mounted on the end of that pole barn that dropped significantly. Right now it’s varying between -88 dBm and -95 dBM.

So there are two issues, here: 1. Why, when sited approximately equally, was the Nighthawk consistently beating the Pro 5G in upload–by as much as 4x and 2. Why, when the Maritime 40G antenna was installed on the end of that pole barn did it perform more poorly?

I think I know the answer to #2. I’m hoping somebody with at least as much knowledge as I in the fields of propagation, antennas, and transmission line theory will chime-in with their thoughts.

The answer to #1: I’ve no clue. But I find it… curious that, of all the various tests I/we ran, in two separate locations, the Pepwave product performed worse on upload.

Cross-polarization may be a solution. The dense foliage probably results in no end of diffraction, refraction, and scatter.

Thanks for the pointers, Eric. I doubt my friend is going to be interested in what may well be throwing good money after bad in the hope of guessing right on an antenna, but I’ll keep those in mind.

DId you ever test the Peplink device with just the included paddle antennas attached? Away from the barn.

2 Likes

Hello @James_Seymour,
Concurring with our colleagues here in the forum, we always recommended that buyers consult with an experienced locally based Peplink Partner when acquiring the antenna hardware solutions? They can also advise you of the most suitable local settings for your region with the router.

Certified Peplink partners have access to a very extensive antenna selection guide within the Peplink Forum Partner Section, covering many scenarios and solutions.

As @erickufrin noted, a directional antenna solution is usually the correct way for a fixed installation. Some excellent quality antennas have been mentioned already in the thread; we will add the Telco Antennas XPOL 4x4 MIMO into the mix ( https://risingconnection.com.au/product/4pol-700-3800pl/ ) that we have been deploying with exceptional results for LTE & 5G.

Peplink has developed an excellent 4x4 MIMO + GPS antenna (we have had it field-tested in Australia); they need to get enough commercial interest to go into production, and we hope they will get that interest.

Happy to Help,
Marcus :slight_smile:

2 Likes

@erickufrin, yes. as I noted in my OP:

We actually got better results with the stock antennas, fastened to the back of the modem, with the modem sitting on a bucket, about a foot off the ground, but still never better than 2Mb/s up.

And, btw, the bucket was sitting nearly between the barn and cabin at the time. The SE tower might have been “visible,” and possibly the W tower.

We never thought to test the paddle antennas with the modem sitting right next to the Nighthawk on that outside stand-up table. I suppose because it never occurred to us they’d out-perform the Maritime 40G on an eight foot poll six feet away.

Admittedly: Much of what I’ve seen would never have occurred to me, so there’s that.

Thanks for the follow-up, @mldowling.

As I noted in my OP:

Upon our Peplink vendor’s recommendation we acquired the hardware noted above.

Perhaps I should have written that “Upon the recommendation by the certified channel partner to whom Peplink referred us?”

In one conversation with said certified partner he allowed as how they’d sold solutions to others in our region in the past.

Re: Directional antenna: As I noted in my OP: There are at least four AT&T towers that are more-or-less equidistant from his location. ISTM a directional antenna would limit him to one, possibly two of them, depending upon its radiation pattern. This would prevent the one, possibly two, towers’ ability to hand him off to a less-congested tower.

Yes: I realize not all four of them may be usable. Quite possibly even two “adjacent” towers. Short of buying a directional antenna and pointing it: I don’t know how we’d know.

I do know congestion occurs up there. When we were testing the Elsys Amplimax it at first did well enough on download (interestingly: it also did not do well on upload, either). Then the end of the workday hit and download performance went right in the gutter.

Re: Antenna suggestions: I suppose it’s not particularly important, his being a fixed location, but I’d prefer not to lose the GPS function.

What kind of cable length do you have for the antenna? And what kind of distance between your antenna and the cell towers?

Sounds to me the Netgear is connecting to other Bands, try disabling Band 66 and see how the Peplink performs.

2 Likes

Re: GPS - the antenna which is providing cellular need not also contain a GPS antenna. The 15ft GPS antenna that comes with the unit could be positioned near a window by where the router sits inside. This is what I do in my campervan. I actually cut the GPS antenna lead off one of my roof antennas because I did not want to run that cable through the roof. I then attached the included GPS antenna and affixed it inside the van. It gets GPS just fine without having clear view of the sky. FWIW the GPS functionality of these devices is really intended for “fleet management”. In a fixed location there is essentially no benefit. You could just leave GPS unattached.

2 Likes

Thanks for the follow-up, @padaco-daniel!

The Peplink Maritime 40G comes with 6 foot cables. To those we added 15 foot (LMR 200) extensions. According to our vendor, that fifteen feet of LMR 200 added only 0.8dB loss. (I don’t know if that includes the SMA connector insertion loss, though.)

We tried it on at least bands 2, 12, and 66. Band 2 (1900MHz) was horrible. I disabled it. Band 12 (700MHz) yields the best signal strength, but, predictably, poorer speeds. Band 66 (1700/2100MHz) had “adequate” signal strength when the antenna was near the ground and away from the pole barn, and yielded the best down/up speeds. (Still sub-par upload speeds, though.)

@erickufrin, yeah, I know he doesn’t need the GPS. I understand its purpose. Just seems a shame to lose it.

The inside GPS antenna is a non-starter. Metal roof, metal walls, no windows anywhere.

Btw: I’m looking at those Panorama antennas you recommended. Nowhere do they give the antennas’ polar plots–or even the directional beam width. So: It’s directional, but how directional? Seems kind of strange to me.

Have you checked the specs of the Maritime 40G?

  • 3.2dBi: 410-470MHz
  • 3.8dBi: 617-960MHz (band 12)
  • 5.3dBi: 1710-2700MHz (band 2, 66)
  • 6.5dBi: 3400-4200MHz
  • 6.7dBi: 4900-6000MHz

Cable Loss of the by default connected cable, without extension:

  • 0.33dB/m@900MHz
  • 0.49dB/m@2000MHz
  • 0.55dB/m@2500MHz
  • 0.87dB/m@5800MHz

So even without the extension cable you have the following Gain left per frequency/band:

  • 3,14dB@900Mhz (band 12)
  • 4,32dB@2000Mhz (band 2, 66)

The specs of the included small black antennas:

  • 3dBi: 600-960MHz
  • 4.5dBi: 1710-2700MHz
  • 2.5dBi3400-3800MHz

So in the bands you are using, the Maritime 40G with the extension cable has less signal than the small antennas that are included with the unit…

As stated before, the Maritime 40G might not be the best fit for this deployment, as the name suggests it’s mainly suitable for Maritime deployment.

We have good experience with the Poynting XPOL-2-5G (v3) antennas. As the antenna Gain is a lot higher, the longer cable (5m/16,4ft) is less a problem, and because it’s a directional antenna, it will make an even bigger difference. It doesn’t have GPS, but you can use the included GPS antenna which has a 5m cable by default. Or get the Peplink IoT 20G, which is specced comparable with the XPOL-2-5G in the mid bands you are using.

With the CAT18 modems you could use just Main1/Aux1 ports for the Low-Mid bands, only 2600Mhz and higher used the Main2/Aux2 ports. But not sure how the division is with the 5G modems which have A/B/C/D ports. So far the 5G models we have always setup with the Mobility 42G for mobile deployments or 2x XPOL-2-5G for more demanding installations.

As someone else suggested, try the BR1 Pro 5G with the supplied antennas directly next to the Netgear and see how it performs. And if possible retry the signal/speeds on the spot where you put the Maritime 40G but than with the small antennas and the unit outside (just for testing).

Have you checked cellmapper.net to see what the radio towers should support in terms of bands/frequencies? Sadly it’s not up-to-date for us in NL (EU), but we have a website that has current data for just our country.

And can you check what the distance is from your installation spot and the radio towers?

2 Likes

@padaco-daniel, I looked at the Maritime 40G’s numbers only after the fact. Here’s where I went wrong, and I’ve only myself to blame: Transmission line loss. And I think it’s much greater than the numbers you stated. It’s certainly much greater than what my vendor’s rep. claimed it was.

I went to Times Microwave Systems’ site, looked up LMR-200 (that’s what the 15 foot extensions are made of and, I assume, the six foot cables on the antenna). Here they are:

Time Microwave Systems LMR-200 Coax Attenuation
(N.B.: Does not include connector insertion loss.)
                                         Loss (dB)
AT&T Band    Frequency (MHz)    6 Feet    15 feet    21 feet
    2           1900              1.9       4.5        6.4
    5            850              1.2       2.9        4.1
   12            700              1.1       2.7        3.7
   14           2300              2.1       5.0        7.0
   66           1700              1.8       4.3        5.9
                2300              2.1       5.0        7.0

N.B.: The numbers in the 21-foot column do not match the addition of the six and fifteen foot columns due to rounding error.

TMS’ specs for LMR-200 list about double the attenuation/meter of what you listed, above. I’ve more experience with Belden cable than TMS, but I find it hard to believe anybody has a coax of the same type that has half the loss of LMR-200. I’m more than happy to be proven wrong, though.

Ok, so now we have part of the story: Why the 40G on the ground performed no better than, and, in some respects, worse than the paddle antennas attached directly to the back of the modem.

Ok, it’s been more than twenty-four hours since I posted the thread, and no antenna gurus have shown up to chime-in, so here, I believe, is why the Maritime 40G antenna performs worse when it’s twenty-five feet in the air, clearing all nearby structures, than it was merely eight feet off the ground.

Two words: “Ground plane.”

Here’s a little basic “antenna geek” for background:

A dipole in free space has a gain of 3dB with an omni-directional radiation pattern normal to the axis of the dipole. Looked at from above, along the axis of the dipole, it looks like a donut. A monopole antenna in free space has the same radiation pattern, but needs a ground plane underneath it to create a “virtual” other half of the dipole. (The “donut,” btw, only virtually extends below the ground plane. In reality: It does not exist.)

Then we have monopole antennas like the Maritime 40G, which does not need a ground plane against which to work to create the “illusion” of the bottom half of the dipole being present. That’s accomplished with a bit of electronic wizardry that, TBH, I’ve forgotten. (I haven’t done radio stuff for a good many years, now.)

And that kind of thing works, even if there is a ground plane under it, because the design makes the antenna more-or-less… uhm… Immune to the existence of it.

End of antenna geek explanation :slight_smile:

It works, that is, until you try to do what we tried to do: Put a big, really big, metal ground plane under the antenna–directly under it and all on one side.

What happens to a simple monopole when you do that is you take what would otherwise be an omni-directional antenna and turn it into a unidirectional antenna–with the direction being the direction of the massive ground plane.

I strongly suspect that’s what’s happening atop that barn. That big metal roof, being directly under one side of that Maritime 40G, is “dragging” it’s radiation pattern almost entirely off to one side of the antenna. And, Murphy being Murphy, it’s precisely the opposite direction we need it to be.

@padaco-daniel,

I’m coming to the conclusion that a pair of Poynting XPOL-2-5G (v3) cross-polarized, directional antennas may be the best fit for my friends installation. It’s got the best gain numbers of the antennas mentioned so far in this thread, and it’s the only one that’s cross-polarized.

Unfortunately, with those antennas I may have to have to completely replace two network cables and the extended power cable for the modem, but that’s just grunt work and not particularly challenging. (I left generous service loops on everything, so maybe not.)

I’ll almost certainly be going back up there this weekend to run additional tests–including pulling the modem down, putting its stock antennas on it, and evaluating it right next to the Nighthawk.

The cell towers are all ±3 to ±4 miles away.

Hello @James_Seymour,

Don’t I count? Our colleagues here in the forum have been covering the information for you quite well, so there is no need for us to duplicate it further. I know several of the Peplink team have been following, though it is the same with them; when colleagues and community members have already chipped in, there is no need to duplicate information.

If you have not been given the correct information from your Peplink supplier, please get in touch with Peplink through their main website (use the contact form or via a support ticket) and let them know so that the partner can get additional support, training and guidance, please do not name them here as everyone occasionally makes a human mistake in their advice.

In reading the thread, I suggest that you continue with your weekend plans and then update us with your findings for the community to review.

FYI. I go out to the site with tens of thousands of dollars worth of RF test equipment; many things are impossible without the higher-end equipment to isolate and develop a rock-solid solution.

We only use omnidirectional antennas on fixed locations when there is a need for the multipath signal between different towers and the installation location, plus the bandwidth requirement is low. For everything else, we use directional antennas or custom design an antenna solution for the installation (we do fixed building and mobile moving kits, including for emergency services and maritime antenna solutions).

I’m based in Australia (& travel the world) doing this sort of work; however, there will be someone locally based that has the level of expertise we have; they may not be a Peplink Partner, though one of the local Peplink Partners should know of someone, suggest rinning around your local partners to see who they recommend.

Happy to Help,
Marcus :slight_smile:

3 Likes