26 dBi ultra wide-band parabolic grid antenna

[xdavidx]

Has anyone tried one of these?

26dbi_ultra_wide-band_parabolic_grid_antenna.jpg

[xdavidx]

On a related note, does anyone have any experience with any other high gain antennas that are wide-band (700 to 2700 MHz, minimum), yet have a narrow horizontal beam width?

[swwifty]

I don't know of anyone that has.

Those two things don't coincide. If you want a higher gain antenna it is not typically going to be a wide bandwith. Especially when you get on the lower frequencies. A high gain 700mhz antenna is huge.

[xdavidx]

So do you think the claims of that 26 dBi antenna are far fetched?

[swwifty]

So do you think the claims of that 26 dBi antenna are far fetched?

Yes, and if they are true its most surely for the highest frequency that antenna supports, which is probably 2700mhz.

[xdavidx]

Claims are:

14-17 dBi
600-960 MHz

20-23 dBi
1700-2700 MHz

23-26 dBi
3000-6500 MHz

[swwifty]

Claims are:

14-17 dBi
600-960 MHz

20-23 dBi
1700-2700 MHz

23-26 dBi
3000-6500 MHz

That doesn't seem too far off, but I'd go with something that supports MIMO, does that antenna support it? If I recall in the past that thing was only one polarization?

[xdavidx]

That doesn't seem too far off, but I'd go with something that supports MIMO, does that antenna support it? If I recall in the past that thing was only one polarization?

It says vertical or horizontal polarization.

[swwifty]

It says vertical or horizontal polarization.

You'd want both for LTE, not just one.

[xdavidx]

You'd want both for LTE, not just one.

I was wondering about their use of "or" instead of "and". I am not exactly sure what they mean to say. I haven't been able to find any mention of mimo. However, given the price and size of the antenna, and challenge in mounting in a non-vertical way, it may be more of a rarity vs truly not supporting it. I'll contact the company. I believe they are also located in Texas.

[tgoodwin]

I'm just getting into LTE and purchased that antenna before I found this site. I'm not sure what all the numbers mean yet but this is what I get with a WE826-T and MC7455.

Baseline with the little 3dB omni antennas

standard antenna.PNG

WIth the parabolic antenna

antenna only.PNG

With the parabolic antenna and amp

antenna and amp.PNG

[swwifty]

Interesting results. It looks like though with the parabolic antenna you aren't getting a diversity signal? You can see that on the RxD RSSI line as compared to the omni antennas.

Does the parabolic antenna have two RF cables (like N connectors) or just one?

[xdavidx]

Very cool results. Thanks for sharing!

Did you leave the secondary omni antenna attached and just replace the primary with the grid antenna? That's what it looks like from the results.

It would be really interesting to see what it shows with a 2nd grid antenna attached, but I don't suppose you have access to another one?

I would be very interested in speed test results for the same 3 scenarios: omni, grid, grid with amp.

Regarding what the numbers mean, I'll give a high level overview. You may already know some of this. I'll go over a few main concepts. Only meaning to help, so just ignore the parts you already know.

The RSRP stands for Reference Signal Receive Power. It represents how strong the signal is. In audio, it would be analogous to you hearing a whisper vs a loud voice. With just the omni antennas attached, the primary antenna, represented by the RxM (M = main) lines, isn't as "loud" as the secondary antenna, represented by the RxD (D stands for Diversity) lines. And there are 2 sets of lines, because you are receiving data on 2 LTE chunks of frequency, Band 4 (a 5 MHz wide chunk in this case), and another chunk of Band 4 (10 MHz wide). Each "band" represents a certain minimum and maximum width of frequencies. In some cases it even represents a couple segments of frequencies, typically one for downloading and one for uploading.

Putting that all together, you have an RSRP value of -118 on the primary antenna for the first Band 4 chunk and a value of -119 for the second Band 4 chunk. You have an RSRP value of -113 on the secondary antenna for the first Band 4 chunk and a value of -115 for the second Band 4 chunk.

As I mentioned, your primary omni antenna isn't hearing the signal as loudly as the secondary antenna for some reason. You might want to tighten down the connector for the primary antenna. These values can vary from one second to the next, but there is a big enough gap, that it seems like something is off. It could be how it is pointed at the tower too, but I wouldn't expect omni antennas to vary for that reason much. Every 3 points on that scale is double the power. For the first chunk of Band 4, your antennas are separated by 5 points and for the second chunk, by 4 points. 6 points would be quadruple the power difference. 10 points is 10 times the power difference. So 5 and 4 points is a significant difference. I'm mainly mentioning all this to help you understand the numbers.

The smaller the RSRP value (closer to 0), the better. -113 is better than -118, for example. And you going from -118 to -81 is spectacular. That's a difference of 37, which is over 5000 times the power! You'll also amplify noise, so that's why I'm curious what effect it has on the resulting speeds. There are also diminishing returns on signal strength where getting a stronger signal is no longer the limiting factor and it doesn't help the speeds as much as it did with lower values.

The MC7455 is a 2 carrier aggregation modem. That means it can take 2 chunks of bandwidth and pull data simultaneously from both. Other modems allow for greater aggregation. The EM7565, for instance, is a 3 carrier aggregation (3CA) modem and can utilize 3 chunks (carriers) simultaneously. The width of each chunk is a factor in how much speed you can get out of it. How the data is modulated (encoded/decoded) is another factor. In general, you'll get more data out of a 20 MHz chunk than a 5 MHz chunk, all other things being equal. You can think of that like the dial on a radio. If you tuned into a single station, you'd get one song. If you could tune into 2 stations, using more of the FM or AM bands, you would get 2 songs, or twice the "data". There are more details, but that's a high level view.

Your 2 carriers happen to be coming from the same Band 4. Normally you have to check the stats during a speed test in order to see all the bands being used. If the carriers are in the same band, I guess it maintains a connection to multiple, even when not transmitting data.

Now, why doesn't it just show 15 MHz of Band 4 being used? That's because there is a gap between frequencies of the chunks of Band 4 you are being given. If they were contiguous chunks, then they would show up as 15 MHz. With the gap, you have intra-band, non-contiguous carrier aggregation. If they were contiguous, then that 15 MHz would only use up 1 of your 2 chunks that the modem can aggregate and *if* the modem could talk to the tower on another band, you'd also get some bandwidth from that band and increase your speeds. Alternatively, if you are able to point to another tower, then you might get a different breakdown of bands and chunks and have a different potential speed.

The Cell ID is a unique ID for the equipment you are talking to on the cell tower. All 3 of your tests were talking to the same tower and the same equipment on that tower. A tower can have multiple Cell IDs, generally facing different directions.

SINR stands for Signal to Interference and Noise Ratio. In other words, how much clean signal are you getting vs Interference and Noise in the frequencies you are communicating to the tower on. Higher values (away from zero) are better.

RSRQ stands for Reference Signal Received Quality. As the acronym states, it is a representation of the quality of the signal. Numbers closer to zero are better quality. Swwifty found some research showing that the number isn't accurate if the modem is transmitting. In your screenshots, it was only transmitting in your first test, with the mimo antennas. Tx Power was non-zero. Either something connected to the modem was doing some communicating on the internet or you were running a speed test at the time.

Putting SINR and RSRQ together, if you have a high SINR and low (closer to zero) RSRQ, your speeds will generally be better than a low SINR and a high (larger negative number) RSRQ. You can have the best RSRP power in the world and still have slow speeds if you have poor SINR and RSRQ values. In general though, if interference is a constant (which it certainly isn't, unless you are in a lab), then a better RSRP value will end up giving you a better SINR value and higher speeds.

RSRQ and SINR will jump around a lot, depending on intra and inter-tower usage, whether it is an odd second of the minute, and whether you got out of bed on your left or right foot.

What we see as a "connection" is really a highly orchestrated system of the radios constantly (at the millisecond level) shifting around between different frequencies at the micro level, adjusting modulation schemes, and attempting to overcome interference. This allows many mobile devices (or routers) to share the limited resources provided by the towers.

The previous discussion was about whether this particular highly directional grid antennas supports MIMO (Multiple Inputs Multiple Outputs). Without going into all the details, the basic idea is that if you can increase the number of antennas, then you can increase the throughput. Put a tunnel parallel under a road (or loooong bridge) and you can get twice the number of cars through. The MC7455 supports 2 antennas, and if the tower and the modem support that, and the conditions are right, then you'll get roughly double the throughput vs 1 antenna, with all things being equal.

Your 2 omni antennas most likely are performing in this way. Whether the grid antennas would perform in that way is the question. With enough separation, your grid antenna and one of the omni antennas might be able to perform in that way as well. There is a big difference in antenna gain in swapping out one of the omni antennas, so it isn't perfectly easy to determine the cause of a speed difference. It could be due to all the data flowing through the grid antenna much better than it could with both of the omni antennas or it could be due to it working in concert (MIMO) with the other omni antenna. The speed tests would be cool to see and analyze.

Anther test would be to test the speed with only the primary omni antenna screwed in and the second port not having anything plugged in. And also testing the speed with the primary port using the grid antenna and the second port not having anything plugged in. And the same, but with the amplifier being used. This would be a total of 9 tests. The closer in time the tests are run, the better, because other variables that change over time can really affect results.

Thanks again for posting!

[tgoodwin]

xdavidx that is exactly what I was looking for! This should be a sticky somewhere here. Thanks so much for taking to time to spell it out for me.

The parabolic antenna only has a single RF connection. It is the primary and the dinky omni is my secondary.

There are only 4 towers to pick from ranging from 2.6 to 5.8 miles. Surprisingly my elevation is higher than the towers and has no obstructions, no LOS because of trees. Before I locked to B4 (I'm on Verizon) it would sometimes go to B13 where the speeds are terrible.

I'll run the tests tomorrow and post the results.

[swwifty]

sometimes lower gain antennas work better as the signal is scattered coming through the trees.

MIMO is critical in LTE, and I would highly suggest using an antenna that supports it (or two physically separate ones for that matter).

[xdavidx]

You'll find different charts like this online with different values, since it is a bit subjective. It does help to give you an idea of what's good vs bad for the different metrics. As you can see, the first one kind of messed up the RSRQ good/fair rating by giving them the same values and skipping between -5 and -6 and -10 and -11. But that aside, the ranges are a lot different. The third one is similar to the second, but without the RSSI being able to be read on the same lines, and the RSRQ a little better represented.

lte_signal_values.png

lte_signal_values_2.jpg

lte_signal_values_3.jpg

[xdavidx]

There are only 4 towers to pick from ranging from 2.6 to 5.8 miles. Surprisingly my elevation is higher than the towers and has no obstructions, no LOS because of trees. Before I locked to B4 (I'm on Verizon) it would sometimes go to B13 where the speeds are terrible.

Do you have the omni antenna inside during the test and the directional antenna outside?

Regarding the 4 towers, have you connected to each of them by pointing the directional antenna at them? If you look at the CID, you'll be able to see which one you are on. Sometimes it takes a while for it to switch even if you have a better connection to the new one vs what it was previously connected to, or unless you reboot. By pointing the directional antenna at different points on the 360 degree path and watching RSRP it will give you a pretty good idea of when you are pointing at the tower (RSRP value closest to zero for that CID).

The fun thing is that you can sometimes point a little bit away from the tower and get better SINR/RSRQ values, because you are also pointing away from another nearby tower that is causing issues.

Tilting the antenna forward and backward will have an effect too. I'm guessing you'll see the strongest RSRP values with your antenna pointing down a little bit, since you are higher than the towers.

I'll run the tests tomorrow and post the results.

Looking forward to seeing them. In my experience, with other towers nearby, and variations in users hopping on/off, downloading/not downloading, etc., the speeds can vary quite a bit from one minute to the next and definitely from one hour to the next. Hard to rely on just one speed test for a given set-up. Taking a few and averaging, or throwing out outliars are a couple ways to deal with this. There isn't a perfect way to do it. Normally it doesn't matter that much, but if you are trying to test one setup vs another, the natural variation in speed tests can lead to conclusions that are the opposite of reality.

Having a clean test system helps too. Ethernet is less variable than wifi, but having a Windows machine connected over ethernet allows for Windows to do things in the background that sucks up bandwidth and messes with tests. If you are running the test on Windows and don't want to go through the hassle of blocking all outbound traffic from other apps, running the task manager and watching the network activity can help determine if Windows is doing a sneaky update in the background or some other app is bogging down things. Disconnecting all other clients helps avoid variation too.

I was doing some tests one day and then found out that my daughter's cell phone had been connected to the router during a big chunk of that time! I forgot I had used it to do some testing on another day and had given it the wifi password. That type of thing could affect DSL or Cable too, but cellular has a lot more built in variation from test to test than the other platforms.

[xdavidx]

Look what I found:

26dbi_ultra_wide-band_parabolic_grid_antenna_45_degree.jpg

No guarantee that this means MIMO, but very intriguing that they have 2 of them at a 90 degree angle to each other and 45 degree to the vertical.

[tgoodwin]

Before I get to the results I'll answer the questions first.

The grid antenna is on the roof while the omni is inside. I have 20' of LMA400 between the antenna and amp. I did connect to each tower and ran speed tests to find the one that worked the best. I would use the compass on the iPhone to get it close then I would watch RSRP (I didn't know why there were 4 at the time, I would look at the top most one) while moving it in small amounts. When it would drop I'd move it back a tad and stop. This was before I knew about looking for the CID. So, I could very well have been on a different tower.

I also drove to each tower and used my phone to run a speed test. There is a huge difference as to what I get at the tower and what I get at the house. One tower gave me over 100Mbps down yet the best I've seen here is ~30. Again this was when I just got into it. I didn't check to see what band the phone was using.

If I do add another antenna to the secondary connection (and have a setup like the picture) won't the modem do MIMO? Or does the antennas have something to do with it?

The tests were run on a Windows 10 machine (auto update disabled) wired to the router and the router's wifi turned off. I ran at least 2 tests for each configuration to make sure they were close to each other. The amp is a Wilson 4G 460119. Today is not the best day since it is rainy and overcast.

Now for the results.

Primary: omni, Secondary: omni, No amp

at-omni-omni-no.PNG

speed-omni-omni-no.PNG

[tgoodwin]

Primary: omni, Secondary: none, No amp

at-omni-none-no.PNG

speed-omni-none-no.PNG

Primary: omni, Secondary: omni, with amp

at-omni-omni-yes.PNG

speed-omni-omni-yes.PNG

Primary: omni, Secondary: none, with amp

at-omni-none-yes.PNG

speed-omni-none-yes.PNG

[tgoodwin]

Primary: grid, Secondary: omni, no amp

at-grid-omni-no.PNG

speed-grid-omni-no.PNG

Primary: grid, Secondary: none, no amp

at-grid-none-no.PNG

speed-grid-none-no.PNG

Primary: grid, Secondary: omni, with amp

at-grid-omni-yes.PNG

speed-grid-omni-yes.PNG

Primary: grid, Secondary: none, with amp

at-grid-none-yes.PNG

speed-grid-none-yes.PNG

[swwifty]

Interesting results. That grid antenna really helps with upload speeds.

I did some research into that amplifier, it looks like it only takes one input though, and you can't hook up both antennas to it?

Also, if AT&T is in the area, I'd recommend trying them. It looks like Verizon's network is pretty busy there. (based on RSRQ values)

[xdavidx]

Nice set of results. Thanks for doing that. It really helps to have hard data.

I need to go fix my parents' wifi network. Their dsl router, which is also their wireless access point, was reset. :-/

When I get some time, I'll analyze the results more and see what insights they hold. At a glance, the system was switching between 2 different CIDs. They are most likely on the same tower, because they only differ by one number, ending in 256 vs 257. They obviously overlap in coverage and you are in that overlapping area. That's unfortunate for the testing, because one CID might have a very different set of variables going on than the other, at the same exact time. You could, in the extreme, have one CID to yourself and be sharing the other with many other people. Or the individual micro blocks of frequency that you are given on one CID could be clean, whereas what you are given on the other CID could have interference from other CIDs that are using those same micro blocks for other connections, or interference from other noise sources.

Having said that, I would view you being in coverage of 2 CIDs from the same tower as a good thing. With a highly directional antenna, you are putting all your eggs in one basket -- the tower you are pointing at. By having access to 2 CIDs on that tower, the tower might switch you from one CID to the other during times of congestion. And if one CID gets knocked out for some reason, you have a fallback (assuming the whole tower doesn't have an issue).

CIDs aside, that amp certainly helps real world speeds. MIMO status is unclear. Grid with omni vs grid with no omni performed a little better, but on different CIDs. Grid with omni with amp performed a little less than grid without omni with amp. :-/ And omni omni with amp performed very well too, which is interesting. Need more time to look and ponder.

Questions: Did you use web or windows app version of speedtest.net? Are you able to look in your results history to se if you were connected to the same speedtest server for each test? Just more variables to consider.

Thanks again!

[tgoodwin]

The antenna does a decent job. I'm wondering if adding another or switching to a pair of the 15dB panels would perform better.

I picked Verizon because they have the best coverage for phones out here. I haven't a clue what AT&T would be like with an antenna. There are 3 towers that are in the same radius as the Verizon towers. How can you tell how busy the tower is by RSRQ?

The amp has a single antenna input.

I used the web version not realizing there was a Windows app.

[xdavidx]

The windows app version doesn't have ads. I am not convinced the web version is more variable, but I went with it anyway. My AT&T connection thinks I am in Chicago (I am near Minneapolis, MN). So it picks speedtest sites in IL. My guess is that the AT&T connection goes over fiber to Chicago and out to the internet from there. So I do still have it pick a server in IL. However, I marked speedtest.net in Chicago as my favorite and I always watch to make sure that is what I am connected to. Using the same server is less variable than using different servers and I have found that some non-speedtest.net servers are a lot more variable than spedtest's own servers. I am sure there are those out there that are consistent, but I figure speedtest is in the business of testing speeds, so they might manage things better than some others.

RSRQ is challenging to analyze. If you had no interference or noise, -10.8 would mean all the slots in your CID, for that band, are in use (measured when your TX power is zero -- thanks again swwifty). So that is one type of congestion.

When dealing with a single tower, it has a limit on bandwidth to the internet. That backhaul connection may be 1 to 2 Gbps fiber, or it could be microwave or even DSL. The future might have them upgrading to more fiber, low earth orbit satellite or even 5G from tower to tower. I believe most towers have enough backhaul capacity to handle all the LTE they can deliver to consumers. Towers in rural areas and less populated areas? Who knows? Backhaul limits could be more of a limiting factor when those towers aren't maxed out on the RF side of things.

5G changes everything. Either they need to upgrade their backhaul connection, or the promise of 5G speeds won't be realized once more than a small number of people are sucking down the bandwidth. I guess the good news is that most people won't be running with the pedal to the metal. Streaming a 4K movie on a 1+Gbps 5G connection will mean just a blip of use every 10 seconds or so to buffer the next bit of video. So a lot of people can be streaming 4K at the same time without maxing out the backhaul.

Anyway, tower backhaul capacity can, in theory, be a limiting factor. That won't show in RSRQ.

This is all with one CID on one tower. Add in other towers and external noise, and now RSRQ gets harder to analyze. Does the -11 mean your CID is fully loaded and another tower is interferring a little bit? Or does it mean your CID is hardly used and another tower is hyper used and interferring a lot?

If you have a single tower, that's one thing. Have 3 or 4 withing earshot and things get more difficult. A hyper directional antenna is one way to avoid some of that interference from the other towers.

Am I correct that you are wondering, why even have the grid antenna if you get 22 mbps with an amped omni? If so, that's a valid question, but there is still the question of why are you only getting 22 mbps? It could be due to tower congestion. Other antennas aren't going to help that. Trying the tests in the middle of the night will help you narrow it down a bit.

The other thing I can think of is that your horizontal aim is more important with the grid antenna. You did the nudge, check RSRP, nudge, check, etc. It might be worth doing that again and looking at the CID to make sure you have what you think you have. Vertical aim matters too. You could try adjusting that.

My best guess is you are dealing with a busy CID and/or interfrence from other Verizon towers. One way to find out the max you can get out of a CID is to bring the router and antennas to the tower and test there. Run the stuff off an inverter plugged into your car or a battery. I did this. Doing it a few blocks away, with a clear line of sight is better than sitting right under the tower.

If you do this, run a *lot* of tests. I got 40 mbps for some periods of time, 60 mbps other periods, and over 100 mbps other times. Pretty sure that is either normal tower congestion or my AT&T service being deprioritized due to some amount of congestion.

Oh, and you are dealing with 15 MHz total (5 + 10) on Band 4. You can only get so much out of that. If you had 30 MHz or 40 or 60, you would have a greater potential speed. I am going to make a wild guess that you might get 70+ mbps, max, right up near the tower, when usage is low, with that 15 MHz.

Regarding AT&T and phones vs routers... It is difficult to compare phones to the modems we are using. Are they using modems that support the same bands, do they support the same carrier aggregation level, do they aggregate the same combinations of bands, and does the software make use of all that the hardware can do? Having said that, if you compare an AT&T phone next to an AT&T tower and your verizon phone next to your verizon tower, if you get great speeds out of the AT&T tower and not the verizon tower, then maybe AT&T is a better choice.

[tgoodwin]

This is turning into a little hobby. It sucked me in and I want to suck down bits as fast as possible. I'll pick up another antenna to travel around with and visit all the towers again. Now that I know a little more about what I'm looking at.

Thanks again for all the help/feedback. Hopefully I was able to help answer your questions about the grid antenna, the actual thread topic.

[xdavidx]

This is turning into a little hobby. It sucked me in and I want to suck down bits as fast as possible. I'll pick up another antenna to travel around with and visit all the towers again. Now that I know a little more about what I'm looking at.

Thanks again for all the help/feedback. Hopefully I was able to help answer your questions about the grid antenna, the actual thread topic.

I can relate to that!

I took my little omni antennas up to the towers. Don't need directional antennas when you are close to the tower.

Are you saying you are going to get another grid antenna for traveling to the towers?

I'm putting together something to help analyze your results. Should be able to post it soon.

[xdavidx]

Oh boy. We've got some variables I didn't see initially and almost missed again, even with a thorough look.

Here are the results in a tabular format:

grid_antenna_analysis.png

Going left to right:

1. Orange - Switched from CID ending in 257 to CID ending in 256 and then back again. This caused all sorts of variables that make it difficult to analyze the results.
2. Purple - With the switch to CID 256, you now get 10 MHz for the first carrier, instead of 5 MHz. This is good. This will improve upload speeds, because the first carrier is used for uploads. I almost missed this and assumed it was 5 + 10 for all the tests.
3. Red - CID 257 is not only giving you 5 MHz for the first carrier, which slows down upload speeds, it isn't allowing for carrier aggregation, so you are *only* getting 5MHz for download too! This is one of the things I almost missed.
4. Cyan - Download speeds about the same or less than upload speeds. This indicates a congested tower. More people download than upload, generally, so the download often decreases with tower congestion more than uploads do.
   
   The upload speeds are very good. The Grid really helped with that on both CIDs. I'm not sure exactly why, since the RSRP's of Omni with Amp and Grid without Amp were about the same (or even worse in one case). Hard to tell what's going on with RSRQ, but maybe it avoided interference from other towers or noise from other sources a lot better than the Omnis (which is what we would hope/expect).
5. Yellow - Why did the upload speeds decrease with the Grid and Amp vs Grid and no Amp? Because you switched to the other tower again and only have 5 MHz to work with for the first carrier that is used for uploads.
   
   Also, why are the upload speeds for those two rows less than the download speeds? You have 5 MHz for download and 5 MHz for upload. B4 uses 1700 MHz for uploads and 2100 MHz for downloads. Maybe 1700 is having some sort of interference issues or maybe the tower is limiting things in some other way (different modulation, giving you fewer micro blocks to work with, etc.).
   
   *Update: I just now realized, as I was going back through these points, that you are using a Category 6 modem. That can only do 16QAM modulation for uploads (and only 64QAM for downloads), so it can only do 2/3rds the speed of the download *without* MIMO and if you do have 2x2 MIMO, then the difference is even greater, because LTE doesn't do MIMO for uploads, only downloads. This, the difference between 16QAM and 64QAM, and that the downloads are doing MIMO (see below), explains why the Yellow upload speeds are less than the download speeds.
6. No color - MIMO seems present on the Omnis, judging by the drop from 15.31 to 7.99. Going from 1.25 to 1.43 Mbps upload with/without the second Omni is a little weird. Natural variation or a busy CID possibly. *Update: As stated above, no MIMO for uploads, so uploads won't be affected by disconnecting the second antenna.
   
   If an antenna isn't plugged in, the port is still acting as a poor antenna, so you may still get some benefit from it. You won't necessarily see half the speed without vs with, hence "Partial". If you were sitting right next to the tower, removing the antenna won't matter as much as doing it at home, with the router inside. With -140 RSRP, it probably isn't contributing much.
   
   If you amplify the first antenna, or use a Grid antenna, you'll create an inequality between the first and second antenna, with the first transmitting more data, hence "Unequal" (or "Greatly Unequal", when you have the Grid *and* you are amplifying it).
   
   Why "Maybe" for the Grid tests? The download and upload speeds did go down with Grid/Nothing vs Grid/Omni, especially the upload speed. However, the download speed went down from Omni/Omni/No Amp to Grid/Omni/No Amp, even with only having 5 MHz for the first and 15 MHz for the second. Why? Well, you are dealing with a different CID and the second CID (256) seems congested. Hard to be conclusive with those variables present.
   
   Why did you get slightly faster download and upload speeds for Grid/Nothing/Amp vs Grid/Omni/Amp?? Maybe it isn't using MIMO?
   The tower certainly looks like it was doing MIMO with the Omni antenna as the primary antenna earlier. Now that you have a stronger signal, is it cutting you off and not doing MIMO anymore at the tower? That doesn't seem likely, but who knows?
   
   Maybe CID 257 is congested too, and the speed tests were too variable to let us see what is going on?
   
   *Update: Ding, ding, ding!! As I was going back over everything and noted above that you are using a Category 6 modem, I also realized that getting those download speeds of 22.52 and 23.91 *prove* that you are using MIMO. It does assume that the modem is accurately reporting that it isn't doing Carrier Aggregation. If that is true, and you are only using 5 MHz of bandwidth, then the max you could theoretically get, with SISO, is 18.75 Mbps down and 12.5 Mbps up. Go up to 2x2 MIMO for the download and you can get 37.5 Mbps down and 12.5 Mbps up. You can't get more than the theoretical speed for SISO, so you *have* to be doing MIMO with that grid antenna.
   
   My "Maybe"s should be changed to definitely, assuming you were only given 5 MHz total for downloads.
   
   Before realizing the above, the biggest indicator I saw that would lead me to believe the Grid is doing MIMO is that the upload for Grid/Omni/No Amp went down significantly when we switched to Grid/Nothing/No Amp. You lost almost half your upload speed, which is what I would expect from unplugging the secondary antenna. And that secondary antenna was only an Omni, not even a directional antenna. *Update: WRONG! There is no MIMO for uploads, so you shouldn't lose anything (for uploads) by unplugging it. It was probably just natural variation on a congested tower that saw the upload speed go down.
7. Mauve - The stats were most likely collected during a speed test (or something else in the background transferring data), since the modem was transmitting. That makes the RSRQ/SINR values suspect, according to what swwifty found.
   
   *NOTE: The modem will only show ACTIVE for carrier aggregation if you are running a test at the same time you collect the stats. At least that is the way it works when the modems are doing inter-band carrier aggregation. I'm assuming it works the same way for intra-band CA.
8. Green - Related to the above, we can see that the SINR is highest and the RSRQ is best with the transmit power at 0. This one was probably captured before or after the speed test.

Looking at things at a higher level, and after going through the details multiple times, I think I can say this: You were getting about a max of 22-24 Mbps for downloads across both CIDs. For 5 MHz, that might be the best you can expect. That's over 60% of theoretical for 5 MHz, 64QAM, with 2x2 MIMO, on CID 257.

For CID 256, where you have 15 MHz for downloads, my best guess is that you had a congested CID that was limiting what you could get. I said in an earlier message that you might be able to get 70+. That was without remembering that the modem can only do 64QAM. I think you'd be lucky If you could hit 55 Mbps with 15 MHz, with everyone in the area sleeping.

One thing you could try is Grid/Omni/No Amp and Grid/Omni/Amp with a tool that runs speed tests automatically for you. I haven't messed with this one too much, but it does provide for automatic testing: https://testmy.net/auto

Setting it for both upload and download and running it every 30 minutes for 50 times will cover a whole day. That should show if tower congestion or inter-tower interference/noise certain times of the day is limiting you.

Testing up close to the tower won't help with congestion. It might give you a different CID than what you can get at your house, so you'll have to watch for that. Pointing at it from the same direction as your house sits will avoid that issue. Moving around will allow you to test both the CIDs you can get at your house. Watching for the bandwidth size you are given for the primary and secondary carriers will tell you if it is acting differently up close than at your house, as well as watching whether carrier aggregation is active or inactive. It may give you different bands entirely up close than it does far away.

I am guessing that CID 256 is your best chance for max speed from that tower, but only when it isn't as congested. 257 might be your best chance for max speed during heavy usage times. It is a trade-off. You can try pointing the grid antenna slightly away from the tower one way to have a little better chance of only getting one CID and slightly away the other direction to have a little better chance of only getting the other CID. No idea if it will work though. It seems to work for less focused antennas when you want to avoid another tower.

You seem to have very similar signal strengths to each CID. It is possible that 256 was congested, so it threw you back into 257 part way through the testing. It might have said "Hey, you're making a bad situation worse! Go play over on this tiny 257 CID again for a while!"

I'm not sure what your speed expectations are, but you want as much bandwidth as you can get. It is possible that there are other bands on those CIDs and you aren't seeing them, because your modem can only do 2 carrier aggregation. A 3 or 4 carrier aggregation modem may provide more bandwidth on the same tower. Difficult to know what the tower has and what it is going to give to you without a modem that can aggregate on more bands.

[xdavidx]

I'm not sure what your speed expectations are, but you want as much bandwidth as you can get. It is possible that there are other bands on those CIDs and you aren't seeing them, because your modem can only do 2 carrier aggregation. A 3 or 4 carrier aggregation modem may provide more bandwidth on the same tower. Difficult to know what the tower has and what it is going to give to you without a modem that can aggregate on more bands.

*Update: Actually, there *is* a way. You can block Band 4 so that the modem *has* to pick another band. If no other bands are available, then you'll get no connection at all. If they are available, then you may even see 2 new bands, since your modem can do 2 carrier aggregation (as long as you run the stats during a speed test). *If* you get one or two new bands when you do this, then a 3 carrier or 4 carrier aggregation modem might help you. There is no guarantee that the tower will give you the extra bands on a new modem if you re-enable Band 4. It might withhold the other bands until the tower isn't congested, or withhold them regardless.

However, *if* the other bands provide more bandwidth than Band 4, then you might end up with faster speeds without Band 4 being used at all. This can be tested without getting another modem. If you block Band 4 and get more bandwidth with a new band/bands, then you can either:

A) Keep band 4 blocked and continue using the MC7455 on the other bands.

B) Buy a new modem that does 3 or 4 CA and see if the tower will give you both chunks of Band 4 *and* the other band/bands, simultaneously. If it doesn't, then you would have to block Band 4 again. In this situation, a 3/4 CA modem doesn't just help with carrier aggregation, however. With the increased category ranking, you also get the possibility of:

- 256QAM modulation for downloads (instead of the 64 that you have now)
- 64QAM modulation for uploads (instead of the 16 that you have now)
- Carrier aggregation for uploads
- The chance for other bands that the MC7455 doesn't support. For example, the MC7455 supports bands: B1, B2, B3, B4, B5, B7, B8, B12, B13, B20, B25, B26, B29, B30, B41. The EM7565 supports bands: B1, B2, B3, B4, B5, B7, B8, B9, B12, B13, B18, B19, B20, B26, B28, B29, B30, B32, B41, B42, B43, B46, B48, B66. There are other modems, of course, but that gives you an idea of band support differences.

C) Wait for the 5G modems to come out. Even if you aren't using 5G, you'll get more carrier aggregation, higher QAM and support for a higher number of antennas (and more MIMO). Sierra Wireless's first offering may not support everything that the Qualcomm chip has, but just to give you an idea, the Qualcomm Chip (X55), with Category 22 LTE, has:

- 7 Carrier Aggregation downloads
- 4x4 MIMO downloads
- 1024QAM downloads
- 3 Carrier Aggregation uploads
- 256QAM uploads (I'm guessing on this)

And if you can find a tower with that much bandwidth and support for all this, the theoretical speeds are 2.5 Gbps down and 316 MBps up!

Let's say you can find only 80 MHz of bandwidth (instead of 140), 4x4 MIMO, 256QAM download, 3 carrier uploads (3x20), and 64QAM uploads. Your theoretical would be over 1500 Mbps down and 300 Mbps up. Take half that for real world speeds, and you're at 750 down and 150 up.

Even without more bandwidth, the MIMO and more antennas will double existing download speeds. And as people start using 5G, which is more efficient with spectrum, and/or jump onto other frequencies, maybe the 4G bands will be less congested.

[JimHelms]

This is turning into a little hobby. It sucked me in and I want to suck down bits as fast as possible.

We have termed this as "The Need for Speed" (hence, The Wireless Haven's slogan - "Your Wireless Connection for Speed"). Haha

[xdavidx]

I heard back from the manufacturer. I asked them if the antennas support MIMO. I got the following response back:

"Thank you for contacting us.

Yes, you can use 2 of these antennas connecting to a single LTE modem as long as you use the correct connector based on your situation."

I tried again:

"Thanks for the reply. Does the mount allow for turning each antenna to be angled 90 degrees from each other for optimizing MIMO?

And just to be clear, you are saying 2 being used together will support spatial multiplexing in order to double the speed?"

Their response:

"Does the mount allow for turning each antenna to be angled 90 degrees from each other for optimizing MIMO?
- The mountig will allow you to adjust the direction of each antenna - you would to purchase an actual mount to go along with the antenna.

you are saying 2 being used together will support spatial multiplexing in order to double the speed?
- Not certain. Utilizing 2 antennas in our conventional application allow us to pull from 2 different towers into our system. Depending on your application the antennas can dial into the 2.5-5 GHz range and, in theory, allow you to achieve a better connection."

I tried again:

"-- Is there some way to find out this information from someone else at Bolton? If I decide to buy 2 antennas, it would be due to them working for MIMO Spatial Multiplexing on a single LTE modem, with each antenna getting signals from a single tower. I'm working in the 700 to 2300 MHz LTE spectrum space. If they support MIMO Spatial Multiplexing, through 90 degree polarization, then the speed should be X when one is plugged in and about 2X when both are plugged in, when both antennas are near each other and angled 90 degrees relative to each other.

There are other people who are interested in the answer as well who I'm in contact with."

I'll post when I get a response again.

In theory, if they are separated by enough space, then it shouldn't matter. The polarized antennas can be close to each other, because they are polarized. Antennas that aren't polarized differently from each other should be able to get uncorrelated signals if they are far enough apart. tgoodwin's test would seem to show that the one grid antenna is benefiting from spatial multiplexing when paired with an omni antenna. The antennas are spaced far apart.

My personal take is that one or two grid antennas are only cost effective if:

- You are a great distance away from a tower, with LOS, and want to pull in multiple bands (we haven't proven they are equally good for multiple LTE bands yet, only Band 4 has been empirically tested). If you want to lock onto a tower far away on a single band, then you can get Yagi antennas priced more reasonably (see: swwifty's setup). And if you aren't too far away, then there are other, more reasonably priced antennas, that will pull in multiple bands and still give you plenty of increased gain (like the 15 dBi panel antennas).

- You want to lock onto one tower out of multiple that are in a relatively close radius and possibly weed out some of the interference from the other towers too. Going with a single grid antenna as the primary would allow locking onto a specific tower and the secondary antenna, as long as it can hear the tower in question, will tag along. The interference improvement would only help the streams on the primary antenna, not the secondary. Going with 2 grid antennas would allow for both the locking benefit and the interference reduction on both primary and secondary connections.

Whether extra hundreds of dollars is cost effective for X Mbps improvement is also a personal decision.

I *might* consider getting one of these in the future and pairing it with one of the 15 dBi panel antennas that I already purchased. And if I get a 5G modem that supports 4x4 MIMO, I might get another 15 dBi panel antenna, for a total of 4 antennas, with the primary being the grid antenna and the other 3 being the panel antennas.

[tgoodwin]

I have the antenna mounted on a Channel Master CM-3090. It allows adjustment on 2 axes. The antenna itself has the standard U-bolt that clamps around a pole. There is no way to rotate it 45 degrees.

CM-3090-4.jpg

[xdavidx]

Ah, I took his statements to mean that it didn't even have the U-bolt with it. I suppose you can tilt the channel master left or right? But maybe that only provides 20 degrees or so of tilt in those directions?

[tgoodwin]

Yeah. The tilt is enough to level it from the roof pitch. The bracket is to wide for the fascia to mount it level. So I mounted it parallel then rotated it.

I should have looked earlier. See what you were looking for in the instructions? It's a pro tip.

[xdavidx]

Pro tip!! Good find!

For anyone who doesn't see it, it says to use 2 grids for MIMO.

I was thinking of the left to right adjustment grooves on the J-Mount, rather than the ones that allow you to adjust the mast forward and backward.

I'm not able to visualize the parallel/rotated fascia mount.

Thanks for posting that document for the antenna. Am I correct that you can only tilt the antenna up or down by changing which hole one of the u-bolts goes into in their bracket? I guess with an adjustable J-mount, you could move the mast forward and backward to do that too, but it won't necessarily be in line with the tower, so would cause side tilt of the antenna too. Maybe that side to side adjustment on the J-mount could be used to compensate for that. Not an easy adjustment for aiming the antenna up/down. It would be nice if the antenna bracket had a slot instead of just a fixed number of holes.

[JimHelms]

Here is a diagram that we prepared for a customer showing the two separate MIMO configurations for our Grid antennas.

Grid MIMO.jpg

[xdavidx]

Very nice. We still aren't clear on whether the Bolton antennas are polarized correctly and whether they can be that close together and still get uncorrelated signals for spacial multiplexing MIMO. They are too darn expensive to buy and test! Good to know the The Wireless Haven antennas are polarized properly.

I wasn't even aware The Wireless Haven had grid antennas. That looks like a nice solution for anyone needing that frequency range. And this is the proper way to supply an LTE mount (quoted from the The Wireless Haven website): "The antenna is supplied with a 60 degree tilt and swivel mast mount kit which allows installation at various degrees of incline for easy alignment. It can be adjusted up or down from 0 to 60deg."

The link to that antenna: https://thewirelesshaven.com/shop/antennas/wifi-a ... -n-female/

[xdavidx]

Final reply from Bolton, in response to these previous replies:

"-- Is there some way to find out this information from someone else at Bolton? If I decide to buy 2 antennas, it would be due to them working for MIMO Spatial Multiplexing on a single LTE modem, with each antenna getting signals from a single tower. I'm working in the 700 to 2300 MHz LTE spectrum space. If they support MIMO Spatial Multiplexing, through 90 degree polarization, then the speed should be X when one is plugged in and about 2X when both are plugged in, when both antennas are near each other and angled 90 degrees relative to each other.

There are other people who are interested in the answer as well who I'm in contact with."

"We checked with the RF engineer from the manufacturer and he said yes if set up correctly."

[coldknob]

Sorry to be so late to this party but I believe I have found the ultimate "need for speed" solution.
Time won't allow me to go into all of the details of my journey, but needless to say I have amassed quite a collection of antennas, amplifiers, cables etc.

My home is in the mountains, closest towers 6 mi. away. Predominately band 12, 5mhz. but 2 towers with band 2, 15mhz which provide much better speeds. Trouble is I couldn't pick up the band 2 signal at all.

Fast forward several months and I settled on a parabolic grid wifi antenna which gave me consistent speeds of 25 down on band 2 and effectively blocked the band 12 signal. A few months of tinkering with the cables, antenna placement, etc. got my top speed up to 45 down, 15 up which is more than I need.

However, I always wondered what I could get if I had a mimo setup but 2 grid antennas at 45 degree angles is a larger footprint than I wanted and I was satisfied...until I came across the existence of a mimo feed horn that replaces the one on your present antenna. Found a Chinese supplier on eBay and 1 month and $55 later got it and bolted it on.

MVIMG_20190708_101435541.jpg

The results were pretty amazing. 70 down 25 up consistently, 85+ down frequently. This is from a tower more than 7 miles away. $40 for 2.4mhz parabolic grid antenna, $55 for 1710-2710mhz feed horn = 24X2 gain for less than $100.

[swwifty]

nice setup! Based on the bands you're using it sounds like you're using AT&T?

Mind sharing links to the antenna and feed horn you got?

[xdavidx]

FYI, I contacted the seller to see if they have any feed horns that go down to 700 MHz too. I'll report back if they reply.

They sell different models, but the one that @coldknob bought is the widest frequency spread. That's the only problem I see with it is that you can't get any 700 MHz bands, or 600 MHz for that matter, if someone was using T-Mobile, or 800 MHz for Sprint. But 700 is common to AT&T, T-Mobile and Verizon.

[xdavidx]

@coldknob: Are you saying you are getting a peak of 85+ Mbps from a single carrier Band 2, 15 MHz wide?

[coldknob]

Links to feed horns:
https://www.alibaba.com/product-detail/ ... 18598.html

https://www.ebay.com/itm/LTE-Fixed-Para ... wL8xcf4iEz

The grid is any generic 2.4ghz wifi antenna. I believe this is the one I have.
https://www.altelix.com/2-4-GHz-24-dBi- ... g24g24.htm

This link is where I first ran across this setup. Check out the feedback!
http://www.uplift.ph/Skywave/Products/SkyWave-Zensei-Ultra-MIMO

[coldknob]

@coldknob: Are you saying you are getting a peak of 85+ Mbps from a single carrier Band 2, 15 MHz wide?

With an ethernet connection I get 85+ with the mimo setup, 45 max with a single antenna connection. I'm on ATT, very rural, lightly used tower. Although the tower is 7 miles away I have clear line-of-sight, not so much as a leaf between my antenna and the tower.

[xdavidx]

Those are some crazy setups!

How would you like to climb this one? http://i.imgur.com/5IYr8t7.jpg

The one in the top of the tree was fun too: https://i.imgur.com/FHxPVMb.jpg

[xdavidx]

With an ethernet connection I get 85+ with the mimo setup, 45 max with a single antenna connection. I'm on ATT, very rural, lightly used tower. Although the tower is 7 miles away I have clear line-of-sight, not so much as a leaf between my antenna and the tower.

That is darn good for 15MHz. You have to be pretty happy with that.

So it only has Band 2? No other bands available?

Which modem are you using?

[coldknob]

That is darn good for 15MHz. You have to be pretty happy with that.

So it only has Band 2? No other bands available?

Which modem are you using?

Previously I had satellite internet and with no other options you could say I'm pretty happy. Streamed a 4k movie the other day and it didn't buffer once.

Band 12/17 is available but only at 5mhz. Signal strength is generally better than band 2 but the highest speed I could
attain was 15 down with 2 antennas.

My experimentation and setup was done with a we826/MC7455 but I recently hooked up a zte mf279 (ATT home base) and was surprised to get even better speeds. I think the mf279 has a cat12 modem which may explain the increase.

[xdavidx]

Very nice.

So you don't see band 12 at all on this feed horn, or you are manually blocking it in the modem?

Maybe the mf279 is picking up some other bands. Are you able to see which bands it gets?

[coldknob]

Very nice.

So you don't see band 12 at all on this feed horn, or you are manually blocking it in the modem?

Maybe the mf279 is picking up some other bands. Are you able to see which bands it gets?

Well I can't verify exactly what the mf279 is doing because it only shows signal strength and 4g or 4gLTE. However it performs similar to the we826 so I assume it's using the same band.

The we826 would aggregate bands 2 and 12 when set on all bands but I never saw any improvement in speed versus being locked on band 2. Band 12 is only 5mhz in my area.
The problem with the all bands setting was that sometimes the modem would select band 12 as the primary band (even with a 1900mhz yagi) which would kill my speeds. Locking on band 2 solved that problem.
I haven't had that problem yet with the mf279, even though I can't lock bands. I'm hoping that my current signal strength on band 2 (-92dbm) will eliminate that occurrence.

I have no idea what algorithm AT&T uses to select what band you're on but I suspect it's based solely on signal strength which makes sense when you're driving around in order to seamlessly switch between towers.
For data use from a fixed location bandwidth is the key. -115db @ 15mhz=good speeds, -90db@5mhz=crappy speeds.

I should note that I'm in a very rural, mountainous location. Many hills and valleys and many dead zones. Also I'm located on the edge of the National Radio Quiet Zone where cell phone signals are prohibited.
Every area is different but this is what works for me in my specific location.

Also keep in mind these antennas are extremely directional (we're talking millimeters making a difference). I can't actually see the tower I'm connected to but I know it's exact location and elevation in relation to my house by using google earth and other resources. It's akin to aiming a satellite dish, you get a good signal or you get nothing.

[xdavidx]

I heard back from the ebay seller. He doesn't have any horns that go down to 700MHz.

[swwifty]

Well I can't verify exactly what the mf279 is doing because it only shows signal strength and 4g or 4gLTE. However it performs similar to the we826 so I assume it's using the same band.

The we826 would aggregate bands 2 and 12 when set on all bands but I never saw any improvement in speed versus being locked on band 2. Band 12 is only 5mhz in my area.
The problem with the all bands setting was that sometimes the modem would select band 12 as the primary band (even with a 1900mhz yagi) which would kill my speeds. Locking on band 2 solved that problem.
I haven't had that problem yet with the mf279, even though I can't lock bands. I'm hoping that my current signal strength on band 2 (-92dbm) will eliminate that occurrence.

I have no idea what algorithm AT&T uses to select what band you're on but I suspect it's based solely on signal strength which makes sense when you're driving around in order to seamlessly switch between towers.
For data use from a fixed location bandwidth is the key. -115db @ 15mhz=good speeds, -90db@5mhz=crappy speeds.

I should note that I'm in a very rural, mountainous location. Many hills and valleys and many dead zones. Also I'm located on the edge of the National Radio Quiet Zone where cell phone signals are prohibited.
Every area is different but this is what works for me in my specific location.

Also keep in mind these antennas are extremely directional (we're talking millimeters making a difference). I can't actually see the tower I'm connected to but I know it's exact location and elevation in relation to my house by using google earth and other resources. It's akin to aiming a satellite dish, you get a good signal or you get nothing.

Nice work on this. Band selection is indeed done based on signal, but they always try to push you to the higher frequency bands, as the lower frequency bands are more valuable due to range and penetration.

I'm tempted to try this setup my self, but I suspect my wife wouldn't be too happy about an antenna that large on the roof, lol. The tower I connect to for AT&T service is about 5 miles away NLOS.

I had never heard of the National Radio Quiet Zone, super fascinating after reading up on it online.

[coldknob]

I heard back from the ebay seller. He doesn't have any horns that go down to 700MHz.

My understanding of parabolic physics is limited but I believe for a grid antenna to give 24dbi gain at 700mhz it would have to be more than 6 ft. in diameter. That's probably why you don't see them.

[xdavidx]

My understanding of parabolic physics is limited but I believe for a grid antenna to give 24dbi gain at 700mhz it would have to be more than 6 ft. in diameter. That's probably why you don't see them.

Good point. Generally, not as much gain is needed for the lower frequencies, since they travel farther. The antenna this thread started with has the following gain for the low frequencies:

14-17 dBi
600-960 MHz

That would be fine. I'm just not quite ready to give up on the dream of having the lower frequencies available for carrier aggregation too. It might be a foolhardy dream, however, since as @swwifty pointed out, they like to withhold those bands. In my case, I'm not given the 700MHz bands unless it is late at night, when the towers aren't busy.

I even found a tower that has at least 4 bands: B2 20MHz, B30 10MHz, B12 10 MHz and B66 10MHz. It won't give me B12 until late at night. It will give me B2 and B30 all day long, and then adds B12 at night. The only time I got B66 is when I aimed away from the tower enough that it couldn't pick up the higher frequency B30. Then it would give me B66! I haven't tried locking out B30 to see if it will give me B66 when aimed correctly, but that's on my list of things to try. But what is weird is that it will give me B12 if I am physically right next to the tower.

My only guess at their algorithm is that it it will give 3 bands if you are close and have really good signal strength, since they want to maximize the resources for people who can make the most out of them. The rules change when you don't have as strong of a signal. In that case they seem to want to reserve the lower frequency bands for people farther out (except for B2, which is always the primary and has the most bandwidth, with 20 MHz).

In theory, the towers can use your distance in (estimated by a round trip ping time), to determine which resource blocks to give you vs other users on the same band from another, approximately equi-distant tower, in order to reduce interference for those users in the overlapping coverage areas between the towers. In practice, I'm really not sure how much that is utilized.

Thinking about this more, *if* they are only using signal strength and not using the estimated distance, then if I can get my signal strength high enough, it might treat me like I'm right next to the tower and give me the lower frequency bands, even during prime time usage hours. If it will do that, then it might make sense for me to try a grid antenna, if it could get my signal strength high enough. My problem isn't being too far away from the towers. My problem is having too many towers around and too many people. So even if I only get the signal strength high enough for the primary antenna, that might be good enough, since they probably only look at that, not at some average of the two. My panel antennas get good enough signals for reasonable speeds. I just "need" (want) more bands, so that I have more bandwidth, hence more speed.

Lots of "ifs" here, but if the above works out, I *could* use a cheaper grid antenna that doesn't go down to 700MHz as my primary and, either use one of my panel antennas as secondary, or use the second MIMO connection of the horn style antenna that @coldknob is using. Since I only have a 3CA modem currently, B2 + B30 + B66 would give me as much as I can use anyway, and I wouldn't need B12.

It is a price difference of about $150 between the wide band grid antenna and the feed horn variant. Of course, if someone doesn't have secondary antennas, or needs a very high gain for the secondary antennas to get MIMO, then the feed horn antenna is a greater bargain vs buying 2 of the wide band grid antennas.

This is another MIMO feed horn that is priced less, but it is made for parabolic dishes. Not sure if it could be made to work with a grid: https://www.ebay.com/itm/Dual-polarity- ... 2797598911

I'll add to my list of tests to see at what signal strength, physically getting farther and farther from the tower, it stops giving me Band 12 and 66. Using the gain specs on the panel antennas vs the grid antennas might allow me to do the math to figure out if I'll be able to get to a high enough signal strength to fake out the tower into giving me the lower frequency bands. I'll also need to see if it will give me B30 and B66 at the same time if I lock out Band 12 and I'm close to the tower. They could have some other rule about not giving those 2 bands to the same user, or not support all the allowable CA combinations (of which B2+B30+B66 is one).

The variables never end with this stuff.

[coldknob]

The variables never end with this stuff.

Especially when the carriers give out zero information on this stuff and you have to try and figure out what they're doing,
I know where all the towers are in my area (they're all on top of mountains) but do you think AT&T would tell me which towers were theirs and what bands they use? I had to spend a month driving around with 2 cellphones figuring it out for myself. The cell mapper websites/apps were pretty useless because a lot of them are based on user info and I guess there are not many users in my area.
The best I found was Network Cell Info Pro.

[xdavidx]

The cell mapper websites/apps were pretty useless because a lot of them are based on user info and I guess there are not many users in my area.

I have plenty of people around and they were still pretty useless -- at least regarding location of the towers. I didn't do a ton of comparing the bands they show vs reality.

[coldknob]

Nice work on this. Band selection is indeed done based on signal, but they always try to push you to the higher frequency bands, as the lower frequency bands are more valuable due to range and penetration.

I'm tempted to try this setup my self, but I suspect my wife wouldn't be too happy about an antenna that large on the roof, lol. The tower I connect to for AT&T service is about 5 miles away NLOS.

I had never heard of the National Radio Quiet Zone, super fascinating after reading up on it online.

Thanks, my location has a lot of similarities to yours and reading your posts verified a lot of my own experiences. Thanks for sharing your expertise.
When I first got my z700a home base I thought I was in heaven because I could get 5 down, 1 up on HSPA+. Little did I know then what I was getting myself in to.
My wife thinks I'm obsessed and she's probably right but its been fun.

[swwifty]

Thanks, my location has a lot of similarities to yours and reading your posts verified a lot of my own experiences. Thanks for sharing your expertise.
When I first got my z700a home base I thought I was in heaven because I could get 5 down, 1 up on HSPA+. Little did I know then what I was getting myself in to.
My wife thinks I'm obsessed and she's probably right but its been fun.

It literally is a rabbit hole for sure! So glad to hear I was able to help in some way.

My wife thinks I'm obsessed too, but all of this has lead me to start a WISP where I live, so I think it's a good thing

[xdavidx]

My wife thinks I'm obsessed too, but all of this has lead me to start a WISP where I live, so I think it's a good thing

Now *that* is an idea!

Wonder how long it would take for the cable company to get me shut down if I tried that. Our house and the handful around ours can't get cable, but a lot of other people on the street can. I'm sure the cable company would wonder what is going on when all their customers drop, since DSL is *horrible* here.

[swwifty]

Now *that* is an idea!

Wonder how long it would take for the cable company to get me shut down if I tried that. Our house and the handful around ours can't get cable, but a lot of other people on the street can. I'm sure the cable company would wonder what is going on when all their customers drop, since DSL is *horrible* here.

well I'm not trying to resell my internet. I'm doing a full blown ISP with a carrier grade whole sale connection

DSL is the only option for most people around here, and the average speed is about 3-5mbps. I'm fortunate and actually get about 20mbps where I live, but not the case for most folks as I live close to the high way and the DSLAM.

You could setup some Ubiquiti radios to give your neighbors internet that have LOS for "free" and they give you something else for "free" every month (You don't hear that idea from me )

[xdavidx]

well I'm not trying to resell my internet. I'm doing a full blown ISP with a carrier grade whole sale connection

I was saying the cable company could try to create problems (even if there weren't any FCC rules being broken) on the RF front through the long range wifi network being set up. And it wouldn't be reselling of a cable connection. I'm sure they'd *really* love that!

How are you getting the carrier grade wholesale connection? Cellular or some form of wired?

You could setup some Ubiquiti radios to give your neighbors internet that have LOS for "free" and they give you something else for "free" every month (You don't hear that idea from me )

Trust me, that has been discussed in the past with one neighbor. Then they got their DSL upgraded. Mine was upgraded a few days after that. It worked for a month with fewer issues than I had previously. Then it all fell apart. The phone company tech couldn't figure out how to fix it, so I had to have them put me back to a slower speed that has issues now and then.

[swwifty]

I was saying the cable company could try to create problems (even if there weren't any FCC rules being broken) on the RF front through the long range wifi network being set up. And it wouldn't be reselling of a cable connection. I'm sure they'd *really* love that!

How are you getting the carrier grade wholesale connection? Cellular or some form of wired?


Trust me, that has been discussed in the past with one neighbor. Then they got their DSL upgraded. Mine was upgraded a few days after that. It worked for a month with fewer issues than I had previously. Then it all fell apart. The phone company tech couldn't figure out how to fix it, so I had to have them put me back to a slower speed that has issues now and then.

The cable company is setting up a wireless network? That seems odd.

Carrier grade connections are almost always fiber. I'll have a 500-1G connection at first, and then add multiples for redundancy upstream.

If you give it away for free, shouldn't be a contract violation, just sayin.

[swwifty]

Good point. Generally, not as much gain is needed for the lower frequencies, since they travel farther. The antenna this thread started with has the following gain for the low frequencies:

14-17 dBi
600-960 MHz

That would be fine. I'm just not quite ready to give up on the dream of having the lower frequencies available for carrier aggregation too. It might be a foolhardy dream, however, since as @swwifty pointed out, they like to withhold those bands. In my case, I'm not given the 700MHz bands unless it is late at night, when the towers aren't busy.

I even found a tower that has at least 4 bands: B2 20MHz, B30 10MHz, B12 10 MHz and B66 10MHz. It won't give me B12 until late at night. It will give me B2 and B30 all day long, and then adds B12 at night. The only time I got B66 is when I aimed away from the tower enough that it couldn't pick up the higher frequency B30. Then it would give me B66! I haven't tried locking out B30 to see if it will give me B66 when aimed correctly, but that's on my list of things to try. But what is weird is that it will give me B12 if I am physically right next to the tower.

My only guess at their algorithm is that it it will give 3 bands if you are close and have really good signal strength, since they want to maximize the resources for people who can make the most out of them. The rules change when you don't have as strong of a signal. In that case they seem to want to reserve the lower frequency bands for people farther out (except for B2, which is always the primary and has the most bandwidth, with 20 MHz).

In theory, the towers can use your distance in (estimated by a round trip ping time), to determine which resource blocks to give you vs other users on the same band from another, approximately equi-distant tower, in order to reduce interference for those users in the overlapping coverage areas between the towers. In practice, I'm really not sure how much that is utilized.

Thinking about this more, *if* they are only using signal strength and not using the estimated distance, then if I can get my signal strength high enough, it might treat me like I'm right next to the tower and give me the lower frequency bands, even during prime time usage hours. If it will do that, then it might make sense for me to try a grid antenna, if it could get my signal strength high enough. My problem isn't being too far away from the towers. My problem is having too many towers around and too many people. So even if I only get the signal strength high enough for the primary antenna, that might be good enough, since they probably only look at that, not at some average of the two. My panel antennas get good enough signals for reasonable speeds. I just "need" (want) more bands, so that I have more bandwidth, hence more speed.

Lots of "ifs" here, but if the above works out, I *could* use a cheaper grid antenna that doesn't go down to 700MHz as my primary and, either use one of my panel antennas as secondary, or use the second MIMO connection of the horn style antenna that @coldknob is using. Since I only have a 3CA modem currently, B2 + B30 + B66 would give me as much as I can use anyway, and I wouldn't need B12.

It is a price difference of about $150 between the wide band grid antenna and the feed horn variant. Of course, if someone doesn't have secondary antennas, or needs a very high gain for the secondary antennas to get MIMO, then the feed horn antenna is a greater bargain vs buying 2 of the wide band grid antennas.

This is another MIMO feed horn that is priced less, but it is made for parabolic dishes. Not sure if it could be made to work with a grid: https://www.ebay.com/itm/Dual-polarity- ... 2797598911

I'll add to my list of tests to see at what signal strength, physically getting farther and farther from the tower, it stops giving me Band 12 and 66. Using the gain specs on the panel antennas vs the grid antennas might allow me to do the math to figure out if I'll be able to get to a high enough signal strength to fake out the tower into giving me the lower frequency bands. I'll also need to see if it will give me B30 and B66 at the same time if I lock out Band 12 and I'm close to the tower. They could have some other rule about not giving those 2 bands to the same user, or not support all the allowable CA combinations (of which B2+B30+B66 is one).

The variables never end with this stuff.

I think you have it a bit reversed. LTE is a very complex protocol, and the UE (Mobile phone or modem) and the Cell are always providing feedback in a loop to make very complex decisions about band connections/transmission modes and such. Resource blocks in LTE are assigned based on feedback as well and utilized OFDMA (I recommend reading about this very interesting) It's not only signal strength, but that is for sure a big factor. I'd recommend reading the following article, it's quite technical, but will help you understand one aspect of how complex LTE really is:

http://www.sharetechnote.com/html/Handbook_LTE_TransmissionMode.html

Carriers almost always want the UE to connect to the higher frequency bands for a few reasons.

1. They are typically wider bandwith, and less congested.
2. Preserve usage on the lower bands, so they can be used for customers with devices with poor antennas (think cell phones) or devices that are far away or shadowed by obstructions such as inside a building and such.
3. They paid a lot of money for the lower frequency bands, so the less utilized they are the better. This means less network density (they don't have to build out as many towers) and customers are not complaining about coverage and performance.

Typically I don't want to do CA with lower bands cause they are almost always congested especially in rural areas because they cover large areas and theres typically a lot of obstructions. This is why I originally used some band 2 specific yagis to lock only onto that band. I found later that band 2 had a lot of multipathing that was causing poor performance, so I opted to swap out for some antennas that covered 700-2700mhz, so I could band 12 signal in better to help CA with that. Fortunately where I live AT&Ts band 12 isn't super congested, so I can still get 40-50mbps just on that band. Combined with Band 2 i typically get 65-80mbps down.

I recommend now almost always using wide band antennas so you can use all available bands, and let the logic in the LTE modem / cell decide what you should use. Antenna diversity is a good thing which LTE provides, but frequency diversity through different bands is also good as well. It's what makes LTE work quite well in very difficult RF situations.

[xdavidx]

The cable company is setting up a wireless network? That seems odd.

No. What I am saying is that in my situation, if a competitor sprung up that was taking away cable customers, they could face harrassment from the cable company. The form of that harrassment could be trying to create legal problems on the RF front. That doesn't mean they would have anything legitimate to complain to the FCC or local government about, but they could still do it. They've (cable companies) been known to aggressively target competitors before.

Carrier grade connections are almost always fiber. I'll have a 500-1G connection at first, and then add multiples for redundancy upstream.

That is great you are able to get that type of connection in the mountains. I figured you meant fiber, but didn't know you could get that to your house. Unless you are getting it to some other location and setting up the wifi equipment there? Sounds like the start of a fun adventure.

If you give it away for free, shouldn't be a contract violation, just sayin.

The policies are different per provider. Some are no sharing and some are no reselling and no sharing. I know AT&T has a no reselling policy. I'm not sure of the sharing policy. But what is policy and what people do aren't necessarily the same thing. Either way, my main neighbor has a fast enough connection now, even though it is only 22 Mbps, that he wouldn't be interested in setting up a wireless link between houses. There are many who are stuck with 6 Mbps or less DSL that has reliability issues too. It will be a happy day when that crooked phone company (Frontier Communications) loses out to some other, new technology (cellular or low earth orbit satellite).

I have considered approaching the city to see about setting up some sort of coop for a wisp. That's about as far as I've ever gotten. That, and looking into the equipment a bit. My main concern, since it is a populated area (semi-rural, but still enough affected people), is interferring with personal wifi networks. Strategically placed links that go through area with fewer houses might mitigate some of this. Setting up towers or finding existing structures to use (no mountains here) seems problematic as well.

It will probably never happen, but fun to think about when your DSL service is out.

[xdavidx]

I think you have it a bit reversed.

All I can say is that theory is fine, but it doesn't trump what is done in practice. If you are saying they won't give out lower frequency bands to those users close to the towers (those with strong signal strengths), that isn't the case here. I have physically been to the towers and been given low frequency bands, during prime time.

How providers dole out resources is not a standard baked into LTE. It is up to the providers to implement resource allocation in whatever way they want. And that can vary from one area to another, even with the same provider. So what happens with the towers near me may not happen for other people.

I am not claiming that everyone will be able to use higher gain antennas and band locking in order to get their towers to give them extra bands. I am stating that my towers appear to give out lower frequency bands to high signal strength users and low signal strength users, but not give them out to medium signal strength users, during times of congestion. I intend to empirically test this further, and if testing proves it out, I may opt for using one or more higher gain antennas.

Regarding not wanting to use lower frequency bands... I would rather have 2 higher frequency bands aggregated with a congested lower frequency band than only have 2 higher frequency bands. I am dealing with a 3CA modem, not 2CA.

If someone has to pick between a lower frequency band and a higher frequency band of the same bandwidth, then it would generally make sense to *pick* the higher frequency band, as you said. That assumes there aren't other users camped out on the higher frequency band, with LTE modems, downloading torrents 24x7. There will always be exceptions.

But *adding* a low frequency band where no other band exists, for a higher carrier count, with carrier aggregation, can only increase speeds.

At the end of the day, the name of the game is to get the highest average number of resource blocks assigned to your modem during daytime hours, assuming that is when most people generally want to use their lte modems. And, in general, the most likey way to maximize resource blocks is to have access to more aggregate bandwidth, with carrier aggregation.

There is a cost to higher gain antennas and tweaking modem settings, both in terms of price of the equipment and time investment. Most people will opt for and be overjoyed with one of the options you laid out in your great reviews and comparisons of antennas and leaving all bands open in the modem. What we are discussing in this thread is hot-rodding.

[coldknob]

What we are discussing in this thread is hot-rodding.

Hot-rodding! I like it.

You're right that the vast majority could care less what their highest speed is as long as they can stream Netflix, etc.

[swwifty]

Hot-rodding! I like it.

You're right that the vast majority could care less what their highest speed is as long as they can stream Netflix, etc.

And honestly now days thats all we really need. Video conferencing software typically uses less than 5mb/sec.

The average household only uses about 10mb/sec at peak times.

[swwifty]

No. What I am saying is that in my situation, if a competitor sprung up that was taking away cable customers, they could face harrassment from the cable company. The form of that harrassment could be trying to create legal problems on the RF front. That doesn't mean they would have anything legitimate to complain to the FCC or local government about, but they could still do it. They've (cable companies) been known to aggressively target competitors before.


That is great you are able to get that type of connection in the mountains. I figured you meant fiber, but didn't know you could get that to your house. Unless you are getting it to some other location and setting up the wifi equipment there? Sounds like the start of a fun adventure.


The policies are different per provider. Some are no sharing and some are no reselling and no sharing. I know AT&T has a no reselling policy. I'm not sure of the sharing policy. But what is policy and what people do aren't necessarily the same thing. Either way, my main neighbor has a fast enough connection now, even though it is only 22 Mbps, that he wouldn't be interested in setting up a wireless link between houses. There are many who are stuck with 6 Mbps or less DSL that has reliability issues too. It will be a happy day when that crooked phone company (Frontier Communications) loses out to some other, new technology (cellular or low earth orbit satellite).

I have considered approaching the city to see about setting up some sort of coop for a wisp. That's about as far as I've ever gotten. That, and looking into the equipment a bit. My main concern, since it is a populated area (semi-rural, but still enough affected people), is interferring with personal wifi networks. Strategically placed links that go through area with fewer houses might mitigate some of this. Setting up towers or finding existing structures to use (no mountains here) seems problematic as well.

It will probably never happen, but fun to think about when your DSL service is out.

heh, I wish I could get fiber to my house, but that would cost tens of thousands of dollars.

I'll have fiber at a central location in downtown and backhaul it wireless to various towers in the area, which will then serve customers off of that.

While starting a WISP is not easy, I'm learning very quickly its not a technology problem. The main issue is spectrum, and competeing with the big carriers that want it all. Unfortunately, the government only makes this problem worse than it is buy making it hard for WISPS to effectively serve customers and be cost effective.

[swwifty]

All I can say is that theory is fine, but it doesn't trump what is done in practice. If you are saying they won't give out lower frequency bands to those users close to the towers (those with strong signal strengths), that isn't the case here. I have physically been to the towers and been given low frequency bands, during prime time.

How providers dole out resources is not a standard baked into LTE. It is up to the providers to implement resource allocation in whatever way they want. And that can vary from one area to another, even with the same provider. So what happens with the towers near me may not happen for other people.

I am not claiming that everyone will be able to use higher gain antennas and band locking in order to get their towers to give them extra bands. I am stating that my towers appear to give out lower frequency bands to high signal strength users and low signal strength users, but not give them out to medium signal strength users, during times of congestion. I intend to empirically test this further, and if testing proves it out, I may opt for using one or more higher gain antennas.

Regarding not wanting to use lower frequency bands... I would rather have 2 higher frequency bands aggregated with a congested lower frequency band than only have 2 higher frequency bands. I am dealing with a 3CA modem, not 2CA.

If someone has to pick between a lower frequency band and a higher frequency band of the same bandwidth, then it would generally make sense to *pick* the higher frequency band, as you said. That assumes there aren't other users camped out on the higher frequency band, with LTE modems, downloading torrents 24x7. There will always be exceptions.

But *adding* a low frequency band where no other band exists, for a higher carrier count, with carrier aggregation, can only increase speeds.

At the end of the day, the name of the game is to get the highest average number of resource blocks assigned to your modem during daytime hours, assuming that is when most people generally want to use their lte modems. And, in general, the most likey way to maximize resource blocks is to have access to more aggregate bandwidth, with carrier aggregation.

There is a cost to higher gain antennas and tweaking modem settings, both in terms of price of the equipment and time investment. Most people will opt for and be overjoyed with one of the options you laid out in your great reviews and comparisons of antennas and leaving all bands open in the modem. What we are discussing in this thread is hot-rodding.

What kind of antenna were you using up close at the base of the tower? I've noticed while switching antennas, that just the internal UFL cables and SMA connectors act as an antenna that is good enough to pickup the 700mhz band, heh.

[xdavidx]

What kind of antenna were you using up close at the base of the tower? I've noticed while switching antennas, that just the internal UFL cables and SMA connectors act as an antenna that is good enough to pickup the 700mhz band, heh.

I saw this both with some small omni antennas and the 15 dBi panel antennas. I got 3CA with both types, with one of the carriers being Band 12.

[swwifty]

I saw this both with some small omni antennas and the 15 dBi panel antennas. I got 3CA with both types, with one of the carriers being Band 12.

oh, but the PCC wasn't Band 12, right?

[xdavidx]

oh, but the PCC wasn't Band 12, right?

Nope. I've never gotten that. I haven't tried blocking all bands except 12, though. That might do it.

But that's not my goal. I just want Band 12 or 66 as a 3rd carrier, to go along with Band 2 and 30, or along with Band 2 and 4 on another tower.

[swwifty]

Nope. I've never gotten that. I haven't tried blocking all bands except 12, though. That might do it.

But that's not my goal. I just want Band 12 or 66 as a 3rd carrier, to go along with Band 2 and 30, or along with Band 2 and 4 on another tower.

Gotcha that makes more sense. I thought earlier you were saying that band 12 was a PCC when you were closer to the tower.

Sorry I'm trying to do too much right now, lol, working 2 full time jobs.

[xdavidx]

Gotcha that makes more sense. I thought earlier you were saying that band 12 was a PCC when you were closer to the tower.

No, I generally get Band 2 or Band 66 as primary. I do have one tower where my notes show a possibly Band 12 as primary, but I could have written that down in the wrong order.

Sorry I'm trying to do too much right now, lol, working 2 full time jobs.

I'm sure that WISP is a *lot* of work. Hopefully a fun type of work.

[swwifty]

I'm sure that WISP is a *lot* of work. Hopefully a fun type of work.

Yeah it is. Hopefully it'll be a bit more fun once I start building out the network and hook up customers who are going to be thrilled.

Right now it's a lot of business planning. Although, I am looking forward to doing some long range testing with a few various potential radios I'll be using this weekend.

[xdavidx]

Yeah it is. Hopefully it'll be a bit more fun once I start building out the network and hook up customers who are going to be thrilled.

Customers are a nice thing for a business to have!

Right now it's a lot of business planning. Although, I am looking forward to doing some long range testing with a few various potential radios I'll be using this weekend.

I bet.

So are these towers owned by the cellular companies? The rates they charge for antenna space must be pretty reasonable to not put you in the red?

How many wireless hops do you have to take to get from the fiber up to the first distribution node?