Slow 4G connection only, but only sometimes -- SOLVED

New data:

Wife got home with ere Galaxy GS9 and I tested with it - turns out - as you said - cellmapper is out of date which is why I could not find the cell towers the T-Mobile CAN was finding.

They are ALL on the same physical tower - the 0.7 miles away tower through a TON of heavy woods - but then only Line-of-sight (except for woods) cell tower I have anywhere near.

in addition to the 12,66 it says in cellmapper, it has b2, b41 and b71.

So that accounts for every connection the can has made from anywhere in the house. and the PCI’s match and the cellmapper cell identifier when I could get a phone to give it to me.

Again, to describe my situation - imagine a canyon (Atlanta is hilly), and I live 2/3 way up the canyon - and down the canyon is basically straight west.

And I have NO west windows.

So the North and Sorth, west corner windows are the best available and where I get the best signal - EVEN in the basement, (which again, surprises the heck out of me).

so - if I come to a point of needing it - mounting a MIMO on the west side of the house - 3 stories up - should actually give me a SUPER signal … I should probably do that sometime before I can’t climb a 32 ladder anymore ...

as for other matters - not quite technical ...

I agree totally on the “tone”, iTineralot - in fact, I have ALWAYS for well more than a decade got prompt and helpful response from t-mobile service.  And I try not to comment on forums (and I don’t do social media), unless I have something to “add” to the conversation.  I have not bothered t-mobile this time, since I knew this was basically brand new with them - and the poor service call folks were probably covered up. 

I don’t comment “too” much about the limited S/W -because I am hoping that gets better - and the cell towers get better, etc. 

I don’t have an Apple anything - almost everyone else in my extended family does - and I am so fed up with M$ - that as soon as my current Dell Laptop finally dies (it is only 2.5 years old and the screen is “separating”), I think I will finally get a MacBook.

The problem is ALL the S/W I have built up (and written) over the years for Windoze.

Actually we have a covered screened porch on the Southside I could put the antenna on the side of the house on the South Side while standing on the roof of that porch. I had been thinking of that after discussing it with my wife. But I was wondering if I would be better off with the yagi versus the panel. Since I now know that I'm getting everything from one particular point.

I certainly hear you on the lightning arrestor. I actually have two 12 ft copper ground wires on a 12ft to part on the south side of the house grounding the house wiring. Because I found out that crazy builders had not even grounded the interior grounds. I also use a one shot lightning arrester on my ham antenna.  And PHYSICALLY unscrew the coax cable outside when not using it.  We have a LOT of lightning here in Atlanta.

iTinkeralot - Wow - nice.  Looks like a place I would like to live.

And looks like a place one of my brothers lives in TN.

The other two also live outside of small towns. 

BTW - on lightning - I am a REAL believer in it.

Neighbor on one side - lightning hit tree on the other side of the yard next to the other sides fence and exploded bark all the way across the fence in between and into our yard. from about 15 ftup the tree.

Houses across in the front and diagonal in the front - lightning hit trees and came in on cable line and from there fried just about every piece of electronics in the house - including garage door opener.

Both houses - separate times.

We had a big oak tree in the front hit - and one side of it from a split about 20 feet up - just died totally.

With that and having 4 different houses no more than 2 houses away have trees fall on them from near tornado winds over the years - I had all trees that might split the house cut down in the last 10 years.

A number of other folks have too.

This is a 45-50 year old neighborhood and one of the largest trees I had cut - within 50 foot of house - was an oak over 120 ft tall.  the base where they cut measured 5.5 feet wide - and I counted tree rings putting it back to about 1870 - soon after Sherman came through and burned nearly everything in Atlanta.

We also live in what is called tornado alley - or at least GA’s version.  We have had 3 tornados within 1 miles of house in the 37 years we have lived here. The alley stretches from mid AL to to here.

the other reason not to have large trees ….

The B66 is probably the primary signal and is a 4G LTE band 4 extension band operating on the 1700-2100 Mhz frequency. The n41 is a 5G NR channel operates at 2.5 GHz and is commonly seen on the secondary signal. it may likely be the tower the router associates with only has the n41 radio currently but there might be, or could be plans to also deploy an n71 5G NR there. The way to know the location of the tower and know what radios are there is to goto www.cellmapper.net and select T-Mobile as the carrier and then look for both 4G and 5G NR towers in your area. If you are going to target for the connection with an external MIMO antenna you need to confirm the PCI listed in the T-Mobile web UI with those towers you locate in the area. Then you can obtain the best results with less effort when making the connection. If you have a T-Mobile SIM and iPhone you can put it in “field test mode” *3001#12345#* and get pretty good information. Android phones have some locator apps available you can download for the task. It really does help. Using the information at cellmapper.net is very good to have as well. Don’t make assumptions. With just a little homework you can have the data in hand to make an informed decision as to where to locate the router and an external antenna. I spent considerable time with this router here and through multiple sources, including confirming on a call with T-Mobile support the tower coordinates and location. Using the cellular statistics, the overview information and status information from the router web UI over time recording values you can profile the operation of the router. If you are going to use an external antenna from waveform talk to those guys. They are helpful and the guide they have for the Nokia router is very good. 

I had a similar situation where my options for internet were limited to satellite or T-Mobile home internet. There was no contest there. I commonly get 10X the speed that I had with DSL from the service we had in CA and at less cost. It was a no brainer. I found the WIFI from this Nokia router/gateway actually works pretty well. I get very good results here in our 3300 sq ft home with just the Nokia WIFI. If the clients have good radios they do well. Only my MacBook Pro gave me grief so I wire it in with CAT 6 cable and it is good to go. It does take a little time to dial in the router location in some places to get the best results but it is well worth the time and effort to get the facts and experiment. Good luck, I hope it turns out to be as good a solution for you as it has for me. 

007BondM16 have you identified the tower connections 100% i.e. you know the PCI the router reports for both the 4G and 5G NR channels? Use http://www.cellmapper.net/ and confirm the tower(s) that are used as the PCI, physical cell ID can be determined and checked agains what the router reports. In my testing early on with this router I made incorrect assumptions and burned way too much time before I used cellmpapper to have 100% verification. I had a call with T-Mobile support they told me the tower coordinates, of the tower, and I confirmed it all using cellmapper.net to be clear about both the 4G and 5G NR connections. With that information you can know the radios that are present on the towers and see the 4G LTE cell identifications. Since the radios are not commonly omnidirectional it provides more clues as to what cell is sending in the location where you are. It takes a bit of time to pick it apart but is well worth the effort. Use that information and the reporting from the Nokia router to have the facts. One suggestion was made to use Opensignal to locate the tower but it does NOT provide sufficient detail nor does the app differentiate between carrier. It also does not currently support 5G NR signaling. Using www.cellmapper.net with a local computer takes a little to get used to but is very informative. When I open the web page it has me in the Atlantic ocean off the coast of Africa. If everything is blue zoom out till you can see where you are on the map. Then orient yourself to where you need to be and dial it in. You would find using the T-Mobile USA - 310260 with a 5G tower search to start is less cluttered than using 4G LTE towers as a beginning. There are fewer 5G NR towers reported so it is not as cluttered and you can better see the details of the map. Going from one to the other is a simple and quick selection. In 5-10 minutes you will clearly know what T-Mobile towers are around you and be able to identify the one(s) that you connect to via the PCI reported. The cell phone and router may or may not latch onto the same tower. An iPhone in field test mode will identify the PCI just as the router identifies the PCI for each signal. Start with confirmation of the actual tower the router links to. If you have speeds over 100 Mbs down then your RSRP and RSRQ and SNR values should be decent. 

Yet another datapoint for those who may be testing.

(BTW, iTinkeralot, I am definitely going to look into the external antenna - but quite busy at the moment, it may be 2-3 months).

I have continued monitoring my unit and doing speed test BECAUSE it switches among cell tower bands.

I have traced them ALL to the same tower I mentioned above - hidden by LOTS of trees but theoretically near line of site 0.7+ miles away.

What I have found is very interesting:

Speed measured with browser on laptop connected via ethernet cable:

Primary  Secondary  Down   Up

b2          n71             112.3    14.6
                                 177.2    20.2
                                 171.0    17.8
                                 126.6    18.7
                                 174.5    23.1
                                 143.0    15.0
                                 147.7    19.6

                                 124.5    19.5
b2          n41             165.9     7.4
                                 265.3     6.5

                                 170.3     4.64
b66        n41             186.5     9.13
 

As you can see the download speed increases, but the upload speed to to poor whenever n41 is in use,

whether b66 is primary or b2.

My guess is the n41 (2.5GHz) is scattered more by the trees and the n71 (600MHz) can make it through even though theoretically it is slower speed - but only, of course, given equal signal strengths.

It could also be that the n41 power of the T-Mobile CAN is insufficient to make it through the trees, for upload, but the cell tower has sufficient power to make it through the trees for reception by the CAN.

So you might keep that in mind - that the n71 will both go farther and through more obstacles than n41.

(And forget n260 or n261 - unless you live next to them with line of sight! - I think they are for mounting on buildings only.)

Another thing to consider - I think the cell towers are never omni-directional.  I.E. they have beam antennas that cover various beams depending on the need.

For example, one near me is pointing AWAY from me and is a VERY narrow beam designed to go straight down an incredibly heavy traffic road.

So even if you have 5 bands on a tower (I have at least that many on the one mentioned), They may or may not be pointing in your direction.

So it could be that n71 is more in my direction than the n41 one.

one thing you learn as an amateur radio operator - radio wave propagation is extremely complicated when you through in all the variables.

itinkeralot, from that photo, has a near perfect line of sight - assuming that antenna has a beam pointing in his direction.

Yes rotation even very little can make a big diff.

I have not found an online source of tower data that is really even 75% up to date. At least in my area half of the towers are not even in the correct lat long. The 5G is almost not at all up to date as to which tower has what. I read that many towers and building are lower so they don't have to register the details with the FCC. Anyways I will have some more time to play in a few days had to do some pay the bills work lately.

But past few days I have not moved my can and speeds vary wide from 1 to 150 various times of the day. When I check the can signal always the same so seems to be congestion.

I do love all the in depth details you are going into I think it helps.

I hooked up an external antenna and it helped not one bit. I still just get 2 bars on 5G as a secondary connection to go along with my 3 bars of 4G coverage, Walking around doesn’t seem to improve the connection but it’s hard to tell since the meter on the cylinder just shows the best 3bar connection and not specifically 5G. I don’t have a 5G phone to test with.

Looking at the Cellmapper web site I don’t see the only 5G connection I can get; N41 / PCI 7.

Supposedly my area has “5G Ultra Capacity” coverage but the towers to support that don’t show up on cellmapper.net; mostly just along the freeways. I guess it only has what users submit.

Maybe I should hook up the external antenna to the 4G line and try to get more than 3 bars.

Usually the 4G is much better than what I used to have but sometimes it’s spotty with even poorer performance than my old 10 megabit DSL.I suspect that the router doesn’t do that great of a job handling the drastically shifting speeds available so I get lots of bufferbloat and whatnot. Upload speeds are often < 1 megabit with tests often failing to even complete when there are slowdowns.

Edit: I still get 2 bars on my antenna for 5G even with the antenna disconnected so maybe I didn’t get it connected right. The again, it was the same with the internal antenna so perhaps I just can’t get a freaking 5G signal at my house despite being in a “5G ultra capacity” zone. :|

Metric is not CID, but CGI. 

Just correcting the record. 
 

Dennis

Depends on combination of bands as what download speed is. B2 with n71 I average over 50 Mbps download. But a lot more buffering trying to watch tv. B66 with n71 getting 15 to 40 Mbps has much less buffering. The horrible scenario is no secondary band. I too often get B71. Two bars but only a few kbps. Usually if I get any at all on B71 it’s 100 - 150 kbps!

One thing to be aware of with wireless signals — and this applies to both the OP’s experience and what @tjweller just posted — is that it’s important to understand how obstructions affect signal quality, AND the physics of how signals are obstructed. One of the most counterintuitive aspects is that if a signal has to pass through a solid object (a wall, window pane, etc.), the closer you get to that obstruction, the more of the signal it blocks.

The reason is simple trigonometry: When you pass through a solid object at an angle, it appears thinnest when you take an exactly perpendicular path (pass through it head-on). If you approach it at an angle, the farther from perpendicular you get, the thicker the object appears.

Consider this diagram:

The red arrow represents a perpendicular path, the thickness of the object is the same as triangle side ‘a’. But if we take the blue path, the length of side ‘c’ represents the apparent thickness of the object (in terms of the wireless signal). The length of ‘c’ can be computed in terms of ‘a’ and the angle at ‘x’. (See formulas. The last one just solves from Pythagoras: c² = a² + b².)

When x is 45°, c is roughly 1.4×a. When it’s 30°, c is 2×a. At 15°, c grows to nearly 4×a.

Now, a wireless signal is nothing like a straight arrow; that’s a massive oversimplification. In truth, the properties of wifi signals are almost impossibly complicated, and calculating or predicting them is a science unto itself.

But the basic premise here still holds: A solid obstruction like a wall will obscure more of the signal being received at steep angles, when you’re right next to it.