Thermal Issues With Home Internet Gateway/Router as a Cause of Dropped/Slow Internet or Wi-Fi Connection


Badge

I am a Newbie here and a fairly new T-Mobile Home Internet customer. But my background is electrical engineering and I've been around a while. Like many community members I have experienced frustrating degradation of my internet connectivity after a period of operation, which is temporarily fixed by cycling the Gateway's power (otherwise known as a soft reset, not the paperclip nuclear option). The standard fix per T-Mobile Tech Support is to just perform that reset, which isn't a “fix” at all. Though many others have noted the thermal connection, I feel it’s time we get organized on this.

 

Without going into great detail and after surfing a number of forum threads and through my own direct experience, I have become convinced that the primary cause of the various complaints expressed by posters is heat buildup in the Gateway's enclosure and a subsequent compromise in performance in one or several of the Gateways numerous devices. There could be multiple devices affected, manifesting in different failure modes, but one primary cause – heat.

 

What I'm asking community members to do is to test this theory by keeping a record of the time from device reset (powering down, then back up) to failure or degradation of either internet or wi-fi connectivity or any other failure mode, and then to apply some means of enhanced airflow through the Gateway's enclosure, whether it be placing a muffin fan on top of the Gateway (blowing up, not down) or simply sitting the Gateway atop an A/C vent in your home with constant airflow and recording any difference in the time between reset and the failure. Repeating this sequence several times and then posting your results here, on this thread, will be most helpful.

 

If, as I strongly suspect will be the case, a correlation between enhanced airflow and increased “up-time” is documented in one place (here) by multiple forum members, we will have a powerful tool to get T-Mobile's management to fix this problem. Please don’t post here unless you have performed the requested action; there are many other threads to use.

 

What do you say?


19 replies

Userlevel 3
Badge +4

Been there.  Done that.  Here’s my public record of troubleshooting over the last few months.

I summarized my findings for easy reference here.

And I will add that my current uptime is only 3 days and 16 hours.  The last reboot I documented in my Weekly Reboot thread was probably around a week ago during a period of heavy bidirectional transfer.  But then I had to reboot a few days ago when I wasn’t even on my computer.

Rebooting is annoying but I’d rather do that once a week as opposed to the issues I was having with my previous provider where a reboot did nothing.  I’m in more control now.  I’m only using T-Mobile for home internet and my phones have been on Verizon for years.  If Verizon starts offering a similar service in my area I will be quick to try it out and switch if it’s more stable.

Badge

Having that same issue my connectivity started to become an issue and I’ve only had it for 2 months. Was very good WiFi strength to now currently weak WiFi strength. I’ve restarted the home gateway but I’ve never gotten back to very good WiFi strength even though I have coverage. Lately even my cell phone (iPhone 11 pro max) has been experiencing terrible cellular coverage having only one bar. 

Badge

I’m having to update my post here as the edit function doesn’t seem to exist.

IMPORTANT: To reboot, use the power switch; power off for 30 seconds and then power back up using the switch. If you attempt a reset by unplugging the AC adapter a portion of the modem will remain powered by the battery. I’m not sure you can get an effective reset by using that method.

Badge

I had same problem was constantly losing internet during May.  It went down 8 times in 30 days, always came back right back on after power down and back on with button on back of 5G tower.  Listened to advice on forum and ordered USB fan to set 5G tower on, that was on June 10th, it is now July 11th and haven’t had a single day with needing to do the power down.  Definitely a overheating problem in their 5G tower design.  

Badge

Great, Smoky, this is exactly the kind of documentation we need. Now if you could do one more thing: Take the fan away, set the tower up like it was when it was losing connection and see how long it runs without needing a reset. It would also be helpful if you can verify that the extra airflow is the only thing that changed, that is, you didn’t move the gateway to a new spot, etc.

We had the same problem. To me the root cause was the advice given to us subscribers to place the unit near window.

I moved the unit place with clear view via double glass sliding patio door that is 18 feet away. The modem is sitting on a top of 5 feet high cabinet.

No problems anymore; both up and down load speeds are excellent.

No outage over the last 2 months. Bar count is absolutely stable.

Our roof (including the porch) is metal.  No problems either (no Faraday cage effect).

Badge

@ Stary, Thank you. Do you have any sense of whether the original window location was receiving more direct sunshine than the second location? T-Mobile’s installation instructions cautions customers not to place the unit in direct sunlight, which of course translates to increased temperatures.

Badge

The existing design of the 5g gateway alows for a 1/4 inch area at the bottom for incoming airflow. Besides adding the fan at the top, I also added 1/4 inch high self adhesive feet to the underside of the unit to double the intake flow area. Unfortunately, my gateway always runs slow so that it is unusable. T-mobile told me that they are working on my nearby tower (for three months now).

been moving it. and it over heating  with 2 device. both passive connection.

Badge

I have now not had a loss of internet since June 10th.  The fan solved the problem so far.  
 

The 5G cylinder sits in the exact same place as it did during the many May failures.  It is just raised about 1 inch by sitting it on top of the USB fan. I made no other changes to the setup, not even rotating the 5g cylinder.  
 

I went from seemingly random failures to none.
 

I recently added t-mobile’s YouTube TV offer and no problems.   I was concerned it wouldn’t show local channels with the home internet jumping around, but the location is determined by your cellphone. Very pleased so far.  

Userlevel 2
Badge +1

I have not seen a single user report the actual hardware version. This may be an important factor in the equation. The HTML reporting will show the hardware version. I know mine is 3TG00739AAAA and I do not see heat related issues. Is it important? I cannot say for sure but it is a data point that should be considered and posted. I agree with Farjohn 100% the more facts people report in the thread the better the profile of the facts.

The point made about using the power button is an excellent point. If the router is unplugged and the power switch is still on then it stands to reason the battery can still provide power to the LTE and 5G radios but NOT the WIFI radios.

Has anyone with heat issues looked into the top of the can with a flashlight and examined the heat sinks? I recall seeing two in opposition to one another while working with my router. Heat sinks can fall off when the adhesive is not applied properly so look at the heat sinks. If they are not secure the chips are not getting proper heat transfer away. Shake the can and see if it rattles. If it rattles there may be a heat sink that has fallen off the board. In manufacturing there can be equipment runs that are made with materials that are not to spec and this can lead to batch failures. The layout is sort of bilateral with the antennas in a radial pattern around the outside of the internal board. There are (4) WIFI and (4) 5G alternating about the inside of the router. It is common that serial numbers have details for model and other aspects of the appliance in addition to its individual part. If T-Mobile were to provide information on the part(s) of the serial numbers that could be relevant to expose a serial run of components it might be a problem due to components or heat sink glue that is not to spec for a build run that can clarify why some people have head related issues. Adding a fan is a work around. Identification of the root cause is what needs to happen. Data. Not what we think but what is KNOWN. As many facts as possible.

  • hardware version
  • software version
  • serial number of units, (sequential pattern could indicate bad manufacturing run)
  • heat sinks secure/present
  • fan used or not (air transport bottom up)
  • environmental factors (warm, cool, sunny, shade etc…
  • number of wired and/or wireless devices recorded on the Overview page
  • details recorded (bullet the facts) make things obvious and quick to consume
  • alarms reported on the LED panel in the top of the can (yes/no)
  • indication of periodic element or total random
  • Report the upload and download of cell data and primary/secondary channels
  • record the “Running Time” if possible with issues present from the HTML page (192.168.12.1)
  • Record the “Data Sent” & “Data Received” values from the Status/WIFI reporting (this may help relate the traffic loading over the uptime period) 
  • SHARE this collection of data points with T-Mobile support engineers
  • Add any other data points, facts that you feel may be important
  • Discuss common factors in the thread with others
Userlevel 2
Badge +1

Thank you Farjohn for reminding me of all the obvious things I know from working with enterprise trouble shooting for 22 years. The facts shared and reported for others and for T-Mobile support will help lead to finding the root cause. If they see the pattern they can take the gear, pull it apart and put them in heat chambers to confirm with testing that the components behave according to spec or not. I am sure if the hard facts get back to the Nokia engineers they can identify the problem and fix it.

I have not seen a single user report the actual hardware version. This may be an important factor in the equation. The HTML reporting will show the hardware version. I know mine is 3TG00739AAAA and I do not see heat related issues. Is it important? I cannot say for sure but it is a data point that should be considered and posted. I agree with Farjohn 100% the more facts people report in the thread the better the profile of the facts.

The point made about using the power button is an excellent point. If the router is unplugged and the power switch is still on then it stands to reason the battery can still provide power to the LTE and 5G radios but NOT the WIFI radios.

Has anyone with heat issues looked into the top of the can with a flashlight and examined the heat sinks? I recall seeing two in opposition to one another while working with my router. Heat sinks can fall off when the adhesive is not applied properly so look at the heat sinks. If they are not secure the chips are not getting proper heat transfer away. Shake the can and see if it rattles. If it rattles there may be a heat sink that has fallen off the board. In manufacturing there can be equipment runs that are made with materials that are not to spec and this can lead to batch failures. The layout is sort of bilateral with the antennas in a radial pattern around the outside of the internal board. There are (4) WIFI and (4) 5G alternating about the inside of the router. It is common that serial numbers have details for model and other aspects of the appliance in addition to its individual part. If T-Mobile were to provide information on the part(s) of the serial numbers that could be relevant to expose a serial run of components it might be a problem due to components or heat sink glue that is not to spec for a build run that can clarify why some people have head related issues. Adding a fan is a work around. Identification of the root cause is what needs to happen. Data. Not what we think but what is KNOWN. As many facts as possible.

  • hardware version
  • software version
  • serial number of units, (sequential pattern could indicate bad manufacturing run)
  • heat sinks secure/present
  • fan used or not (air transport bottom up)
  • environmental factors (warm, cool, sunny, shade etc…
  • number of wired and/or wireless devices recorded on the Overview page
  • details recorded (bullet the facts) make things obvious and quick to consume
  • alarms reported on the LED panel in the top of the can (yes/no)
  • indication of periodic element or total random
  • Report the upload and download of cell data and primary/secondary channels
  • record the “Running Time” if possible with issues present from the HTML page (192.168.12.1)
  • Record the “Data Sent” & “Data Received” values from the Status/WIFI reporting (this may help relate the traffic loading over the uptime period) 
  • SHARE this collection of data points with T-Mobile support engineers
  • Add any other data points, facts that you feel may be important
  • Discuss common factors in the thread with others


Great list, I think the only other thing I would add is ambient temperature. You do mention environmental factors (warm, cool, sunny, shade etc…) that can be helpful say if its in front of a window or not. Then again my windows don't shot much of a change in temp with the double glass UV prevention coatings.
So if your in a climate controlled environment the ambient temp may also play a part in this troubleshooting.

Userlevel 2
Badge +1

Thank you for the response. I was just thinking out loud and sharing. Right ambient temperature can be a factor. It would be helpful if T-Mobile support was more proactive and asked such questions OR better yet would leverage the software to automate the collection of key data points and use a template of questions to better profile customer problems. Hopefully T-Mobile will be able to improve their 5G deployments for more consistent service.

Userlevel 2
Badge +1

Just an observation I have made regarding the drops. I can see service here tends to be delivered from two different towers. The tower at 5.3 miles 360 degrees, North, is where the secondary channel is delivered. The 4G LTE “primary” signal appears to be delivered from a tower 270 degrees, NW roughly 2.5 miles away. When the 5G secondary channel drops this becomes clear as traffic appears to be interrupted and of course the routing tends to be slow to recover and of course more latency and slower traffic. It is rather difficult to locate a router where the primary and secondary signals originate from two different locations and have reliable results. It appears to me most of the signal disconnects I have seen have been related to the loss of the secondary 5G channel and I am pretty sure this is mostly due to T-Mobile doing work on the tower equipment. It does make it difficult to know when T-Mobile is not transparent about service work and service delivery. The system may be designed to do traffic hand off but the process does not appear to be seamless. 

Badge

I have now not had a loss of internet since June 10th.  The fan solved the problem so far.  
 

The 5G cylinder sits in the exact same place as it did during the many May failures.  It is just raised about 1 inch by sitting it on top of the USB fan. I made no other changes to the setup, not even rotating the 5g cylinder.  
 

I went from seemingly random failures to none.
 

I recently added t-mobile’s YouTube TV offer and no problems.   I was concerned it wouldn’t show local channels with the home internet jumping around, but the location is determined by your cellphone. Very pleased so far.  

I now have gone over two months without a reset since adding the fan below the unit.  Plus my usage has gone up as we now have YouTube TV in use regularly.  I get 80 to 120 down and 20 to 50 uploads.  One thing I forgot to report is that I have the WiFi turned off.  I go out the Ethernet port to an eero mesh system to supply WiFi coverage in my house. 
 

It is obvious to me there is a heat problem at a minimum in the design of the 5g cylinder.  

Userlevel 2
Badge +1

There may be a heat related issue with “some” of the routers. I have had mine since early January 2020 and I really cannot say with any certainty that it has had any heat related problem. I do believe there may be some of the routers that are from a given manufacturing run that for some reason are more prone to over heating. It could be some component(s) out spec. These probably should have had some sort of fan included in the unit but I see no evidence of one. There are two fair sized heat sinks but still that might not be sufficient. I put a small fan under mine a week or so ago but I can’t say it makes any difference. I am more inclined to believe the majority of the disruptions I have had are related to maintenance work on the tower where the 5G signal is broadcasted. From all indications the primary signal comes from the tower north of here and the 4G LTE signal comes from another tower off to the NW at 270 degrees. When the 5G, secondary signal, falls and the router has to use the 4G LTE alone the router always experiences a period of disruption. These stationary routers should not be treated as phones are with respect to signal delivery it seems. They don’t appear to handle channel transitions very well. When both the 4G LTE and 5G signals are present it works pretty well. The 4G LTE delivery is just meh… to blah

Badge +1

Yes, I don’t think a fan is necessary unless you have the gateway in a windowsill in the sun or something like that. People think heat emanating from the top is bad but actually that’s a sign that the big heatsync inside is doing it’s job and moving the heat out from the internals. 

Userlevel 2
Badge +1

If T-Mobile support engineers consistently requested specific information about the routers and recorded it so they could establish a thorough profile of the population of routers where customers are reporting behavior that appears to be heat related they might just identify a batch run of routers that have a specific related fault or a common out of spec component that results in the behavior. The need for a reboot seems more related to memory corruption or a buffer issue. If the issue “is” heat related I would be inclined to believe a cool down period would be required to improve behavior for any amount of time. The T-Mobile engineers need to gather the data, provide it to engineering, and focus on resolution not temporary workarounds that do not resolve the issue. With remote access to the routers the T-Mobile engineers should be able to gather data to build a clear profile of the routers involved. 

Reply