One legitimate concern with the new Amazon Fire TV Stick 2 is how well it handles heat dissipation. Cramming in a quad-core CPU into such a tiny package, without any forced airflow, could be a recipe for disaster if heat isn’t managed correctly. Many 1st-generation Fire TV Stick owners know too well that it can get quite toasty, which is likely why Amazon under-clocked its dual-core CPU to 1Ghz, even though it was manufactured to run at 1.2Ghz. To determine how well the 2nd-gen Fire TV Stick handles heat, I’ve taken internal and external temperature measurements of the 1st and 2nd-gen Fire TV Sticks, both at idle and under a 100% CPU load stress test. I’ve concluded that, while the 2nd-gen Fire TV Stick’s CPU runs at a notably higher temperature than the 1st-gen Fire TV Stick’s CPU, the 2nd-gen Fire TV Stick dissipates heat better, which results in lower external temperatures.
My temperature test was fairly straightforward. I left both the 1st and 2nd-gen Fire TV Sticks off in the same room for an extended period of time to ensure they were both at room temperature, which was right around 80 °F. I powered both devices on and let them sit idle on their home screens for 5 minutes, then I started a stress test app which maxed out their CPUs at 100% usage.
I measured temperatures both internally at the CPU and externally on the exterior housing. For internal temperature measurements, I wrote a script that polled each device once per second and recorded the CPU temperature reported by the OS. For external temperature measurements, I manually measured and recorded the temperature every 10 seconds during the stress test, using a digital laser infrared thermometer. I also measured the external temperature just before powering on, 1 minute after booting up, and 5 minutes after booting up, to establish an idle baseline temperature.
As you can see from the graph above, idle internal temperatures reported by the CPU are about 10 degrees apart, with the 2nd-gen Fire TV Stick running at a higher temperature. However, external temperature measurements for the 2nd-gen Fire TV Stick are about 8 degrees lower than the 1st-gen Fire TV Stick at idle. With an internal versus external temperature difference of about 25 degrees for the 1st-gen Fire TV Stick, and a difference of about 40 degrees for the 2nd-gen Fire TV Stick, it appears the newer device does a much better job of dissipating its heat away from the CPU.
At 100% CPU load during the stress test, the temperature gap between the 1st and 2nd-gen Fire TV Sticks grows greater than at idle. I suspect both devices are thermal throttling at some point during the test, but the larger gap could be an indication of the 1st-gen Fire TV Stick thermal throttling earlier. I ended the stress test earlier than I originally planned because the 1st-gen Fire TV Stick disconnected from its remote and became unresponsive. Temperatures for both devices had not plateaued at that point.
Overall I’m pleased with the performance of the 2nd-gen Fire TV Stick in this test. I expected its much more powerful CPU to run hotter, but it’s pleasantly surprising to see that doesn’t translate into a device that feels hotter than the 1st-gen Fire TV Stick. Good heat dissipation is a serious concern for these devices, both for longevity and performance, since they tend to get crammed behind already hot electronics. There’s no point in a powerful CPU if it’s just going to throttle down its performance because it gets too hot. These test results indicate that heat should be less of a problem with the 2nd-gen Fire TV Stick than with the 1st-gen device.
I purchased 3 new fire sticks and had to return 1 because it overheated and shut down and restarted with temperture high on screen
Ricardo – Do you remember what you were doing that causes the stick to overheat?
i bought a refurbish fire tv stick 1st gen with voice remote and it overheats, i get the message after running any app. i had to place 2 heat sinks off a pc motherboard on both sides of the stick to cool it down, its now works without the shutdown message. Amazon uses rubber heat sinks on the inside of their sticks but i believe they should create ventilation slots on the casing like the roku 1 stick or use metal heat sinks, these things get really hot.
I’m probably confused, but wouldn’t it be the opposite? I think I’d rather see the internal temp less, and the exteral(i.e. away from the SoC) hotter. Wouldn’t the hotter the external casing be, means more heat drawn away from the internal components?
I bought two of these, and so far have only used one. I noticed the outside stays fairly cool compared to other streaming devices. I took that to mean that it generated less heat, but looking at your graph it would appear more heat is being trapped internally.
No your not confused, but correct.
You do not want the heat trapped inside where there is almost no chance of cooling, but rather you want the heat to be conducted to the external casing where it can be cooled by outside air flow. The better transfer of heat from inside to outside via paste, heat-sink, etc the cooler the inside will become.
Replace its current packaging with aluminum or copper (which transfers/conducts heat even better) and paste the SoC right to it, watch how hot the casing becomes, it dissipating/conducting heat from the SoC to the casing, now cool that casing using convention (or for more fun liquid nitrogen) to move/remove heat and the result will be the SoC becoming cooler.
Sorry been an Engineer for 22+ years and got a little carried away.
BTW Before everyone yells at me about Wi-Fi, Bluetooth, etc I know putting it in aluminum or copper would mess with the transmission of signals, the above was just an example of what would allow better heat management.
I realize now that I left out an important point in describing my measurement procedure that helps explain the reasoning for my conclusion. Before starting the test, I turned the devices on and ran the IR thermometer across the entire housing to locate the single hottest point. (This is why I needed to let them sit for a few hours to bring them back down to room temperature.) I marked that location with a dot so I could measure the same spot every time for consistency. The exterior temperature varied greatly, by as much as 20 degrees, depending on where I pointed the IR thermometer. The 2nd-gen Fire TV Stick exterior temperatures at different points did not vary nearly as much as the 1st-gen Fire TV Stick. This is why I concluded that the 2nd-gen Stick was dissipating heat more efficiently, across the entire housing, where as the 1st-gen Stick was letting heat out less efficiently across a smaller surface.
My conclusion could absolutely be wrong. I’m not saying it isn’t. I just want everyone to have all the facts and decide for themselves.
Thank Elias. While I am not sure I interprete the results the same way, it is great information.
Do you feel confident that the OS is reporting the correct internal temps. I can’t remember if it was the old stick or the FTV box, but I could swear Kodi used to have some really odd tempurature readings. Maybe that is a Kodi issue though, as I am not sure if it gets it’s readings direct from OS.
On one of my boxes, can’t remember which, I have even seen negative temps reported.
No, I’m not entirely confident with the OS’s CPU temperature readings. I was mostly interested in the temperature changes relative to each device, which is why I still published this article, even though the actual temperature value might be wrong. Meaning, if you remove the Y-axis values in the chart above, it doesn’t really effect the article much because I’m more interested in temperature changes than exact readings.
Kodi definitely does not report temperature correctly. On the 1st-gen Stick for example Kodi just displays a question mark for temperature, so I wouldn’t trust it’s reading on any device. There are several different temperatures reported by the OS. I tested with them all and went with the one that seemed to be correlated with the CPU.
For those curious, temperature readings I used were from:
1st-gen = sys/class/hwmon/hwmon1/device/temp1_input
2nd-gen = sys/class/thermal/thermal_zone0/temp
To see the temperate value via ADB, run the command:
adb shell cat [PATH]
For a constantly updating readout of temperatures, on Mac/Linux, run the command:
while true; do adb shell cat [PATH]; sleep .9; done
Also 200 F seems really hot. I wouldn’t be surprised to see a future software update that underclocks these new sticks.
200 degrees is very very hot. That seems like internal limiters aren’t kicking in.
I’m no expert on the matter, but I don’t think I have ever heard of any chip running that hot. I wonder if the software/OS is not reporting the temp correctly. Over the years, I have seen instances where streaming devices display clearly incorrect temps.
I had installed a couple apps but did not even get to play a video on it
And it overheated. I think it was missing thermal paste because it just happen really fast without out really using it.
Yeah, probably just an odd defect then.
Another galaxy note 7 repeat, soon to catch on fire?
There is no battery, so probably not, but >200 degrees sounds pretty HOT!
Just had one over heat on me showed a thermometer on screen then went all yellow pixelated screen and rebooted and it was real hot to the touch
That’s two now. Not a good sign.
Omar – Was it like Ricardo’s, where is over-heated right away, or was it a stick you had been using for a few days?
I wonder if defective batch went out without thermal paste or something.
I was loading It with software like Kodi and restoring a backup
I am not going to touch one of the new sticks until this heat issue improves. I sideload sticks for friends and sure would not want to have any part of a stick setting fire to their house. Fraught with peril.
I went to Amazon and searched the 1,134 review for “heat”. No reviews mention the word. Therefore, I am hoping just a small defect batch slipped out, rather than this being a wide spread issue.
I have a second, unopened stick that I was sitting on. I wonder if I need to open it now, just to make sure it is ok.
Can you test PS Vue app on Fire Stick 2? It was so slow on original stick.