NES-101 Heat Research
I had a customer ask me about heat build-up.
“I’m wondering if you’ve ever put vent holes leading into the cartridge bay? as it can get very warm in there. I don’t know if it gets hot enough to worry about but, I thought I’d ask anyway.”
I had wanted to put my thermal cam on it for a while, now I had an excuse to do some research.
I made a short video of the cam taking pics of a top loader board and put it up on youtube. Just to give you an idea of how the pics were made.
Here are pics of the NES-101 and game cartridge after leaving it powered ON for 1 hour with a game in it, not playing it, no r/f or controllers hooked up either.
The numbers are hard to read once the pics are shrunk, but the first pic shows most of the heat on top of the console is above the voltage regulator’s heat sink, no surprise, 96 °F, but notice some in front of the cart too, strange. The second pic is from the rear and we see three different areas of intense heat, one from the rectifier area: 114.5 °F, one from the voltage regulator area: 116.8 °F, and a surprisingly high amount from the r/f box: 111.4 °F. The third pic is looking down through the throat at the cart slot, 119.6 °F. The last pic of the cart shows 109.3 °F. The connector and cart temps are pretty high, but i’m curious if that’s just because of their close proximity to the regulator heat sink.
So far the only conclusion I can make is that it hasn’t been a notable problem yet and i’m not about to drill more vent holes in it if its not causing a problem.
In order to really see where the heat is created the shell had to come off. The console is not modded, the r/f is outputting video, i’m letting the SMB intro loop. I used the same board for all 3 tests, but was sure to give it plenty of time in between test to completely cool.
I used 3 different power adapters, the original 9Vac nes adapter and two universal DC adapters of 9 and 12 volt output, just to see if there was a difference.
9 volts AC, initial:
9 volts AC, 20 minutes runtime:
9 volts DC, initial:
9 volts DC, 10 minutes runtime:
12 volts DC, initial:
12 volts DC, 30 minutes runtime:
This doesn’t really tell us much, no matter what supply all the components get hot and are relatively the same temps.
However, it does give us a baseline, something to reference when trying to improve performance.
Just for fun I removed the heat sink, the 7805 voltage regulator jumped to 205 °F in a matter of seconds! But stayed there…
I have heard that if left off the regulator will hit thermal overload and shut down the console.
The Power Research page and this section are closely related. There are a couple thermal pics on that page for the switching wall cube mod.
Switching regulators and wall cubes are better technology, but do they produce less heat?
These two pics are with the 7805 replaced with a drop in replacement switching regulator I bought from Mouser for $10.
Runtime 20 minutes, SMB intro looping through r/f output, nes-002 9Vac power supply.
Seemed to be a small drop in cpu and ppu temps. The regulator itself did get hot, but not as hot as the 7805 without a heat sink, but not ZERO heat as some promised.
I put the shell back on and ran the same test, but for a full hour, compare with pics at the top of the page.
The shell did actually stay cooler, but no surprise, the a/c showed the same temps.
I then decided to retest with a 9V DC adapter. Initial:
30 minutes runtime:
CPU and PPU look hotter than on 9vac and the switch reg looks cooler than on 9vac but again no surprise the 9vdc adapter is about the same as other tests.
All in all didn’t learn much, but I don’t think i’ll drop an extra $10 to replace the 7805 unless I need the space the 7805’s heat sink takes up.
I also ordered all the parts needed to build my own switching regulator for about $4. I followed the regulator’s recommendations for external parts so maybe they will show better performance or I can at least identify what components are creating all the heat. But I haven’t got it done yet, so stay tuned…
CPU and PPU Heat Sinks:
The cpu and ppu are important and get hot, I wondered if I could just slap a sink on it of some kind and get the heat off them quicker and get them to run cooler.
The heat is very central to each chip so i didn’t need a big sink and happened to find a couple small ones for cheap to experiment with.
The sinks are just sitting on top of each IC with thermal grease in between.
These pics show the progression of heat, the thermal cam automatically changes its temp range as needed and that is when I took a new pic, often at first then tapering off.
Finally some progress! Over an hour of runtime and the CPU only gets to 102.3°F and the PPU only gets to 100.9°F! 10-12 degrees cooler than without sinks, see initial test pics above.
Then I figured out a way to make the sinks semi permanent.
Then I put it all together and retested. Otherwise unmodded console, SMB intro looping through r/f output, 1hour runtime.
The first pic makes me want to redo the pics at the top of the page as those don’t show any heat above the cpu and ppu, but 81°F is not hot by any means, just hotter than the surrounding plastic of the console. Maybe all I did was direct the heat to a more confined area, could retest with the ported shield but no sinks…
I was curious if the video mod did anything to the heat, no major change.
You would think speeding up the CPU would cause it to heat up, but it didn’t. Remember this is only a 32% speed increase though… First pic is after 5 minutes in over-clock mode with SMB running, second pic is after 30 minutes.
The Bridge Rectifier: This is a big source of heat that I haven’t looked at too hard, but I did order 4 Schottky diodes to replace original rectifier and test, just haven’t gotten to it yet… I’ve read there are much more efficient diode bridge designs using mosfets and control circuits, but haven’t researched enough to see if anyone offers a drop in replacement like the switch reg or if I have to build one or if its even doable…
If you have ideas, contact me!