I'm glad that you re-ran the active cooling cases. That was a lot of work, but the responsible thing to do. I have learned to count on you to do the responsible thing.
These videos always put a smile on my face. It's just the perfect combination of great technology, a simple and clean video design and of course, the most charming host
I managed to get one degree cooler in my Flirc case. I think the only difference between your setup and mine was that I initially bought a Canakit which came with 3 heat sinks (CPU, USB and, I think, Network chips) so when I got the Flirc case I only removed the CPU heat sink leaving the other two since they did not interfere with the case. I am very happy with the Flirc and haven't seen a single thermal throttle to date. I loved the THX 1138 reference! I really appreciate the subtle bits of humor you sneak into your viedeos. temp=38.0'C temp=47.0'C temp=48.0'C temp=49.0'C temp=50.0'C temp=51.0'C temp=53.0'C temp=53.0'C
Hi, thanks for your video! I've both used the Flirc case as well as the 'heat sink as a case' style (exactly the same model as you use). I'm using my Rpi4 as a 'Stellarmate', a tool for astrophotography. What I've found is that when using the Flirc case, when I use my Rpi4 in hotspot mode ánd use bluetooth, it loses the hotspot. With the 'heat sink as a case' the Wifi transmitter is less obstructed (actually visible from the outside) and doesn't cause this problem of losing the hotspot. I suggest it's some kind of interference. It's maybe interesting to run a test on that with your better understanding of computers. Cheers! Bart
Long time subscriber and also a tinkerer of Raspberry Pi. Just wanted to say thanks for your informative videos as I've learnt a lot from them and passed this information on to others as and when they ask. Keep up the good work Mr Barnatt.
I kinda genuinely fell in love with the Raspberry Foundation. Consumes less power than some peoples light bulbs, yet is a super affordable computer, if you want it to be. Awesome!
Thank you Chris, i am a firm(ware) believer in passive cooling for RASPI that sit a coner just doing datalogging and behind equipment performing various minor media tasks such as Kodi servers and the like in a quiet man cave or she shed!
Active cooling solutions test clearly shows that 7 sysbench iterations are enouch for CPU temperature to stabilize. But those big passive heatsinks have a lot higher heat capacity than active coolers, and the test might not be able to heat up the CPU to the maximum temperature (which can be seen from the results, the temperature only increases for passive coolers and never decreases). In fact, in both passive coolers tests the last reading is higher than the previous one that shows the temperature will rise more if the test continues.It would be more interesting and informative to see the resulting temperature after it truly stabilizes (it takes only about 11 minutes for the current version of the test to complete, so increasing the time to 30-40 minutes might be enough). And, by the way, I suspect those heatsinks might not be that safe to touch after the test.
Yep. I'd guess it's going to max out in the upper 50s but you should at least run enough iterations to get some duplicate peak readings before stopping the test.
I did exactly that with the FLIRC case and the temperature finally stabilized at 58 Celsius after 30 minutes. The test ran 60 minutes. The case got warm but was still safe to touch.
I ran ffmpeg for 2 days on a raspberry pi 3 with this case. I never made it over 60F. Ffmpeg uses all available processors, and will use 100% of the cpu. I suggest using "nice" with it, unless you really don't plan to be using the computer for anything else.
Valid point. With passive cooling only, it takes at least 30 minutes of 100% CPU to reach equilibrium. Measurements after only 10 minutes are very much influenced by the initial conditions and should be discarded.
Ice Tower and Noctua fan just outperforms all cooling methods, but these passive cooling solutions are also performing well. Thanks for another video. I eagerly wait for these to arrive.
Hot Flash (pun). MagPi Magazine 88 Pg 74 suggests orienting the Pi 4 vertically, with the USB ports up. This is said to result in 2°C cooler running at idle. Also, do a "sudo apt full-upgrade -y", then "sudo apt -r shutdown" (the magazine shows this in error as "sudo apt - r shutdown" but there should not be be a space between - and r).
Donald Porter or at the very least he could do “Explaining Politics.Com”. Of course Stanley the Knife and Mr. Scissors would be helpful showing examples of how one could “cut red tape”.
I have an old Raspberry PI but after watching this I am going to get the Kodi case and the PI 4 to replace my core 2 nuc (DVI output only). Thanks for this.
I suspect then, that official case will now smash the 90 degrees bomb scare score now too? Well done the Pi lads! Thats a top bit of software engineering. Very impressed. Thanks Chris, for doing them all again. It deserves to be highlighted. Well done.
What a cool idea for a video. I am always overwhelmed with the selection of cases and cooling solutions. Now I have a good idea of what to buy. Thank you.
that aluminum heatsink case is called the raspberry pi armor case. The bottom heatsink is carried over from the previous versions where the pi had the ram on the bottom. it's still there for ease of use and design and just makes it look damn good.
I put the heatsink from a quad-xeon server on my RasPi4. Never gets more than 2 degrees above ambient now. It's not pretty mind you and it weighs over a kilo now but it is at least 100% silent. :)
i have several of the "Aluminium Heatsink Case" for my 3 and 3b's and i ordered one for my new 4. I order them at aliexpress and are so cheap and work excellent. Good to see some tests and comparison.
Thanks for this, appreciated. I am planning a 1-hour documentary on the history and future of computing for 2020. I'm not sure yet if it will be here on RUclips, or on Amazon's Prime Video Direct. It depends on what happens here on RUclips after January! :O
@@mdtaylor2274 Almost certainly. Providing I am still here on RUclips! :) There are big RUclips changes happpening in January. I will discuss this in more depth in an update video in two weeks time.
Raspberry Pi noob here. Just wanted to say that, aside from creating desktop shortcuts and other tweaks, recreating and running your temp_test script was the first useful project that I actually completed.
Thank you for the experience, watching on a Pi 3 with and 8.99 case , acrylic top and bottom with fan, after two of your videos, has a 53 degree internal monitored temp, sd card is from a Pi 4 and has the latest update. Always learn something when I watch your videos.
Hi, interesting video, as usual. The last summary table missed the no-cooling test results. They would be useful to check how the update performs in your same ambient conditions. I used the same passive heatsink case as you, but by looking inside the fully mounted thing I saw the cpu sticky pad wasn't touching the case log. So, after verifying the ram chip doesn't heat considerably, I placed two sticky pads over the cpu and I got a better result. Bye from Italy.
Of the two tested I prefer the Aluminium Heatsink Case, simply because you can get to the GPIO pins. I suspect you could add a fan with some self-tapping screws that gripped between the cooling fins on the top. Obviously choosing screws that were not too long. The Ice Tower remains the one to beat, so top marks to them once again.
What a difference in ℃ after updating π. Great job π foundation! And of course, excellent video Chris! Sunday delight! R2D2 & 3PO love TPH002! As for me, I'm all active cooling Burj Khalifa towers. 😁
Hi Elvira. :) I remain amazed how effective the new firmware is in improving power management. It makes you wonder what firmware changes could do in other systems.
I was thinking about making some little server on Raspberry Pi, that would even run at night and would not disturb me. Now I know a good solution against overheating. Thanks for the video!
Very impressed with this video, it shows progress, I think I might have been one of the first with a flirc case, it arrived to me on one of the hottest days of the year, I put it together, and started testing it out, before too long the pi was getting very hot, the red sign was on the pi and and the pi was giving off a dreadful smell, pleasant to touch it was not, like handling a kettle or a toaster, I had several flirc cases in the past, which were fantastic on the pi 3, I was so disappointed with this one, especially as I had it shipped in and paid extortionate taxes on it for me and I got one for my brother, which he wasn't going to want, but now these raspberry pi 4 run nice and cool, even with an overclock, good work pi foundation, though perhaps the pi 4 was rushed out, maybe things like this needed ironing out first and probably still do, I have had two raspberry pi 4 with display and sound issues after running beta berryboot, hats off to the official raspberry pi outlets who posted out a replacements straight away, so the raspberry pi stays my favourite SBC.
Nearly 2M sold - if we had waited until everything was perfect, we would not have sold any yet and everyone would still be waiting for many more months. And of course, no product is perfect.
I have a pair of Pi4s with Flirc cases and I agree that the temps stay low, especially with the updated firmware. I couldn't justify a passive cooler for these, and the silent operation is a plus. Now I just need true USB booting and the Pi4 will be perfect for my needs.
Thank you for taking the extra time to do redo the active with the firmware updates. I'd kinda be curious to see a row of the Pis in the heatsink cases fair inside a square tube with a fan on either end pushing and pulling the air through...
There is a better passive cooler case sold now by Geekpi which has a contact similar to the Flirc case, but the contact is solid metal and not hollow. It also contacts the power supply chip. I have had results in the low to mid thirties using this cooler. I will note that I used a heat sink pad which covers the entire bottom of the board and sinks the heat to the lower case heat sink. The case sells for $14.99 US and is far better than those shown in your video.
I should probably copy your testing script and see how well the case with fan I bought does. I'm glad you re-tested the previous cooling solutions, as it truly highlights the impact of those updates. I'm curious how the official case fares. Hopefully, it would stay below throttling temperatures.
The Heatsink case are a little bit higher on heat but cooling 3 chips instead of 1. It must cooler the memory and graphics chip cooler than without with other solutions.
I think I'd prefer the two-part aluminum case as it allows access to the GPIO pins. Edit: How well would the two-part case work if you managed to mount a 40mm fan on it?
@@cyrfung The fans drop the temperature by about 10 degrees, and replacing the alleged "thermal pads" with 0.5mm thick copper shims and thermal grease drops the temperature again to about 40 degrees.
My own follow-up, after I upgraded the firmware, aluminium heatsink, with 2 fan 25mm (they are under 20dB, so virtually silent), and with managed speed: temp=47.0'C temp=58.0'C temp=60.0'C temp=61.0'C temp=62.0'C temp=62.0'C temp=62.0'C temp=61.0'C The starting temperature is high because the RPi has been sitting in the non-aircond room all day long, so it was over 30C at that time and even right now, the aircond is set at 27C.
Thermal mass of this passive coolers bigger than of small radiators. Temperature didn't stabilized in 10 min run and still rising. For more comprehensive comparison passive coolers need to be run much longer, maybe 2-3 times more.
i've the impression that the kodi case needs a longer test period for the simple reason that it has a big mass and it might not reached peak temperature. In fact I belive the second passive case has a proper heatsink on it and should exchange heat with the ambient faster although being of small mass it reaches the peak temperature faster.
Amazingly great information Christopher. I didn't know about this thermal update. This makes the RPi4 a lot more usable. I wonder if the performance is still the same when overclocked. I'll have to do some testing too. But so busy with the Jetson Nano. Also the USB problem with the 4GB RPi4 model in Ubuntu arm64 is fixed with this update. You still need to use the trick of lowering to 3gb to install, but once rpi-update is performed it can be removed again. I love my metal case for the NanoPi M4V2 too. It does use a fan, but it's a 12V fan, and at 5V you can't hear it and it still keeps it cool enough. I hate loud fans, now I don't use a x86 pc anymore I can hear the sbc fans a lot better. Have a great day, it was a great piece of investigation. I love to see the effort you went through to make this video. Greetings, NicoD.
Hi Nico! :) Thanks for this. I know you will appreciated the time involved when this video evolved on me. I did not anticipate doing so much testing of different setups.
The disadvantage of the Alu-Armor heat sink is that the 40-pin header connector cannot be attached to it. The milling needs either a B-undercut due to the milling cutter radius when milling from above (ugly) or one milling from the side, i.e. a second milling spindle or laser cut (expensive) or reclamping (labor-intensive). So you can only remove the radius by hand with the flat file - the question void the guarantee of the board if it was sold as a kit. Despite all the trouble, the cooling for continuous operation works perfectly without noise and very importantly without power consumption for the cooling. The heavy heat sink is problematic for flight operations.
Thank you for an excellent video. The video does an excellent job of explaining the process for the testing the cases and providing potential consumers with results that are trustworthy.
I have pi 1 model B, 2xpi 2, and 4x pi 3. I wouldn't buy a 4 b/c I didn't want active cooling. This video puts the pi 4 on my list. I skipped the 3+. I like my droid hc1 and was hoping for similar case idea with case as heat sink. I like the 2nd one better than the flirt case, but might just get 2 pi's and both cases.I'm just very opposed to active cooling for SBCs. Great video, Chris as always!
FYI, when we test prototype boards at engineering labs, what we are most interested in is the temperature delta - i.e. the rise in temperature under different conditions. For digital electronics, assuming you're operating within normal operating ranges, the "delta T" is the temperature rise we expect for a given HW / SW configuration. It is the difference between ambient room (or thermal chamber) temperature, and the temperature of the component under test. If "delta T" is 27 degrees, then if the room is 0 degrees, then the component would be at 27 deg. If room temp is 50 degrees, we can expect the device to be at 77 deg. If delta T is the die temp, (as it would be for a chip with an internal temperature sensor,) we can subtract delta T from the max junction temp (Tjmax) for that device and determine the maximum ambient temperature for that configuration. This also makes it easier to compare heat sink performance as very often the ambient temperature for tests aren't the same from day to day, but the delta T will be (within measurement limits). Thus, the system with the lowest delta T gives you the best performance. For the best performance per dollar (or pound or Euro) you can divide delt T by $ to see which is the most cost effective. I enjoy these videos - keep 'em coming!
Flirc Raspberry Pi 4 Case (Kodi Edition) $21 Canadian. This a nice case for low temperatures but not good for adding a camera or access to the GPO pins. The heatsink case is the winner for providing full access to GPO and camera. Thank Chris for all your hard work lots a testing going on and rebuilding the Pi busywork 😳
ExplainingComputers / back to you Chris. Your going to need a good Christmas holiday and take some much needed rest. Most viewer don't realise the amount of work that's required to hammer out all these videos, that a the research needed to have a good informative show. So have a good Christmas !
For those who is going to try heatsink case with sticky pads, - word of caution - it is afterwards very hard to remove the top heatsink as it attaches itself in three places very firmly with those pads and the sink is quite large- I.e. you can’t twist it or something and just tearing it off from the board has a risk of breaking PCB. I recommend using proper CPU heatsink paste instead of sticky pads, which also improves thermal transfer and gives better (cooler) results. Case then can also be very simply taken apart if needed.
@@bartoszri paste cant evaporate, it usually gives better thermal contact with heatsink (obviously if there isnt too little paste). Same as we use for CPUHeatsink in conventional PCs.
With the Aluminum heat sink case, I've found if you raise it off the bench top, on a stand that allows free airflow around it, that makes a significant difference to how cool it gets. I'm using part of an old PC (used to cage hard drives, an open mesh of metal about one centimeter off the desk) to rise the Pi off up. I found without doing this the case would get uncomfortable to touch. With it raised it ends up being warm to the touch.
The Flirc cases are great. However, they have a larger mass, which can absorb more energy. This means they take longer to reach equilibrium. You need to run the tests for longer. Active cooling will reach equilibrium quicker. All tests should be ran until equilibrium is reached, ie full load until no temperature change.
I really liked that last case -- it had a very "industrial" quality that I thought looked very cool, and I liked the way the top piece fit behind the USB and Ethernet ports. And it was the least expensive of all the solutions in this video. I think that's the one I'd buy if I were in need of a cooling solution for a Raspberry Pi 4.
We all Appreciate your Efforts of Conducting the Tests again with Active Cooling Systems after the Firmware Update.
Greetings!
Thanks!
Not only is your Pi "fed up" with all the cooling changes, I'm sure it's "fed up" with factoring numbers. Great video, lots of work.
I'm glad that you re-ran the active cooling cases. That was a lot of work, but the responsible thing to do. I have learned to count on you to do the responsible thing.
This is the best video of the Raspberry pi 4 since the launch of the Raspberry pi 4.
best tech show on youtube. always good information, and a funny 80's style, reminds me of Commodore 64 shows, lol.
And without annoying background music.
Can you still smell the BBC micro computers in your head
I guess you could call this passive aggressive cooling
Clever! 😆
Clever indeed! ruclips.net/video/6DC9xez5pxU/видео.html
🤣
That's hilarious!
So good.
These videos always put a smile on my face. It's just the perfect combination of great technology, a simple and clean video design and of course, the most charming host
Superb host, I will agree.
I managed to get one degree cooler in my Flirc case. I think the only difference between your setup and mine was that I initially bought a Canakit which came with 3 heat sinks (CPU, USB and, I think, Network chips) so when I got the Flirc case I only removed the CPU heat sink leaving the other two since they did not interfere with the case. I am very happy with the Flirc and haven't seen a single thermal throttle to date.
I loved the THX 1138 reference! I really appreciate the subtle bits of humor you sneak into your viedeos.
temp=38.0'C
temp=47.0'C
temp=48.0'C
temp=49.0'C
temp=50.0'C
temp=51.0'C
temp=53.0'C
temp=53.0'C
Thanks for sharing your results -- and appreciating THX 1138. :)
We also appreciate the EXTRA time you took to run the previous tests too :)
Well, I WAS going to get the Official Case but after seeing that chart, I'll be getting a FLIRC case. What a difference!
I love how these videos sound like the educational videos from my childhood, but are about modern technology
I've used the Flirc case for years so I had to get a new kodi edition for my Pi 4 and it looks amazing.
I have the aluminum heatsink case and it's fantastic, love it! Highly recommend it!
Hi, thanks for your video!
I've both used the Flirc case as well as the 'heat sink as a case' style (exactly the same model as you use). I'm using my Rpi4 as a 'Stellarmate', a tool for astrophotography. What I've found is that when using the Flirc case, when I use my Rpi4 in hotspot mode ánd use bluetooth, it loses the hotspot. With the 'heat sink as a case' the Wifi transmitter is less obstructed (actually visible from the outside) and doesn't cause this problem of losing the hotspot. I suggest it's some kind of interference.
It's maybe interesting to run a test on that with your better understanding of computers.
Cheers! Bart
Long time subscriber and also a tinkerer of Raspberry Pi. Just wanted to say thanks for your informative videos as I've learnt a lot from them and passed this information on to others as and when they ask. Keep up the good work Mr Barnatt.
Thanks.
I kinda genuinely fell in love with the Raspberry Foundation. Consumes less power than some peoples light bulbs, yet is a super affordable computer, if you want it to be. Awesome!
Love your impromptu, couch conversations. Even this was on the road. Good job, sounded very cool!
Thank you Chris, i am a firm(ware) believer in passive cooling for RASPI that sit a coner just doing datalogging and behind equipment performing various minor media tasks such as Kodi servers and the like in a quiet man cave or she shed!
Active cooling solutions test clearly shows that 7 sysbench iterations are enouch for CPU temperature to stabilize. But those big passive heatsinks have a lot higher heat capacity than active coolers, and the test might not be able to heat up the CPU to the maximum temperature (which can be seen from the results, the temperature only increases for passive coolers and never decreases). In fact, in both passive coolers tests the last reading is higher than the previous one that shows the temperature will rise more if the test continues.It would be more interesting and informative to see the resulting temperature after it truly stabilizes (it takes only about 11 minutes for the current version of the test to complete, so increasing the time to 30-40 minutes might be enough). And, by the way, I suspect those heatsinks might not be that safe to touch after the test.
This, exactly
Yep. I'd guess it's going to max out in the upper 50s but you should at least run enough iterations to get some duplicate peak readings before stopping the test.
I did exactly that with the FLIRC case and the temperature finally stabilized at 58 Celsius after 30 minutes. The test ran 60 minutes. The case got warm but was still safe to touch.
I ran ffmpeg for 2 days on a raspberry pi 3 with this case. I never made it over 60F. Ffmpeg uses all available processors, and will use 100% of the cpu. I suggest using "nice" with it, unless you really don't plan to be using the computer for anything else.
Valid point. With passive cooling only, it takes at least 30 minutes of 100% CPU to reach equilibrium. Measurements after only 10 minutes are very much influenced by the initial conditions and should be discarded.
I really like the way the heatsink case(2 pieces) look.
Thank you for testing. Passive cooling is the best.
Another masterpiece. And nice to meet a new family member: Allen the Key. Cheers from Spokane!
Ice Tower and Noctua fan just outperforms all cooling methods, but these passive cooling solutions are also performing well.
Thanks for another video. I eagerly wait for these to arrive.
Thanks professor Christopher .the ice tower still at the top and the cheaper passive design looks good nice on the new firmware .
I got the FLIRC case for my RPI4 4GB and it's beautiful and works great.
Love your 1970s OU Broadcast Style...waiting for Massive Shirt collars...lol
All that's missing is a B&W telly an a Kipper tie.
Hot Flash (pun). MagPi Magazine 88 Pg 74 suggests orienting the Pi 4 vertically, with the USB ports up. This is said to result in 2°C cooler running at idle. Also, do a "sudo apt full-upgrade -y", then "sudo apt -r shutdown" (the magazine shows this in error as "sudo apt - r shutdown" but there should not be be a space between - and r).
I wish you would run for public office in the US. Somebody I could actually TRUST!
Donald Porter or at the very least he could do “Explaining Politics.Com”. Of course Stanley the Knife and Mr. Scissors would be helpful showing examples of how one could “cut red tape”.
I've always said that Firmware Engineers are the finest people in the world.
Agreed. And they get so little credit.
I have an old Raspberry PI but after watching this I am going to get the Kodi case and the PI 4 to replace my core 2 nuc (DVI output only). Thanks for this.
I suspect then, that official case will now smash the 90 degrees bomb scare score now too? Well done the Pi lads! Thats a top bit of software engineering. Very impressed. Thanks Chris, for doing them all again. It deserves to be highlighted. Well done.
I have never seen a video on the Explaining Computers channel where pressing the "like" button felt like adequate praise. This one is no exception.
What a cool idea for a video. I am always overwhelmed with the selection of cases and cooling solutions. Now I have a good idea of what to buy. Thank you.
I just realized you don't have a PATREON!!! I love your videos. I would love to be able to contribute directly to you work.
Your channel just keeps getting bigger! Good for you!
That was a lot of work retesting, good job!
It was! The video took a lot longer to make than I expected!
Very impressive that the software update could make that big of a difference in temperature. Also, “Allen the key” - loved it!
I recomend putting it vertical, the heat transfer by natrual conveccion will be a lot better, and the board will be cooler
that aluminum heatsink case is called the raspberry pi armor case. The bottom heatsink is carried over from the previous versions where the pi had the ram on the bottom. it's still there for ease of use and design and just makes it look damn good.
There are lots of names for the case -- the castings are sold generically by many companies. :)
The copper-colored heatsinks have a smooth bottom.
huh, i didn't know about the smooth bottom varient. guess i need to hunt one specifically to find it
I put the heatsink from a quad-xeon server on my RasPi4. Never gets more than 2 degrees above ambient now. It's not pretty mind you and it weighs over a kilo now but it is at least 100% silent. :)
i have several of the "Aluminium Heatsink Case" for my 3 and 3b's and i ordered one for my new 4.
I order them at aliexpress and are so cheap and work excellent.
Good to see some tests and comparison.
Well Done Sir. You certainly do a thorough test. We can always trust you to do so.
great information. thanks for taking the time to rerun all those tests
I can see this dude doing documentaries. He's better than 99% of the youtubers.
Thanks for this, appreciated. I am planning a 1-hour documentary on the history and future of computing for 2020. I'm not sure yet if it will be here on RUclips, or on Amazon's Prime Video Direct. It depends on what happens here on RUclips after January! :O
@@ExplainingComputers would it be possible to host on both platforms?
@@mdtaylor2274 Almost certainly. Providing I am still here on RUclips! :) There are big RUclips changes happpening in January. I will discuss this in more depth in an update video in two weeks time.
@@ExplainingComputers yeah RUclips and their Swine fingers! Either way, I'll follow you like white on rice ❤️
@@mdtaylor2274 Thanks.
Raspberry Pi noob here. Just wanted to say that, aside from creating desktop shortcuts and other tweaks, recreating and running your temp_test script was the first useful project that I actually completed.
Great to hear! :)
Especially with passive cooling, the mounting position is crucial.
Ahh passive cool my old friend
I got tired just thinking about how much work it was to produce this terrific video.
Thank you for the experience, watching on a Pi 3 with and 8.99 case , acrylic top and bottom with fan, after two of your videos, has a 53 degree internal monitored temp, sd card is from a Pi 4 and has the latest update. Always learn something when I watch your videos.
Thank you for slow maner of telling
It help to teach English.
Hi, interesting video, as usual. The last summary table missed the no-cooling test results. They would be useful to check how the update performs in your same ambient conditions. I used the same passive heatsink case as you, but by looking inside the fully mounted thing I saw the cpu sticky pad wasn't touching the case log. So, after verifying the ram chip doesn't heat considerably, I placed two sticky pads over the cpu and I got a better result. Bye from Italy.
Of the two tested I prefer the Aluminium Heatsink Case, simply because you can get to the GPIO pins. I suspect you could add a fan with some self-tapping screws that gripped between the cooling fins on the top. Obviously choosing screws that were not too long. The Ice Tower remains the one to beat, so top marks to them once again.
Allen the key joins the family..... yeyyyy!!!!!
Absolutely fantastic review Chris. Thank you.
Thanks! I do prefer a silent passive cooling system. Some fans are a complete pain, especially if you want to stream high quality sound.
If you find cooling fan to be a pain. You should not stick your sensitive bit into to fan.😬😇😈
@@dadillen5902 indeed!!
What a difference in ℃ after updating π.
Great job π foundation!
And of course, excellent video Chris!
Sunday delight!
R2D2 & 3PO love TPH002!
As for me, I'm all active cooling Burj Khalifa towers. 😁
Hi Elvira. :) I remain amazed how effective the new firmware is in improving power management. It makes you wonder what firmware changes could do in other systems.
@ExplainingComputers
Exactly my thoughts!
Makes me wonder what changes would we get on other systems just by firmware changes... 🤔
ExplainingComputers
Hey why did you use
“sudo apt upgrade” instead of
“sudo apt full-upgrade”
I was thinking about making some little server on Raspberry Pi, that would even run at night and would not disturb me.
Now I know a good solution against overheating.
Thanks for the video!
For hifi use, in a streamer for example, passive is the best choice.
Nice. The PI 4 is like a mini sun so no small heat sink will do the job. For PI 4 with 8GB of RAM and heavy use I will recommend the active heat sink.
you illustrate that the Raspberry Pi 4 is one tough unit ,excellent results both
Very impressed with this video, it shows progress, I think I might have been one of the first with a flirc case, it arrived to me on one of the hottest days of the year, I put it together, and started testing it out, before too long the pi was getting very hot, the red sign was on the pi and and the pi was giving off a dreadful smell, pleasant to touch it was not, like handling a kettle or a toaster, I had several flirc cases in the past, which were fantastic on the pi 3, I was so disappointed with this one, especially as I had it shipped in and paid extortionate taxes on it for me and I got one for my brother, which he wasn't going to want, but now these raspberry pi 4 run nice and cool, even with an overclock, good work pi foundation, though perhaps the pi 4 was rushed out, maybe things like this needed ironing out first and probably still do, I have had two raspberry pi 4 with display and sound issues after running beta berryboot, hats off to the official raspberry pi outlets who posted out a replacements straight away, so the raspberry pi stays my favourite SBC.
Nearly 2M sold - if we had waited until everything was perfect, we would not have sold any yet and everyone would still be waiting for many more months. And of course, no product is perfect.
I have a pair of Pi4s with Flirc cases and I agree that the temps stay low, especially with the updated firmware. I couldn't justify a passive cooler for these, and the silent operation is a plus. Now I just need true USB booting and the Pi4 will be perfect for my needs.
Wonderful review !
Can't wait to get my hands on that fine machine.
Thank you for taking the extra time to do redo the active with the firmware updates. I'd kinda be curious to see a row of the Pis in the heatsink cases fair inside a square tube with a fan on either end pushing and pulling the air through...
There is a better passive cooler case sold now by Geekpi which has a contact similar to the Flirc case, but the contact is solid metal and not hollow. It also contacts the power supply chip. I have had results in the low to mid thirties using this cooler. I will note that I used a heat sink pad which covers the entire bottom of the board and sinks the heat to the lower case heat sink. The case sells for $14.99 US and is far better than those shown in your video.
Do you have a link to where you can order from?
I should probably copy your testing script and see how well the case with fan I bought does. I'm glad you re-tested the previous cooling solutions, as it truly highlights the impact of those updates. I'm curious how the official case fares. Hopefully, it would stay below throttling temperatures.
The Heatsink case are a little bit higher on heat but cooling 3 chips instead of 1. It must cooler the memory and graphics chip cooler than without with other solutions.
Pretty sure I just found my next Pi case, thanks EC!
Man the new update is awesome
Thanks a lot for this information! In this video you answered several questions that I've had on my mind for a while + I learned new things :)
I think I'd prefer the two-part aluminum case as it allows access to the GPIO pins.
Edit: How well would the two-part case work if you managed to mount a 40mm fan on it?
Eric Carver Same here
If you don't need access to the board the Flirc case is the way to go.
ninline2000 but if you need to connect lcd s for example you can use the 2 part aluminium heat sink case
There is a version of the heat sink case with two small fans. Time for more testing!
@@cyrfung The fans drop the temperature by about 10 degrees, and replacing the alleged "thermal pads" with 0.5mm thick copper shims and thermal grease drops the temperature again to about 40 degrees.
My own follow-up, after I upgraded the firmware, aluminium heatsink, with 2 fan 25mm (they are under 20dB, so virtually silent), and with managed speed:
temp=47.0'C
temp=58.0'C
temp=60.0'C
temp=61.0'C
temp=62.0'C
temp=62.0'C
temp=62.0'C
temp=61.0'C
The starting temperature is high because the RPi has been sitting in the non-aircond room all day long, so it was over 30C at that time and even right now, the aircond is set at 27C.
El mejor programa tecnológico en RUclips . Saludos desde México. 😁👍
Thanks, appreciated.
I bought the second one, and it works really well for me. it's on the pi I'm watching this video on.
Very good video, Chris. Can't wait to buy myself a raspberrry pi 4B
Wonderful video! I learn something new and, more importantly, useful every week!
The 'Stanley The Knife vs Mr Scissors' rivalry just got more complicated.
Now we also have Alan the Key
Thermal mass of this passive coolers bigger than of small radiators. Temperature didn't stabilized in 10 min run and still rising. For more comprehensive comparison passive coolers need to be run much longer, maybe 2-3 times more.
i've the impression that the kodi case needs a longer test period for the simple reason that it has a big mass and it might not reached peak temperature. In fact I belive the second passive case has a proper heatsink on it and should exchange heat with the ambient faster although being of small mass it reaches the peak temperature faster.
Wonderful and thoughtful testing! Well done!
Thank you very much!
Amazingly great information Christopher. I didn't know about this thermal update. This makes the RPi4 a lot more usable. I wonder if the performance is still the same when overclocked. I'll have to do some testing too. But so busy with the Jetson Nano. Also the USB problem with the 4GB RPi4 model in Ubuntu arm64 is fixed with this update. You still need to use the trick of lowering to 3gb to install, but once rpi-update is performed it can be removed again.
I love my metal case for the NanoPi M4V2 too. It does use a fan, but it's a 12V fan, and at 5V you can't hear it and it still keeps it cool enough. I hate loud fans, now I don't use a x86 pc anymore I can hear the sbc fans a lot better. Have a great day, it was a great piece of investigation. I love to see the effort you went through to make this video. Greetings, NicoD.
Hi Nico! :) Thanks for this. I know you will appreciated the time involved when this video evolved on me. I did not anticipate doing so much testing of different setups.
Thanks for another great video Chris. These comparisons are great value, and thanks for the pricing info also.
The disadvantage of the Alu-Armor heat sink is that the 40-pin header connector cannot be attached to it. The milling needs either a B-undercut due to the milling cutter radius when milling from above (ugly) or one milling from the side, i.e. a second milling spindle or laser cut (expensive) or reclamping (labor-intensive). So you can only remove the radius by hand with the flat file - the question void the guarantee of the board if it was sold as a kit. Despite all the trouble, the cooling for continuous operation works perfectly without noise and very importantly without power consumption for the cooling. The heavy heat sink is problematic for flight operations.
Should give the Argon One Raspberry PI case a test, it combines a heatsink case + fan
Thank you for an excellent video. The video does an excellent job of explaining the process for the testing the cases and providing potential consumers with results that are trustworthy.
Very nice video, as always. The R Pi peeps seem to have made a very wise decision when they put programmable firmware on the Pi.
Very well executed and presented, Chris! Interesting!
Thanks Don.
Another rasbperry pi related video from my favourite channel.
This was very informative and helpful, thanks for making the video!
Very Nice the case's are getting better
I have pi 1 model B, 2xpi 2, and 4x pi 3. I wouldn't buy a 4 b/c I didn't want active cooling. This video puts the pi 4 on my list. I skipped the 3+. I like my droid hc1 and was hoping for similar case idea with case as heat sink. I like the 2nd one better than the flirt case, but might just get 2 pi's and both cases.I'm just very opposed to active cooling for SBCs. Great video, Chris as always!
FYI, when we test prototype boards at engineering labs, what we are most interested in is the temperature delta - i.e. the rise in temperature under different conditions. For digital electronics, assuming you're operating within normal operating ranges, the "delta T" is the temperature rise we expect for a given HW / SW configuration. It is the difference between ambient room (or thermal chamber) temperature, and the temperature of the component under test.
If "delta T" is 27 degrees, then if the room is 0 degrees, then the component would be at 27 deg. If room temp is 50 degrees, we can expect the device to be at 77 deg. If delta T is the die temp, (as it would be for a chip with an internal temperature sensor,) we can subtract delta T from the max junction temp (Tjmax) for that device and determine the maximum ambient temperature for that configuration.
This also makes it easier to compare heat sink performance as very often the ambient temperature for tests aren't the same from day to day, but the delta T will be (within measurement limits). Thus, the system with the lowest delta T gives you the best performance. For the best performance per dollar (or pound or Euro) you can divide delt T by $ to see which is the most cost effective.
I enjoy these videos - keep 'em coming!
Flirc Raspberry Pi 4 Case (Kodi Edition) $21 Canadian. This a nice case for low temperatures but not good for adding a camera or access to the GPO pins. The heatsink case is the winner for providing full access to GPO and camera. Thank Chris for all your hard work lots a testing going on and rebuilding the Pi busywork 😳
Hi Dale.
ExplainingComputers / back to you Chris. Your going to need a good Christmas holiday and take some much needed rest. Most viewer don't realise the amount of work that's required to hammer out all these videos, that a the research needed to have a good informative show. So have a good Christmas !
Exceptional performance from Alan the key.
I just hope he's not upset that spelled his name incorrectly, it's Allen not Alan. As in Allen Millyard. Bob
@@r1273m You're right, I was overcome by the way he lined himself up for a screw, sheer poetry!
Great video as usual! In this case has helped me make a definitive decision on the cooling I am going to use. Thanks !
For those who is going to try heatsink case with sticky pads, - word of caution - it is afterwards very hard to remove the top heatsink as it attaches itself in three places very firmly with those pads and the sink is quite large- I.e. you can’t twist it or something and just tearing it off from the board has a risk of breaking PCB. I recommend using proper CPU heatsink paste instead of sticky pads, which also improves thermal transfer and gives better (cooler) results. Case then can also be very simply taken apart if needed.
Hi, if I apply a thermal paste instead thermal pads it will be good and stable contact with heatsing or is risk of evaporate of paste?
@@bartoszri paste cant evaporate, it usually gives better thermal contact with heatsink (obviously if there isnt too little paste). Same as we use for CPUHeatsink in conventional PCs.
With the Aluminum heat sink case, I've found if you raise it off the bench top, on a stand that allows free airflow around it, that makes a significant difference to how cool it gets.
I'm using part of an old PC (used to cage hard drives, an open mesh of metal about one centimeter off the desk) to rise the Pi off up. I found without doing this the case would get uncomfortable to touch. With it raised it ends up being warm to the touch.
Good tip!
Really great video Sir. I want to buy a RPi 4, and thanks to your video now I know what type of cooling to purchase.
The Flirc cases are great. However, they have a larger mass, which can absorb more energy. This means they take longer to reach equilibrium. You need to run the tests for longer. Active cooling will reach equilibrium quicker. All tests should be ran until equilibrium is reached, ie full load until no temperature change.
Great video, thank you for your works presenting all this in this nice format.
I really liked that last case -- it had a very "industrial" quality that I thought looked very cool, and I liked the way the top piece fit behind the USB and Ethernet ports. And it was the least expensive of all the solutions in this video. I think that's the one I'd buy if I were in need of a cooling solution for a Raspberry Pi 4.