Pin me Chris! Fellow friends, let's all hit that like button and share the videos. I absolutely love this channel and it'll be excellent to see Chris get 1 million subscribers. Already half way there! 👍🏼
ExplainingComputers I love this simple 3D printed case & cooling solution, but am always concerned with the extra height that having the fan on top, causes. I would be very interested in a solution that had the fan positioned at the SD-card end, either pushing or pulling air over the heatsinks (I’ve always wondered which direction was better... an idea for a quick video?). Also, with the ability to create sexy compound curves using 3D printing, I reckon you could induce an awesome airflow, over the heatsinks, & through the fan, with some fancy “ported” baffles (something only an engine head porter could dream of ua-cam.com/video/iejDWSQEsqI/v-deo.html). I’d DIY, but my arms don’t work and typing this with my thumb is about all I’m up to.
Since many have asked in the comments what the results would be with official case open but without heatsink, I decided to test it. Installed newest Raspbian to SD card, put my Pi 4 4GB to official case but removed the top part and ran the same test script. Here are the results: temp=65.0'C temp=81.0'C temp=82.0'C temp=81.0'C temp=82.0'C temp=81.0'C temp=82.0'C temp=81.0'C Test took 12 minutes and 55 seconds to finish, so 2 minutes longer than with heatsink. Ambient temperature room tempeerature was about 26 degrees but Pi was on the floor where it was perhaps few degrees cooler.
Very nice comparison of different cooling methods. The video clearly demonstrates the Pi's heating issues. It seems the official case is only useful to store the Pi 4 when not in use and disconnected.
The Pi Foundation may come to regret the "official" case for the Pi 4. That temperature profile's asking for hardware problems, not to mention the performance hit. At the very least, they should have provided ventilation holes in the top, and possibly a slot (maybe a knock-out) over the GPIO pins. An interesting experiment might be a copper block the size of the SoC, tall enough to reach the top of the ports, connected to a copper mesh the size of the board; the idea being to get convection flow. Might cause wifi problems though, if it acts as a Faraday cage. Copper mesh is commercially available online, but the block might be hard to find.
The best value setup is to get the 9-layer cooled case from ebay ($7) and pair it with this ebay heatsink ($6) which is the same as the Pimoroni heatsink in this video but half the price: www.ebay.com/itm/273958255795
I hope not as i am keeping with the Raspberry Pi 3B due to not needing a heat sink nor any cooling for normal desktop tasks for anything more than simple internet browsing i go down stairs and use a full size windows properly built windows machine ..... even though the only thing it has over the Raspberry Pi 4 is storage space, CPU speed and a x86 processor
@@yumri4 I have a Pi 3B also, but I don't use it for anything too heavy. It's primary tasks involve basic terminal applications like Wordgrinder. I prefer that for writing because on my old matte Dell display it's not too harsh on the eyes when I want to write. I have a lot of computers for more power hungry tasks, but that's in the only place I wanted something silent and with a low power consumption for lightweight computational tasks.
There is a cooler available for RPi that very much looks like a Hyper 212 EVO, albeit a smaller version: ua-cam.com/video/RyUXC3886Ic/v-deo.html It's called "The Ice Tower".
Funny you mention that. There is actually a tower cooler for the raspberry pi 4 that works great. ETA Prime did a video on it. I think the tower was even good for no-fan operation too.
A clear and uncluttered video, goes straight to the point and gives a solution. If the Pi 4 gets that hot it might even make a cup of tea - a hitch hikers Brownian motion producer.
I've done the test on a Pimoroni case with the fan shim. Here are the results (ambient temp. 21'C): temp=60.0'C temp=64.0'C temp=67.0'C temp=63.0'C temp=61.0'C temp=65.0'C temp=63.0'C temp=66.0'C
The temps you get when using the Fan SHIM depend on how you configure the fan control script. Mine is set for the fan to activate when temp reaches 55, so I get a range of 54-57 when running the temp_test script.
Great ! This is the video that is so important for users of this new SBC ! Going back to the Pi 3B and 3B+, at least a heatsink if not an active cooling solution IS needed for real-time activities such as SDR ( software-defined radio ), which I run on a 3B. Quite often without a heatsink, the SBC "stutters" ( cuts out for part of a second ) , or "goes into the red" as regards overheating. So YES, your previous "active-cooling" work HAS been very valuable in those situations. Your final solution with the Pi 4 is obviously the one to go for, for anyone doing more than the most basic work - people dreaming of running two HDMI screens, and using the Pi for streaming video, or indeed SDR as I do, really need to opt for you final solution to "stay cool". Overheating reduces the life of any computer systems, and throughout me career over 30 years in the computer industry, cooling was a mandatory part of computer installations. It certainly needs to be with the Pi 4B. What worries me is the number of school classrooms who will be installing these SBCs over the school holidays. I just HOPE that the installers will install then with active cooling systems. Children have very sensitive hands as compared to adults, and we don't want children burned by the heat, especially when they take the lids off the SBCs, as children are prone to do. Thanks for this very valuable video, Chris, and please ALL educators take note of Chris's findings !
Very cool Chris, thanks! My only objection is: During the 2nd test (when you first added the heat sink), was the improved result due to the added heat sink or due to the increased ventilation? You changed two parameters at once, so it's difficult to say which had the biggest effect. Would it be possible to include results from running a bare Pi4 out of the case first, before adding the heat sink?
Exactly. And how does the standard small heat sink perform in combination with the fan. The difference between the small and large heat sinks was not that big.
@@edwinschaap5532 It won't be that big because it is limited by convection (or rather conduction). The heat load should be slower to build, as you have a bigger slab of metal to absorb heat but after that things will level out. In a more real world situation you would probably use heat pipes to move the heat from the heatsink to outside the case where it might get some convection heat transfer (because of breeze) but still be technically passive. Kit's can be bought on eBay although they aren't that cheap to the extent the Noctua would be a better choice cost wise at the introduction of moving parts and energy costs.
The reason for not putting a heatsink on while in the case is because that would only delay the inevitable. Without ventilation, the heat has nowhere to go, so adding the heatsink will only provide a slightly larger "space" for the heat to go. Of course, a heatsink is always better than no heatsink, simply because it makes for a thermal buffer that will smooth out the heat generated by spikes in power usage. This is the same for M.2 drives and RAM; the heatsink is not primarily for cooling, but for thermal buffering. And of course with the amount of power that the Pi 4 draws you'd want active cooling anyways, such as the S2Pi ice tower
@Anthony Spendlove. Yes I realise that. My original question was about checking cooling performance outside a case (i.e, good ventilation), but without a heatsink. A logical test protocol would be: 1. Inside case, 2. Outside case, 3. Outside case with heatsink. He skipped no 2.
In South Africa, during the hot summer months, it is essential that you make use of a fan and decent heat sink. I must say that I love your Noctua fan fitting kit.
Gandalf: "Open the script Frodo." Frodo: vi temp_test.sh Gandalf: What can you see? Can you see anything?" Frodo: "Nothing. There's nothing... wait... there are markings. It's some form of script, I can't read it." Gandalf: "There are few who can. The language is that of Bash, which I will not utter here.
July 21, 2019 - Great COOL tech video, Chris. In electrical engineering classes, we spent a great deal of time about heat management of solid state devices (and nearby passive devices too, such as power resistors). I think today's tech does a near lousy job of moving hot air through earphone jack holes and speaker holes. When using a heat sink, it is also important to provide cool air across the heat sink fins. And then channel the heat flow out. Motorcycle engines do this very well. A coat of paint on the fins degrades a heat sink dramatically. Your new vid proves a basic yet overlooked point: Manage the pathway of semi-conductor produced heat. Details matter. Thank you, Chris!
Not a lot of heat sink fins on motorcycles these days. They are watercooled, sometimes oil cooled, rarely aircooled. The fins such as they are will be cosmetic
I've seen this done (online)with PCs back in the early 2000s. Think massively overclocked early Xeons. ;) ..oh, and it's all over youtube as well. I guess it's a meme, now.
The video we've been waiting for! The new version of the Raspberry Pi needed a thorough analysis of its operating temperatures: now we have it! Thank you very much! It will be interesting to follow in the coming months the different offers of third parties manifacturers concerning passively and actively ventilated cases. Personally, I will wait until these offers are available before purchasing my next RPi4 card. Excellent educational work on your part Chris, once again!
Thank you! So many of these tests forget to say what the local "ambient temperature" is. You mentioned it it at about 2:12, as being 25C. Some places here in Australia can reach over 40C in the summer, this needs to be factored into any local setup.
Great video, watched on my raspberry pi with noctua active cooling, ceramic heat sinks and an over clock to boot, all this inside a snespi case, it never over heats, look forward to your next video keep up the good work.
Hi Chris. great video, bought a Raspberry Pi 4 with a Vilros case. It did not explain what pins to connect the fan to, but your video did. Now with the fan and heat sinks, I can keep my Pi cool, Thanks.
If fan noise is an issue, one could connect the fan to the 3.3v pin. It wouuld, of course, be less efficient at the same time, perhaps two fans at 3.3v would be quiet and efficient. Thanks for another great video.
Thank you so much for this. I was assuming that the manufacturer provided the best options overall and our additions made a nominal difference. Thank you for showing me it is all on us!
Chris, Just commented yesterday, have 3 more thoughts 1. Best cooling solution is always creating less heat, another firmware update like the USB controller. 2. With a fan solution, even a smaller one, millions of creative users and companies out there designing/using "hats", will require a 4" (sorry 10 cm) connector extension and a huge case. Ideally the case should provide independent air intake and exhaust so its not just recirculating hot air. 3. If the PCB was designed with the SoC on the bottom side there would be no obstruction from adjacent components so you could cool passively with a copper or aluminum plate the full size of the PCB using the 4 mounting holes, and yes, your feet or some spacer. This would mess with the classic case design but could eliminate 1 of the 2 case pieces.
Great video -- I considered the Ice Tower fan from your later RPi4 setup but Amazon reviews indicated that the mounting brackets obstruct some of the GPIO headers. This is a huge help for rolling my own solution once I actually use my Pi 4 for workloads that passive cooling can't deal with.
Really well done video The Raspbery Pi 4 is no longer safe for children. I believe a warning should be included with the Raspbery Pi 4 regarding the extreme temperatures.
From your tests we can say that almost no throttling occours already with passive heatsink (11mins vs 10 mins) and that minimal heatsink is already effective as the big one. Considered that modern silicon is typically designed for working reliably (10+ years) at 100+ degrees it means that a minimal passive heatsink is already doing the job (well maybe it's better saying it's now mandatory to have it) and active cooling is only for having low temperatures for paranoids...
I purchased the raspberry Pi Devastator Robot and the wireless keyboard after watching one of your your videos. I had a few days off work and spent the time building the robot and connecting it to the Pi Zero and Pi 4. I had lots of fun. I had problems having the Pi boot up by itself and running the Python program but stuck at it and got it to work. Lot of great support on the internet when there are problems. I will follow your other video and connect it up to th camera when I get some more days off work. Thanks
Would have been better if there was a bare no case run I agree. Just the air able to move away from the chip surface should make a difference vs no airflow when in the case.
I understand his logic for not doing your version of the control. He states at the end of the video that he believes a cooling solution of some sort is absolutely necessary fro the Pi 4, which I think most people will agree with.
I pillaged the 40X40mm heatsink and fan off my 3B+ (I think the cooler was featured in one of your extreme cooling pi vids). After the install, my temps while stressed went down to between 88°F and 100°F (31-37°C) before it would be between 140°-180° without the cooler. Which I thought was a really good result. I have since made the Pi4 my official new HTPC running Libreelec. And it runs perfectly.
@@alvallac2171 no sorry. It came out wrong. I just read that again. After installing the heatsink and fan, my temperatures when stressed stay between 88° and 100° F. Before the heatsinks it was in the range of 140°-180°F.
Nice video! I think that the raspberry pi 4 has come with some great updates, but it is a shame that the raspberry pi foundation didn't invest much effort into the thermal management of the board. Maybe the unexpected release of the new board was a precipitous act ...
Yet another brilliant video thanks for showing me the result's, it clearly shows that the Raspberry Pi needs to have a better cooling system than what is on offer right now.
*video. Thanks *results. It (plural) Apostrophes are for possession or contraction, not for pluralizing (except for single letters). Also, that was a run-on sentence and should be written as three separate sentences.
Excellent video, thank you for sharing. Please do the final test again "without" the heatsinks and just the Noctua fan. I suspect that the results will be very similar, and save you from having to spend money on heatsinks, or having to cobble the heatsinks as well.
Thanks as always Chris! The one thing I would have liked to see was the RPi 4 with no heatsink, and no enclosure, to see how much the enclosure added to the heat.
Thank you, just done the Test temp=36.0'C temp=49.0'C temp=51.0'C temp=50.0'C temp=50.0'C temp=50.0'C temp=51.0'C temp=50.0'C with the normal Pi Cooler and a 5V Fan
The Fan is a 4010 Cooling Fan Cooler Radiator 40x40x10 mm form the well known Chinese Market Place for less than 1€. The room temp was 23°C ,.. and I placed the Pi in a printed Case from Thingiverse. Nothing very spectacular :-) I was suprised, that this works so well with the small heat sink
Lovely video and just in time. I've put my new Pi 4 through its paces for the first time today and was surprised at how hot everything gets. The USB and ETH ports (not just the chips, the ports as well) get very hot - and all this running out in the open with air condition on. The active cooling would probably cool down those jacks as well
Raspberry Pi Trading has been very diligently deleting my comments on their social media, so I'll try here. The Pi 4 overheats and throttles in the official case while just idling at desktop. Yeah, you read that right. I connected it to a 4k display, booted it up and deliberately did absolutely nothing, just left it siting at desktop. In an hour the temperature reached 81 °C and it started throttling. Ambient temperature was 18 °C. It went 63 °C over ambient by just idling! Maybe I'm old-fashioned, but to me that is completely unacceptable for an official accessory. I have to wonder if anyone at RPT even looked at a prototype of the case before greenlighting mass production, because this design flaw is pretty much impossible to miss.
Great video as always. The heat concern of the Pi 4 is yet another indication that the board was released early and without much consideration for real world use. I’m glad to own a Pi 4. But this one has given me a moment of pause regarding what the future may hold for the platform.
Hasn't anyone invented a 'Windowsill' pi cooler?! I won't give up the day job. I'm amazed how hot the pi 4 runs though, I'll be sticking with my 3 B+ for now.
I tested with an old 80mm 12V Noctua fan from a computer. Either hooking it up to the 3.3V or the 5V fan worked fine. Running stress test without any heat sinks I got 70C (3.3V) and 65C (5V) 60 min test. Idle was below 50C. The big advantage was that the fan was running completely without audible sound. Which was my goal as I like RPI as a small desktop computer because it is silent.
Excellent video. I should have waited for this video as I ordered and started testing a 12v Noctura 40mm fan instead of the 5v you suggested (I had assumed it would work with the 3-to-2 pin adapter, but I was wrong). Thank you for your video(s).
Would have been interesting to compare with a bare Raspberry Pi (with no heat sink) out of the case. I am left wondering how much of the cooler temperatures with the passive heat sinks are due to simply not having the Pi inside the standard fully enclosed case.
Great video. I already use a case with a fan on my 3b+. It has been up for 365 days and counting. I use it for my Unifi Controller. Looking forward to get a pi 4.
Which version of the VLI Firmware were you running. The old 00013701 firmware was causing the VLI to produce a lot of heat. Can you run the tests with fan and no heat sink. Having a heat sink doesn't appear to lower temps much.
Tried this with my own rpi4 with a 40mm noctua and got the same results. 56-57 sustained peak. Absolutely needs active cooling for good performance. Cheers.
Another great video. I would like to see some testing with a flirc aluminum heat sink case to see if t hat is good enough to keep the pi4 below throttle under load w/o a fan.
The entire SBC is getting very hot to the touch! Everything! Every chip is scorching hot (even with heat sinks on every chip) and so are the USB and LAN port(s) and even the SD Card reader! I'm waiting for the Pi 4 version of the GeeekPi full cover heat sink before even attempting to use the Pi 4 for anything beyond testing.
@@Loundre3 No, only versions for Pi 3B and 3B+. I asked them about the 4B version and they said it would be available within the next couple of days. I am anxiously waiting for this heat sink. Honestly, passive cooling is the reason why I decided to run my home automation system on a Raspberry Pi. Als prior iterations worked fine (I started on a Pi 2 and have upgraded regularly). By now my home automation system has outgrown the 1GB RAM limit and I had started to look at other SBCs with more RAM (like the up board). Then, just at the right time, the Pi 4 with 4GB RAM was launched. I ordered mine within minutes of the launch and have spent the last couple of weeks trying to figure out how to get it to run cool enough to to replace the 3B+ of my live system.
Great video, and I agree that this Pi needs cooling. I would have liked to have seen the small heatsink performance as this is the easiest active solution for most folks.
I always felt that cooling would be an issue for SBC's but it opens up a massive avenue for testing and getting the best out of them with overclocking and cooling projects. What would've been nice to see would have been some thermal imaging shots from the first test to show surface heat and a thermometer to show the real world temperature within close proximity of the board. It's always nice to see projects like this as factual data is something that can appreciated on many levels and expanded on until someone reaches the ACME stage of Pi cooling! I was happy that the Pi3+ I got my daughter had 2 heat sinks and a case with functional and decorative slots that allowed the heat to dissipate quicker.
I've got my 3Bs inside Flirc cases, as i'm a fan of passive cooling whenever possible. I'd be quite interested in seeing how the 4B performs inside the 4B-specific Flirc case, when it ships.
I just got the FLIRC 4B case in the mail. I ran this same script/test and here are my results with a Pi 4b 4GB model ------------------------- (Ambient) 24.4'C 47.0'C 56.0'C 58.0'C 59.0'C 60.0'C 61.0'C 61.0'C 63.0'C Finished in about 10 minutes with no throttling. ------------------------- I also ran it 3 times back to back (without any cool down period) and the case seems to reach steady state (dissipate heat as fast as the SOC can create it) at 68.0'C (at 24.4'C ambient at least).
Impressive what the small heatsink could achieve. Temps are still high, but there was only a bit throttling. Only with a fan you could save roughly 10% of the time needed to perform the test. I like the idea to have it completely passive cooled. If you put a fan there I would make it temperate controlled. Because most of the time the fan can be off with just the heatsink.
Comparing the first and second run (in case+no hs vs outside+small hs): how much of the difference comes from adding the hs and how much from removing the case?
I couldn't help noticing that you always use your heat sinks in conjunction with those tapes to stick the heat sink on the sbc but recently I saw a video about testing cpu cooling with various components, including thermal pads and -of course- the classical thermal paste with the result that the paste brings some degrees advantage compared to the other methods. Mentioning this I would really enjoy to watch those cooling tests rerun using thermal paste to attach those heat sinks to the sbc, not only with the Raspberry Pi 4 but also with other manufacture's sbc.
I havae once used thermal paste in a Pi rig -- ua-cam.com/video/1AYGnw6MwFM/v-deo.html -- works very well, but makes it harder to hold things together! But I may well try what you suggest.
Thank you, Chris. Well done. Not a user of the RPi ecosystem, but ... gives some ideas for other situations, and worth filing away for future reference should I decide to experiment with Pi.
It's too bad someone at the Pi foundation didn't think ahead on this one, and put a nice *slim* fan in the "official" case. There's *just* enough room to mount a small one flush if there's no heatsink, or surface mount if one is present. And if a good fan was chosen, no heatsink would be needed anyway. One caveat.... Yes, the case would cost more. But I don't believe that would matter. If I were buying a Pi4B I'd be wanting a fan-based cooling solution anyway.
Pi4 SoC got ready almost one year ahead of schedule. I can only imagine the foundation chose to rush things up, in order to reduce time to market. Even so, one can barely use it with official casing. Also, if a company advertises dual monitor 4k 60fps capabilities, and then explains that only works for desktop, it looks like gross misleading. That's not cool, that's not right. If codecs are not ready at deploy time, than the board is not capable yet. It should _not_ be announced otherwise.
This was so interesting. From looking at the results any aftermarket make of cooling systems must look closely at this. I would say it must have an active heat sink.
To my shame, I just realized I never subscribed to your channel. Edit: I think that 90 Celsius is ludicrous. Edit 2: switching from the Kintaro passive aluminium heatsink for my RPi3B to a flatter custom aluminium on top of small copper discs in hope to accommodate it better into a portable frame. I'm seriously considering adding a PWM-driven 5v fan with power circuitry at the side. Unfortunately this is going to add a little bit more strain on the batteries, but I don't think there is much to do if I want some good cooling: either widening the frame or drawing more current.
I don't have a Pi4 yet but I'm on the list to get one. On my Pi 3B+ I use the C4Labs Zebra Virtue Fan Case Black Ice version. It is a little on the pricey side but it seems to be the best off the shelf case I have found. Running Ubuntu Mate I force over clock to 1520 525 on the RAM GPU and I stay about 10 degrees above ambient air temps. I do have a copper heat sink on it that is a bit larger than comes with the kit though. I had the heatsink for many years and I think it came out of a GPU cooling kit. Happy Cooling.
I had heard that it runs hotter so bought the Pimoroni Fan Shim. A little fiddly to set up but works well with the pibow coupe and looks good. Very quiet and seems to keep it very cool running stress. It is also smart in that you can set what temperature it turns on at. Given that there will be more fan solutions for the 4 maybe next video should be a comparison.
Thank you, really nice video! Some notes: - maybe the larger heatsink without the top "case" would perform better, but still probably not good enough; -there are already 3d printable available open cases that allow various fan mount(from 40 to 70 mm) . Thanks to the people who shared them (let's tip them); - also the RAM next to the CPU is likely going to run hot. I really liked the idea to cut your own heatsinks, thank you!
Chris, Great Video as always, however your conclusion that active cooling is required may have an exception. I just got the FLIRC passive cooling case for RasPi4 in the mail. I ran this same script/test and here are my results. ------------------------- (Ambient) 24.4'C 47.0'C 56.0'C 58.0'C 59.0'C 60.0'C 61.0'C 61.0'C 63.0'C Finished in about 10 minutes with no throttling. ------------------------- I also ran it 3 times back to back (without any cool down period) and the case seems to reach steady state (dissipate heat as fast as the SOC can create it) at 68.0'C (at 24.4'C ambient at least).
WOW idling at 74 degrees! Although I''m a big 'fan' of the Raspberry Pi 4, after watching your video I realize that I just won't have enough 'huff and puff' within me to keep it cool. Looking forward to your next video!
I love passive cooling too, I often use my Pi1b if I can as it is silent and doesn't use much power. It is enclosed in a plastic case and at 800MHz only gets to 56C running flat out, this was just today converting some of my old .WMA to .MP3..slowly edit: wav to wma, always get those mixed up.
@@Kenobi5001 I don't have any Raspberry but I have four x86 PCs with M3-7Y30, N3450, Z8500 and Z8350 CPUs + another cheap Chinese TV Box with ARM S905X CPU. All of them are passively cooled & are dealing well with heavy duties. No reason to change it for Raspberry unless it's going to work 24/7 in hard conditions.
i love your videos on SBC's and 3D printing, and Arduino projects. I'm looking forward to obtaining a few of these for my nieces and nephews for Christmas this year. It would be nice if you had a video out comparing the best buys for young minds (they're in their preteen and early teen years) around the end of November, also I like your new channel Explainingthefuture.com Amazing!, keep up the good work.. and here's another thumbs up 👍
It's worth noting that there is a 10mm thick version of that fan, and the fan being half the height might be useful when the pi is so small. Combine with shorter standoffs to get closer to the heatsink and you could shave 15mm off the height. Also, it would have been nice to see temperatures for the bare pi with no heatsink and no case, since we have no idea how much of the small heatsink improvement is from the heatsink and how much is from being open-air.
Pin me Chris! Fellow friends, let's all hit that like button and share the videos. I absolutely love this channel and it'll be excellent to see Chris get 1 million subscribers. Already half way there! 👍🏼
Pinned! :)
@@ExplainingComputers better than sweet apple pie ❤️
ExplainingComputers I love this simple 3D printed case & cooling solution, but am always concerned with the extra height that having the fan on top, causes. I would be very interested in a solution that had the fan positioned at the SD-card end, either pushing or pulling air over the heatsinks (I’ve always wondered which direction was better... an idea for a quick video?). Also, with the ability to create sexy compound curves using 3D printing, I reckon you could induce an awesome airflow, over the heatsinks, & through the fan, with some fancy “ported” baffles (something only an engine head porter could dream of ua-cam.com/video/iejDWSQEsqI/v-deo.html). I’d DIY, but my arms don’t work and typing this with my thumb is about all I’m up to.
@Michael Taylor
Agree. 😁👍🏻
Agree
Since many have asked in the comments what the results would be with official case open but without heatsink, I decided to test it. Installed newest Raspbian to SD card, put my Pi 4 4GB to official case but removed the top part and ran the same test script.
Here are the results:
temp=65.0'C
temp=81.0'C
temp=82.0'C
temp=81.0'C
temp=82.0'C
temp=81.0'C
temp=82.0'C
temp=81.0'C
Test took 12 minutes and 55 seconds to finish, so 2 minutes longer than with heatsink. Ambient temperature room tempeerature was about 26 degrees but Pi was on the floor where it was perhaps few degrees cooler.
This needs more upvotes. Science!
Very, very helpful. Thanks greatly for sharing this data.
Next up liquid nitrogen. Nothing like a frozen raspberry pi.
time for Bullzoid to do some LN2 runs.
Submerged in mineral oil with a bubbler stone and air pump. :D
OC to 3.2 Ghz!!
I like freeze dried fruit
Ohmwrecker nah how bout kingpin's roboclocker
Very nice comparison of different cooling methods. The video clearly demonstrates the Pi's heating issues. It seems the official case is only useful to store the Pi 4 when not in use and disconnected.
LOL! :D
The Pi Foundation may come to regret the "official" case for the Pi 4. That temperature profile's asking for hardware problems, not to mention the performance hit. At the very least, they should have provided ventilation holes in the top, and possibly a slot (maybe a knock-out) over the GPIO pins.
An interesting experiment might be a copper block the size of the SoC, tall enough to reach the top of the ports, connected to a copper mesh the size of the board; the idea being to get convection flow. Might cause wifi problems though, if it acts as a Faraday cage. Copper mesh is commercially available online, but the block might be hard to find.
The best value setup is to get the 9-layer cooled case from ebay ($7) and pair it with this ebay heatsink ($6) which is the same as the Pimoroni heatsink in this video but half the price: www.ebay.com/itm/273958255795
By the time we get to a Raspberry Pi 10 you'll need a Hyper 212 EVO cooler or an AIO liquid cooling solution to keep up.
I hope not as i am keeping with the Raspberry Pi 3B due to not needing a heat sink nor any cooling for normal desktop tasks for anything more than simple internet browsing i go down stairs and use a full size windows properly built windows machine ..... even though the only thing it has over the Raspberry Pi 4 is storage space, CPU speed and a x86 processor
@@yumri4 I have a Pi 3B also, but I don't use it for anything too heavy. It's primary tasks involve basic terminal applications like Wordgrinder. I prefer that for writing because on my old matte Dell display it's not too harsh on the eyes when I want to write.
I have a lot of computers for more power hungry tasks, but that's in the only place I wanted something silent and with a low power consumption for lightweight computational tasks.
@@sbrazenor2 that is where the raspberry pi shines low power silent operation that works for light compute tasks
There is a cooler available for RPi that very much looks like a Hyper 212 EVO, albeit a smaller version:
ua-cam.com/video/RyUXC3886Ic/v-deo.html
It's called "The Ice Tower".
Funny you mention that. There is actually a tower cooler for the raspberry pi 4 that works great. ETA Prime did a video on it. I think the tower was even good for no-fan operation too.
A clear and uncluttered video, goes straight to the point and gives a solution. If the Pi 4 gets that hot it might even make a cup of tea - a hitch hikers Brownian motion producer.
I've done the test on a Pimoroni case with the fan shim.
Here are the results (ambient temp. 21'C):
temp=60.0'C
temp=64.0'C
temp=67.0'C
temp=63.0'C
temp=61.0'C
temp=65.0'C
temp=63.0'C
temp=66.0'C
Not bad results -- thanks for sharing.
The temps you get when using the Fan SHIM depend on how you configure the fan control script. Mine is set for the fan to activate when temp reaches 55, so I get a range of 54-57 when running the temp_test script.
@@mikejcross thanks for the feedback. My fan shim is set for the default temperatures of 65C with a hysteresis of 5C.
Came here to ask if you'd run it with the fan shim as well, thanks for this.
Great ! This is the video that is so important for users of this new SBC ! Going back to the Pi 3B and 3B+, at least a heatsink if not an active cooling solution IS needed for real-time activities such as SDR ( software-defined radio ), which I run on a 3B. Quite often without a heatsink, the SBC "stutters" ( cuts out for part of a second ) , or "goes into the red" as regards overheating. So YES, your previous "active-cooling" work HAS been very valuable in those situations. Your final solution with the Pi 4 is obviously the one to go for, for anyone doing more than the most basic work - people dreaming of running two HDMI screens, and using the Pi for streaming video, or indeed SDR as I do, really need to opt for you final solution to "stay cool". Overheating reduces the life of any computer systems, and throughout me career over 30 years in the computer industry, cooling was a mandatory part of computer installations. It certainly needs to be with the Pi 4B. What worries me is the number of school classrooms who will be installing these SBCs over the school holidays. I just HOPE that the installers will install then with active cooling systems. Children have very sensitive hands as compared to adults, and we don't want children burned by the heat, especially when they take the lids off the SBCs, as children are prone to do. Thanks for this very valuable video, Chris, and please ALL educators take note of Chris's findings !
Very cool Chris, thanks! My only objection is: During the 2nd test (when you first added the heat sink), was the improved result due to the added heat sink or due to the increased ventilation? You changed two parameters at once, so it's difficult to say which had the biggest effect. Would it be possible to include results from running a bare Pi4 out of the case first, before adding the heat sink?
Exactly. And how does the standard small heat sink perform in combination with the fan. The difference between the small and large heat sinks was not that big.
I would also love to see active cooling without heat sink (outside of the standard case)
@@edwinschaap5532 It won't be that big because it is limited by convection (or rather conduction). The heat load should be slower to build, as you have a bigger slab of metal to absorb heat but after that things will level out. In a more real world situation you would probably use heat pipes to move the heat from the heatsink to outside the case where it might get some convection heat transfer (because of breeze) but still be technically passive. Kit's can be bought on eBay although they aren't that cheap to the extent the Noctua would be a better choice cost wise at the introduction of moving parts and energy costs.
The reason for not putting a heatsink on while in the case is because that would only delay the inevitable. Without ventilation, the heat has nowhere to go, so adding the heatsink will only provide a slightly larger "space" for the heat to go.
Of course, a heatsink is always better than no heatsink, simply because it makes for a thermal buffer that will smooth out the heat generated by spikes in power usage. This is the same for M.2 drives and RAM; the heatsink is not primarily for cooling, but for thermal buffering.
And of course with the amount of power that the Pi 4 draws you'd want active cooling anyways, such as the S2Pi ice tower
@Anthony Spendlove. Yes I realise that. My original question was about checking cooling performance outside a case (i.e, good ventilation), but without a heatsink. A logical test protocol would be: 1. Inside case, 2. Outside case, 3. Outside case with heatsink. He skipped no 2.
In South Africa, during the hot summer months, it is essential that you make use of a fan and decent heat sink. I must say that I love your Noctua fan fitting kit.
If only Frodo had the raspberry Pi 4. It would melt the ring in no time
Gandalf: "Open the script Frodo."
Frodo: vi temp_test.sh
Gandalf: What can you see? Can you see anything?"
Frodo: "Nothing. There's nothing... wait... there are markings. It's some form of script, I can't read it."
Gandalf: "There are few who can. The language is that of Bash, which I will not utter here.
@@il2xbox so good lmao
@@il2xbox :')
@@il2xbox Oh my gosh.... if ever a pinned post was deserved, this is it!
The original Pentium 4 works better.
Hmm. Theres a pattern here - P4, Pi4...
July 21, 2019 - Great COOL tech video, Chris. In electrical engineering classes, we spent a great deal of time about heat management of solid state devices (and nearby passive devices too, such as power resistors). I think today's tech does a near lousy job of moving hot air through earphone jack holes and speaker holes. When using a heat sink, it is also important to provide cool air across the heat sink fins. And then channel the heat flow out. Motorcycle engines do this very well. A coat of paint on the fins degrades a heat sink dramatically. Your new vid proves a basic yet overlooked point: Manage the pathway of semi-conductor produced heat. Details matter. Thank you, Chris!
Not a lot of heat sink fins on motorcycles these days. They are watercooled, sometimes oil cooled, rarely aircooled. The fins such as they are will be cosmetic
@@twmbarlwmstar I appreciate the update. I was thinking about H-Ds from the 80s. You are correct. Thanks!
How to cook an egg with a RPI,looking forward to the tutorial.
Very true! Wow it was hot in that case after the first test.
I've seen this done (online)with PCs back in the early 2000s. Think massively overclocked early Xeons. ;)
..oh, and it's all over youtube as well. I guess it's a meme, now.
Best option might be a metal case with the case thermally coupled to the SOC. It will still be warm but a great place to keep you coffee mug warm.
@@jasonwilson8312
Sounds like the case that the pi foundation made for the ISS: astro-pi.org/about/mission/
@@jasonwilson8312 Found a better link: astro-pi.org/about/hardware/flight-case/
I swear your channel and you look like you came straight out of the 80ties and i love that aesthetic
What a difference the fan makes, eh? I did already question the Pi factory enclosure so this answered THAT question quite nicely.....
Back in the groove This channel is Absolutely Fabulous
Quite a beefy cooler for such a miniscule device, I must admit. But the results - well they speak for themselves! Well done!
The video we've been waiting for! The new version of the Raspberry Pi needed a thorough analysis of its operating temperatures: now we have it! Thank you very much! It will be interesting to follow in the coming months the different offers of third parties manifacturers concerning passively and actively ventilated cases. Personally, I will wait until these offers are available before purchasing my next RPi4 card. Excellent educational work on your part Chris, once again!
That's awesome. This is the sort of stuff that I subscribed for. You, it turns out, are the cooling king.
Thank you! So many of these tests forget to say what the local "ambient temperature" is.
You mentioned it it at about 2:12, as being 25C.
Some places here in Australia can reach over 40C in the summer, this needs to be factored into any local setup.
Stable low temperatures mean you're not overclocking it enough :D
I like to see the last one without the heat sinks. Only the fan. That would be very interesting how/if they are necessary! Thanks for you work here !!
Yes, I should have tried this. I suspect the results would be pretty good.
Great video, watched on my raspberry pi with noctua active cooling, ceramic heat sinks and an over clock to boot, all this inside a snespi case, it never over heats, look forward to your next video keep up the good work.
*video. Watched
*overclock
*boot. All
*case. It
*overheats. I look
*video. Keep
www.chompchomp.com/terms/commasplice.htm
Hi Chris. great video, bought a Raspberry Pi 4 with a Vilros case. It did not explain what pins to connect the fan to, but your video did. Now with the fan and heat sinks, I can keep my Pi cool, Thanks.
If fan noise is an issue, one could connect the fan to the 3.3v pin. It wouuld, of course, be less efficient at the same time, perhaps two fans at 3.3v would be quiet and efficient.
Thanks for another great video.
The Noctua NF-A4x10 i used inside the unity-case and within its rubbers is hard to hear from 1meter...
Thank you so much for this. I was assuming that the manufacturer provided the best options overall and our additions made a nominal difference. Thank you for showing me it is all on us!
Yeah that small hs should be included with the pi 4 because idle temps shouldn't be in the 70's.
Agreed.
Canakit sells a heat sink set (and a power supply) with its Pi4 boards
a vented case as well, a heatsink does little good if the case is enclosed
AliExpress has really nice heatsinks for 2-4 dollars. Buy a pack of these and get a high quality fan from something like noctua.
@@coffee-vz1nr I just picked up a 40 piece kit of amazon (large and small) for 10.99. parts for 10 pi (easycargo kit)
Chris, Just commented yesterday, have 3 more thoughts
1. Best cooling solution is always creating less heat, another firmware update like the USB controller.
2. With a fan solution, even a smaller one, millions of creative users and companies out there designing/using "hats", will require a 4" (sorry 10 cm) connector extension and a huge case. Ideally the case should provide independent air intake and exhaust so its not just recirculating hot air.
3. If the PCB was designed with the SoC on the bottom side there would be no obstruction from adjacent components so you could cool passively with a copper or aluminum plate the full size of the PCB using the 4 mounting holes, and yes, your feet or some spacer. This would mess with the classic case design but could eliminate 1 of the 2 case pieces.
Wow! How cool is that! I think you can create a channel called explainingPi.
"How cool is that". Way cooler then a RPi4 running in the original case. :-)
Great video -- I considered the Ice Tower fan from your later RPi4 setup but Amazon reviews indicated that the mounting brackets obstruct some of the GPIO headers. This is a huge help for rolling my own solution once I actually use my Pi 4 for workloads that passive cooling can't deal with.
Really well done video
The Raspbery Pi 4 is no longer safe for children. I believe a warning should be included with the Raspbery Pi 4 regarding the extreme temperatures.
This is a very good point.
Well, you could heat even the previous models to these temperatures if you try hard.
So many choice, so many variations you can do with the options you did. Plus there's always other custom jobs others can do too.
From your tests we can say that almost no throttling occours already with passive heatsink (11mins vs 10 mins) and that minimal heatsink is already effective as the big one. Considered that modern silicon is typically designed for working reliably (10+ years) at 100+ degrees it means that a minimal passive heatsink is already doing the job (well maybe it's better saying it's now mandatory to have it) and active cooling is only for having low temperatures for paranoids...
I purchased the raspberry Pi Devastator Robot and the wireless keyboard after watching one of your your videos. I had a few days off work and spent the time building the robot and connecting it to the Pi Zero and Pi 4. I had lots of fun. I had problems having the Pi boot up by itself and running the Python program but stuck at it and got it to work. Lot of great support on the internet when there are problems. I will follow your other video and connect it up to th camera when I get some more days off work. Thanks
Great to hear -- this is what the Pi and UA-cam are all about. :)
Control should have been without any case or heat sink.
Exactly. How do you know how much cooling is due to no case vs adding heatsink
@@codycast Or two controls. With and without case.
Would have been better if there was a bare no case run I agree. Just the air able to move away from the chip surface should make a difference vs no airflow when in the case.
I understand his logic for not doing your version of the control. He states at the end of the video that he believes a cooling solution of some sort is absolutely necessary fro the Pi 4, which I think most people will agree with.
I pillaged the 40X40mm heatsink and fan off my 3B+ (I think the cooler was featured in one of your extreme cooling pi vids). After the install, my temps while stressed went down to between 88°F and 100°F (31-37°C) before it would be between 140°-180° without the cooler. Which I thought was a really good result. I have since made the Pi4 my official new HTPC running Libreelec. And it runs perfectly.
Great to hear.
Did you mean the other way around? Otherwise, you're saying that it _increased_ your temperatures.
@@alvallac2171 no sorry. It came out wrong. I just read that again. After installing the heatsink and fan, my temperatures when stressed stay between 88° and 100° F. Before the heatsinks it was in the range of 140°-180°F.
Nice video! I think that the raspberry pi 4 has come with some great updates, but it is a shame that the raspberry pi foundation didn't invest much effort into the thermal management of the board. Maybe the unexpected release of the new board was a precipitous act ...
Yet another brilliant video thanks for showing me the result's, it clearly shows that the Raspberry Pi needs to have a better cooling system than what is on offer right now.
*video. Thanks
*results. It (plural)
Apostrophes are for possession or contraction, not for pluralizing (except for single letters). Also, that was a run-on sentence and should be written as three separate sentences.
Excellent video, thank you for sharing. Please do the final test again "without" the heatsinks and just the Noctua fan. I suspect that the results will be very similar, and save you from having to spend money on heatsinks, or having to cobble the heatsinks as well.
The active cooling solution worked so well I had to put on a sweater! Great video!
Chilling results...
Thanks as always Chris! The one thing I would have liked to see was the RPi 4 with no heatsink, and no enclosure, to see how much the enclosure added to the heat.
Thank you, just done the Test
temp=36.0'C
temp=49.0'C
temp=51.0'C
temp=50.0'C
temp=50.0'C
temp=50.0'C
temp=51.0'C
temp=50.0'C
with the normal Pi Cooler and a 5V Fan
Good results. Thanks for sharing. What is your ambient? And size of fan?
The Fan is a 4010 Cooling Fan Cooler Radiator 40x40x10 mm form the well known Chinese Market Place for less than 1€.
The room temp was 23°C ,.. and I placed the Pi in a printed Case from Thingiverse. Nothing very spectacular :-)
I was suprised, that this works so well with the small heat sink
Lovely video and just in time. I've put my new Pi 4 through its paces for the first time today and was surprised at how hot everything gets. The USB and ETH ports (not just the chips, the ports as well) get very hot - and all this running out in the open with air condition on. The active cooling would probably cool down those jacks as well
Raspberry Pi Trading has been very diligently deleting my comments on their social media, so I'll try here.
The Pi 4 overheats and throttles in the official case while just idling at desktop. Yeah, you read that right. I connected it to a 4k display, booted it up and deliberately did absolutely nothing, just left it siting at desktop. In an hour the temperature reached 81 °C and it started throttling. Ambient temperature was 18 °C. It went 63 °C over ambient by just idling! Maybe I'm old-fashioned, but to me that is completely unacceptable for an official accessory.
I have to wonder if anyone at RPT even looked at a prototype of the case before greenlighting mass production, because this design flaw is pretty much impossible to miss.
Great video as always. The heat concern of the Pi 4 is yet another indication that the board was released early and without much consideration for real world use. I’m glad to own a Pi 4. But this one has given me a moment of pause regarding what the future may hold for the platform.
✨How about doing a water cooled Raspberry Pi 4 solution and test!
@Quantum Unity Tea anyone?
Mineral oil bath time?
Absolutely love the Noctua fans! Note there is also a 5v variant that is not so high (10mm).
Hasn't anyone invented a 'Windowsill' pi cooler?! I won't give up the day job. I'm amazed how hot the pi 4 runs though, I'll be sticking with my 3 B+ for now.
I tested with an old 80mm 12V Noctua fan from a computer. Either hooking it up to the 3.3V or the 5V fan worked fine. Running stress test without any heat sinks I got 70C (3.3V) and 65C (5V) 60 min test. Idle was below 50C. The big advantage was that the fan was running completely without audible sound. Which was my goal as I like RPI as a small desktop computer because it is silent.
" You need to bake the pi(e) to use it!" LOL
Excellent video. I should have waited for this video as I ordered and started testing a 12v Noctura 40mm fan instead of the 5v you suggested (I had assumed it would work with the 3-to-2 pin adapter, but I was wrong). Thank you for your video(s).
Would have been interesting to compare with a bare Raspberry Pi (with no heat sink) out of the case. I am left wondering how much of the cooler temperatures with the passive heat sinks are due to simply not having the Pi inside the standard fully enclosed case.
Very useful demo. We really need a fan with a 4B! Thanks!
Another great video Chris! Thank you!
Great video. I already use a case with a fan on my 3b+. It has been up for 365 days and counting. I use it for my Unifi Controller. Looking forward to get a pi 4.
Which version of the VLI Firmware were you running. The old 00013701 firmware was causing the VLI to produce a lot of heat.
Can you run the tests with fan and no heat sink. Having a heat sink doesn't appear to lower temps much.
I had applied the update.
Tried this with my own rpi4 with a 40mm noctua and got the same results. 56-57 sustained peak. Absolutely needs active cooling for good performance. Cheers.
Another great video. I would like to see some testing with a flirc aluminum heat sink case to see if t hat is good enough to keep the pi4 below throttle under load w/o a fan.
thats a very good cooling solution because my raspberry pi 4B is useing 3 heat sinks and a cooling fan as well
Wonderful test..significant cooling solution..
Good job..thanks..
I loved that! I really like people improving their hardware with a little thinking and getting these amazing results!
Liked and Subscribed!
Thanks for the sub -- welome aboard!
The entire SBC is getting very hot to the touch! Everything! Every chip is scorching hot (even with heat sinks on every chip) and so are the USB and LAN port(s) and even the SD Card reader!
I'm waiting for the Pi 4 version of the GeeekPi full cover heat sink before even attempting to use the Pi 4 for anything beyond testing.
That heatsink sure looks like one hell of a beast for cooling the Raspberry Pi 3 and 4.
But don't they have them already for sale?
@@Loundre3 No, only versions for Pi 3B and 3B+. I asked them about the 4B version and they said it would be available within the next couple of days. I am anxiously waiting for this heat sink.
Honestly, passive cooling is the reason why I decided to run my home automation system on a Raspberry Pi. Als prior iterations worked fine (I started on a Pi 2 and have upgraded regularly). By now my home automation system has outgrown the 1GB RAM limit and I had started to look at other SBCs with more RAM (like the up board). Then, just at the right time, the Pi 4 with 4GB RAM was launched. I ordered mine within minutes of the launch and have spent the last couple of weeks trying to figure out how to get it to run cool enough to to replace the 3B+ of my live system.
Wait are you saying the actual physical USB and Ethernet sockets are getting hot to the touch? That's hilarious in a way.
Just get an old South bridge heatsink from an LGA 775 motherboard, it's looks like it should fit, if not just hack a bit off.
@@MrMrsirr yes, that's what I'm saying. Hot enough that you can't put your finger on them for more than a couple of seconds.
A big thank you, from Nice France. Keep up the good job. Each week I'm eager to see your new video....
Greetings from the UK. :)
u should put together a kit and sell it as a whole like companies like eleduino do
Great video, and I agree that this Pi needs cooling.
I would have liked to have seen the small heatsink performance as this is the easiest active solution for most folks.
Would be interesting to see what the temperatures would be with the fan and no heatsinks.
I always felt that cooling would be an issue for SBC's but it opens up a massive avenue for testing and getting the best out of them with overclocking and cooling projects. What would've been nice to see would have been some thermal imaging shots from the first test to show surface heat and a thermometer to show the real world temperature within close proximity of the board.
It's always nice to see projects like this as factual data is something that can appreciated on many levels and expanded on until someone reaches the ACME stage of Pi cooling!
I was happy that the Pi3+ I got my daughter had 2 heat sinks and a case with functional and decorative slots that allowed the heat to dissipate quicker.
I've got my 3Bs inside Flirc cases, as i'm a fan of passive cooling whenever possible. I'd be quite interested in seeing how the 4B performs inside the 4B-specific Flirc case, when it ships.
I just got the FLIRC 4B case in the mail. I ran this same script/test and here are my results with a Pi 4b 4GB model
-------------------------
(Ambient) 24.4'C
47.0'C
56.0'C
58.0'C
59.0'C
60.0'C
61.0'C
61.0'C
63.0'C
Finished in about 10 minutes with no throttling.
-------------------------
I also ran it 3 times back to back (without any cool down period) and the case seems to reach steady state (dissipate heat as fast as the SOC can create it) at 68.0'C (at 24.4'C ambient at least).
Impressive what the small heatsink could achieve. Temps are still high, but there was only a bit throttling. Only with a fan you could save roughly 10% of the time needed to perform the test.
I like the idea to have it completely passive cooled. If you put a fan there I would make it temperate controlled. Because most of the time the fan can be off with just the heatsink.
It would be interesting to see how the Pi 4 would overclock with a tiny aio cooler.
An updated video on RPi 4B cases examining relative cooling would be much appreciated. Love the 3D printed mounts, though! Great content.
I am working on such a video! :) Contenders are already arriving.
It would be awesome to have thermal vision on the pi while the test are running.
I've seen some on the web, it gets hot everywhere except the USB sockets.
on that note, I wonder where the temperature sensor(s) are placed on the board.
Increasingly bulletproof and brilliant... as always... thanks Chris.
Comparing the first and second run (in case+no hs vs outside+small hs): how much of the difference comes from adding the hs and how much from removing the case?
I think it is about 50 per cent case and 50 per cent fan.
Thanks for the info. Also great video, as usual.
@@davidgriffin79 I don't think he had it in a case when he was using it with the small heat sink.
I couldn't help noticing that you always use your heat sinks in conjunction with those tapes to stick the heat sink on the sbc but recently I saw a video about testing cpu cooling with various components, including thermal pads and -of course- the classical thermal paste with the result that the paste brings some degrees advantage compared to the other methods.
Mentioning this I would really enjoy to watch those cooling tests rerun using thermal paste to attach those heat sinks to the sbc, not only with the Raspberry Pi 4 but also with other manufacture's sbc.
I havae once used thermal paste in a Pi rig -- ua-cam.com/video/1AYGnw6MwFM/v-deo.html -- works very well, but makes it harder to hold things together! But I may well try what you suggest.
Is there a test that'll max out the CPU and the GPU? People who run emulators would be interested in that.
Thank you, Chris. Well done.
Not a user of the RPi ecosystem, but ... gives some ideas for other situations, and
worth filing away for future reference should I decide to experiment with Pi.
I can not look at your face without singing Country Roads
Just ordered a Raspi 4 4GB, Heatsinks and a Noctua Fan.
Thank you for your help, let see how it works out! :)
I hope it works out very well! :)
It's too bad someone at the Pi foundation didn't think ahead on this one, and put a nice *slim* fan in the "official" case. There's *just* enough room to mount a small one flush if there's no heatsink, or surface mount if one is present. And if a good fan was chosen, no heatsink would be needed anyway. One caveat.... Yes, the case would cost more. But I don't believe that would matter. If I were buying a Pi4B I'd be wanting a fan-based cooling solution anyway.
Pi4 SoC got ready almost one year ahead of schedule. I can only imagine the foundation chose to rush things up, in order to reduce time to market. Even so, one can barely use it with official casing. Also, if a company advertises dual monitor 4k 60fps capabilities, and then explains that only works for desktop, it looks like gross misleading. That's not cool, that's not right. If codecs are not ready at deploy time, than the board is not capable yet. It should _not_ be announced otherwise.
Get the Pibow case with the fan SHIM
I just finished doing this with this fan before seeing this. Highly recommend this fan. The new one has a smaller hub and moves more air.
Sad results in my opinion. Adding up to the USB-C problem I conclude this Pi release was REALLY rushed.
The results in the Pi official case do raise questions . . . they must have known this at launch.
This was so interesting. From looking at the results any aftermarket make of cooling systems must look closely at this. I would say it must have an active heat sink.
To my shame, I just realized I never subscribed to your channel.
Edit: I think that 90 Celsius is ludicrous.
Edit 2: switching from the Kintaro passive aluminium heatsink for my RPi3B to a flatter custom aluminium on top of small copper discs in hope to accommodate it better into a portable frame. I'm seriously considering adding a PWM-driven 5v fan with power circuitry at the side. Unfortunately this is going to add a little bit more strain on the batteries, but I don't think there is much to do if I want some good cooling: either widening the frame or drawing more current.
I don't have a Pi4 yet but I'm on the list to get one. On my Pi 3B+ I use the C4Labs Zebra Virtue Fan Case Black Ice version. It is a little on the pricey side but it seems to be the best off the shelf case I have found. Running Ubuntu Mate I force over clock to 1520 525 on the RAM GPU and I stay about 10 degrees above ambient air temps. I do have a copper heat sink on it that is a bit larger than comes with the kit though. I had the heatsink for many years and I think it came out of a GPU cooling kit. Happy Cooling.
Thanks for the information. I should have bought a 20p heatsink iso a £5 case then.
🤬
I had heard that it runs hotter so bought the Pimoroni Fan Shim. A little fiddly to set up but works well with the pibow coupe and looks good. Very quiet and seems to keep it very cool running stress. It is also smart in that you can set what temperature it turns on at.
Given that there will be more fan solutions for the 4 maybe next video should be a comparison.
Pimoroni have been in touch this afternoon. The Fan Shim was out of stock for this video . . .
Just run your test on my Pi4 which has a 5V fan but no heatsink. The temperature won't go above 54 C, with ambient at 28 C.
The fan is the key!
Thank you, really nice video! Some notes: - maybe the larger heatsink without the top "case" would perform better, but still probably not good enough; -there are already 3d printable available open cases that allow various fan mount(from 40 to 70 mm) . Thanks to the people who shared them (let's tip them); - also the RAM next to the CPU is likely going to run hot. I really liked the idea to cut your own heatsinks, thank you!
Thermal runaway will turn it into a Raspberry Pie... :)
Chris, Great Video as always, however your conclusion that active cooling is required may have an exception.
I just got the FLIRC passive cooling case for RasPi4 in the mail. I ran this same script/test and here are my results.
-------------------------
(Ambient) 24.4'C
47.0'C
56.0'C
58.0'C
59.0'C
60.0'C
61.0'C
61.0'C
63.0'C
Finished in about 10 minutes with no throttling.
-------------------------
I also ran it 3 times back to back (without any cool down period) and the case seems to reach steady state (dissipate heat as fast as the SOC can create it) at 68.0'C (at 24.4'C ambient at least).
Thanks for sharing this data -- very interesting. That is excellent passive cooling.
You did not do a stress test with it out of the box and without any heatsinks.
I was starting to worry about the poor Pi! :)
WOW idling at 74 degrees! Although I''m a big 'fan' of the Raspberry Pi 4, after watching your video I realize that I just won't have enough 'huff and puff' within me to keep it cool. Looking forward to your next video!
I'm against any active cooling. It kills one of the main advantages of low Wattage PCs which is fanless = noiseless.
I love passive cooling too, I often use my Pi1b if I can as it is silent and doesn't use much power. It is enclosed in a plastic case and at 800MHz only gets to 56C running flat out, this was just today converting some of my old .WMA to .MP3..slowly
edit: wav to wma, always get those mixed up.
@@Kenobi5001 I don't have any Raspberry but I have four x86 PCs with M3-7Y30, N3450, Z8500 and Z8350 CPUs + another cheap Chinese TV Box with ARM S905X CPU. All of them are passively cooled & are dealing well with heavy duties. No reason to change it for Raspberry unless it's going to work 24/7 in hard conditions.
i love your videos on SBC's and 3D printing, and Arduino projects. I'm looking forward to obtaining a few of these for my nieces and nephews for Christmas this year. It would be nice if you had a video out comparing the best buys for young minds (they're in their preteen and early teen years) around the end of November, also I like your new channel Explainingthefuture.com Amazing!, keep up the good work.. and here's another thumbs up 👍
I hoped to see Mr Scissors, but I settled for Chris' sticky feet.
Next week Stanley makes an appearance . . . :)
It's worth noting that there is a 10mm thick version of that fan, and the fan being half the height might be useful when the pi is so small. Combine with shorter standoffs to get closer to the heatsink and you could shave 15mm off the height.
Also, it would have been nice to see temperatures for the bare pi with no heatsink and no case, since we have no idea how much of the small heatsink improvement is from the heatsink and how much is from being open-air.
I can imagine that a company will release an Extreme Cooling Pi 4 case with heat pipes etc that will be 4 or 5 times the price of the Pi xD