Raspberry Pi 5 cooling

[sunnyboy010101]

I've read all the posts in this group, and I saw mention of 'factory cooler' in conjunction with some fitment issues on the build.

My question is based on the fact that the RPi 5 I bought came with no cooling, but the supplier included a few stick-on aluminum heat sinks. Where I live in Canada the only real electronics supply is amazon (.ca) and all the Pi5's are bare boards. 

I did a look on "Raspberry Pi 5 cooling" and see several large coolers that clip on the Pi5, and verified these against the 'official' cooler on the Raspberry pi.org site.

But my question to the group is  - if your pi didn't come with a cooler, what did you get? I see shiny aluminum ones that look like the official pi.org one, as well as some with bigger fans in anodized black aluminum.

If fitment is an issue, what coolers are most recommended by this group?

Thanks.

[terry-...@glaver.org]

I have some 2.5mm spacers coming next week that I'm going to try. I'll report back with how it goes.

[Nick Taylor]

I’m about to order a Pi 5 for this project.

Will the PiDP-10 benefit from 8GB RAM, or should I just go for the 4GB version and save some money?

Thanks

Nick

[wjegr...@gmail.com]

I got the direct-attach cooling fan for mine, it looks like it will fit, but won't know until I try it. They're cheap, you can get one from Amazon. As for ram, I went 8gb. If you want to run more than the basic rPi OS and the 10 stuff, such as a full Ubuntu, I figure it's worth the smallish price difference.

[DR]

My take on it is that Pis seem to last forever and get used in other
things at my place, or swapped out.

I know the extra charge for so little more memory seems odd, but I have
never regretted getting the most memory that I could.  I have a few very
early Pi units from people who've gifted them to me and with small
memory allocations, it 'seems' that they aren't quite as much fun as
equivalent  units with the larger memory choice.

In any event, have fun. Dale

[DR]

Indeed, a concern as I want to do it once.

I ordered a Pi5 when first announced and I got the cooler kit and the
unit has been doing odd job tasks and giving me fun stuff to do all this
time.  My observation is that the Pi5 does generate some heat.  The fan
doesn't always run, but clearly works in cooling.


I know it has been said that IF you put the two big fans on the back,
that would be sufficient.  I would be OK with that if those who've used
that setup for awhile can tell me the precious little Pi won't be baked.

On the other hand, Terry has a thought about making the mount for the
covering board stick out further with using what I assume are those
little hex shaped stand off type of things that have a thread on one end
and the hex body has a threaded nut-like hole to make the board stick
out slightly further. I haven't started on my kit yet (too many
interruptions and I wanted to enjoy the process) but have a nice
selection of those threaded hex stand offs so I can try that before
taking a drill to Printed Circuit board and having it chip or shatter.

With more kits out there, what have people done (in addition to locate
the proper place to make a hole in the board and doing that as was
illustrated with the pics a few days ago) as an alternative to drilling?

Dale

[johntk...@gmail.com]

I use the standard Raspberry Pi cooling fan that clips onto the Pi, and it fits perfectly into the PiDP-10.

I also have an NVME "hat" that attaches to the bottom of the Pi (many attach to the top, these probably won't work).

    

[steve...@gmail.com]

All,

Here is the Pi 5 kit I got from Amazon:

https://www.amazon.com/iUniker-Raspberry-Pi-Starter-Extreme/dp/B0CSYL4ZZ5/ref=sr_1_1_sspa?crid=27IC41163LA1X&dib=eyJ2IjoiMSJ9.Kh690eXtnrTHFwNoo7z4dcUaS4ireIN2TBTJ4oA4QQAdWuWLaJEcDi0RfgxA7zum5iVCX2rQsI1lLj3N1zPgwShzrC6KCUDpB9tKpqQ2eKbeRGhpyqi00vJWw7xBeYlt.mDH2xNBS7re2gqKbVb_-ccrKSqp4hsffJHxAiXYcKKY&dib_tag=se&keywords=iuniker+raspberry+pi+5+starter+kit&qid=1714147816&sprefix=iuniker+raspberry+pi+5+starter+kit%2Caps%2C132&sr=8-1-spons&sp_csd=d2lkZ2V0TmFtZT1zcF9hdGY&psc=1

It came with everything needed, and a little bit more.

I had originally installed the Pi 5 into the PiDP-10 without the fan fearing that it would not fit with. The PiOS temperature monitor continually showed temperatures in the 65C range when idle and up to 85C when the CPUs were loaded. And, by the way, I did install two 120mm fans from Amazon:

https://www.amazon.com/ELUTENG-Computer-Receiver-Compatible-Playstation/dp/B06Y5WWBHH/ref=sr_1_3?crid=WXW7XB7NOWE5&dib=eyJ2IjoiMSJ9.FvqhKZ1ZUSpmFCWZ9QggIXElAGTjJYLdxmGQK7RRFB24ZlQIpZKvSjvMwhdDduY3qt7trwWStFgIAUr2IRWpCVu4PpX9PQUMiwHjk15VBdHq1yBcA_8_JLOQAxFTNxQcpi4hCiHaK5KDX6ZnWWa45g.ImDjOMzVVVBP-srPVxR8DoYijCXs3-41-coYx2qyFyQ&dib_tag=se&keywords=eluteng%2B120mm%2Bfan%2Bwith%2B3%2Bspeed&qid=1714148017&sprefix=elutneg%2B120mm%2Bfan%2Caps%2C167&sr=8-3&th=1

but found them to be entirely ineffective at cooling the Pi. there is no ducting to channel the fan stream at the Pi, there are too many gaps elsewhere, and the distance between the PCB and the decorative backplate is simply too narrow for effective air flow. The fans are much more of an audio enhancement on the kit than anything, IMHO. Well worth it, to be sure, but really just for sound's sake.

On finding that the 120mm fans were ineffective at cooling the pi, I later installed the on-board fan and am quite impressed. It keeps the CPUs at as low as 35C when idle and no more than 45C when loaded.

I'm very happy with my set up, the PiDP-10 is a beautiful machine.

-- steve

[sunnyboy010101]

I got the 8gb one. I always try to max out memory on any computer purchase (to the limit I can afford). I can't say I've ever regretted having as much memory as possible on any computer.

[sunnyboy010101]

Wanting to save a bit of money, I got a "cana kit" version that had only the bare 8gb Raspberry Pi 5 and a 45W power supply. I learned many Pi's ago that you should not skimp on the power supply, especially running Oscar's kits. (I have the PiDP 8 when it first came out and then the PiDP11 when it first came out). 

I also bought a 128MB micro-SD card (Sandisk). Together it came to a lot less than the kits with "all the stuff" by about $50CDN.

Then I had a problem installing Raspbien in a headless format (my old instructions simply didn't work with Bookworm) and bought a micro-HDMI to HDMI cable. Still total cost less than the kits and I have (I think) better quality parts (SD card, HDMI cable). (aside: I bought a BlueRigger HDMI as i;ve already used this brand very successfully - even their 35ft braided HDMI cable is flawless).

I ended up using the Raspberry Pi Imager to configure and install Bookworm on my Pi. Worked perfectly on WIFI in headless configuration. 

I just now bought the GeekPi cooler as it comes sooner than any others for about the same price. Black anodized aluminum and a good looking large fan. 

Waiting for Oscar to give me the "go ahead and order/pay" email. Looks like a great kit.

[steve...@gmail.com]

I agree. the cost differential is really not that much, particularly in terms of the overall cost of a completed PiDP-10. And in my earlier years I would always get the smallest memory to save on cost only to be disappointed later because it turned out to be too limited. "They" always find a way to use more and more of the resources until it doesn't fit on your machine any more. Going big buys you time in the long run.

-- steve

[Jorge Amodio]

This is the official one

https://www.raspberrypi.com/products/active-cooler/

DigiKey, SparkFun and other distributors have them

https://www.sparkfun.com/products/23585?src=raspberrypi

Regards

-Jorge

-Jorge

On Apr 25, 2024, at 23:52, sunnyboy010101 <sunnybo...@gmail.com> wrote:

I've read all the posts in this group, and I saw mention of 'factory cooler' in conjunction with some fitment issues on the build.

--
You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/af70596c-6d64-49c3-bcbf-61f6140f1c78n%40googlegroups.com.

[terry-...@glaver.org]

On Friday, April 26, 2024 at 8:47:44 AM UTC-4 daleea...@gmail.com wrote:

On the other hand, Terry has a thought about making the mount for the
covering board stick out further with using what I assume are those
little hex shaped stand off type of things that have a thread on one end
and the hex body has a threaded nut-like hole to make the board stick
out slightly further.

  I'll be doing a write-up once I have everything set up the way I like.

I did get the spacers and installed them and it is working nicely.

  Another thing I did was use 2 layers of Kapton tape on the USB port 

tops and the Ethernet connector for insulation. Makes things much

nicer-looking than cardboard.

  I replaced the 3 factory IC sockets with Mill-Max ones and added

one more for the MAX232.

  I got some nice snap-in DB-25 male connectors for the rear panel.

DEC traditionally used DB-25 male for the computer end of things -

they didn't switch to DE-9 until after their adventures in MMJ-land

ended, which is well past period-correct for this CPU.

  Clearances in there are VERY tight - I couldn't mount the MAX232

charge pump caps I used on the PiDP-11 as they interfere with the

LED board sitting on the front panel. I have some (physically) smaller

ones on the way.

  Another clearance issue is with the fan, power, key switch and

MAX232 headers. The back cover won't seat fully. I have some

1/4" 6-32 standoffs on order to hopefully deal with that. That will

also let me use real metal screws/washers/nuts  to join the 3

parts of the rear cover together, instead of the self-stripping plastic

ones in the kit.

[Bob Eager]

I am building a write-up with photographs for my (serial 19) PiDP-10. It's been a bit slower than I hoped,
mainly due to lack of time, but also because I keep finding new bits to add, and have to wait for them
to arrive!

On Saturday 27 April 2024 at 05:44:49 UTC+1 terry-...@glaver.org wrote:

On Friday, April 26, 2024 at 8:47:44 AM UTC-4 daleea...@gmail.com wrote:

On the other hand, Terry has a thought about making the mount for the
covering board stick out further with using what I assume are those
little hex shaped stand off type of things that have a thread on one end
and the hex body has a threaded nut-like hole to make the board stick
out slightly further.

  I'll be doing a write-up once I have everything set up the way I like.

I did get the spacers and installed them and it is working nicely.

I decided on this almost from the outset. I have 6mm nylon hex standoffs in stock, so it was no issue.

 Another thing I did was use 2 layers of Kapton tape on the USB port 

tops and the Ethernet connector for insulation. Makes things much

nicer-looking than cardboard.

So did i, although I was paranoid and used 4 layers! This was nothing new, as I did this back on the PiDP-11 as well.

 

 I replaced the 3 factory IC sockets with Mill-Max ones and added

one more for the MAX232.

I added one more (machined/turned pin) and as noted elsewhere I wished I'd replaced the originals too.

 

 I got some nice snap-in DB-25 male connectors for the rear panel.

DEC traditionally used DB-25 male for the computer end of things -

they didn't switch to DE-9 until after their adventures in MMJ-land

ended, which is well past period-correct for this CPU.

I decided to use DE-9 as it was more convenient. I am making some lightweight extension 'tails' from
IDC DE-9 connectors and ribbon cable, to reduce loading on the PCB from those cables.

I decided not to solder any wires to the PCB; everything is headers and sockets.

 

 Clearances in there are VERY tight - I couldn't mount the MAX232

charge pump caps I used on the PiDP-11 as they interfere with the

LED board sitting on the front panel. I have some (physically) smaller

ones on the way.

I used a MAX202, which only needs 0.1uF parts; nonpolarised, ceramic discs.

 

Another clearance issue is with the fan, power, key switch and

MAX232 headers. The back cover won't seat fully. I have some

1/4" 6-32 standoffs on order to hopefully deal with that. That will

also let me use real metal screws/washers/nuts  to join the 3

parts of the rear cover together, instead of the self-stripping plastic

ones in the kit.

I haven't encountered that yet. But I have standoffs if needed. I had a similar issue on the PiDP-11. The solution there was a frame

of 10mm square acrylic that spaces the rear cover away, and some longer screws.

[terry-...@glaver.org]

On Saturday, April 27, 2024 at 4:52:12 AM UTC-4 warp...@gmail.com wrote:

On Saturday 27 April 2024 at 05:44:49 UTC+1 terry-...@glaver.org wrote:

I replaced the 3 factory IC sockets with Mill-Max ones and added

one more for the MAX232.

I added one more (machined/turned pin) and as noted elsewhere I wished I'd replaced the originals too.

I actually decided to remove the supplied sockets after I'd soldered them to the board.

There's something a bit "off" about them, where the chips don't necessarily line up as

expected. That can't happen with the Mill-Max sockets.

Note: I do not suggest anyone else try removing the sockets. While it isn't difficult for

an experienced solderer, many of these kits are assembled by people new(ish) to

soldering (hence Oscar's detailed soldering instructions) and you can damage the PCB

if you overheat things trying to remove the sockets.

[Brian Denley]

I keep hearing about changing the ic sockets for low profile ones but I can’t seem to find out why.  Do the ics get it the way somewhere?

Brian 

Brian Denley

KB1VBF

Sent from my iPad

On Apr 27, 2024, at 1:14 PM, terry-...@glaver.org <terry-...@glaver.org> wrote:

--

You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/4cf64529-5f82-4df4-b1bd-7af74bdfc3a2n%40googlegroups.com.

[terry-...@glaver.org]

On Saturday, April 27, 2024 at 6:20:30 PM UTC-4 briandenley wrote:

I keep hearing about changing the ic sockets for low profile ones but I can’t seem to find out why.  Do the ics get it the way somewhere?

It isn't an issue of height, just a quality issue. Normally I tack two diagonal

corners, re-heat one and push the socket down, then repeat on the other

corner. In this case, the pin popped up in the socket when I tried seating

it.

And there are those of us who have been doing this long enough that we

still have memories of the run of defective TI IC sockets in the late 70's.

With the Mill-Max sockets, you're assured of getting a quality socket and

a gas-tight metal-to-metal connection.

Back in the 80's when I was designing hardware for a living, I used Augat

wire-wrap sockets for the initial prototypes and then a special version of

Augat pins as PCB sockets in the pilot runs. These were individual pins,

but in an IC-sized pin carrier. So you soldered them in and then pulled the

carrier out and discarded it. This made etch cuts and similar inside the

socket footprint effortless.

[DR]

There has been some discussion of substituting brass hex standoffs for
the nylon ones that come with the kit.

If we have them on hand, it seems like something easy enough to
do...until we learn that the nylon was used as an insulator and subbing
brass now makes for connections or grounding where there shouldn't be any.


For those who've built it, does it appear the nylons were chosen for
their insulating qualities and doing what seems to be an upgrade in
materials will just invite trouble?

I hate to seem to be second guessing Oscar on all of this, but upping
quality when it is within the builder's grasp or affordability or just a
great junk box of parts to dive into, seems a reasonable thing to do.

Dale

[The Oracle]

Dale,

Is this what your asking about?

Thanks,

Lynn

[DR]

Yes, nice pix.  Can one freely substitute brass ones without shorting
out something?

[The Oracle]

Dale,

I cant see any reason, not to use brass.

[DR]

Thanks.  Dale

I think I have an assortment box of various lengths and pitch sizes.

[oscarv]

Terry,

If I understand correctly: If you fit the optional MAX232, the solder points will against the screws that mount the optional fan? That is indeed one combination that I did not try.

The solution would indeed be to use some spacers (or O-rings) on the top line of screws that hold the back panel against the case.

Should I put that into the building instructions: if you do both the MAX-232 and the fan option, create some extra space with O-rings?

Kind regards,

Oscar.

[oscarv]

Dale,

The nylon hex spacers used to mount the Pi onto the PCB were chosen purely for cost & weight reasons. The weight reason is particularly important: we're bumping against the 2kg international shipping limit. Above 2kgs, the cost of international shipping often doubles. And for instance, shipping a PiDP-10 to Australia costs us $100 at the moment! So we're fighting for grams to push the parcel weight juuuuuust below 2000 grams. I think it is 2010-2020 grams at the moment. Yikes.

But I've always used the nylon spacers, for the PiDP-8 and -11. I don't think you gain much by using metal ones, the mount is very firm with 4 nylon plus the GPIO connector.

Kind regards,

Oscar.

[Carlos Sancho]

Hello,

Using brass can be dangerous. If you look closely to the holes you can see that some traces pass too close to them. Tightening those brass could peel the trace isulation and make contact and thus make unwanted shorts.

Just my 2 cents.

Regards

> El 28 abr 2024, a las 6:05, DR <daleea...@gmail.com> escribió:
>
> Thanks. Dale

>
> I think I have an assortment box of various lengths and pitch sizes.
>
>

> --
> You received this message because you are subscribed to the Google Groups "PiDP-10" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.

> To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/7c582b35-9ab2-47d3-af9b-f0ea068809c4%40gmail.com.

[Andy]

yeah I've taken to using nylon nuts/bolts etc for smaller applications (often with a dab of epoxy when done), especially where there is no movement or stress. They are surprisingly strong!

[terry-...@glaver.org]

There are a bunch of clearance issues in various places:

1) The caps I originally used for the MAX232 charge pumps were too tall, and prevented the LED board from sitting flush with the front panel artwork board. These are the same caps I used on the PiDP-11 with no issue (but of course, there the LEDs were mounted on spacers so there was more clearance between the LED board and artwork on the PiDP-11). I have some new (physically) smaller ones coming.

2) The factory Pi 5 cooling fan pop pin mounts interfere with the back cover for the Pi. I solved this with M2100-2545-SS standoffs via Digi-Key. Note that I built the Pi based on the illustration in the building instructions - flat head metal screws through the LED board to nylon standoffs on the back of the LED board, then the Pi, then the Digi-Key standoffs, the kit's Pi cover plate, then 2.5mm lockwashers and metal hex nuts. Others have solved this by drilling holes in the kit's Pi cover plate.

3) [The biggie] The back cover does not fit flush with the mounting holes if any of the pads for the power cable, on/off switch, fans, MAX232 I/O are populated, even with low-profile right-angle headers. I'm waiting on some 8717 aluminum hex standoffs (also via Digi-Key) which I will try to thread into the case holes instead of the self-tapping screws. That will give me 1/4" more clearance between the LED board and the back panel, while still keeping the panel recessed from the rim of the case. That will also let me use metal M2.5 screws / nuts / lockwashers to assemble the multiple sections of the back cover instead of the easily-stripped nylon ones, since clearance will no longer be an issue.

[Malcolm Ray]

Hmm. If you increase the clearance between the light panel and the back panel, I think the latter will obstruct the Pi's ethernet port and two USB ports.

You may need to use longer standoffs to mount the Pi, plus a booster header for the GPIO pins.

--

You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/2bd35c94-55b8-499a-8682-6ca134a0cdd4n%40googlegroups.com.

[oscarv]

Terry,

Some good points!

On Sunday, April 28, 2024 at 9:32:06 PM UTC+2 terry-...@glaver.org wrote:

1) The caps I originally used for the MAX232 charge pumps were too tall,

The 100pF ceramics should be fine, but those are for MAX232A and (I believe) MAX202. Should I amend the web page text?

 

2) The factory Pi 5 cooling fan pop pin mounts interfere with the back cover for the Pi.

Yes! This is an issue. The simplest solution so far (from the Google Group) is to drill two holes in the Pi cover panel. I did that yesterday, it solves the problem.

 

3) [The biggie] The back cover does not fit flush with the mounting holes if any of the pads for the power cable, on/off switch, fans, MAX232 I/O are populated, even with low-profile right-angle headers.

Hm! Indeed. The solution would be O-rings (is that what you call them, spacer rings around a bolt or screw?) around the screws between case and back panel, to give it some more room. 

But I think I need to add the the web pages that there's only enough space to solder wires in (which indeed is what I did to test the prototypes...) which is not ideal. Or drill a hole on the back panel to give the connectors some space. Also not ideal. Need to think this through.

Kind regards,

Oscar.

[oscarv]

On Sunday, April 28, 2024 at 10:05:28 PM UTC+2 Sheepless wrote:

Hmm. If you increase the clearance between the light panel and the back panel, I think the latter will obstruct the Pi's ethernet port and two USB ports.

There is a press-out-or-flex-in (there is a new word) pad in front of the Pi's USB ports, so that should accommodate a slightly more distanced back panel without blocking the Pi's connectors.

Hmm. So far, the best solution I can think of is indeed spacer rings around the mount screws. Not entirely satisfactory.

Kind regards,

Oscar

[terry-...@glaver.org]

On Sunday, April 28, 2024 at 4:35:43 PM UTC-4 oscarv wrote:

Terry,

Some good points!

On Sunday, April 28, 2024 at 9:32:06 PM UTC+2 terry-...@glaver.org wrote:

1) The caps I originally used for the MAX232 charge pumps were too tall,

The 100pF ceramics should be fine, but those are for MAX232A and (I believe) MAX202. Should I amend the web page text?

  That sounds like a good idea.

 

2) The factory Pi 5 cooling fan pop pin mounts interfere with the back cover for the Pi.

Yes! This is an issue. The simplest solution so far (from the Google Group) is to drill two holes in the Pi cover panel. I did that yesterday, it solves the problem.

  I'll send pictures of my completed setup when all done.

 

3) [The biggie] The back cover does not fit flush with the mounting holes if any of the pads for the power cable, on/off switch, fans, MAX232 I/O are populated, even with low-profile right-angle headers.

Hm! Indeed. The solution would be O-rings (is that what you call them, spacer rings around a bolt or screw?) around the screws between case and back panel, to give it some more room.

  That would work, but it makes reassembling the case more of a pain. I went through that with the brass standoffs and the threaded inserts in the PiDP-11. I've become a fan of permanently-installed solutions that allow rapid disassembly and reassembly.

 

But I think I need to add the the web pages that there's only enough space to solder wires in (which indeed is what I did to test the prototypes...) which is not ideal. Or drill a hole on the back panel to give the connectors some space. Also not ideal. Need to think this through.

  The silkscreen for the power / switch / fan headers does show right-angle PCB-mount connectors, so that's what I used. I'll report back on my success (or lack thereof) with the standoffs. And I still need to install my case fans.  

[oscarv]

I think that the insufficient space for angled pin headers on the fan connectors and optional external power power circuit is the first snafu on the PiDP-10 that is not so easy to fix. Hmm.

Kind regards,

Oscar.

[Rob Doyle]

The recommended capacitor for the MAX202 and the MAX232A are 0.1uF - not
100pF.

Rob.

> --
> You received this message because you are subscribed to the Google
> Groups "PiDP-10" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to pidp-10+u...@googlegroups.com

> <mailto:pidp-10+u...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/pidp-10/f2d066b8-1265-454f-8e16-47e26832eacfn%40googlegroups.com <https://groups.google.com/d/msgid/pidp-10/f2d066b8-1265-454f-8e16-47e26832eacfn%40googlegroups.com?utm_medium=email&utm_source=footer>.