Does anyone know if this is *fully* isolated from the gubbins?
Measurements with an ordinary digital meter suggests that it is, but I'd like
to be sure.

The unit I'm developing has a grounded diecast alloy case and I'd like to use
this as a heatsink, ideally with a small strip of copper in direct contact with
both, and just a smear of heatsink compound.

--
W J G

On 02-07-2021 23:23, Folderol wrote:
> Does anyone know if this is *fully* isolated from the gubbins?
> Measurements with an ordinary digital meter suggests that it is, but I'd like
> to be sure.
>
> The unit I'm developing has a grounded diecast alloy case and I'd like to use
> this as a heatsink, ideally with a small strip of copper in direct contact with
> both, and just a smear of heatsink compound.

It's a heat spreader, not connected to the internals electrically (of
course?). What you plan to do is what all the alloy cases do that act as
heatsinks; you'll be safe. Hardest part will be to get the distance just
right: not too close or you'll bend the board and the connections, not
too far or it won't transfer heat.

Ref.:
https://www.raspberrypi.org/blog/raspberry-pi-3-model-bplus-sale-now-35/

On Sat, 3 Jul 2021 11:15:05 +0200
"A. Dumas" <alexandre@dumas.fr.invalid> wrote:

>On 02-07-2021 23:23, Folderol wrote:
>> Does anyone know if this is *fully* isolated from the gubbins?
>> Measurements with an ordinary digital meter suggests that it is, but I'd like
>> to be sure.
>>
>> The unit I'm developing has a grounded diecast alloy case and I'd like to use
>> this as a heatsink, ideally with a small strip of copper in direct contact with
>> both, and just a smear of heatsink compound.
>
>It's a heat spreader, not connected to the internals electrically (of
>course?). What you plan to do is what all the alloy cases do that act as
>heatsinks; you'll be safe. Hardest part will be to get the distance just
>right: not too close or you'll bend the board and the connections, not
>too far or it won't transfer heat.

Thanks for the confirmation. I was pretty sure that was the case.
My intention is to mount the Pi upside down with long bolts and springs
pushing it down (it will be fully enclosed) so that should keep everything safe.

--
W J G

On 03/07/2021 13:04, Folderol wrote:
> On Sat, 3 Jul 2021 11:15:05 +0200
> "A. Dumas" <alexandre@dumas.fr.invalid> wrote:
>
>> On 02-07-2021 23:23, Folderol wrote:
>>> Does anyone know if this is *fully* isolated from the gubbins?
>>> Measurements with an ordinary digital meter suggests that it is, but I'd like
>>> to be sure.
>>>
>>> The unit I'm developing has a grounded diecast alloy case and I'd like to use
>>> this as a heatsink, ideally with a small strip of copper in direct contact with
>>> both, and just a smear of heatsink compound.
>>
>> It's a heat spreader, not connected to the internals electrically (of
>> course?). What you plan to do is what all the alloy cases do that act as
>> heatsinks; you'll be safe. Hardest part will be to get the distance just
>> right: not too close or you'll bend the board and the connections, not
>> too far or it won't transfer heat.
>
> Thanks for the confirmation. I was pretty sure that was the case.
> My intention is to mount the Pi upside down with long bolts and springs
> pushing it down (it will be fully enclosed) so that should keep everything safe.
>
smear of heatsink compound will help as well

<https://www.amazon.co.uk/Arctic-Silver-Thermal-Compound-3-5g/dp/B0087X728K/ref=dp_prsubs_1?pd_rd_i=B0087X728K&psc=1>
--
“Progress is precisely that which rules and regulations did not foresee,”

– Ludwig von Mises

On Sat, 3 Jul 2021 11:15:05 +0200
"A. Dumas" <alexandre@dumas.fr.invalid> wrote:

> On 02-07-2021 23:23, Folderol wrote:
> > Does anyone know if this is *fully* isolated from the gubbins?
> > Measurements with an ordinary digital meter suggests that it is,
> > but I'd like to be sure.
> >
> > The unit I'm developing has a grounded diecast alloy case and I'd
> > like to use this as a heatsink, ideally with a small strip of
> > copper in direct contact with both, and just a smear of heatsink
> > compound.
>
> It's a heat spreader, not connected to the internals electrically (of
> course?). What you plan to do is what all the alloy cases do that act
> as heatsinks; you'll be safe. Hardest part will be to get the
> distance just right: not too close or you'll bend the board and the
> connections, not too far or it won't transfer heat.
>
> Ref.:
> https://www.raspberrypi.org/blog/raspberry-pi-3-model-bplus-sale-now-35/

There are various flexible thermally conductive materials, such as:

https://uk.farnell.com/w/c/cooling-thermal-management/thermal-interface-materials/thermally-conductive-materials?conductive-material=silicone|silicone-elastomer&sort=P_PRICE

--
Joe

Folderol <general@musically.me.uk> wrote:
> Thanks for the confirmation. I was pretty sure that was the case.
> My intention is to mount the Pi upside down with long bolts and springs
> pushing it down (it will be fully enclosed) so that should keep everything safe.

Make sure to have a close look at all the Pi 4 thermal images you can find.
The SoC is a good contact point but there are a few others that get hot,
too. If you don't connect to those and keep the board upside down, you
might trap the heat.

Also, I'd be wary of using springs because they will probably bend the
board. That might damage the traces and all those minuscule solder points.
Maybe not a problem if they're very slack.

On Sat, 03 Jul 2021 14:42:09 +0200
A. Dumas <alexandre@dumas.fr.invalid> wrote:
>The SoC is a good contact point but there are a few others that get hot,
>too. If you don't connect to those and keep the board upside down, you
>might trap the heat.

Good point! Thanks.

--
W J G

On 03/07/2021 22:00, Folderol wrote:
> On Sat, 03 Jul 2021 14:42:09 +0200
> A. Dumas <alexandre@dumas.fr.invalid> wrote:
>> The SoC is a good contact point but there are a few others that get hot,
>> too. If you don't connect to those and keep the board upside down, you
>> might trap the heat.
>
> Good point! Thanks.
>
Ventilation is important: I once designed a small audio equaliser in a
sealed 19" rack case - it was only about 15 watts total draw, but after
5 hours continuous usage it packed up - inside temperatures rose to over
100°C!

WE drilled a few holes in the case and that fixed it.

Airflow is better than transmission through even metal cases. The
problem is that with no air movement the air inside becomes an effective
thermal insulator.

Furthermore *forced* air cooling can multiply heat loss by up to a 5
times or more.

--
“The urge to save humanity is almost always only a false face for the
urge to rule it.”
– H. L. Mencken

On 02/07/2021 22:23, Folderol wrote:
>
> Does anyone know if this is *fully* isolated from the gubbins?
> Measurements with an ordinary digital meter suggests that it is, but I'd like
> to be sure.
>
> The unit I'm developing has a grounded diecast alloy case and I'd like to use
> this as a heatsink, ideally with a small strip of copper in direct contact with
> both, and just a smear of heatsink compound.
>
Would the RPi-400 be relevant in your usage?

--
Cheers,
David
Web: http://www.satsignal.eu

On Sun, 4 Jul 2021 08:41:30 +0100
The Natural Philosopher <tnp@invalid.invalid> wrote:

>On 03/07/2021 22:00, Folderol wrote:
>> On Sat, 03 Jul 2021 14:42:09 +0200
>> A. Dumas <alexandre@dumas.fr.invalid> wrote:
>>> The SoC is a good contact point but there are a few others that get hot,
>>> too. If you don't connect to those and keep the board upside down, you
>>> might trap the heat.
>>
>> Good point! Thanks.
>>
>Ventilation is important: I once designed a small audio equaliser in a
>sealed 19" rack case - it was only about 15 watts total draw, but after
>5 hours continuous usage it packed up - inside temperatures rose to over
>100°C!
>
>WE drilled a few holes in the case and that fixed it.
>
>Airflow is better than transmission through even metal cases. The
>problem is that with no air movement the air inside becomes an effective
>thermal insulator.
>
>Furthermore *forced* air cooling can multiply heat loss by up to a 5
>times or more.

Indeed. I been reading up on all the reports various people have made, and
come to the conclusion I was heading down the wrong path - note past tense :)

It now looks I'll get best results with normal mounting, a finned heatsink
stuck on top with a 30mm fan placed alongside so it blows through the fins.

I want to avoid external ventilation as there is mains inside, so there would
need to be all sorts of additional protection barriers to deal with small hands
and/or brains!

--
W J G

On Sun, 4 Jul 2021 08:53:49 +0100
David Taylor <david-taylor@blueyonder.co.uk.invalid> wrote:

>On 02/07/2021 22:23, Folderol wrote:
>>
>> Does anyone know if this is *fully* isolated from the gubbins?
>> Measurements with an ordinary digital meter suggests that it is, but I'd like
>> to be sure.
>>
>> The unit I'm developing has a grounded diecast alloy case and I'd like to use
>> this as a heatsink, ideally with a small strip of copper in direct contact with
>> both, and just a smear of heatsink compound.
>>
>Would the RPi-400 be relevant in your usage?
>
No. I looked at it and immediately rejected it - it's just completely wrong.

--
W J G

On Sun, 04 Jul 2021 08:41:30 +0100, The Natural Philosopher wrote:

> On 03/07/2021 22:00, Folderol wrote:
>> On Sat, 03 Jul 2021 14:42:09 +0200 A. Dumas
>> <alexandre@dumas.fr.invalid> wrote:
>>> The SoC is a good contact point but there are a few others that get
>>> hot,
>>> too. If you don't connect to those and keep the board upside down, you
>>> might trap the heat.
>>
>> Good point! Thanks.
>>
> Ventilation is important: I once designed a small audio equaliser in a
> sealed 19" rack case - it was only about 15 watts total draw, but after
> 5 hours continuous usage it packed up - inside temperatures rose to over
> 100°C!
>
> WE drilled a few holes in the case and that fixed it.
>
> Airflow is better than transmission through even metal cases. The
> problem is that with no air movement the air inside becomes an effective
> thermal insulator.
>
> Furthermore *forced* air cooling can multiply heat loss by up to a 5
> times or more.

Have you any comparable figures for passive 'chimney effect' cooling,
i.e. use a metal contact heat sink on the chip(s) but fit, say, a 50mm x
400mm cardboard or plastic tube rather than a fan on the enclosure. A
benefit is that chimney effect is quiet and can't go wrong (unless you
count dead moths etc accumlating inside it, but I'm curious how effective
it would be compared with a fan.

--
Martin | martin at
Gregorie | gregorie dot org

On Sun, 4 Jul 2021 12:27:31 -0000 (UTC)
Martin Gregorie <martin@mydomain.invalid> wrote:

>On Sun, 04 Jul 2021 08:41:30 +0100, The Natural Philosopher wrote:
>
>> On 03/07/2021 22:00, Folderol wrote:
>>> On Sat, 03 Jul 2021 14:42:09 +0200 A. Dumas
>>> <alexandre@dumas.fr.invalid> wrote:
>>>> The SoC is a good contact point but there are a few others that get
>>>> hot,
>>>> too. If you don't connect to those and keep the board upside down, you
>>>> might trap the heat.
>>>
>>> Good point! Thanks.
>>>
>> Ventilation is important: I once designed a small audio equaliser in a
>> sealed 19" rack case - it was only about 15 watts total draw, but after
>> 5 hours continuous usage it packed up - inside temperatures rose to over
>> 100°C!
>>
>> WE drilled a few holes in the case and that fixed it.
>>
>> Airflow is better than transmission through even metal cases. The
>> problem is that with no air movement the air inside becomes an effective
>> thermal insulator.
>>
>> Furthermore *forced* air cooling can multiply heat loss by up to a 5
>> times or more.
>
>Have you any comparable figures for passive 'chimney effect' cooling,
>i.e. use a metal contact heat sink on the chip(s) but fit, say, a 50mm x
>400mm cardboard or plastic tube rather than a fan on the enclosure. A
>benefit is that chimney effect is quiet and can't go wrong (unless you
>count dead moths etc accumlating inside it, but I'm curious how effective
>it would be compared with a fan.

Chimney heat pipes do work, I know from my TV servicing days. However, the unit
can be used upside down, or even on one edge so that's not really an option!

I'll get a fan, and try it first with no heatsink, and blowing across the
top of the Pi. If that's cools it enough, then I won't bother to go further.

As it stands, the absolute hottest I've managed to get it when fully loaded up
was 74C, so even that is usable, though obviously it's not a good idea!

--
W J G

On 04/07/2021 13:27, Martin Gregorie wrote:
> On Sun, 04 Jul 2021 08:41:30 +0100, The Natural Philosopher wrote:
>
>> On 03/07/2021 22:00, Folderol wrote:
>>> On Sat, 03 Jul 2021 14:42:09 +0200 A. Dumas
>>> <alexandre@dumas.fr.invalid> wrote:
>>>> The SoC is a good contact point but there are a few others that get
>>>> hot,
>>>> too. If you don't connect to those and keep the board upside down, you
>>>> might trap the heat.
>>>
>>> Good point! Thanks.
>>>
>> Ventilation is important: I once designed a small audio equaliser in a
>> sealed 19" rack case - it was only about 15 watts total draw, but after
>> 5 hours continuous usage it packed up - inside temperatures rose to over
>> 100°C!
>>
>> WE drilled a few holes in the case and that fixed it.
>>
>> Airflow is better than transmission through even metal cases. The
>> problem is that with no air movement the air inside becomes an effective
>> thermal insulator.
>>
>> Furthermore *forced* air cooling can multiply heat loss by up to a 5
>> times or more.
>
> Have you any comparable figures for passive 'chimney effect' cooling,
> i.e. use a metal contact heat sink on the chip(s) but fit, say, a 50mm x
> 400mm cardboard or plastic tube rather than a fan on the enclosure. A
> benefit is that chimney effect is quiet and can't go wrong (unless you
> count dead moths etc accumlating inside it, but I'm curious how effective
> it would be compared with a fan.
>
>
Not very, or we wouldn't use fans ;-)

--
No Apple devices were knowingly used in the preparation of this post.

The Natural Philosopher wrote:

>> Have you any comparable figures for passive 'chimney effect' cooling,
>> i.e. use a metal contact heat sink on the chip(s) but fit, say, a 50mm x
>> 400mm cardboard or plastic tube rather than a fan on the enclosure. A
>> benefit is that chimney effect is quiet and can't go wrong (unless you
>> count dead moths etc accumlating inside it, but I'm curious how effective
>> it would be compared with a fan.
>>
>>
> Not very, or we wouldn't use fans ;-)

may be in a server room where there is natural air flow caused by
ventilation - but the fan multiplies the cooling factor for sure

On Sun, 4 Jul 2021 15:54:57 +0100
The Natural Philosopher <tnp@invalid.invalid> wrote:

> On 04/07/2021 13:27, Martin Gregorie wrote:

> > Have you any comparable figures for passive 'chimney effect' cooling,
> > i.e. use a metal contact heat sink on the chip(s) but fit, say, a 50mm x
> > 400mm cardboard or plastic tube rather than a fan on the enclosure. A
> > benefit is that chimney effect is quiet and can't go wrong (unless you
> > count dead moths etc accumlating inside it, but I'm curious how
> > effective it would be compared with a fan.
> >
> >
> Not very, or we wouldn't use fans ;-)

You can always move more air with a fan, but sometimes a chimney
moves enough and when it does it is quieter, more reliable and larger than
a fan.

--
Steve O'Hara-Smith | Directable Mirror Arrays
C:\>WIN | A better way to focus the sun
The computer obeys and wins. | licences available see
You lose and Bill collects. | http://www.sohara.org/

"The Natural Philosopher" <tnp@invalid.invalid> wrote in message
news:sbronb$p04$1@dont-email.me...
> On 03/07/2021 22:00, Folderol wrote:
>> On Sat, 03 Jul 2021 14:42:09 +0200
>> A. Dumas <alexandre@dumas.fr.invalid> wrote:
>>> The SoC is a good contact point but there are a few others that get hot,
>>> too. If you don't connect to those and keep the board upside down, you
>>> might trap the heat.
>>
>> Good point! Thanks.
>>
> Ventilation is important: I once designed a small audio equaliser in a
> sealed 19" rack case - it was only about 15 watts total draw, but after 5
> hours continuous usage it packed up - inside temperatures rose to over
> 100°C!
>
> WE drilled a few holes in the case and that fixed it.
>
> Airflow is better than transmission through even metal cases. The problem
> is that with no air movement the air inside becomes an effective thermal
> insulator.
>
> Furthermore *forced* air cooling can multiply heat loss by up to a 5
> times or more.

My Pi4 is on 24/7 because I use it as a PVR for recording TV programmes. The
recordings are written to an external USB HDD (ie not to the SD card "system
disc"). The Pi reports a CPU temperature of around 60 deg C which I
understand is within safe limits. The Pi is in a standard plastic case (as
sold by Raspberry Pi Foundation) with no fan; the case feels warm but not
hot to the touch. The Pi (plus a Pi 3 for logging data from a weather
station) and the external HDD are kept in a small cupboard of a TV cabinet,
which only has a hole cut in the back to take cables. The temperature inside
the TV cabinet is about 27 deg C - so a little warmer than normal room
temperature.

One thing I noticed about the Pi4 compared with the Pi3 is that its CPU
temperature remains a lot more constant; the Pi3's CPU is normally about 40
deg C but rises to about 80 deg C as soon as I run Firefox (so I don't run
FF on it - partly because of the temperature rise and partly because it runs
like a wounded snail on the Pi3).

NY <me@privacy.invalid> wrote:
> The Pi reports a CPU temperature of around 60 deg C which I
> understand is within safe limits.

Easily!

> The Pi is in a standard plastic case (as
> sold by Raspberry Pi Foundation) with no fan;

Still, if you want it lower than 60, get just about *any* other case. Those
"official" cases are the worst. Do not get the fan for that case because
reviews say it whines loudly. These are good & fairly cheap
https://shop.pimoroni.com/products/aluminium-heatsink-case-for-raspberry-pi-4?variant=29430673178707
(also at lots of other vendors). I have one for an always on Pi 4 which is
about 45-48 C.

A. Dumas <alexandre@dumas.fr.invalid> wrote:
> about 45-48 C.

To be clear: at idle.

On Sun, 04 Jul 2021 19:41:11 +0100, Ahem A Rivet's Shot wrote:

> On Sun, 4 Jul 2021 15:54:57 +0100 The Natural Philosopher
> <tnp@invalid.invalid> wrote:
>
>> On 04/07/2021 13:27, Martin Gregorie wrote:
>
>> > Have you any comparable figures for passive 'chimney effect' cooling,
>> > i.e. use a metal contact heat sink on the chip(s) but fit, say, a
>> > 50mm x 400mm cardboard or plastic tube rather than a fan on the
>> > enclosure. A benefit is that chimney effect is quiet and can't go
>> > wrong (unless you count dead moths etc accumlating inside it, but I'm
>> > curious how effective it would be compared with a fan.
>> >
>> >
>> Not very, or we wouldn't use fans ;-)
>
> You can always move more air with a fan, but sometimes a chimney
> moves enough and when it does it is quieter, more reliable and larger
> than a fan.

Yep, pretty much what I've understood, but I was hoping that somebody
would have numbers to quantify the difference between fan and chimney
cooling.

FWIW the only Pi I'm currently running continuously is an old 512MB Pi
2B, mounted with PCB on its side in a hexagonal Salzburger Mozartkugeln
transparent plastic box with its lid on to keep dust out. The box is warm
to the touch, but the Pi isn't showing any temperature or other problems
after several years of operation.

--
Martin | martin at
Gregorie | gregorie dot org

On Sun, 4 Jul 2021 21:53:53 +0100
"NY" <me@privacy.invalid> wrote:

>"The Natural Philosopher" <tnp@invalid.invalid> wrote in message
>news:sbronb$p04$1@dont-email.me...
>> On 03/07/2021 22:00, Folderol wrote:
>>> On Sat, 03 Jul 2021 14:42:09 +0200
>>> A. Dumas <alexandre@dumas.fr.invalid> wrote:
>>>> The SoC is a good contact point but there are a few others that get hot,
>>>> too. If you don't connect to those and keep the board upside down, you
>>>> might trap the heat.
>>>
>>> Good point! Thanks.
>>>
>> Ventilation is important: I once designed a small audio equaliser in a
>> sealed 19" rack case - it was only about 15 watts total draw, but after 5
>> hours continuous usage it packed up - inside temperatures rose to over
>> 100°C!
>>
>> WE drilled a few holes in the case and that fixed it.
>>
>> Airflow is better than transmission through even metal cases. The problem
>> is that with no air movement the air inside becomes an effective thermal
>> insulator.
>>
>> Furthermore *forced* air cooling can multiply heat loss by up to a 5
>> times or more.
>
>My Pi4 is on 24/7 because I use it as a PVR for recording TV programmes. The
>recordings are written to an external USB HDD (ie not to the SD card "system
>disc"). The Pi reports a CPU temperature of around 60 deg C which I
>understand is within safe limits. The Pi is in a standard plastic case (as
>sold by Raspberry Pi Foundation) with no fan; the case feels warm but not
>hot to the touch. The Pi (plus a Pi 3 for logging data from a weather
>station) and the external HDD are kept in a small cupboard of a TV cabinet,
>which only has a hole cut in the back to take cables. The temperature inside
>the TV cabinet is about 27 deg C - so a little warmer than normal room
>temperature.
>
>One thing I noticed about the Pi4 compared with the Pi3 is that its CPU
>temperature remains a lot more constant; the Pi3's CPU is normally about 40
>deg C but rises to about 80 deg C as soon as I run Firefox (so I don't run
>FF on it - partly because of the temperature rise and partly because it runs
>like a wounded snail on the Pi3).
>

All very interesting info folks.

Some numbers here, all deg C:
Ambient temp. 22
Pretty much idle for 2 hours 60
Moderate load for further 30min 68
Max load for further 15 min 73

This in the enclosed metal box, with no other treatment.

What I'll try first is a 12V 30mm fan running at 7V blowing across the Pi
This will be in about a weeks time as I'll off to the new Forest Folk Festival
on Wednesday :))

I'm not after arctic conditions - apart from anything else, cool the surface
too much and you get into the thermal stress realms.

--
W J G

On Sun, 4 Jul 2021 22:41:11 -0000 (UTC)
Martin Gregorie <martin@mydomain.invalid> wrote:

> On Sun, 04 Jul 2021 19:41:11 +0100, Ahem A Rivet's Shot wrote:
>
> > You can always move more air with a fan, but sometimes a chimney
> > moves enough and when it does it is quieter, more reliable and larger
> > than a fan.
>
> Yep, pretty much what I've understood, but I was hoping that somebody
> would have numbers to quantify the difference between fan and chimney
> cooling.

Really hard to do - fans are easy(ish) you can rate their flow in
free air, max static pressure and measure the noise and pretty much compare
them from that. Chimneys are quite sensitive to design and conditions - a
bit like the transfer ports of a two-stroke engine - a small change in
geometry can make a large difference to the airflow.

--
Steve O'Hara-Smith | Directable Mirror Arrays
C:\>WIN | A better way to focus the sun
The computer obeys and wins. | licences available see
You lose and Bill collects. | http://www.sohara.org/

On Mon, 05 Jul 2021 11:14:43 +0100, Ahem A Rivet's Shot wrote:

>
> Really hard to do - fans are easy(ish) you can rate their flow in
> free air, max static pressure and measure the noise and pretty much
> compare them from that. Chimneys are quite sensitive to design and
> conditions - a bit like the transfer ports of a two-stroke engine - a
> small change in geometry can make a large difference to the airflow.
>
Good example: compare output from, say, Shuriken 06, AD06 and Cyclon 06 -
all running on 25% nitro and with the same 7x3.5 prop!

--
Martin | martin at
Gregorie | gregorie dot org