EDAboard.com | EDAboard.de | EDAboard.co.uk | WTWH Media

removing heat with thermal tape and small heatsinks

Ask a question - edaboard.com

elektroda.net NewsGroups Forum Index - Electronics Design - removing heat with thermal tape and small heatsinks

Goto page Previous  1, 2, 3, 4, 5, 6, 7  Next

George Herold
Guest

Tue Feb 05, 2019 1:45 am   



On Monday, February 4, 2019 at 3:05:17 PM UTC-5, John Larkin wrote:
Quote:
On 4 Feb 2019 11:38:17 -0800, Winfield Hill <hill_at_rowland.harvard.edu
wrote:

John Larkin wrote...
Phil Hobbs wrote:

So what are you using exactly? 6 W/m/k is quite decent.

TW-T600-2MM from 3G Shielding. I start with 2mm material
and squash it down to 1mm, which doesn't take a lot of force.

I've tested it and the thermal conductivity really is
close to 6, compressed.

I got some T600 samples last week. I wonder, does the
thermal conductivity have full inverse proportionality
to squashed thickness? They don't address compression
on the datasheet, although thickness is in the formula.

https://www.dropbox.com/s/bt8jrz77m159327/3G_TW_3.JPG?dl=0

Those measurements are obviously pretty rough.

Of course, squashing reduces both specific thermal conductivity and
thickness. That's a double whammy on theta.

The dielectric strength is 6 KV/mm.

It's fun stuff, strongly resembling used chewing gum.
Yuck, how long does it last?


Can I ask what you put on top, and how
you hold it down? More and more I think
a single screw/ clamp is best.

(more screws is asking for production errors.)

George H.
Quote:




--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com


John Larkin
Guest

Tue Feb 05, 2019 2:45 am   



On Mon, 4 Feb 2019 16:35:55 -0800 (PST), George Herold
<gherold_at_teachspin.com> wrote:

Quote:
On Monday, February 4, 2019 at 3:05:17 PM UTC-5, John Larkin wrote:
On 4 Feb 2019 11:38:17 -0800, Winfield Hill <hill_at_rowland.harvard.edu
wrote:

John Larkin wrote...
Phil Hobbs wrote:

So what are you using exactly? 6 W/m/k is quite decent.

TW-T600-2MM from 3G Shielding. I start with 2mm material
and squash it down to 1mm, which doesn't take a lot of force.

I've tested it and the thermal conductivity really is
close to 6, compressed.

I got some T600 samples last week. I wonder, does the
thermal conductivity have full inverse proportionality
to squashed thickness? They don't address compression
on the datasheet, although thickness is in the formula.

https://www.dropbox.com/s/bt8jrz77m159327/3G_TW_3.JPG?dl=0

Those measurements are obviously pretty rough.

Of course, squashing reduces both specific thermal conductivity and
thickness. That's a double whammy on theta.

The dielectric strength is 6 KV/mm.

It's fun stuff, strongly resembling used chewing gum.
Yuck, how long does it last?


Yuck indeed, but I have no idea how long it lasts.

Quote:

Can I ask what you put on top, and how
you hold it down? More and more I think
a single screw/ clamp is best.

(more screws is asking for production errors.)


Here's the water-cooled aluminum baseplate. A PC board screws down
onto that with five 2-56 screws. The board bows a tiny bit, from
compressing the gunk. Lots of copper pours and thermal vias.

https://www.dropbox.com/s/6g80f5yw695e8ah/Baseplate.jpg?dl=0

It looks a little ratty, but that stuff really works.


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

George Herold
Guest

Tue Feb 05, 2019 2:45 am   



On Monday, February 4, 2019 at 7:51:11 PM UTC-5, John Larkin wrote:
Quote:
On Mon, 4 Feb 2019 16:35:55 -0800 (PST), George Herold
gherold_at_teachspin.com> wrote:

On Monday, February 4, 2019 at 3:05:17 PM UTC-5, John Larkin wrote:
On 4 Feb 2019 11:38:17 -0800, Winfield Hill <hill_at_rowland.harvard.edu
wrote:

John Larkin wrote...
Phil Hobbs wrote:

So what are you using exactly? 6 W/m/k is quite decent.

TW-T600-2MM from 3G Shielding. I start with 2mm material
and squash it down to 1mm, which doesn't take a lot of force.

I've tested it and the thermal conductivity really is
close to 6, compressed.

I got some T600 samples last week. I wonder, does the
thermal conductivity have full inverse proportionality
to squashed thickness? They don't address compression
on the datasheet, although thickness is in the formula.

https://www.dropbox.com/s/bt8jrz77m159327/3G_TW_3.JPG?dl=0

Those measurements are obviously pretty rough.

Of course, squashing reduces both specific thermal conductivity and
thickness. That's a double whammy on theta.

The dielectric strength is 6 KV/mm.

It's fun stuff, strongly resembling used chewing gum.
Yuck, how long does it last?

Yuck indeed, but I have no idea how long it lasts.


Can I ask what you put on top, and how
you hold it down? More and more I think
a single screw/ clamp is best.

(more screws is asking for production errors.)

Here's the water-cooled aluminum baseplate. A PC board screws down
onto that with five 2-56 screws. The board bows a tiny bit, from
compressing the gunk. Lots of copper pours and thermal vias.

https://www.dropbox.com/s/6g80f5yw695e8ah/Baseplate.jpg?dl=0

It looks a little ratty, but that stuff really works.

Well back to back stacks of bellville washers, if money is
no object. :^)

George H.
Quote:


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com


George Herold
Guest

Tue Feb 05, 2019 2:45 am   



On Monday, February 4, 2019 at 7:51:11 PM UTC-5, John Larkin wrote:
Quote:
On Mon, 4 Feb 2019 16:35:55 -0800 (PST), George Herold
gherold_at_teachspin.com> wrote:

On Monday, February 4, 2019 at 3:05:17 PM UTC-5, John Larkin wrote:
On 4 Feb 2019 11:38:17 -0800, Winfield Hill <hill_at_rowland.harvard.edu
wrote:

John Larkin wrote...
Phil Hobbs wrote:

So what are you using exactly? 6 W/m/k is quite decent.

TW-T600-2MM from 3G Shielding. I start with 2mm material
and squash it down to 1mm, which doesn't take a lot of force.

I've tested it and the thermal conductivity really is
close to 6, compressed.

I got some T600 samples last week. I wonder, does the
thermal conductivity have full inverse proportionality
to squashed thickness? They don't address compression
on the datasheet, although thickness is in the formula.

https://www.dropbox.com/s/bt8jrz77m159327/3G_TW_3.JPG?dl=0

Those measurements are obviously pretty rough.

Of course, squashing reduces both specific thermal conductivity and
thickness. That's a double whammy on theta.

The dielectric strength is 6 KV/mm.

It's fun stuff, strongly resembling used chewing gum.
Yuck, how long does it last?

Yuck indeed, but I have no idea how long it lasts.


Can I ask what you put on top, and how
you hold it down? More and more I think
a single screw/ clamp is best.

(more screws is asking for production errors.)

Here's the water-cooled aluminum baseplate. A PC board screws down
onto that with five 2-56 screws. The board bows a tiny bit, from
compressing the gunk. Lots of copper pours and thermal vias.

https://www.dropbox.com/s/6g80f5yw695e8ah/Baseplate.jpg?dl=0

It looks a little ratty, but that stuff really works.

OK I have no idea, but at least think about having to replace it
in five years.

George H.
(total thread bend)
I've been using these 20V zeners as noise sources for years (~15)
the latest batch is suddenly a lot quieter. (crap!) I've ordered
100 from other suppliers... or I'll have to redo the circuit...
faster opamp maybe.
GH
Quote:


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com



Guest

Tue Feb 05, 2019 3:45 am   



On Tuesday, February 5, 2019 at 10:21:30 AM UTC+11, DecadentLinux...@decadence.org wrote:
Quote:
Robert Baer <robertbaer_at_localnet.com> wrote in news:KYQ5E.234980
$Ap2.158615_at_fx23.iad:

I presume that you know that "thermally conductive" means "slightly
better than nothing".

You say some of the most stupid things ever. Are you competing with
Larkin for a stupid fuck in sed award?


You haven't been paying attention. Baer is definitely dumber than John Larkin (who occasionally shows signs of intelligence, mostly obscured by his lame attempts to look clever).

--
Bill Sloman, Sydney


Guest

Tue Feb 05, 2019 3:45 am   



On Tuesday, February 5, 2019 at 12:10:47 PM UTC+11, George Herold wrote:
Quote:
On Monday, February 4, 2019 at 7:51:11 PM UTC-5, John Larkin wrote:
On Mon, 4 Feb 2019 16:35:55 -0800 (PST), George Herold
gherold_at_teachspin.com> wrote:

On Monday, February 4, 2019 at 3:05:17 PM UTC-5, John Larkin wrote:
On 4 Feb 2019 11:38:17 -0800, Winfield Hill <hill_at_rowland.harvard.edu
wrote:

John Larkin wrote...
Phil Hobbs wrote:

So what are you using exactly? 6 W/m/k is quite decent.

TW-T600-2MM from 3G Shielding. I start with 2mm material
and squash it down to 1mm, which doesn't take a lot of force.

I've tested it and the thermal conductivity really is
close to 6, compressed.

I got some T600 samples last week. I wonder, does the
thermal conductivity have full inverse proportionality
to squashed thickness? They don't address compression
on the datasheet, although thickness is in the formula.

https://www.dropbox.com/s/bt8jrz77m159327/3G_TW_3.JPG?dl=0

Those measurements are obviously pretty rough.

Of course, squashing reduces both specific thermal conductivity and
thickness. That's a double whammy on theta.

The dielectric strength is 6 KV/mm.

It's fun stuff, strongly resembling used chewing gum.
Yuck, how long does it last?

Yuck indeed, but I have no idea how long it lasts.


Can I ask what you put on top, and how
you hold it down? More and more I think
a single screw/ clamp is best.

(more screws is asking for production errors.)

Here's the water-cooled aluminum baseplate. A PC board screws down
onto that with five 2-56 screws. The board bows a tiny bit, from
compressing the gunk. Lots of copper pours and thermal vias.

https://www.dropbox.com/s/6g80f5yw695e8ah/Baseplate.jpg?dl=0

It looks a little ratty, but that stuff really works.
OK I have no idea, but at least think about having to replace it
in five years.

(total thread bend)
I've been using these 20V zeners as noise sources for years (~15)
the latest batch is suddenly a lot quieter. (crap!) I've ordered
100 from other suppliers... or I'll have to redo the circuit...
faster opamp maybe.


That's odd. They are still going to be avalanche diodes.

Using them as noise sources involves picking an average current which is low enough that the multiplication process fails to sustain avalanche from time to time, so the avalanche gets turned off briefly (and intermittently) before another charge carrier shows up to get it going again.

Playing with the standing current might help. The actual avalanche region is very small, and apparently emits photons as the charge carriers multiply (and these photons can generate new charge carrier pairs) and the manufacturer may have changed the production process enough to mess about with some aspect of this.

Changing supplier might also work.

--
Bill Sloman, Sydney

John Larkin
Guest

Tue Feb 05, 2019 4:45 am   



On Mon, 4 Feb 2019 17:10:42 -0800 (PST), George Herold
<gherold_at_teachspin.com> wrote:

Quote:
On Monday, February 4, 2019 at 7:51:11 PM UTC-5, John Larkin wrote:
On Mon, 4 Feb 2019 16:35:55 -0800 (PST), George Herold
gherold_at_teachspin.com> wrote:

On Monday, February 4, 2019 at 3:05:17 PM UTC-5, John Larkin wrote:
On 4 Feb 2019 11:38:17 -0800, Winfield Hill <hill_at_rowland.harvard.edu
wrote:

John Larkin wrote...
Phil Hobbs wrote:

So what are you using exactly? 6 W/m/k is quite decent.

TW-T600-2MM from 3G Shielding. I start with 2mm material
and squash it down to 1mm, which doesn't take a lot of force.

I've tested it and the thermal conductivity really is
close to 6, compressed.

I got some T600 samples last week. I wonder, does the
thermal conductivity have full inverse proportionality
to squashed thickness? They don't address compression
on the datasheet, although thickness is in the formula.

https://www.dropbox.com/s/bt8jrz77m159327/3G_TW_3.JPG?dl=0

Those measurements are obviously pretty rough.

Of course, squashing reduces both specific thermal conductivity and
thickness. That's a double whammy on theta.

The dielectric strength is 6 KV/mm.

It's fun stuff, strongly resembling used chewing gum.
Yuck, how long does it last?

Yuck indeed, but I have no idea how long it lasts.


Can I ask what you put on top, and how
you hold it down? More and more I think
a single screw/ clamp is best.

(more screws is asking for production errors.)

Here's the water-cooled aluminum baseplate. A PC board screws down
onto that with five 2-56 screws. The board bows a tiny bit, from
compressing the gunk. Lots of copper pours and thermal vias.

https://www.dropbox.com/s/6g80f5yw695e8ah/Baseplate.jpg?dl=0

It looks a little ratty, but that stuff really works.
OK I have no idea, but at least think about having to replace it
in five years.

George H.
(total thread bend)
I've been using these 20V zeners as noise sources for years (~15)
the latest batch is suddenly a lot quieter. (crap!) I've ordered
100 from other suppliers... or I'll have to redo the circuit...
faster opamp maybe.
GH


Maybe they changed the junction area. Some people sell the same part
as 1/4 watt, 1/2 watt, 1 watt zeners.


Just like 1N4001...1N4007. Most people only really make two diodes to
cover the whole series.




--

John Larkin Highland Technology, Inc

lunatic fringe electronics

John Larkin
Guest

Tue Feb 05, 2019 4:45 am   



On Mon, 4 Feb 2019 17:14:35 -0800 (PST), George Herold
<gherold_at_teachspin.com> wrote:

Quote:
On Monday, February 4, 2019 at 7:51:11 PM UTC-5, John Larkin wrote:
On Mon, 4 Feb 2019 16:35:55 -0800 (PST), George Herold
gherold_at_teachspin.com> wrote:

On Monday, February 4, 2019 at 3:05:17 PM UTC-5, John Larkin wrote:
On 4 Feb 2019 11:38:17 -0800, Winfield Hill <hill_at_rowland.harvard.edu
wrote:

John Larkin wrote...
Phil Hobbs wrote:

So what are you using exactly? 6 W/m/k is quite decent.

TW-T600-2MM from 3G Shielding. I start with 2mm material
and squash it down to 1mm, which doesn't take a lot of force.

I've tested it and the thermal conductivity really is
close to 6, compressed.

I got some T600 samples last week. I wonder, does the
thermal conductivity have full inverse proportionality
to squashed thickness? They don't address compression
on the datasheet, although thickness is in the formula.

https://www.dropbox.com/s/bt8jrz77m159327/3G_TW_3.JPG?dl=0

Those measurements are obviously pretty rough.

Of course, squashing reduces both specific thermal conductivity and
thickness. That's a double whammy on theta.

The dielectric strength is 6 KV/mm.

It's fun stuff, strongly resembling used chewing gum.
Yuck, how long does it last?

Yuck indeed, but I have no idea how long it lasts.


Can I ask what you put on top, and how
you hold it down? More and more I think
a single screw/ clamp is best.

(more screws is asking for production errors.)

Here's the water-cooled aluminum baseplate. A PC board screws down
onto that with five 2-56 screws. The board bows a tiny bit, from
compressing the gunk. Lots of copper pours and thermal vias.

https://www.dropbox.com/s/6g80f5yw695e8ah/Baseplate.jpg?dl=0

It looks a little ratty, but that stuff really works.
Well back to back stacks of bellville washers, if money is
no object. :^)


The pad stuff is plenty springy.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

John Larkin
Guest

Tue Feb 05, 2019 4:45 am   



On Mon, 4 Feb 2019 23:30:25 +0000 (UTC),
DecadentLinuxUserNumeroUno_at_decadence.org wrote:

Quote:
Winfield Hill <hill_at_rowland.harvard.edu> wrote in
news:q3a4b9032a_at_drn.newsguy.com:

John Larkin wrote...
Phil Hobbs wrote:

So what are you using exactly? 6 W/m/k is quite decent.

TW-T600-2MM from 3G Shielding. I start with 2mm material
and squash it down to 1mm, which doesn't take a lot of force.

I've tested it and the thermal conductivity really is
close to 6, compressed.

I got some T600 samples last week. I wonder, does the
thermal conductivity have full inverse proportionality
to squashed thickness? They don't address compression
on the datasheet, although thickness is in the formula.



Yes, the cross-sectional thickness of the compressed area of the
pad wituated over the element to be sinked matters. Thinner is
better.

The stuff is usually very expensive though, and we had to have
ours dies cut for uniformity in mil spec device mfgr. They came
with a thick foil sheet bonded to one side to interface with the
product enclosure.


A sheet of T600 costs about $1 per square inch. I'm cutting it
manually with an x-acto, but we might have it die cut if product
volume ever picks up.

I don't waste any, just patch together any loose bits to cover the
area in need.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

John Larkin
Guest

Tue Feb 05, 2019 4:45 am   



On Mon, 4 Feb 2019 22:45:22 +0000 (UTC),
DecadentLinuxUserNumeroUno_at_decadence.org wrote:

Quote:
John Larkin <jjlarkin_at_highlandtechnology.com> wrote in
news:f8ef5e53v447p1kqj5v63fc2tjfu50de1q_at_4ax.com:

One square inch if the stuff that I use, compressed to 1 mm thick, is
0.25 K/W.


We had gap pads that were the size of a 3.5 inch hard drive and they
were a quarter inch thick. The compressed cross-sectional thickness
over the areas to be sinked is the most critical because the stuff
simply is not as good as hard, intimate epoxied matings.

The level of heat being produced in his application in not something
a mushy gap pad is going to be comfortable with, even after you
compress it down to a mm thikness.


He's got about 5 watts and maybe 15 square inches of PCB. The power
density is tiny. A few postage-stamp sized pieces of T600 should be
plenty.






--

John Larkin Highland Technology, Inc

lunatic fringe electronics

John Larkin
Guest

Tue Feb 05, 2019 4:45 am   



On Mon, 4 Feb 2019 16:35:55 -0800 (PST), George Herold
<gherold_at_teachspin.com> wrote:

Quote:
On Monday, February 4, 2019 at 3:05:17 PM UTC-5, John Larkin wrote:
On 4 Feb 2019 11:38:17 -0800, Winfield Hill <hill_at_rowland.harvard.edu
wrote:

John Larkin wrote...
Phil Hobbs wrote:

So what are you using exactly? 6 W/m/k is quite decent.

TW-T600-2MM from 3G Shielding. I start with 2mm material
and squash it down to 1mm, which doesn't take a lot of force.

I've tested it and the thermal conductivity really is
close to 6, compressed.

I got some T600 samples last week. I wonder, does the
thermal conductivity have full inverse proportionality
to squashed thickness? They don't address compression
on the datasheet, although thickness is in the formula.

https://www.dropbox.com/s/bt8jrz77m159327/3G_TW_3.JPG?dl=0

Those measurements are obviously pretty rough.

Of course, squashing reduces both specific thermal conductivity and
thickness. That's a double whammy on theta.

The dielectric strength is 6 KV/mm.

It's fun stuff, strongly resembling used chewing gum.
Yuck, how long does it last?


Most of the flavor is gone in a half hour.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

John Larkin
Guest

Tue Feb 05, 2019 4:45 am   



On Mon, 4 Feb 2019 22:52:55 +0000 (UTC),
DecadentLinuxUserNumeroUno_at_decadence.org wrote:

Quote:
John Larkin <jjlarkin_at_highlandtechnology.com> wrote in
news:pmef5ep7i43lkn4lt0qfqic2h9cel11ori_at_4ax.com:

But a PCB has traces and vias and connector pins. Win's board also
has surface-mount parts on the bottom side. A gap-pad conforms to
all that.


You failed to note the level of heat his parts are producing. It is
not that the gap pad would not pull heat away. The problem is that it
would not/will not/can not pull it off fast enough for that
application.


He's well below 0.5 watts per square inch. A square inch of 1 mm T600
would be about 0.25 K/w. So we'd lose about 0.12 degrees C across the
gap pad.

OK, grumble, some hot spots might rise a full degree C.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics


Guest

Tue Feb 05, 2019 5:45 am   



John Larkin <jjlarkin_at_highlandtechnology.com> wrote in
news:hn1i5ehopk7lmckon0j6lml8vgbrg254cj_at_4ax.com:

Quote:
On Mon, 4 Feb 2019 23:00:21 +0000 (UTC),
DecadentLinuxUserNumeroUno_at_decadence.org wrote:

John Larkin <jjlarkin_at_highlandtechnology.com> wrote in
news:u1ff5ehhiqrsnclplf9r0evc2i676ugghq_at_4ax.com:

Hard anodize is reliable up to about 200 volts. It's a pretty
good thermal conductor (because it's thin) but also adds a lot
of capacitance.



You are an idiot, and obviously know nothing or even less than
nothing
about hard anodized Aluminum properties.

OK, supply us some facts. Numbers, please.





500 volts per mil.

John Larkin
Guest

Tue Feb 05, 2019 5:45 am   



On Mon, 4 Feb 2019 23:00:21 +0000 (UTC),
DecadentLinuxUserNumeroUno_at_decadence.org wrote:

Quote:
John Larkin <jjlarkin_at_highlandtechnology.com> wrote in
news:u1ff5ehhiqrsnclplf9r0evc2i676ugghq_at_4ax.com:

Hard anodize is reliable up to about 200 volts. It's a pretty good
thermal conductor (because it's thin) but also adds a lot of
capacitance.



You are an idiot, and obviously know nothing or even less than nothing
about hard anodized Aluminum properties.


OK, supply us some facts. Numbers, please.




--

John Larkin Highland Technology, Inc

lunatic fringe electronics


Guest

Tue Feb 05, 2019 5:45 am   



John Larkin <jjlarkin_at_highlandtechnology.com> wrote in
news:201i5e171cnkt4je5nfve6obg8oq02gjc9_at_4ax.com:

Quote:
A sheet of T600 costs about $1 per square inch. I'm cutting it
manually with an x-acto, but we might have it die cut if product
volume ever picks up.

I don't waste any, just patch together any loose bits to cover the
area in need.



Our stuff was a 1/4 inch thick. Not something a mil contractor hacks
at with an exacto or a more professional scalpel, which I am NOT
surprised you are also in the dark about. Scalpels are far better for
precise industrial carving of things.

Goto page Previous  1, 2, 3, 4, 5, 6, 7  Next

elektroda.net NewsGroups Forum Index - Electronics Design - removing heat with thermal tape and small heatsinks

Ask a question - edaboard.com

Arabic version Bulgarian version Catalan version Czech version Danish version German version Greek version English version Spanish version Finnish version French version Hindi version Croatian version Indonesian version Italian version Hebrew version Japanese version Korean version Lithuanian version Latvian version Dutch version Norwegian version Polish version Portuguese version Romanian version Russian version Slovak version Slovenian version Serbian version Swedish version Tagalog version Ukrainian version Vietnamese version Chinese version Turkish version
EDAboard.com map