Home made heat sink pads

[John]

Is it possible to throw together some home made heat sink pads?

After reading your posts in this thread, I conclude that toilet paper would
serve your purpose.

 

[Eeyore]

That's not *sand paper* though is it ?

I personally like 'Scotch Brite' style material for removing burrs and the like
btw.

Graham

 

[John Larkin]

What's also important is to get it flat, as opposed to smooth. Sanding
may make it shiny, but it's also liable to grind a curve into a
surface, and that will wreck the thermal contact and in the extreme
result in insulator damage. Extrusions tend to be too curvy/wavy for
really good thermal contact. For really good thermal performance,
extrusions should be machined flat, like with a slow pass with an end
mill or a fly cutter.

Grease will fill surface roughness pretty well, but curvature can open
up a gap that will seriously increase theta. Minor things like surface
waviness can throw away a good fraction of a transistor's dissipation
capability. A reasonable goal is 100 micro-inches total roughness plus
flatness across the transistor footprint. You can buy BeO/AlO2/AlN
insulators that are smooth and flat to a few micro-inches.

Power transistors themselves seem to be remarkably flat.

This sort of stuff starts to matter when you want to push transistors
up against their thermal limits, like in fast/RF stuff where you can't
just add more parts in parallel.

John

 

[Rich Grise]

HAHA. What a looser. Someone disagree's with you and you get mad. Just
what I expect from an "expert".

No, you're being a dork, and people are noticing, and starting to
ignore you.

Like I said before, you might eventually find someone who will tell you
what you want to hear, but probably not as long as you're being such an
insufferable, petulant, whining brat.

Good Luck!
Rich

 

[Rich Grise]

The regular Dow Corning thermal grease must have some very small fill
particles. With just modest pressure, I've measured its added spacing at
below 100 microinches (0.1 mils), which is the limit of the micrometer I
had at hand. It flows out of the gap. Right, it's not a very good thermal
conductor, but it does squish out very thin... a sil-pad may be 50-100
times as thick!

Grease over flat, 0.5 mil hard anodize is about as good as it gets.

I've still got a tube of DC-340 (or maybe 320 - I don't remember exactly)
that I lifted when I was in the USAF in the 1970's; I'll swear by the
stuff. Whenever I replace a MB/CPU/heatsink I wipe off the goo that comes
with the heatsink and use my good ol' trusty Dow Corning, that's how much
I trust it.

I wonder if it would make good sunblock? ;-)

Cheers!
Rich

 

[Rich Grise]

In other applications mica is used as a thermal *insulator* !!

Do you have any examples of this? I've seen mica used as an _electrical_
insulator in a high-temp environment, like in toasters and hair dryers
and stuff, but I doubt its intent is thermal insulation there.

Thanks!
Rich

 

[default]

After reading your posts in this thread, I conclude that toilet paper would
serve your purpose.

As long as there isn't a metal burr to puncture the paper and the
device isn't tightened to the point where it can make a compression
cut it should work.

Some old mica windows from kerosene heaters can be cut to make
insulating pads - mica conducts heat well (still need silicone grease
because it isn't perfectly flat).

 

[John Fields]

Quite !



Since when did sanding produce a smooth surface Mr Blow Hard ?

 

[John Fields]

Trying to show your mettle, eh?

(Let me re-phrase that)

Trying to show your metal, eh?

Trying to show your meddle, eh?!


Heheheh :-]

 

[Mike Monett]

Cut a piece off a mylar potato chip bag. Mylar doesn't melt easy.

You're only using 12 volts, so super-thin chip bag mylar is
perfect.

Assuming it really is mylar. A commonly used dielectric. I have
some mylar sheet I use for insulators, but it's thicker than chip
bags. I can tell you, that stuff doesn't melt easy.

I just dissected a Lays Classic 250gm potato chip bag. It measured 2
mil thick and was composed of several layers.

The two outside layers appeared to be mylar, approximately 500
microinches thick. There was a thin layer of paint, a super-thin
layer of metallization, and the rest was a layer of soft
polyethylene that appeared to be 1 mil thick. This corresponds to a
description of barrier layers in food packaging needed to keep out
oxygen:

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Barrier Films - An Overview

Single layer films are generally quite permeable to most gases.

Barrier films are usually multi-layer films that have been designed
to be impervious to gas migration. Some common examples of barrier
films will be found in any grocery store. Bags used for packaging
potato chips and meat products are gas barrier bags, garbage,
freezer, and sandwich bags are not. To understand the difference in
practical terms, compare the longevity of a helium-filled balloon
made of rubber to a balloon made of foil barrier - the rubber
balloon will shrink within hours, but the barrier film balloon will
continue to wish you a "happy birthday", or "get well soon" in
mid-air for weeks.

Construction

Almost all gas barrier films are made of multiple layers. The
outer/top layer will be a strong (usually transparent) plastic with
a relatively high melting point, often made of polyester (Mylar) or
polypropylene. The inner/middle layer will be the gas barrier
material (see next). The inside/bottom layer is almost always low
density polyethylene, which softens at relatively low temperatures.

Barrier layer materials

Most of the cross section of a barrier film is occupied by its inner
(polyethylene) sealing layer. The outer support layer is often
surprisingly thin, and the middle (barrier) layer can sometimes be
of microscopic thickness. (Note that it is the compostion of the
barrier material, and not its thickness that is most important in
determining its permeability).

Metal foil (usually aluminum) is generally the most effective gas
barrier material, closely followed closely by transparent ceramic
barriers (Escal). Sputtered Aluminum, EVOH, barrier nylon, and saran
are transparent plastic barrier layers often used in the food
industry. The barrier layer of Aclar film is similar in composition
to Teflon. Each specific barrier material and formulation has it's
own unique barrier properties.

http://www.keepsafe.ca/barrier.shtml

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The polyethylene layer may mean food wrapping is a poor choice for a
heat sink insulator since it melts at such a low temperature.
However, Christmas wrapping may be more robust. I'm on my way to
WalMart to see if they still have any on sale

Regards,

Mike Monett

Antiviral, Antibacterial Silver Solution:
http://silversol.freewebpage.org/index.htm
SPICE Analysis of Crystal Oscillators:
http://silversol.freewebpage.org/spice/xtal/clapp.htm
Noise-Rejecting Wideband Sampler:
http://www3.sympatico.ca/add.automation/sampler/intro.htm

 

[John Fields]

Have you looked at aluminum nitride? It's almost as good as BeO but
it's not toxic. For high-voltage and high thermal density, I'd go for
a non-insulated copper heat spreader over an AlN insulator, all
greased... if you can stand the capacitance!

I think BeO has the best ratio of thermal conductivity to dielectric
constant, except for diamond. Or flowing distilled water.

---
Flowing distilled water is kinda cheating because water has a high
heat capacity, so even though it's got lousy thermal conductivity,
if it's moving, in a heat exchanger, the area available for heat
transfer goes _way_ up.


Kt Kt/Kd
W/mk Kd (mean)
--------+--------+--------+--------+
H2O (di)| 0.612 | 80 | 0.008 |
--------+--------+--------+--------+
Mica | 0.7 | 4 - 9 | 0.108 |
--------+--------+--------+--------+
BeO | 330 | 6.7 | 49.25 |
--------+--------+--------+--------+
Diamond | 2500 |5.5 - 10| 76.92 |
--------+--------+--------+--------+

 

[John Fields]

That's not *sand paper* though is it ?

 

[[email protected]]

Have you looked at aluminum nitride? It's almost as good as BeO but
it's not toxic. For high-voltage and high thermal density, I'd go for
a non-insulated copper heat spreader over an AlN insulator, all
greased... if you can stand the capacitance!

Hey John,
these guys (http://hypertextbook.com/physics/thermal/conduction/)
report the conductivity of properly oriented graphite as about twice as
good as diamond. It sure would be fun to find a way to take advantage
of that.(*)

(*) Wrongly-oriented graphite, though, is terrible.

I think BeO has the best ratio of thermal conductivity to dielectric
constant, except for diamond. Or flowing distilled water.

The dielectric constant of water's pretty high though, right?

I also wonder about:
a) electroplaning/polishing, for ultra-ultra-flat, smooth surfaces
and
b) microplasmic or other anodic coating (denser than ordinary
anodize)

Cheers,
James Arthur

 

[Eeyore]

Hey John,
these guys (http://hypertextbook.com/physics/thermal/conduction/)
report the conductivity of properly oriented graphite as about twice as
good as diamond. It sure would be fun to find a way to take advantage
of that.(*)

(*) Wrongly-oriented graphite, though, is terrible.

Read Motorola's AN1040.

The best heat sink pad is Grafoil.

Graham

 

[Eeyore]

Do you have any examples of this? I've seen mica used as an _electrical_
insulator in a high-temp environment, like in toasters and hair dryers
and stuff, but I doubt its intent is thermal insulation there.

Most thermal insulators are good electrical insulators and vice versa. The
ceramics are an interesting example of the opposite but have highish K.

Fine mica has been used in blown loft insulation for example.

Graham

 

[Barry Lennox]

As long as there isn't a metal burr to puncture the paper and the
device isn't tightened to the point where it can make a compression
cut it should work.

Some old mica windows from kerosene heaters can be cut to make
insulating pads - mica conducts heat well (still need silicone grease
because it isn't perfectly flat).

I came across a pile of old toasters (well, about 5) last year, and
recycled them into good bits and junk. The mica sheets in there are
perfect for thermally-conductive, insulating material. Why even to buy
a new one at WalMart for $7 gives a lot of mica. With only 12 volts,
the mica can be very thin, and with a bit of practice with a razor
blade, it can be cleaved into very thin sheets.

 

[Mike Monett]

The polyethylene layer may mean food wrapping is a poor choice for
a heat sink insulator since it melts at such a low temperature.
However, Christmas wrapping may be more robust. I'm on my way to
WalMart to see if they still have any on sale

The holiday stuff is gone, but they have rolls of 1 mil X 30" X 144"
(0.0254mm X 76cm X 3.6m) polypropylene to gift wrap baskets, $1.97 +
tax. This might work fairly well. Here's some properties from
various sites:

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Electrical Properties

Dielectric constant @1MHz : 2.2-2.6
Dissipation factor @1MHz : 0.0003 - 0.0005
Surface resistivity : 10^13 Ohm/sq
Volume resistivity : 10^16-10^18 Ohm/cm

Thermal Properties

Coefficient of thermal expansion : 100-180 (x10-6 K-1)
Heat-deflection temperature - 0.45MPa : 100-105C
Heat-deflection temperature - 1.8MPa : 60-65C
Lower working temperature : -10 to -60C
Specific heat : 1700 - 1900 (J K-1 kg-1)
Thermal conductivity @23C : 0.1-0.22 W/m-1 K-1
Upper working temperature : 90-120C

Properties for Polypropylene Film

Dielectric Strength @25um thick : 200 kV/mm (5 kV/mil)
Elongation at Break % : 50-1000
Heat-sealing Temperature : 140-205C
Initial Tear Strength : 18-27 g um-1

http://www.goodfellow.com/csp/active/STATIC/E/Polypropylene.HTML

Melting point Tm : 160-168C
Vicat softening point Tv : 146-150C
Heat deflection temperature THDT : 50-60C
max. continuous service temperature air Tml : 90-120C

http://www.segerfrojd.se/thermal_properties.htm

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Not too bad. 5kV breakdown, usable to 100C. The 1 mil thickness
helps to lower the thermal resistance. I'll give it a try and see
what happens.

Regards,

Mike Monett

Antiviral, Antibacterial Silver Solution:
http://silversol.freewebpage.org/index.htm
SPICE Analysis of Crystal Oscillators:
http://silversol.freewebpage.org/spice/xtal/clapp.htm
Noise-Rejecting Wideband Sampler:
http://www3.sympatico.ca/add.automation/sampler/intro.htm

 

[Eeyore]

I've still got a tube of DC-340 (or maybe 320 - I don't remember exactly)
that I lifted when I was in the USAF in the 1970's; I'll swear by the
stuff. Whenever I replace a MB/CPU/heatsink I wipe off the goo that comes
with the heatsink and use my good ol' trusty Dow Corning, that's how much
I trust it.

The silver loaded greases that come with some CPU heatsinks are worth hanging on
to. There's precious little that can compete with that for sure and you only
need a little dab of it.

Graham

 

[[email protected]]

Read Motorola's AN1040.

http://www.onsemi.com/pub/Collateral/AN1040-D.PDF. Thanks.

Tidbits:

FINISH: They say, when mating to a TO-3 with a 32u-inch finish,
going from 16 to 64 u-inch surface finish on the heat sink made less
than +/- 3% difference.

FLATNESS: really helps, e.g. Fig. 7, which shows 0.3 C/W for 0.002"
flatness over a TO-3 header, but twice that for 0.005" flatness.

Thermally-conductive adhesives can provide a 0.3 C/W interface
through a 0.01" thickness.

The best heat sink pad is Grafoil.

Aavid's Kon-Dux is mentioned (embedded graphite flakes?), but not
tested.

Noted: a) It's not always clear which data AN-1040 collects from
heatsink vendors (hence possibly suspect info), and which represents
original measurements. b) Appears dated, so might not include
significant advances, e.g. in silver-loaded epoxies.

Best,
James Arthur

 

[John Larkin]

Hey John,
these guys (http://hypertextbook.com/physics/thermal/conduction/)

The value for mylar can't be right!

And this is nonsense:

"For many applications a base that was made only of copper would
probably be too effective at transferring heat. Heat applied to a
small region would be transferred so rapidly that it wouldn't have
time to diffuse across the pot's bottom. This would result in uneven
cooking and possibly even local areas of scorching. Capping the base
with stainless steel slows the immediate rate of heat transfer from
the burner or heating element, but once this heat enters the copper
core its high conductivity would spread the heat rapidly and evenly to
all parts of the base."

report the conductivity of properly oriented graphite as about twice as
good as diamond. It sure would be fun to find a way to take advantage
of that.(*)

(*) Wrongly-oriented graphite, though, is terrible.

If it's not an insulator, using it has to hurt, no matter how high the
thermal conductivity. If it inhibits lateral heat spreading, as the
graphite will, it could make things a lot worse.

The dielectric constant of water's pretty high though, right?

At low frequencies, yes. But 1 LPM of water flow has an effective
theta of about 0.014 K/W. You can run 1 LPM through some long, skinny
plastic tubes and have a tiny effective capacitance (and kilovolts of
isolation, if you keep the water clean.)

I also wonder about:
a) electroplaning/polishing, for ultra-ultra-flat, smooth surfaces
and

Machining is probably good enough. The baseplate of a power transistor
is, what, 50 mils thick maybe, so if you machine the heatsink to 50
uinch flatness, that's 1000:1.

b) microplasmic or other anodic coating (denser than ordinary
anodize)

Hard anodize is pretty good, up to a couple of hundred volts maybe. We
really need a mil of diamond. It's sad that nature is so stingy with
diamonds.

John