Plastic screws

[whit3rd]

On Wednesday, July 17, 2019 at 6:25:17 AM UTC-7, Phil Hobbs wrote:

... I need to cool a multi-pixel photon counter (MPPC)
down to about -10C.

I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws. Switching from #2
stainless to #6 nylon reduces the heat leak by 20x. However, nylon
isn't good in a vacuum, because it outgasses,

So, maybe fiberglass-reinforced plastic rod/tube with Tinnerman nuts to hold it
together? There's G10 fiberglass materials with adequate vacuum
performance, and good strength.
Even a simple hoop of fiberglass sheet material, tightened with wedges,
could hold your surfaces together. Can't really cut screw threads
in this kind of layered material and keep any strength, though.

 

[Joseph Gwinn]

On Jul 17, 2019, Phil Hobbs wrote
(in article<94ednTJ5EJEHtbLAnZ2dnUU7-QnNnZ2d@supernews.com&gt:

On 7/17/19 9:41 AM, Joseph Gwinn wrote:
On Jul 17, 2019, Phil Hobbs wrote
(in article<RKudnVUn6NouvLLAnZ2dnUU7-VfNnZ2d@supernews.com&gt:

For that SEM cathodoluminescence gizmo I mentioned in the distributed

attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C. It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design. Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws. Switching from #2
stainless to #6 nylon reduces the heat leak by 20x. However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart. (At least it's reasonably cheap--a nickel a screw or thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just forget
Vespel or any of the other plastics you really want.
I think you will have to machine the non-nylon hardware. Delrin is a good
choice, if it´s OK in vacuum. Note that unlike most plastics, one can
press-fit delrin, so threads may not be needed.

For vacuum, I´d go to ceramic spacers with stainless steel hardware keeping
them in compression.

Or, if thermal insulation is the only reqt, a stack of shiny steel plates
with staggered metal separators (which can be spot-welded), so the leakage
paths are very long, and radiation coupling between plates dominates (there
been no air to support convection).

Joe Gwinn

Ceramic is a surprisingly decent thermal conductor,
unfortunately--better than glass, and much better than plastic. SEMs
don't run in UHV, so FR-4 is okay. At lower delta-T you can just have
the board hang ten over the TEC and use the FR-4 for insulation.

How much vacuum does a SEM need? I was thinking bake-out being needed.

If FR-4 is good enough for that vacuum, then use FR-4 stepped washers, steel
hardware, and belleville spring washers to maintain tension despite thermal
variation.

Or use steel washers, and stagger them to make the thermal path long and
windy.

Note that delrin is very slippery, and so delrin nut and bolt will tend to
unscrew itself. A SS nut forced to thread onto a slightly oversize delrin rod
should stay put. As will a press fit into an undersize hole.

The speed nuts are called Tinnerman nuts, as someone else also mentioned.
Tinnerman invented this kind of nut.

Joe Gwinn

 

[Phil Hobbs]

On 7/17/19 10:00 AM, John Larkin wrote:

On Wed, 17 Jul 2019 09:25:04 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C. It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design. Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws. Switching from #2
stainless to #6 nylon reduces the heat leak by 20x. However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart. (At least it's reasonably cheap--a nickel a screw or thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just forget
Vespel or any of the other plastics you really want.

Anybody got a source of plastic screws other than nylon?

Cheers

Phil Hobbs

Got a sketch of the geometry?

It's just a thin, thermal-via-infested PCB of about 15 mm tall x 20 mm
wide.

It'll sit on an aluminum spreader, which sits on an 9 x 11 mm TEC
(Marlow NL1025T).


The issue is how to maintain a compressive preload on the TEC that will
help prevent damage due to shock or fat fingers.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[John Larkin]

On Wed, 17 Jul 2019 21:54:34 -0400, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

On 7/17/19 10:00 AM, John Larkin wrote:
On Wed, 17 Jul 2019 09:25:04 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C. It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design. Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws. Switching from #2
stainless to #6 nylon reduces the heat leak by 20x. However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart. (At least it's reasonably cheap--a nickel a screw or thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just forget
Vespel or any of the other plastics you really want.

Anybody got a source of plastic screws other than nylon?

Cheers

Phil Hobbs

Got a sketch of the geometry?



It's just a thin, thermal-via-infested PCB of about 15 mm tall x 20 mm
wide.

It'll sit on an aluminum spreader, which sits on an 9 x 11 mm TEC
(Marlow NL1025T).


The issue is how to maintain a compressive preload on the TEC that will
help prevent damage due to shock or fat fingers.

Cheers

Phil Hobbs

How about plastic screws and either coil springs or o-rings to control
the squeezee force?


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[Jasen Betts]

On 2019-07-17, George Herold <gherold@teachspin.com> wrote:

On Wednesday, July 17, 2019 at 9:25:17 AM UTC-4, Phil Hobbs wrote:
For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C. It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design. Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.
Well a bunch of silly ideas...
The Delrin sounds OK, maybe a machine shop can cut a thread on the end
of a rod for you. (There are a few places on the web listing delrin
screws, but no prices.)

How about soldering/ or epoxy and then squeeze it some other way.
(Fishing line/ zip ties, something else?)

Fishing line and zip-ties are nylon, which is no good, by maybe polyester
or linen thread would work

You need my beer can cooler on the hot plate :^)
Has anyone done a distributed two stage TEC, Some small TEC in
vacuum and then a bigger one outside the vacuum going to a
heatsink?

It would seem easier to just use a heat-pipe (if you can find a suitable
fluid) and put all the TEC in atmosphere.

--
When I tried casting out nines I made a hash of it.

 

[Chris Jones]

On 17/07/2019 23:25, Phil Hobbs wrote:

For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C.  It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design.  Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws.  Switching from #2
stainless to #6 nylon reduces the heat leak by 20x.  However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart.  (At least it's reasonably cheap--a nickel a screw or thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just forget
Vespel or any of the other plastics you really want.

Anybody got a source of plastic screws other than nylon?

Cheers

Phil Hobbs

Crazy idea:
sandwich the TEC between big NdFeB magnets. Lots of compressive force
and no heat leak. OTOH I find those magnets sometimes like to
disintegrate to powder after a few years. I guess it might be due to
poor plating quality. I presume that for the better suppliers it is a
solved problem or they wouldn't use them in hard drives.

You could put just a piece of iron on the side of the TEC facing inside
the vacuum, and a big magnet on the outside of the vacuum chamber.

 

[Chris Jones]

On 18/07/2019 16:26, Jasen Betts wrote:

On 2019-07-17, George Herold <gherold@teachspin.com> wrote:
On Wednesday, July 17, 2019 at 9:25:17 AM UTC-4, Phil Hobbs wrote:
For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C. It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design. Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.
Well a bunch of silly ideas...
The Delrin sounds OK, maybe a machine shop can cut a thread on the end
of a rod for you. (There are a few places on the web listing delrin
screws, but no prices.)

How about soldering/ or epoxy and then squeeze it some other way.
(Fishing line/ zip ties, something else?)

Fishing line and zip-ties are nylon, which is no good, by maybe polyester
or linen thread would work

You need my beer can cooler on the hot plate :^)
Has anyone done a distributed two stage TEC, Some small TEC in
vacuum and then a bigger one outside the vacuum going to a
heatsink?

It would seem easier to just use a heat-pipe (if you can find a suitable
fluid) and put all the TEC in atmosphere.

At least with the TEC in a good vacuum you'd get no moisture condensing
on/inside it.

I was wondering if one can buy peltiers with the edges hermetically
sealed (not with silicone rubber) and a vacuum inside. I'm guessing that
sealing the edges with thin stainless foil brazed to the ceramic might
work, and that sealing the edges with just frit and ceramic might be not
flexible enough to withstand differential thermal expansion of the two
ceramic plates.

 

[Bill Sloman]

On Thursday, July 18, 2019 at 2:05:59 PM UTC+2, Chris Jones wrote:

On 17/07/2019 23:25, Phil Hobbs wrote:
For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C.  It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design.  Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws.  Switching from #2
stainless to #6 nylon reduces the heat leak by 20x.  However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart.  (At least it's reasonably cheap--a nickel a screw or thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just forget
Vespel or any of the other plastics you really want.

Anybody got a source of plastic screws other than nylon?

Cheers

Phil Hobbs


Crazy idea:
sandwich the TEC between big NdFeB magnets. Lots of compressive force
and no heat leak. OTOH I find those magnets sometimes like to
disintegrate to powder after a few years. I guess it might be due to
poor plating quality. I presume that for the better suppliers it is a
solved problem or they wouldn't use them in hard drives.

You could put just a piece of iron on the side of the TEC facing inside
the vacuum, and a big magnet on the outside of the vacuum chamber.

SEMs use magnetic fields to focus (and scan) the electron beam.

Two big NdFeB magnets on either side of a TEC won't have a lot if external field, and - since it would be a dipole field, it would decay rapidly with distance - so it might be worth a try, but I wouldn't be too optimistic.

--
Bill Sloman, Sydney

 

[dcaster@krl.org]

On Thursday, July 18, 2019 at 8:55:35 AM UTC-4, George Herold wrote:

How about soldering/ or epoxy and then squeeze it some other way.
(Fishing line/ zip ties, something else?)

Fishing line and zip-ties are nylon, which is no good, by maybe polyester
or linen thread would work
One could use spiderwire, which is carbon fiber?
(but silly ideas)

George H.

Some fishing line is uhmw pe.

Dan

 

[dcaster@krl.org]

On Wednesday, July 17, 2019 at 11:03:16 AM UTC-4, Martin Brown wrote:

Or for a one off buy some PEEK rod and thread it yourself.

https://www.directplastics.co.uk/peek-rod

--
Regards,
Martin Brown

For a one off, I could probably get what ever plastic you want to try and try using a die to cut threads on it.

I am in Delaware and there is a MSC and Grainger outlet near by so no shipping charge and no sales tax.

A friend has machine shop and we could try making a injection mold. He has a job machining PEEK so lots of free non virgin PEEK available.

Dan

 

[George Herold]

On Thursday, July 18, 2019 at 2:31:12 AM UTC-4, Jasen Betts wrote:

On 2019-07-17, George Herold <gherold@teachspin.com> wrote:
On Wednesday, July 17, 2019 at 9:25:17 AM UTC-4, Phil Hobbs wrote:
For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C. It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design. Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.
Well a bunch of silly ideas...
The Delrin sounds OK, maybe a machine shop can cut a thread on the end
of a rod for you. (There are a few places on the web listing delrin
screws, but no prices.)

How about soldering/ or epoxy and then squeeze it some other way.
(Fishing line/ zip ties, something else?)

Fishing line and zip-ties are nylon, which is no good, by maybe polyester
or linen thread would work

One could use spiderwire, which is carbon fiber?
(but silly ideas)

George H.

You need my beer can cooler on the hot plate :^)
Has anyone done a distributed two stage TEC, Some small TEC in
vacuum and then a bigger one outside the vacuum going to a
heatsink?

It would seem easier to just use a heat-pipe (if you can find a suitable
fluid) and put all the TEC in atmosphere.

--
When I tried casting out nines I made a hash of it.

 

[Phil Hobbs]

On 7/18/19 8:05 AM, Chris Jones wrote:

On 17/07/2019 23:25, Phil Hobbs wrote:
For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C.  It's on the end of a piece of aluminum with about
5 K/W thermal resistance, so I need to keep the heat leak to a minimum
or I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design.  Silver epoxy is a
good option, because it eliminates the heat leak through the screws,
but I really want to put a compressive preload on the TEC, because
that improves its shear strength and shock resistance.

One traditional approach is to use nylon screws.  Switching from #2
stainless to #6 nylon reduces the heat leak by 20x.  However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart.  (At least it's reasonably cheap--a nickel a screw or
thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just
forget Vespel or any of the other plastics you really want.

Anybody got a source of plastic screws other than nylon?

Cheers

Phil Hobbs


Crazy idea:
sandwich the TEC between big NdFeB magnets. Lots of compressive force
and no heat leak.

Right, a 1T magnet inside an SEM chamber. That'd go over real big.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Phil Hobbs]

On 7/17/19 6:05 PM, Joerg wrote:

On 2019-07-17 06:25, Phil Hobbs wrote:
For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C.  It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design.  Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws.  Switching from #2
stainless to #6 nylon reduces the heat leak by 20x.  However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart.  (At least it's reasonably cheap--a nickel a screw or
thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just forget
Vespel or any of the other plastics you really want.

Anybody got a source of plastic screws other than nylon?


Can you use rubber rivets?

https://www.amazon.com/Mounting-Screws-Rivets-Silicone-Computer/dp/B01M4LVBY2


They could also provide tension.

Interesting idea, I don't know.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Phil Hobbs]

On 7/17/19 7:59 PM, George Herold wrote:

On Wednesday, July 17, 2019 at 7:32:32 PM UTC-4, George Herold wrote:
On Wednesday, July 17, 2019 at 4:02:14 PM UTC-4, John Larkin wrote:
On Wed, 17 Jul 2019 09:25:04 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C. It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design. Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws. Switching from #2
stainless to #6 nylon reduces the heat leak by 20x. However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart. (At least it's reasonably cheap--a nickel a screw or thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just forget
Vespel or any of the other plastics you really want.

Anybody got a source of plastic screws other than nylon?

Cheers

Phil Hobbs

Amazon has acetal (delrin) screws for a few cents each.
McMaster -Carr had threaded rod, but the smallest was
10-24 (or something)

Back in grad school I made threaded rod with the right size rod,
(maybe with help of a lathe) and a die.
I once just ran a die down a piece of rod,
(~1/8" Tellerium copper) by hand...
the thread got all wiggley-woggley after a while.
(I needed to hold the die perpendicular to the
rod axis... I was much younger then.)

George H.
Oh Phil, I wanted to ask how many screws, you are
using to hold down the TEC. (I've used 2 and 4,
I can make arguments for 1 and 3.)
Curious minds want to know,
George :^)

At the moment I'm leaning towards a larger plastic piece held down by SS
screws, clamping the PCB to the TEC.

Turns out, after quite a lot of math, that the best TEC for the job is
an 80-junction Marlow one, the NL1025T. At 9x11 mm, it's quite a bit
smaller than the board, so something more stable will be required to
keep the preload even.

Cheers

Phil Hobbs


--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Phil Hobbs]

On 7/17/19 4:02 PM, John Larkin wrote:

On Wed, 17 Jul 2019 09:25:04 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C. It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design. Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws. Switching from #2
stainless to #6 nylon reduces the heat leak by 20x. However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart. (At least it's reasonably cheap--a nickel a screw or thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just forget
Vespel or any of the other plastics you really want.

Anybody got a source of plastic screws other than nylon?

Cheers

Phil Hobbs

Amazon has acetal (delrin) screws for a few cents each.


Very nice, thanks!

Amazon swallowed Small Parts Inc. some years ago, and still seems to be
in the biz.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Phil Hobbs]

On 7/18/19 7:55 AM, Chris Jones wrote:

On 18/07/2019 16:26, Jasen Betts wrote:
On 2019-07-17, George Herold <gherold@teachspin.com> wrote:
On Wednesday, July 17, 2019 at 9:25:17 AM UTC-4, Phil Hobbs wrote:
For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C.  It's on the end of a piece of aluminum with
about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design.  Silver epoxy is a
good option, because it eliminates the heat leak through the screws,
but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.
Well a bunch of silly ideas...
The Delrin sounds OK, maybe a machine shop can cut a thread on the end
of a rod for you. (There are a few places on the web listing delrin
screws, but no prices.)

How about soldering/ or epoxy and then squeeze it some other way.
(Fishing line/ zip ties, something else?)

Fishing line and zip-ties are nylon, which is no good, by maybe polyester
or linen thread would work

You need my beer can cooler on the hot plate :^)
Has anyone done a distributed two stage TEC, Some small TEC in
vacuum and then a bigger one outside the vacuum going to a
heatsink?

It would seem easier to just use a heat-pipe (if you can find a suitable
fluid) and put all the TEC in atmosphere.

At least with the TEC in a good vacuum you'd get no moisture condensing
on/inside it.

I was wondering if one can buy peltiers with the edges hermetically
sealed (not with silicone rubber) and a vacuum inside. I'm guessing that
sealing the edges with thin stainless foil brazed to the ceramic might
work, and that sealing the edges with just frit and ceramic might be not
flexible enough to withstand differential thermal expansion of the two
ceramic plates.

It's easier than that. You just put a MEMS pressure sensor on the
board, and when the pressure comes on-scale, you turn off the HV to the
detector and run the cold plate up to room temperature.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Phil Hobbs]

On 7/18/19 2:26 AM, Jasen Betts wrote:

On 2019-07-17, George Herold <gherold@teachspin.com> wrote:
On Wednesday, July 17, 2019 at 9:25:17 AM UTC-4, Phil Hobbs wrote:
For that SEM cathodoluminescence gizmo I mentioned in the
distributed attenuator thread, I need to cool a multi-pixel
photon counter (MPPC) down to about -10C. It's on the end of a
piece of aluminum with about 5 K/W thermal resistance, so I need
to keep the heat leak to a minimum or I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design. Silver epoxy
is a good option, because it eliminates the heat leak through the
screws, but I really want to put a compressive preload on the
TEC, because that improves its shear strength and shock
resistance.
Well a bunch of silly ideas... The Delrin sounds OK, maybe a
machine shop can cut a thread on the end of a rod for you. (There
are a few places on the web listing delrin screws, but no prices.)

How about soldering/ or epoxy and then squeeze it some other way.
(Fishing line/ zip ties, something else?)

Fishing line and zip-ties are nylon, which is no good, by maybe
polyester or linen thread would work

The design is for a customer, so they probably wouldn't appreciate the
Betsy Ross approach.

You need my beer can cooler on the hot plate :^) Has anyone done a
distributed two stage TEC, Some small TEC in vacuum and then a
bigger one outside the vacuum going to a heatsink?

Probably. Two-stage TECs are sort of disappointing because all but the
last stage run at fairly high Q-dot.

It would seem easier to just use a heat-pipe (if you can find a
suitable fluid) and put all the TEC in atmosphere.

I'm leery about using a heat pipe in a SEM--if it started leaking with
the column valve open, the resulting glow discharge could damage the
gun, which is super big bucks to fix.

Looks like OFHC copper will be required, but will be good enough.

Cheers

Phil Hobbs


--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Phil Hobbs]

On 7/17/19 10:08 PM, Joseph Gwinn wrote:

On Jul 17, 2019, Phil Hobbs wrote
(in article<94ednTJ5EJEHtbLAnZ2dnUU7-QnNnZ2d@supernews.com&gt:

On 7/17/19 9:41 AM, Joseph Gwinn wrote:
On Jul 17, 2019, Phil Hobbs wrote
(in article<RKudnVUn6NouvLLAnZ2dnUU7-VfNnZ2d@supernews.com&gt:

For that SEM cathodoluminescence gizmo I mentioned in the distributed

attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C. It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design. Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws. Switching from #2
stainless to #6 nylon reduces the heat leak by 20x. However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart. (At least it's reasonably cheap--a nickel a screw or thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just forget
Vespel or any of the other plastics you really want.
I think you will have to machine the non-nylon hardware. Delrin is a good
choice, if it´s OK in vacuum. Note that unlike most plastics, one can
press-fit delrin, so threads may not be needed.

For vacuum, I´d go to ceramic spacers with stainless steel hardware keeping
them in compression.

Or, if thermal insulation is the only reqt, a stack of shiny steel plates
with staggered metal separators (which can be spot-welded), so the leakage
paths are very long, and radiation coupling between plates dominates (there
been no air to support convection).

Joe Gwinn

Ceramic is a surprisingly decent thermal conductor,
unfortunately--better than glass, and much better than plastic. SEMs
don't run in UHV, so FR-4 is okay. At lower delta-T you can just have
the board hang ten over the TEC and use the FR-4 for insulation.

How much vacuum does a SEM need? I was thinking bake-out being needed.

Nah, SEMs work in the low microtorr range. Most don't even use a
turbopump, because diffusion pumps are good enough and don't cause
vibration.

UHV (<~ 1 nanotorr) is another beast altogether.

If FR-4 is good enough for that vacuum, then use FR-4 stepped washers, steel
hardware, and belleville spring washers to maintain tension despite thermal
variation.

Washers don't provide enough insulation, because they're short and fat.

Or use steel washers, and stagger them to make the thermal path long and
windy.

Hard to get a factor of 100 that way.

Note that delrin is very slippery, and so delrin nut and bolt will tend to
unscrew itself. A SS nut forced to thread onto a slightly oversize delrin rod
should stay put. As will a press fit into an undersize hole.

The speed nuts are called Tinnerman nuts, as someone else also mentioned.
Tinnerman invented this kind of nut.

Thanks!

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[david eather]

On 17/07/2019 11:25 pm, Phil Hobbs wrote:

For that SEM cathodoluminescence gizmo I mentioned in the distributed
attenuator thread, I need to cool a multi-pixel photon counter (MPPC)
down to about -10C.  It's on the end of a piece of aluminum with about 5
K/W thermal resistance, so I need to keep the heat leak to a minimum or
I'll run out of Q-dot.

So I'm sort of breathing on the cold plate design.  Silver epoxy is a
good option, because it eliminates the heat leak through the screws, but
I really want to put a compressive preload on the TEC, because that
improves its shear strength and shock resistance.

One traditional approach is to use nylon screws.  Switching from #2
stainless to #6 nylon reduces the heat leak by 20x.  However, nylon
isn't good in a vacuum, because it outgasses, dries out, and falls
apart.  (At least it's reasonably cheap--a nickel a screw or thereabouts.)

I looked at other plastic screws at McMaster-Carr, and the price
difference is fairly shocking--even screws made of some totally
garden-variety plastic such as PVC cost dollars apiece, and just forget
Vespel or any of the other plastics you really want.

Anybody got a source of plastic screws other than nylon?

Cheers

Phil Hobbs

Are M3 polycarbonite screws of any use?

 

[Martin Brown]

On 18/07/2019 14:11, dcaster@krl.org wrote:

On Wednesday, July 17, 2019 at 11:03:16 AM UTC-4, Martin Brown
wrote:


Or for a one off buy some PEEK rod and thread it yourself.

https://www.directplastics.co.uk/peek-rod

-- Regards, Martin Brown

For a one off, I could probably get what ever plastic you want to try
and try using a die to cut threads on it.

I am in Delaware and there is a MSC and Grainger outlet near by so no
shipping charge and no sales tax.

A friend has machine shop and we could try making a injection mold.
He has a job machining PEEK so lots of free non virgin PEEK
available.

A piece of rod or offcut PEEK with a threaded blind hole in it jammed on
the end of a normal stainless steel bolt might be good enough.

--
Regards,
Martin Brown