Casting PCBs in plastic resin
Richard
directly into a solid plastic resin block. The power and I/O would be
accessible, and the resin would effectively be the "case".
Has anyone here dabbled in this area? A couple questions come to
mind...
1) In general, is plastic resin compatible with such an application
(chemically, electrically, thermally, etc.)
2) Depending on #1, is conformal coating an effective solution to some
of these problems?
I'm curious to hear about any experiences like this - especially if
you've encased PCBs with SMD chips (e.g., QFP).
FWIW, this application is an 8MHz microcontroller with an ISA peripheral
chip.
Pctchrisgibson
Balmar make an alternator controller called the "max-charge" that is built into
an aluminium extrusion then filled with clear resin. I'm sure you'll find a
piccie of one if you search the web.
Chris
Mike Harrison
than others.
Encapsulation is not always as simple as it wounds, due to issues like
mechanical stress buildup during curing.
Polymorph
as 5%. Imagine putting your PCB in a vise and squeezing it 5% and you
see the problem.
So I'd consider that the number one priority. Finding a resin that
doesn't shrink an appreciable amount on curing.
Check out these lists for info and links on casting:
http://groups.yahoo.com/group/nwpropmasters
http://groups.yahoo.com/group/casting
Richard <rh...@azglobal.com> wrote in message news:<3D1D4370...@azglobal.com>...
Mark Lerman
had the molds made and does the encapsulation, so I'm not sure of the resin
used, but I'll see if I can find it for you.
Mark
"Richard" <rh...@azglobal.com> wrote in message
news:3D1D4370...@azglobal.com...
Richard
thanks!
Yes, please - I would be very interested to know all the specs you can
share (mfr & type of resin, curing method, etc.).
Do you happen to have photos online of your project? What kind of
circuit was it? Why did you choose to use resin?
Cheers,
Richard
Richard
> One very big caveat: many types of resin shrink on hardening! As much
> as 5%. Imagine putting your PCB in a vise and squeezing it 5% and you
> see the problem.
An excellent point - thanks for raising it. This PCB will be quite
small (25mm x 70mm), so shrinking will not be visible but could still be
significant enough to lift pads.
Originally I was also thinking about thermal compatibility in terms of
dissipation, but I suppose that heat generated during the curing process
should also be considered (though I doubt it'd reach solder's melting
point). I do know this varies by the curing technique & speed.
> Check out these lists for info and links on casting:
> http://groups.yahoo.com/group/nwpropmasters
> http://groups.yahoo.com/group/casting
Thanks for the pointers. So far, I've been poking around
http://www.smooth-on.com, a resin manufacturer.
no_one
Expansion (CTE) between the chip and the resin; if they expand at different
rates you will fail due to thermo-mechanical cycles over time.
"Richard" <rh...@azglobal.com> wrote in message
news:3D1E04BE...@azglobal.com...
Gareth
Go to www.maplin.co.uk and do a "Code Search" for "AZ92A"
-----------------------------------------------------------------------
To reply to me directly:
Replace the text after the@symbol with: totalise DOT co DOT uk
"Richard" <rh...@azglobal.com> wrote in message
news:3D1D4370...@azglobal.com...
The Bear
stress smt parts, do not generate any exothermic heat on curing, and very
interesting to mix up. They cure in about 10 minutes to a firm jelly
finish, depending on which one you use.
You would need a potting box to hold your board, then just fill it up.
Brill Pappin
protection, and be water-proof. It would also reduce the possibility of
stress on the PCB due to the setting of the "encasing" material.
as far as I know, Silicon moulding is vary cheap, and easy to do. I've been
thinking of using that method myself for some long-term outdoor devices, but
I haven't actually tried it yet.
- Brill Pappin
"Richard" <rh...@azglobal.com> wrote in message
news:3D1D4370...@azglobal.com...
Doug Dwyer
Two part resins have free ions during the setting stage.
I had a fright when a high impedance circuit failed to function 10 hours
after apparent setting; after 3 days it was OK this is probably due to
some conductivity during the setting process.
In article <wEnT8.2610$FG5.2...@newsread2.prod.itd.earthlink.net>,
no_one <no_...@hotmail.com> writes
--
Doug Dwyer
Richard
> You might try silicon casts instead... they would
> also provide shock protection, and be water-proof.
> It would also reduce the possibility of stress on
> the PCB due to the setting of the "encasing"
> material.
I've come across some silicon gel products that looked interesting, but
not in a hardness that I need (so far). The gel products provide
similar benefits to resin, but you can also insert a probe through them
and the hole self-heals.
In this case, the goal is for the resin to be the case for the product,
not just an inner protective layer. The hardness really just needs to
be a "solid", nothing elaborate.
FWIW, I spoke yesterday to a contract firm that claims to be capable of
doing this. Also, I've come across one resin product that claims to fit
this application (p/n UR5048) - and it's clear too:
http://www.electrolube.com/product/selector/resindat.asp
As an aside, has anyone worked with Electrolube in the UK? They look
like a good resource for PCB coatings, etc.
> as far as I know, Silicon moulding is vary cheap,
> and easy to do. I've been thinking of using that
> method myself for some long-term outdoor devices,
> but I haven't actually tried it yet.
Seems to be a popular option for marine applications, so you'd probably
have good results. The process is pretty simple, and Smooth-on offers a
USD$25 starter kit (http://www.smooth-on.com). The only expensive
hardware is a vacuum kit if simple pour molding won't work for you.
Richard
> No one has mentioned a further problem...
> Two part resins have free ions during the setting
> stage.
> I had a fright when a high impedance circuit failed
> to function 10 hours after apparent setting; after
> 3 days it was OK this is probably due to some
> conductivity during the setting process.
Interesting - an excellent tip, thank you!
Can you share some details about your project?
Were you potting your PCB or fully encasing it in a block? Do you
recall the type of resin and curing method?
And to earlier points made here... have you experienced later failures
due to differing thermal expansion rates, or did you get a resin that
accounted for this?
Sir Charles W. Shults III
material that semiconductors are made of. You will not be potting anything in
silicon due to its extremely high melting point.
You are obviously referring to "silicone", a rubbery or gel-like material
that is made from silicon and carbon among other things. For newbies, the two
are hard enough to correctly identify. It's best to eliminate the confusion up
front.
Speaking of up front, I recently received an email from some porno site
touting their "silicon" babes. Who wants to be in bed with a girl who will
shatter at the first minor impact?
Cheers!
Chip Shults
My robotics, space and CGI web page - http://home.cfl.rr.com/aichip
Pctchrisgibson
> This might sound trivial, but "silicon" is an element- a glassy, gray-black
>material that semiconductors are made of. You will not be potting anything
>in
>silicon due to its extremely high melting point.
> You are obviously referring to "silicone", a rubbery or gel-like material
>that is made from silicon and carbon among other things. For newbies, the
>two
>are hard enough to correctly identify. It's best to eliminate the confusion
>up
>front.
> Speaking of up front, I recently received an email from some porno site
>touting their "silicon" babes. Who wants to be in bed with a girl who will
>shatter at the first minor impact?
>
Although it brings new meaning to the phrase "the system's gone down again"
Chris
John Woodgate
<aic...@cfl.rr.com> wrote (in <DCGT8.100531$db.17...@twister.tampabay.
rr.com>) about 'Casting PCBs in plastic resin', on Sun, 30 Jun 2002:
>Speaking of up front, I recently received an email from some porno site
>touting their "silicon" babes.
Sound cool! (;-) P or N-type? Doping levels should be high, to ensure
low resistivity. Absolutely NO requirement for gold bonding!
--
Regards, John Woodgate, OOO - Own Opinions Only. http://www.jmwa.demon.co.uk
Interested in professional sound reinforcement and distribution? Then go to
http://www.isce.org.uk
PLEASE do NOT copy news posts to me by E-MAIL!
Tim Hubberstey
>
> Brill Pappin wrote:
> > You might try silicon casts instead... they would
> > also provide shock protection, and be water-proof.
> > It would also reduce the possibility of stress on
> > the PCB due to the setting of the "encasing"
> > material.
>
> I've come across some silicon gel products that looked interesting, but
> not in a hardness that I need (so far). The gel products provide
> similar benefits to resin, but you can also insert a probe through them
> and the hole self-heals.
>
> In this case, the goal is for the resin to be the case for the product,
> not just an inner protective layer. The hardness really just needs to
> be a "solid", nothing elaborate.
Perhaps you could consider a two-step process of potting first in some
kind of silicone and then pot that assembly into your hard resin. I
would think that would eliminate the mechanical stress on the PCB and
still give you your hard shell. There may still be some pressure applied
to the assembly if the resin shrinks but as long as you have no voids in
any of the components it shouldn't be an issue. The silicone might also
provide some measure of shock absorption as an added benefit. I've never
tried this so I don't know how well (or even if) it will work.
--
Tim Hubberstey, P.Eng. . . . . . Hardware/Software Consulting Engineer
Marmot Engineering . . . . . . . VHDL, ASICs, FPGAs, embedded systems
Vancouver, BC, Canada . . . . . . . . . . . http://www.marmot-eng.com
John Fields
>
> I'm exploring creative ideas, the latest being to encase a project PCB
> directly into a solid plastic resin block. The power and I/O would be
> accessible, and the resin would effectively be the "case".
---
http://www.masterbond.com/
---
John Fields
Professional circuit designer
http://www.austininstruments.com
email: jfields at texas dot net
Richard
> Perhaps you could consider a two-step process of
> potting first in some kind of silicone and then pot
> that assembly into your hard resin. I would think
> that would eliminate the mechanical stress on the
> PCB and still give you your hard shell.
Yep... that thought has recently crossed my mind. I question how
transparent the clear silicone (not silicon ;-) gel would be. Perhaps
it wouldn't need to be very thick to buffer the thermal expansion.
If I go to an opaque resin, then the options broaden, such as using
normal RTV potting compound covered by a colored hard shell. But then
the case wouldn't be quite as nifty.
So far, I've come across a 2-3 hard compounds that claim to be suited
for PCB encapsulation. Keeping my fingers crossed. Thanks!
Ian Stirling
>
> Tim Hubberstey wrote:
>> Perhaps you could consider a two-step process of
>> potting first in some kind of silicone and then pot
>> that assembly into your hard resin. I would think
>> that would eliminate the mechanical stress on the
>> PCB and still give you your hard shell.
>
> Yep... that thought has recently crossed my mind. I question how
> transparent the clear silicone (not silicon ;-) gel would be. Perhaps
> it wouldn't need to be very thick to buffer the thermal expansion.
You can get perfectly clear optical silicone.
Look at the various makers websited.
--
http://inquisitor.i.am/ | mailto:inqui...@i.am | Ian Stirling.
---------------------------+-------------------------+--------------------------
"The theory of everything falls out trivially." -- Etherman, sci.physics kook.
Bob Wilson
In general, potting of PCBs is a pain in the backside. There are many resin
families that can be used. The most common are:
- Epoxies. In particular, polyamide-based epoxies are often favoured
because they they have a variable mixture ratio (resin to hardener) that can
result in a wide ratio of hardnesses. Thus it is possible to ensure the
result is not rock hard, thus avoiding most thermal mismatch problems. The
main problem with polyamide epoxies is that they absorb some moisture from
the air, resulting in reduced bulk resistivity. As far as the "free ion"
thing mentioned earlier, this is not true. NO potting compound is ionic!
What is often thought to be an "ionic" problem is generally due to moisture
that has been absorbed in the amide component before mixing. Use of new
resin and proper storage of old resin will eliminate this problem.
- Polyurethanes. Also used as a potting compound. Some are lightning fast
to cure, with only moderate exotherm. But they tend to be very hard.
There are 2 parts: an isocyanate and a polyol. The polyol is hygroscopic,
and absorbs moisture from the air, causing bubbling when hardening,
reduced bulk resistance (also incorrectly thought by some to be a result of
"ionic" contamination). They are not commonly used.
- Silicones. I'm not talking about "RTV" and other one-part moisture cure
silicones, which are totally unsuitable for potting. Silicones are available
in 2-part formulations (typical ratios are 10:1 up to 100:1). Very
expensive, but resist moisture absorption very well. They tend not to stick
well to components, so are easy to peel away.
But generally speaking, potting is a real porduction bottleneck. Potting
resins typically take up to 24 hours to cure, meaning that you need a LOT of
moulds to get any throughput. Mixing is really messy and subject to ratio
errors unless you have automated mixing equipment. Finally, they are
EXPENSIVE! It is not uncommon for the largest singel cost in a potted
assembly, to be the potting compound itself! Potting tends to be one of
those processes that amateurs think is a great idea, and production people
try to avoid like the plague!.
Bob.
Brill Pappin
finally get something you like...
It that kind of project that we will all need at some obscure point... For
me it'll be remote/outdoor robotics...
Come to think of it, you might also want to consider heat retention and
absorption (for other than the circuit itself)... i.e. for an outdoor
system, I wouldn't want to use a black (or dark) material, as it would suck
up heat quickly and hold onto it.
- Brill Pappin
"Richard" <rh...@azglobal.com> wrote in message
news:3D1F8FE2...@azglobal.com...
Richard
> But generally speaking, potting is a real porduction
> bottleneck. Potting resins typically take up to 24
> hours to cure, meaning that you need a LOT of moulds
> to get any throughput. Mixing is really messy and
> subject to ratio errors unless you have automated
> mixing equipment. Finally, they are EXPENSIVE! It is
> not uncommon for the largest singel cost in a potted
> assembly, to be the potting compound itself!
Excellent words of advice - I'll take them to heart. Thanks for the
insight! I can only imagine the startup costs for a decent number of
molds...
> Potting tends to be one of those processes that
> amateurs think is a great idea, and production
> people try to avoid like the plague!.
I can see how this may be. Even past the technical issues of making it
work, it certainly seems to be the sort of thing that's not suited for
production scale. Hopefully I can find a solution that works well
enough in a hobbyist setting to allow for a low volume of novel
prototypes.
N. Thornton
> I'm exploring creative ideas, the latest being to encase a project PCB
> directly into a solid plastic resin block. The power and I/O would be
> accessible, and the resin would effectively be the "case".
>
> Has anyone here dabbled in this area? A couple questions come to
> mind...
I have used this method, but it wasn't for electronics. Little clear
plastic balls were incorporated for visual effect - they look like
bubbles - and it occurred to me you could possibly halve your resin
cost by pouring balls of low cost material in there with the resin.
Even chopped scrap plastic would possibly do. This approach will
reduce strength to some extent, and reduce cost quite a bit.
Possible fillers - I havent tried these:
chopped plastic
polystyrene balls
chopped cardboard - caution re moisture on that one.
small sealed air filled plastic bags
etc
Regards, NT
Al Adrian
Al...
"N. Thornton" <big...@meeow.co.uk> wrote in message
news:a7076635.0207...@posting.google.com...
Richard
> Not only the shrinking during curing, there is also
> a Coefficient of Thermal Expansion (CTE) between
> the chip and the resin; if they expand at different
> rates you will fail due to thermo-mechanical cycles
> over time.
OK... I've done some homework, and I'm finding quite a range within just
the products designed for electronics encapsulation. CTEs as low as 12
and as high as 120.
I gather the CTE needs to be as close as possible to the PCB's CTE
(which presumably is fairly uniform amongst its components?)...
So is there a "standard" value for the CTE of a populated PCB?
I'll take a leap and suppose that a range of acceptable CTE values can
be calculated depending on the thermal cycling range? (i.e., stress of
large delta @ small thermal cycle == stress of small delta @ large
thermal cycle?)
In this case, the cycle is probably only 10 degrees Celsius with less
than 100 cycles per lifetime. Though I suppose some extremes should be
factored for shipping, storage, and geographic temp differences.
N. Thornton
> Cheap fillers will float...
yup, though not all. The filler I used was pieces of clear plastic. They sank :)
Regards, NT
> > Possible fillers:
> >
> > chopped plastic
Doug Dwyer
The potting material was a 2 part and black sold in plastic sachets
similar to "Robnorganic" from Maplin UK.
Pretty sure that moisture was not the problem.
I imagine that a two part exothermic reaction will have free radicals
during the polymerisation process that I describe as giving rise to
ionic conduction, my terminology may not be correct perhaps an electro
chemist can advise.
The circuit was an op amp circuit with 1M resistors.
with a In article <3D1FF47C...@azglobal.com>, Richard
<rh...@azglobal.com> writes
--
Doug Dwyer
Bob Wilson
says...
>
>RE: change in conductivity.
>The potting material was a 2 part and black sold in plastic sachets
>similar to "Robnorganic" from Maplin UK.
>Pretty sure that moisture was not the problem.
>I imagine that a two part exothermic reaction will have free radicals
>during the polymerisation process that I describe as giving rise to
>ionic conduction, my terminology may not be correct perhaps an electro
>chemist can advise.
>The circuit was an op amp circuit with 1M resistors.
You need to draw the distinction between "Ionic" and "Free Radical". It is
ionic contamination that causes conductivity (in the presence of
moisture). Without moisture, even ionic contamination will not be
conductive, which is why pure anhydrous salt (NaCl) is such an excellent
insulator.
Nonetheless, epoxies and polyurethanes, at least do not involve free
radicals during polmerization. The catalyst usually used for polyesters
(generally Benzoyl Hydroperoxide) does initiate the polymerization reaction
via free radicals, but the free radicals are oxygen radicals (which are not
going to promote conduction of any sort). Besides, no one in his right mind
would pot a circuit in polyester resin anyway!
Organic free radicals do not promote conductivity. After all, how would one
expect non-ionic (non-polar) radicals like epoxide radicals, amides, or
siloxanes to have any effect on conductivity? They are all large, complex,
and non-polar (the latter being manditory to act as a conducing medium).
As I mentioned, the problem has at least something to do with moisure
absorption (something that many of the momomers used in potting systems do
only too well).
Bob.
klmok
klmok
>I'm exploring creative ideas, the latest being to encase a project PCB
>directly into a solid plastic resin block. The power and I/O would be
>accessible, and the resin would effectively be the "case".
I have been following this thread with great interest because I have
the same problem regarding how to cost out the packaging for a low
production run (eg. lots of 1000.) The size and shape of the product
case will fit into a cake of bath soap. PCB with DB-9 connector and
some cables. The electronics project phenolic plastic case is costly
and cutting the holes for the connector and cables isn't practical.
Sheet metal boxes not much better. Injection molded custom cases is
big bucks upfront.
One idea is to use vulcanizing rubber to make a solid rubber block.
The other is to inject plastic which will encase the whole thing as a
solid plastic block. In both methods we are looking at temperatures
of 350 deg F and pressures of a couple of atmospheres when forming.
Probably an impractical solution or else toymakers would have done
this. But any comments if this would work?
Bob Wilson
says...
Never mind the polymer you may be using to pot the thing (RTV silicone is
your ONLY choice here, by the way, since no other type can come close to
withstanding these temperatures). Your problem is going to be your
electronic components. I very much doubt that any commercially available
components will come close to being able to function at this temperature.
Your DB9 connector certainly can't withstand it, and no normal PCB material
is rated for even half this temperature (FR4's glass temperature is around
150 to 200C, after which it becomes useless in short order).
This temperature will destroy any normally avaliable component or PCBs.
Unless you have managed to get some pretty exotic components, your hope of
getting these things to withstand this temperature is pretty much a joke.
Bob.
Brill Pappin
playing around with electronics, I had this habit of encasing the devices in
the plastic that you use with glue guns (most of them were free-formed and I
wanted a way of protecting the circuit)... I would build a mould and squirt
it in... I still have a couple of those devices around (after 15 years)
which work fine and sit on my bench as tools (granted, they are simple
things, like a 555 pulse generator).
I'm sure someone here could tell you if it would degrade the components...
but its simple, and it worked at least to this point.
- Brill Pappin
"N. Thornton" <big...@meeow.co.uk> wrote in message
news:a7076635.0207...@posting.google.com...
Mark Lerman
they (and I) are missing something here, it seems pretty simple to make a
mold and pour. I have the first sample in my hand and the first 50 pieces
are ready to ship. The plastic acts as the case and looks just fine.. The
only production problem they had was the resin getting into the rear of the
piezo element, and they solved it pretty easily. They use a company that
specializes in encapsulation and seemed very familiar with the process. The
biggest problem is that they can only do 1 unit per day per mold. I had 10
molds done (master plus 9 secondary), so they can do 10 per day.
Mark
"klmok" <kl...@shaw.ca> wrote in message
news:3d2f67e8...@shawnews.ed.shawcable.net...
Bob Wilson
says...
>
>I have a product that my assembly company suggested encapsulation of.
Unless
>they (and I) are missing something here, it seems pretty simple to make a
>mold and pour. I have the first sample in my hand and the first 50 pieces
>are ready to ship. The plastic acts as the case and looks just fine.. The
>only production problem they had was the resin getting into the rear of the
>piezo element, and they solved it pretty easily. They use a company that
>specializes in encapsulation and seemed very familiar with the process. The
>biggest problem is that they can only do 1 unit per day per mold. I had 10
>molds done (master plus 9 secondary), so they can do 10 per day.
Ignoring the huge hassle this stuff is to work with (if you want to do it
properly), since you are getting it done by someone who has the proper
equipment (accurate mixing equipment, vaccuum de-airing chamber, and so on),
you need to remember that many of these potting materials can be so hard
that expansion over temperature can cause component failure. this is
especially true at low temperature when many epoxies are as hard as a rock.
Be sure to choose your resin carefully. For a run of 50, this may be a
reasonable choice. For a production run, potting is hopeless!
Bob.
tweeks
hard?
Just curious..
Tweeks
Bob Wilson
>
>What about just using that expanding construction foam that turns rock
>hard?
>
>Just curious..
Because the LAST thing you want is something that is rock hard. Further, the
other thing you want to avoid at all costs is an encapsulation that has air
voids inside! All resins are permeable to water vapour. NONE are "hermetic".
Normally water diffusing through a solid resin is not a problem bcause it
does not exist as a liquid. But if there are voids in the resin block (and a
foam has millions of them!), the water diffusing through can colledt in the
bubbles and then condense to actual LIQUID water. Corrosion city!
Bob.
David Harmon
tweeks <twe...@theweeks.org> wrote:
>What about just using that expanding construction foam that turns rock
>hard?
Probably about the worst thing possible for heat dissipation.
Richard
> [...]
> One idea is to use vulcanizing rubber to make a
> solid rubber block. The other is to inject
> plastic which will encase the whole thing as a
> solid plastic block. In both methods we are
> looking at temperatures of 350 deg F and pressures
> of a couple of atmospheres when forming. Probably
> an impractical solution or else toymakers would
> have done this. But any comments if this would
> work?
It is possible to get resins with a specific CTE range close to the PCB
and components to minimize the thermal problems. I imagine smaller
encapsulations would fare better as well, since the overall
expansion/contraction would be less. The biggest obstacle I've found
thus far is in being particular about the optical clarity.
http://www.masterbond.com has one that's optically clear; samples are en
route. If you're willing to do opaque, there are more options.
Also, consider something on the high end of the Shore A scale (firm
rubber, like the heel of your dress shoe) instead of a Shore D hard
epoxy. These guys offer both types, including clear:
http://www.electrolube.com/product/resin1.asp
Check out this page for a Shore hardness and viscosity reference:
http://www.epoxies.com/convert.htm#Durometer
As you'd expect, there are several techniques for the process. I don't
think injection molding would be a good fit here. More likely, a vacuum
process (where the air is sucked out of the mold while pouring the
resin, eliminating voids and bubbles). Resins also have several curing
methods each; some are more exothermic than others, and some shrink more
than others during curing.
As others have noted, encapsulation may not be worthwhile for larger
production runs, where you can justify the startup costs for other forms
of cases. You can do some resins in your garage, but for the cost of a
good vacuum pump (apparently ~$300), you can pay for the setup costs at
a small encapsulation shop.
Lastly, consider whether you'll need to service this puppy after you
encase it in plastic. ;-) It looks like some of the resins are designed
to be "servicable", where they can be removed in using a solution that's
not damaging to the PCB. No idea how well this works, but it hardly
seems worth the trouble except for failure analysis.
Have fun!
Bob Wilson
You will never get a Tce match, since the Tce of FR4 is different in the
z-axis (as compared to the x and y axes) by a factor of almost 10. That is
why a semi-resilient encapsulant is manditory.
Shin-Etsu makes an optically clear 2-part silicone rubber, but like all
silicones it is expensive. Hysol makes a clear epoxy that remains slighly
(i.e. acceptably) resilient down to -20C or less.
The encapsulents that are "serviceable" are generally Silicones. They do
this by taking advantage of Silicone's poor resistance to common solvents,
and its generally poor adhesion to the PCB and its components. You just
break it off in chunks!
Other resins are simply not soluable in solvents. Some cheap epoxies can be
softened by methylene chloride, but that also attacks the FR4 somewhat.
Urethanes are the encapsulent from hell; extremely hard to remove.
Bob.
klmok
(Bob Wilson) wrote:
>
>You will never get a Tce match, since the Tce of FR4 is different in the
>z-axis (as compared to the x and y axes) by a factor of almost 10. That is
>why a semi-resilient encapsulant is manditory.
>
>Shin-Etsu makes an optically clear 2-part silicone rubber, but like all
>silicones it is expensive. Hysol makes a clear epoxy that remains slighly
>(i.e. acceptably) resilient down to -20C or less.
>
>The encapsulents that are "serviceable" are generally Silicones. They do
>this by taking advantage of Silicone's poor resistance to common solvents,
>and its generally poor adhesion to the PCB and its components. You just
>break it off in chunks!
>
>Other resins are simply not soluable in solvents. Some cheap epoxies can be
>softened by methylene chloride, but that also attacks the FR4 somewhat.
>Urethanes are the encapsulent from hell; extremely hard to remove.
>
As in any packaging the appearance is as important as its mechanical
properties.
The original objective is of course to avoid the cost of making an
injection mold or the labour involved in modifying premade boxes and
yet come up with an attractive package for a limited production run.
Building on the discussion in this thread perhaps one solution is to
first mount the connectors, cables and circuit board on a rigid
aluminium plate, 'L' or 'C' profile, to fix the mechanical
relationship between the parts. Scrap PCB material can probably be
used in place of aluminium. Build this up into a box (cheap and
ugly, nobody sees it anyway.) Then encapsulate the box with hard
polyurethane which has the mechanical properties I desire. The
polyurethane encapsulation forms an attractive, affordable and
mechanically strong packaging for the device. The box isolates the
electronic components from the stresses during the curing of resin or
polymers. The box also keeps the polyurethane away from the components
so that the components can be serviced or reused. To service the
device cut open the encapsulation.
The project I have is electronics for an assistive device for the
disabled. Encapsulatiion seems an advantage in sealing the device
against hostile environment and to put up with rough use. It is more
cost effective to replace a malfuntioning unit than for a technician
to service an occasional unit. Encapsulation would discourage outside
servicemen tinkering with the unit.
Richard
> The original objective is of course to avoid the
> cost of making an injection mold or the labour
> involved in modifying premade boxes and yet come
> up with an attractive package for a limited
> production run.
One option might be to pot the materials in one media (e.g., silicone),
then encase in a harder shell (e.g., urethane). Depends on the size of
your device.
If you're inclined to dabble with general plastics, check out
http://www.smooth-on.com. They've got a $25 starter kit with materials
to make latex molds and cast white resin. You can get small quantities
of their other resins for ~$20 IIRC.
For your application, look also at watertight grommets / strain reliefs
that might make a better transition out of the hard case for your
cables.
Bob Wilson
says...
Potting only keeps the trifers out. It is not really hard to completely
de-pot an assembly.
Your suggestion has one VERY fatal flaw. No resin is hermetic. Water vapour
will slowly diffuse through it. As I mentioned earlier, this is not a
problem is there are no voids in the resin since the water exists only in
the molecular state, not as a liquid. But you are suggesting a honking big
void in the center, within which your circuit resides. Over time, water
vapour will slowly diffuse into this void, and can condensre into liquid
water. This, coupled with the voltage diffentials present will literally
DRIVE a corrosion reaction. Thie is not theory! I have seen it happen.
Bob.
Bob Wilson
>
>klmok wrote:
>> The original objective is of course to avoid the
>> cost of making an injection mold or the labour
>> involved in modifying premade boxes and yet come
>> up with an attractive package for a limited
>> production run.
>
>
>One option might be to pot the materials in one media (e.g., silicone),
>then encase in a harder shell (e.g., urethane). Depends on the size of
>your device.
That is entirely useless! Silicone in incompressible. The rubbery feel it
has results from the material displacing when you squeeze it, NOT from it
actually compressing (which it simply cannot do). When fully encapsulated
with a rigid enclosure (the resin envelope), it will have nowhere to go and
will be as hard as a rock.
Bob.
Sylvan Butler
> In article <3D3582C2...@azglobal.com>, rh...@azglobal.com says...
>>then encase in a harder shell (e.g., urethane). Depends on the size of
> That is entirely useless! Silicone in incompressible. The rubbery feel it
> has results from the material displacing when you squeeze it, NOT from it
> actually compressing (which it simply cannot do). When fully encapsulated
> with a rigid enclosure (the resin envelope), it will have nowhere to go and
> will be as hard as a rock.
Unless you leave voids between the silicon and the harder shell (and
pinhole drains in the shell to keep the voids "dry").
sdb
--
| Sylvan Butler | Not speaking for Hewlett-Packard | sbutler-boi.hp.com |
| Watch out for my e-mail address. Thank UCE. >>>> change ^ to @ <<<< |
They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety. --Benjamin Franklin, 1759
Stop Palladium, TCPA and Senator Fritz Hollings, NOW!
http://www.cl.cam.ac.uk/users/rja14/tcpa-faq.html
Bob Wilson
Znospam+...@hpb13799Z.Zboi.hpZ.com.invalid says...
>
>On Thu, 18 Jul 2002 03:03:06 -0000, Bob Wilson
<rfwi...@intergate.nospam.bc.c
>a> wrote:
>> In article <3D3582C2...@azglobal.com>, rh...@azglobal.com says...
>>>One option might be to pot the materials in one media (e.g., silicone),
>>>then encase in a harder shell (e.g., urethane). Depends on the size of
>
>> That is entirely useless! Silicone in incompressible. The rubbery feel it
>> has results from the material displacing when you squeeze it, NOT from it
>> actually compressing (which it simply cannot do). When fully encapsulated
>> with a rigid enclosure (the resin envelope), it will have nowhere to go
and
>> will be as hard as a rock.
>
>Unless you leave voids between the silicon and the harder shell (and
>pinhole drains in the shell to keep the voids "dry").
Yes, that ought to work, although that might be a trifle difficult to
impliment in a cost effective way.
Bob.
Richard
> Your suggestion has one VERY fatal flaw. No resin
> is hermetic. Water vapour will slowly diffuse
> through it. As I mentioned earlier, this is not a
> problem is there are no voids in the resin since
> the water exists only in the molecular state, not
> as a liquid. But you are suggesting a honking big
> void in the center, within which your circuit
> resides. Over time, water vapour will slowly
> diffuse into this void, and can condensre into
> liquid water. This, coupled with the voltage
> diffentials present will literally DRIVE a
> corrosion reaction. Thie is not theory! I have
> seen it happen.
Excellent info - duly noted.
What's the timeline you've observed - are we talking about product life
expectancy of weeks, months, or years?
Any ideas what catalyzes the diffusion process - does it just occur
naturally out of a need to balance humidity, or is it caused by
controllable factors like thermal cycling or operating temperatures?
Would conformal coating provide an adequate hermetic barrier?
Richard
> That is entirely useless! Silicone in
> incompressible. The rubbery feel it has results
> from the material displacing when you squeeze it,
> NOT from it actually compressing (which it simply
> cannot do). When fully encapsulated with a rigid
> enclosure (the resin envelope), it will have
> nowhere to go and will be as hard as a rock.
Very true.
This thought was along the lines of countering the shear effect of the X
and Y axes (thinking the softer silicone would shift around the
components instead of shearing them off the PCB).
To your point, this would probably be of little benefit in the Z axis,
since the PCB serves as a solid barrier and the silicone would smash the
components onto the PCB as it expands.
Bob Wilson
>
>Bob Wilson wrote:
>> Your suggestion has one VERY fatal flaw. No resin
>> is hermetic. Water vapour will slowly diffuse
>> through it. As I mentioned earlier, this is not a
>> problem is there are no voids in the resin since
>> the water exists only in the molecular state, not
>> as a liquid. But you are suggesting a honking big
>> void in the center, within which your circuit
>> resides. Over time, water vapour will slowly
>> diffuse into this void, and can condensre into
>> liquid water. This, coupled with the voltage
>> diffentials present will literally DRIVE a
>> corrosion reaction. Thie is not theory! I have
>> seen it happen.
>
>
>Excellent info - duly noted.
>
>What's the timeline you've observed - are we talking about product life
>expectancy of weeks, months, or years?
It is all dependent on ambient humidity and temperature (as well as the type
of resin used). I have seen trouble occur in a matter of some months.
>Any ideas what catalyzes the diffusion process - does it just occur
>naturally out of a need to balance humidity, or is it caused by
>controllable factors like thermal cycling or operating temperatures?
Nothing "catalyzes" diffusion! It is a natural process that occurs with
(especially) organic materials. It results from the "spaces" between organic
molecules being much farther apart than with metals or ceramics, thus small
moledules like water can diffuse through. Some are worse than others.
Silicones as a class are by far the most permiable. The worst are RTVs which
by definition MUST be very permeable since they cure by absorbing water
vapour from the air. If they were truly hermetic, once the outside sin
formed, the rest of the resin could never cure!
>Would conformal coating provide an adequate hermetic barrier?
Of course not! A conformal coating is an organic polymer just like any other
polymer. As mentioned, some types are worse than others as a result of their
natural characteristics. The only thing that maks conformal coatings work is
that if they are applied correctly, with no voids or bubbles, water vapour
will diffuse through them, but it is harmless since there are no cavities
for it to collect and condense into droplets of liquid water. In this
regard, it should be pointed out that water vapour is one of the easiest
molecules to diffuse trhough a polymer since it is much smaller than a
molecule of N2, O2 or CO2. Only helium and argon (which are monoatomic) are
smaller.
Bob.
Richard
> The only thing that maks conformal coatings work is
> that if they are applied correctly, with no voids
> or bubbles, water vapour will diffuse through them,
> but it is harmless since there are no cavities for
> it to collect and condense into droplets of liquid
> water.
So, the concern with diffusion is not with encapsulation itself, but
rather the presence of voids in the encapsulation process that allows
for condensation. If so, then preventing voids during encapsulation
would avoid this issue?
IOW...
* Use a low-viscosity resin (to permeate small crevices)
* De-air the resin prior to casting (to remove bubbles)
* Vacuum the mold during casting (to remove voids)
So, water molecules may permeate the resin and come in contact with the
circuit traces, etc. in vaporous form, but with little effect.
(Similar, I would assume, to a properly applied conformal coating, which
it seems water would permeate under normal conditions, yet have no
effect on behavior.)
Bob Wilson
>
>Bob Wilson wrote:
>> The only thing that maks conformal coatings work is
>> that if they are applied correctly, with no voids
>> or bubbles, water vapour will diffuse through them,
>> but it is harmless since there are no cavities for
>> it to collect and condense into droplets of liquid
>> water.
>
>So, the concern with diffusion is not with encapsulation itself, but
>rather the presence of voids in the encapsulation process that allows
>for condensation. If so, then preventing voids during encapsulation
>would avoid this issue?
Precisely. This is why vacuum de-airing and careful pouring of the resin to
avoid entrapment of air during pouring (not easy with thick gudge like
epoxy!) is critical.
>IOW...
>* Use a low-viscosity resin (to permeate small crevices)
>* De-air the resin prior to casting (to remove bubbles)
>* Vacuum the mold during casting (to remove voids)
Low viscosity is nice, theoretically, but in practice most suitable resins
are quite viscous. Those that are non-viscous (e.g. polyester) are generally
unsuitable for this purpose. Of course, there are viscous resins, then there
are REALLY viscous ones! Natually your choice should lean toward the
lesser viscosity if possible. Heating the resin drastically reduces
viscosity, but also reduces pot life. In my experience there can often be
an optimal temperature that reduces viscosity while maintaining an
aceptable pot life. Experimentation is in order. The other points are right
on.
>So, water molecules may permeate the resin and come in contact with the
>circuit traces, etc. in vaporous form, but with little effect.
>(Similar, I would assume, to a properly applied conformal coating, which
>it seems water would permeate under normal conditions, yet have no
>effect on behavior.)
Exactly. Water only acts as "water" when it is a liquid. In molecular (or
vapour) form it is basically just another harmless gas.
Bob.
Richard
> >So, the concern with diffusion is not with
> >encapsulation itself, but rather the presence of
> >voids in the encapsulation process that allows for
> >condensation. If so, then preventing voids during
> >encapsulation would avoid this issue?
>
> Precisely. This is why vacuum de-airing and careful
> pouring of the resin to avoid entrapment of air
> during pouring (not easy with thick gudge like
> epoxy!) is critical.
Good to hear (though probably non-trivial in practice). So far, one of
the materials I've found has a viscosity of 600-700. While not as fluid
as water (0), it's far better than some of the resins I've found with
ratings in the 10,000+ range (sounds like this is closer to syrup).
Actually, I was surprised to see the viscosity scale using water as the
zero- reference. I suppose the scale must go negative, because surely
there are liquids with lower viscosity / surface tension.
> >IOW...
> >* Use a low-viscosity resin (to permeate small
> >crevices)
> >* De-air the resin prior to casting (to remove
> >bubbles)
> >* Vacuum the mold during casting (to remove voids)
>
> Low viscosity is nice, theoretically, but in
> practice most suitable resins are quite viscous.
> Those that are non-viscous (e.g. polyester) are
> generally unsuitable for this purpose.
OK... what's the issue with polyesters? (So far, I haven't been looking
at the composition of the resins, but they're all recommended for
electronic embedment.) Is polyester conductive?
> Heating the resin drastically reduces viscosity,
> but also reduces pot life. In my experience there
> can often be an optimal temperature that reduces
> viscosity while maintaining an aceptable pot life.
Good tip - thanks.
Bob Wilson
>
>Bob Wilson wrote:
>> >So, the concern with diffusion is not with
>> >encapsulation itself, but rather the presence of
>> >voids in the encapsulation process that allows for
>> >condensation. If so, then preventing voids during
>> >encapsulation would avoid this issue?
>>
>> Precisely. This is why vacuum de-airing and careful
>> pouring of the resin to avoid entrapment of air
>> during pouring (not easy with thick gudge like
>> epoxy!) is critical.
>
>Good to hear (though probably non-trivial in practice). So far, one of
>the materials I've found has a viscosity of 600-700. While not as fluid
>as water (0), it's far better than some of the resins I've found with
>ratings in the 10,000+ range (sounds like this is closer to syrup).
10,000 cps isn't just viscous like syrup, it is more like tar! A few years
ago, a co-worker was tasked with selecting a suitable epoxy for potting.
When asked why the one he picked was, in fact, the one he picked, he said it
was because the sales rep told him it was the one most recommended. Nothing
like thinking for one's self! Anyway, it had a viscosity of something over
10,000 cps. I told him that it was totally unsuitable because of its
viscosity. His response was that it MUST be suitable because otherwise why
would the sales rep recommend it so highly (brilliant reasoning). So my
response was to open the can and hold it upside down while at the same time
launching into a 4 or 5 minute diatribe about how one must actually
investigate things one's self, instead of taking the word of a sales rep
(etc. and so on, for 4 or 5 minutes). All the while I was holding the can
upside down in front of him as he watched it nervously. After this time the
stuff still hadn't moved much, I asked him exactly how he expected to
work with the stuff. He got my point.
>> Heating the resin drastically reduces viscosity,
>> but also reduces pot life. In my experience there
>> can often be an optimal temperature that reduces
>> viscosity while maintaining an aceptable pot life.
>
>Good tip - thanks.
Bob.
Sylvan Butler
> In article <3D3B80AD...@azglobal.com>, rh...@azglobal.com says...
>>
>>Bob Wilson wrote:
>>> >So, the concern with diffusion is not with
>>> >encapsulation itself, but rather the presence of
>>> >voids in the encapsulation process that allows for
>>> >condensation. If so, then preventing voids during
>>> >encapsulation would avoid this issue?
>>>
>>> Precisely. This is why vacuum de-airing and careful
>>> pouring of the resin to avoid entrapment of air
>>> during pouring (not easy with thick gudge like
>>> epoxy!) is critical.
>>
>>Good to hear (though probably non-trivial in practice). So far, one of
>>the materials I've found has a viscosity of 600-700. While not as fluid
>>as water (0), it's far better than some of the resins I've found with
>>ratings in the 10,000+ range (sounds like this is closer to syrup).
>
> 10,000 cps isn't just viscous like syrup, it is more like tar! A few years
I think he meant "this" in reference to the 600-700 cps material. :)
> 10,000 cps. I told him that it was totally unsuitable because of its
> viscosity. His response was that it MUST be suitable because otherwise why
> would the sales rep recommend it so highly (brilliant reasoning). So my
> response was to open the can and hold it upside down while at the same time
> launching into a 4 or 5 minute diatribe about how one must actually
> investigate things one's self, instead of taking the word of a sales rep
> (etc. and so on, for 4 or 5 minutes). All the while I was holding the can
> upside down in front of him as he watched it nervously. After this time the
> stuff still hadn't moved much, I asked him exactly how he expected to
> work with the stuff. He got my point.
Sounds about like trying to encapsulate with Bondo (tm)!
rat...@execpc.com
Mix in glass microballoons from your local hobby shop.
Silicone coatings with microballoons are available commercially for
just this application. It may require a little experimentation.
Bob Wilson
says...
>
>
> Mix in glass microballoons from your local hobby shop.
>Silicone coatings with microballoons are available commercially for
>just this application. It may require a little experimentation.
For WHAT application? They are intended for thickening the mix, and/or
allowing the use of less resin. They are rigid (they are glass, after all)
and will not make the resulting cured resin any less rigid (which was what
the post was about). Another common thickening (thixotropic) filler is fumed
silica (aka "Cabosil", or "silica aerogel"), but it will not increase the
resin's resiliance either.
Bob.