Subsea Electronics to 4500psi

On 26 Oct 2003 20:41:05 -0800, snipped-for-privacy@yahoo.com (Harry Conover) Gave us:

Commercial version will have. Mil and or space versions... NOT.

They are not epoxy encapsulated, they are typically hermetically sealed, and that means no trapped air.

This deep sea design should go the long haul, and use only mil spec chips and mil parts in locations of other susceptible components.

On Mon, 27 Oct 2003 10:33:34 GMT, Spehro Pefhany Gave us:

They also make... of all things... Hospital Beds!

At least, that is what the Cincinnati facility makes.

On Mon, 27 Oct 2003 17:27:39 GMT, "Ross Mac" Gave us:

Once the PCBs are made, they are out of the manufacturing control loop. That means that packaging, and shipping, etc. takes no moisture abatement precautions. By the time the customer gets them, they can garner a lot of water. By the time the customer uses them, even more. Especially on unmasked boards, which most HV PCBs are as encapsulant adhesion is critical.

Then most contract manufacturers, and most in house processes use aqueous fluxes, and cleaning methodologies.

So, for ANY critical environmental application, ANY assembly should be baked dry, just before the conformal coating process, or full encapsulation takes place. Especially if HV is involved.

On Mon, 27 Oct 2003 22:20:41 GMT, "daestrom" Gave us:

If the encapsulant is hard epoxy, the voids you mention won't "see" the pressure. Also, the polyurethane version I mentioned is so stiff it stops bullets. Then, one can place the assembly into a hardened container that can handle the number of atmospheres it is to be subjected to with hermetic pass throughs on the conductors or cabling to and from the device. Expansion cavities, and such can be eliminated.

On Mon, 27 Oct 2003 23:32:34 -0500, **THE-RFI-EMI-GUY** Gave us:

Hard epoxy potting, yes. One must choose a material that shrinks minimally during cure, and moves minimally during hot and cold temperature cycling.

On Mon, 27 Oct 2003 03:37:07 GMT, "Ross Mac" Gave us:

How many years ago? PCB technology has likely come a LONG way, since you were in the game. Since you claim to be of millions, and retired.

PCB manufacturing involves several baking operations.

That's true....I was speaking of bare boards, which are very tightly controlled. At least for the major customers...

Well, you may have access to the specs of the epoxy, I don't. How much does it compress under 4500 psi?? If that kind of stress will cause it to deform more than the metal enclosure, then the two of them (epoxy and metal enclosure) must resist the pressure by sharing the forces. If the strain vs. stress capability of the epoxy is the same as the tank wall, the forces would be shared equally. If the epoxy strain vs. stress is a much higher, then the tank wall must be hardened. The thicker the epoxy is, the more the total deflection it will exhibit for a given force. The more the epoxy deflects, the more forces will be left to deform the tank wall.

I got the impression the OP was trying to avoid using a 'hardened' container and wanted to use a lighter weight construction. A 'soft' tank can be used as long as the pressure is kept equalized (hence the oil/fluid).

Stopping a bullet *is* impressive, but 4500 psi on an 8x10 inch board/epoxy means the epoxy has to not deform more than a tiny fraction of an inch with

180 tons on its face. A hardened tank that size might require 2 inch thick walls (or more, I'm not an ME). 306 atmospheres is a tall order for any tank design.

daestrom

That's not what hermetically sealed means.

Bad advice. Many mil-spac chips have ceramic packages with large voids in them. The pressure on the void will crck the chip open.

Why would you think that?

On Wed, 29 Oct 2003 05:08:18 +0000, Guy Macon Gave us:

I didn't say that that was all that it meant. Doh!

On Wed, 29 Oct 2003 05:08:18 +0000, Guy Macon Gave us:

I'm sure it would hold up better than a plastic dip, or epoxy encapsulated die. The die itself is the same relative hardness as the ceramic tablets used to cover them.

They are made SPECIFICALLY for high heat, pressure, vacuum, and other extreme conditions and would likely be the BEST choice. Especially if the completed assembly then gets encapsulated further, as they are best conditioned to conduction cooling, which is what fully potted designs are all relegated to. So, the lifespan would be extended as well. Putting something down deep, one wants the highest possible reliability, and the longest MTBF, cause service calls are a bitch!

On Wed, 29 Oct 2003 05:09:46 +0000, Guy Macon Gave us:

Not to the same degree. Besides, vacuum encapsulation removes all voids. Any left would be devoid of nearly all air, and therefore pressure.

External pressure on a soft, pliable media will apply more force onto such a void than will a hard cured substance, given the same external pressure. Fact. That is why submarines are made of steel instead of play-dough.

What was that you were saying?

All true except the "high pressure" part. Mil-spec chips are made to better withstand other kinds of harsh environments and would otherwise be the logical choice, but mil-spec chips are NOT designed to withstand 4500psi of pressure. In this specific case, an otherwise inferior die attach package (you have seen them if you have opened' any modern electronic toy - it looks like a blob of epoxy on the board) would be better able to withstand the pressure because it doesn't have the nitrogen-filled void that a ceramic mil-spec chip has.

Ah. I see what you are getting at. Do you happen to have any extra mil-spec ICs laying about? If so, please take a hammer and break one in half to see the internal void. Also note that the pressure problem happens when the void has something compressible in it, and that a void filled with vacuum would be compressible. What the vacuum used during encapsulation does (and mil-spec chip packages not vacuum encapsulated) is to remove all voids.

This is an accurate analysis up to the point where the force exceeds the strength of the chip package, submarine hull, or whatever. Alas, 4500psi is much to high for any IC package.

Did you or did you not write "hermetically sealed, and that means no trapped air" ?

Hermetically sealed does NOT mean no trapped air. It means that whatever gas is inside (dry nitrogen in the case of hermetically sealed military specification integrated circuits) is sealed from the outside. In other words, trapped air is an example of being hermetically sealed.

I hope that that wasn't directed at me. As you may have noticed, I have a policy of treating everyone with respect and honor, and if someone insists on making personal comments I simply killfile them and continue the conversation with those who are willing to be civil. I am aware that you insult some other people and that they insult you back, but that's not a game I am willing to play.

Remember those of is who are smart are here to learn, not to impress, and that means admitting it when you are wrong. I think that you are among the smart. It would be a shame if I were to find out that my high opinion of you was misplaced and that you are actually one of the stupid people who never admit that they are wrong and thus never learn anything new.

message

Since you are the EMI guy...what is your opinion on the life span of EMI filters....just a general question....Ross

Mil-spec parts aren't always the best alternative for commercial applications. I had a high-rel requirement for a commercial crypto product and was going to specify Mil-parts, until I got education. Ceramic, particularly the 'C' (IIRC) packages are easily damaged and will leak if the seal is cracked (easy to do with mechanical-insertion). Commercial parts were a better choice for the application.

Subjecting a cavity part to 4500PSI is simply nuts. I don't think I'd specify anything without some serious testing or well known specifications covering my donkey.