Recently I got into a discussion on assembly techniques:  The project was an item requiring a hermetic seal; basically wanting to keep dry.  It was that little computer that lives behind the billboard that tells you how much money you won’t win in the lottery. What is the best (read cheapest least capital intensive) method?

This is the designer/consultant’s Full Employment Act.  No requirements about taking it under a 100 meters of water, but the product would be exposed to temperature, humidity, rain, snow, dirt, spiders etc.  It just sat there receiving phone calls from the lottery’s computer and telling the sign one hundred feet off the ground what numbers to display.

The first consideration is whether this seal has to be permanent or temporary.  Permanent means you put it together and it stays together.  Temporary means either a one time (disposable) use or something you could open and close to repair the inner workings.  In the closed position it had a seal.  Being able to open it might allow repairs.  Because of the internal snazziness of certain components it was two parts that needed to be attached together. There are all sorts of ways to do make it hermetic:

MECHANICAL BONDING
Screws/bolts – this is a great way to make a hermetic seal (that mandates a gasket of some kind along with the appropriate retaining grooves and some heavy algebra calculating the crush on the gasket as the piece is put together.  But it is component and labor intensive.

Snap-fit/crush pins through a gasket.  All the design complexity of the Screws/bolts idea without the screws and bolts.  These can be designed as ‘one way’ meaning if you want to disassemble it you need to break it, or the snaps can be designed to push out of the way so that you can separate the two parts.

CHEMICAL BONDING
Adhesives/solvent welding – good but time intensive and with the concern that the bleed over might mess up the internal components unless the design facilitates someplace for the bleed-over to go. These must be jigged (clamped) in position until whatever you are using as an adhesive has ‘set’.

THERMAL BONDING
Ultra sonic/vibration welding.  This is an excellent technology to make hermetic seals.  However the imparted vibrational energy is not particularly friendly to printed circuit boards or their components.  Since this is a melting technology you also need a place for the bleed-over to go unless it isn’t a cosmetic concern.

Electro-magnetic bonding.  Here we have a ‘gasket’ made from the material but heavily loaded with ferrite.  We put it near and intense alternating magnetic field, the particles vibrate to the point of melting the gasket and the parent material and Viola!  Hermetic seal. Cool!  Very controllable, the darling of the medical device folks because it’s clean.  However just like the vibration welding techniques, it is not particularly friendly to electronics.

Hot plate welding.  This very simply asks you to make to ‘ribs’ on your sealing surface proud.  You then make (have someone make) a contoured heated plate.  Both halves are jigged together and separated only enough to allow the heated plate in.  The parts are then put in contact with the heated plate until the ribs are melted.  The plate is quickly removed and the parts are pushed together.  Your design needs to allow for the bleed-over to go somewhere (usually this is a designed in groove on either side of the ribs OR you just let it smoosh on the inside and outside.  The key here is also the limit you need in any thermal welding.  You need to have the machine designed not only to provide pressure to hold the pieces together but it must also be mounted with stops so that the melt pool isn’t pushed away.

In all thermal bonding techniques it is equally important to get a good melt pool AND keep the parts positioned long enough for the melt to cool.

NOTES:  Mechanical bonding is only as strong as what is holding it together. Under vibration or thermal softening many mechanical techniques fail.  Remember that plastic creeps, you’ll either need to install metal ultra-sonic insets and use some kind of locking screw to keep the screw from backing out or the tri-lobal screws where the plastic will creep back around the screw holding it in place.

Chemical/Thermal bonding is usually considered a one-way assembly technique. To test the quality of the weld, break the unit apart at the welded/glued interface.  An adhesion bond is defined as being able to equally pull the parent material from each of the two halves when it is broken.  This therefore makes the quality of the weld a direct function of cleanliness.  Greasy fingers will leave an ‘interface’ layer where your ultra-sonic/thermal melt pool won’t directly adhere plastic to plastic.  This microscopic seam line will destroy the hermetic nature of the seal.

What did my client end up doing?  Actually kinda funny:  He put the electronics in one half of the plastic part then covered it in Urethane potting compound.  Once it had set up, he then solvent welded the two parts together.  Not only was it waterproof, it could probably survive being hit by a meteor!  I couldn’t convince him either would do the job unless there was a hidden agenda somewhere.  I will concede it passed environmental exposure, water immersion and heat soaking with flying colors.

I had another client who was molding a little ‘rat fink’ load cell in GF Nylon coupled with a micro radio transmitter.  This little one inch cube was fixed to the hub of the tire for a long-haul trailer.  This gizmo would tell whoever read the radio signal, how fast the trailer was going, how long it stopped, acceleration and deceleration.  What fascinated me on this project wasn’t that it was Big Brother looking over the trucker’s driving style, it was how to put a plastic/electronic product on the hub of a truck tire and have it survive the abuse those things see.

In both cases what was important wasn’t the attachment technique.  It was whether the gizmo would survive its use environment for whatever the manufacturer considered an acceptable lifespan (defined in years for the billboard computer and thousands of miles for the truck thingy).  Strong doesn’t mean tough.  Strong means durable or (dare we say it?) Fitness for use.

End use testing is the ‘D’ in R&D that most designers find boring because they are usually not very good at designing real world testing and properly interpreting the (statistical) results.  If you’re in that boat; your friendly community college, university or the American Society of Quality Control can dig up a Geek who can help your design the test and then tell you in small simple words what the results gave you.

= = = = = = =

This article, like all the other ones, is virtual.  You can go to the Engineers and tell them they need a few more line items in their budget to verify their designs (remember FEA is supposed to be based on worst case tests, not theoretical assumptions).  Or you can go to the middle managers and marketeers who’ll tell you it will ruin their profit picture and product recalls are why the company has a legal department.  (you disloyal, nay-saying dummy!) OR you can step up to the plate and do something. OR….. You can read it, nod your head knowingly and feed the printout to the squirrels outside as insulation for the cold winter and go back to sleep.

Your choice.

Trackback URI | Comments RSS