cool dc/dc converters

Depends on the epoxy. Something like SU8, which is fully cross-linked, is totally solvent resistant and very hard to etch, even with an aggressive RIE. It basically has to be removed with a torch.

Cheers

Phil Hobbs

Some epoxies have pretty high glass-transition temps., so you might need a deep fryer (or a solder pot!):

But, hopefully those super-rigid varieties aren't being used for potting. That's component abuse.

I learned a new (for me) epoxy trick recently. Those rubbery, 5-minute epoxies can be cured rock-hard and much stronger by a gentle (60C?) minute with the heat gun.

Cheers, James Arthur

Yup. An hour at 100C improves epoxy remarkably.

Cheers

Phil Hobbs

This particular dc/dc is potted with what looks to be an epoxy-sand mix. It defies a band saw and isn't affected by a hot soldering iron.

I've wondered about mixing. You dump a puddle of A near a puddle of B and mix a bit with a popsicle stick or a toothpick or something. Presumably molecules of A have to get very close to molecules of B to get a reaction. Can a little stirring get all of those molecules all mixed up? Why stir for 5 seconds, and not for an hour? Or a month?

Does set epoxy actually include regions of unmixed, uncured A or B? It doesn't seem to.

Hey, that's nasty. Sounds like you'll need several scut bunnies...

:-)

Cheers, James

I think the cure process must depend on diffusion to some extent. Stirring gets the two components mixed well enough that the diffusion time between a molecule of A, and molecules of B, isn't dreadfully bad.

Heating the mixture would increase diffusion speed and thus speed the curing process.

Dang, there's a scut bunny shortage already.

Mixing is a bit like mille-feuille pastry: if you do it right, the thickness of the layers tries to go down exponentially with the number of folds.

It can also be done wrong.

There was a demo someplace like the Imperial Science Museum with two concentric glass cylinders with oil in between. A narrow stripe of blue-dyed oil bridges the gap. With a crank, you can rotate the inner cylinder, which smears out the blue stripe till it looks uniform.

But then you rotate it back the same number of turns, and the stripe reappears--no mixing actually occurred!

Cheers

Phil Hobbs

-Lasse

Yup, like that.

Cheers

Phil Hobbs

Hmm I think if you left it in the mixed state for some time. (... it depends on the thickness, number of turns. diffusion coef.)

And then "un mixed it" it would be more diffuse, than if you'd let it be.

George H.

Yes, in order for the unwinding to work, the flow must be laminar and diffusion small. That's why the demo I saw earlier used heavy oil, and Lasse's one used corn syrup.

I remember being fascinated by heavy corn syrup as a kid. It's weird non-Newtonian stuff--when you pour it, the bulk equilibrates much faster than the fine surface features made by the last thin dribble. For some reason, honey seems to be more Newtonian. In laminar flow, you don't get the mille-feuilles effect--the layer thickness goes down like 1/shear rather than exp(-k*shear).

Fluids really are surprisingly subtle in some ways. Isn't physics fun?

Cheers

Phil Hobbs

fusion small. That's why the demo I saw earlier used heavy oil, and Lasse's one used corn syrup.

Newtonian stuff--when you pour it, the bulk equilibrates much faster than t he fine surface features made by the last thin dribble. For some reason, ho ney seems to be more Newtonian.

ness goes down like 1/shear rather than exp(-k*shear).

Well of course, (I don't do it for the money :^)

I'm sure I've shared my diffusion story. We made our own bottle of CO2 gas for a CO2 laser (mostly He and N2 with a little CO2.) How long does it take an atom to diffuse ~3 feet at pressure of 2000 psia?

George H.

Weeks, iirc. Too bad you couldn't just use a magnetic stirrer. ;)

Cheers

Phil Hobbs

For extra fun, put a clear bottle of corn syrup between crossed polarizers.