Can an FPGA be operated reliably in a car wheel?

I don't think Tin is viscoelastic. My point is that tin (and other materials) only have this plastic deformation above the yield stress limit. A thin wire stuck out horizontally has a lot of stress. A thick bar of tin isn't going to droop no matter how long you wait. Otherwise solder wire left on the desk would flatten out on its own. The confusion is that Tin and Solder have very low limits compared to other metals, and so it's easy to think the tin flows all the way down to zero stress as it's floppy by comparison with other metals. As you say, different materials have different yield stress limits, and maybe new solder materials have higher limits. The point is that you can work out how much stress the metal can take before it starts to go plastic. And screwing into a terminal block will be above that limit unless you're careful with the screwdriver! :-) Although I suspect the dissimlar metals in a terminal block is a major reason for failure. Cheers, Syms. p.s. Anyone at University wanna pop down to the material science department as ask this question? Is Tin viscoelastic?

Here's a link:- "Antique windowpanes and the flow of supercooled liquids"

Thanks for the link. The information presents some concepts with glass deformation that I wasn't familiar with but the comment "This author believes that the correct explanation lies in the process by which window panes were manufactured at that time" without support for why the thicker end would always be installed on the bottom leaves me with the issue still open. I'll nudge it more toward "maybe."

Apologies to those disturbed by how off topic this got.

So ... the real discussion is how the materials behave at 400G and more.

When solder yields at very low presures and room temp, doesn't that raise a flag when subject to continuous long term 400G radials?

The discussion about Si-O-Si bonds was interesting, and the effects of extra ions in the glass that greatly alter the strength of the glass to stress ... so what about Si doping in the substrate ... will it be stable at 400G for the long term? 400G is a pretty substantial force even on low masses.

Part of the traditional military qualification of ceramic wire-bonded parts (bonding wires essentially unsupported in the cavity) was a centrifuge test at 20,000g (10 000 g for the larger parts.) Plastic encapsulated parts are not tested this way, since the test is considered meaningless (no chance to fail). Peter Alfke

Hi John, I agree it does raise a flag, but unless the engineer does the physics, they're unable to work out whether they'll be ok or not. It seems it's not to do with the acceleration, it's to do with the pressure in the tin. The total force divided by the total cross sectional area of all the BGA balls is what we're interested in. What I'm suggesting is doing the calculation.

So, here's the sum.

Solder ball diameter 3mm. Solder ball area = 7.1E-8 m^2

Let say we're using a 256 ball FPGA. Total xsectional area = 256 * 7.1E-8 m^2

= 1.8E-5 m^2

Ok acceleration = 400g = 4000 m/s^2

If the yield stress limit of tin is 10MPa ref.

to remain under the yield stress limit (YSL)

Sn YSL > accel * mass-of-FPGA / X-sectional area

Therefore the FPGA mass must be less than 10E6*1.8E-5/4000 i.e. less than 45 grams.

To be pedantic, 400g is a pretty substantial _acceleration_ on a mass!

The upshot is, if you want to put an unclamped FPGA in a tyre, I suggest your FPGA has big balls. (With apologies to AC/DC!)

HTH, Syms. :-)

Let's remember that the g-forces have a direction (outward), and it is up to the pc-board designer to take advantage of this. Peter Alfke

Peter, You raise an interesting point. I wonder if the yield stress limit is the same for compression as for expansion? Sod it, just epoxy the damn thing to the board! :-) Cheers, Syms.

Well someone please go turn one upside down, measure it with a micrometer, and make a note of it, so at least our grandkids will be able to settle this.

(make sure it's not near where said grandkids are likely to be playing ball in the years before they gain an appreciation for experimental science)

Symon,

You have to be careful if you have differing thermal coefficients of expansion (if you epoxy it to the pcb).

If the part heats up, and the pcb heats up, you would like the coefficients of thermal expansion to all be the same, so you do not shear the solder bumps off the pcb.

It is a much bigger problem with flip chip, to match all the coefficients and mount the die to the substrate such that it will tolerate many years of thermal cycling without cracking and of the solder bumps.

I suspect this is a wire bond epoxy/plastic based molding compound type of application, where the FPGA package is of similar material to the pcb, so that thermal stresses should be minimal.

Why no use a RF link to send the data from sensors in the wheel to somewhere else?

How the heck does one power a pcb in a wheel? Sounds like there are a ton of problems to solve. How do you communicate with the wheel? "Yo, wheel..."?

Austin

Sym>

On a sunny day (Mon, 08 May 2006 11:11:38 -0700) it happened Austin Lesea wrote in :

You do not want the cellphone to jam yyour brakes....

Rotary transformer (classic solution, can also be used for data transfer, basically 2 halves of a pot core. One could make a mechanical generator with some mass, every time speed changes it would turn, rotate a magnet, like those old automatic mechanical rewind watches. I'd go with the transformer.

Epoxy adds mass. Does the improved tolerance to shear stress (assuming lateral g-force here) outweigh the effects of added mass? (I like the term "outmass" versus "outweigh")

Or, "Lots of Balls" .. :)

-jg

Peter....we've told you a million times not to exaggerate!

Bob

That's what centrifuges are made for. They were invented to separate butterfat from milk, but now they also separate Uranium isotopes from each other. The automobile wheel may spin at 30 revs/sec = 2000 rpm, but centrifuges can do 100 000 rpms... Peter Alfke

I just looked it up:

Austin Lesea asks many questions:

Hi Austin,

I can comment on these for you:

Indeed we do.

With a battery(!)

Oooh yeah - we have fun testing these things!

Like this (apologies for the enormous link!):

Cheers, Martin

I have designed the controls for several small centrifuges, it takes about

I will probably be doing faster ones in the future getting towards

--

Martin,

Thanks!

Aust> Austin Lesea asks many questions: