Gound plane resistance calculator

Haha "real live calculus"! I did some for a project, it felt like the first time in 20 years. And it was that same problem! It was for thermal, but same equations I think - the temperature rise of a circular package soldered to a ground plane. There was definitely a ln(r1/r0) in there :)

I tend to do a lot more math than most other engineers I know, 'cause I end up doing a lot of weirdo systems engineering stuff for which the science needs to be invented as we go along. It's fun, when I'm not beating my head against the wall.

Can't you model it in pspice by modelling the ground plane as a grid with resistors:

Dunno how many nodes Pspice can cope with tough.

Joerg schrieb:

Hello,

if you want to get a nice and precise ohms value, you need to know not only the exact shape of the ground plane and the location of the two points, but you also should know the thickness of the ground plane. Unfortunately, there are large tolerances for the copper thickness, a 35 µm layer may be anything between 45 and 30 µm, for galvanic copper even between 60 and 25 µm. The thickness is even not constant over the whole plane.

Bye

So a "sheet" problem could be solve with Spice... just make up an appropriately shaped mesh? Of course the question is, what values to make a "triangular mesh" equivalent to Ohms-per-square ?:-) ...Jim Thompson

Lots! Make a subcircuit called a "square" ;-) And paste them together. ...Jim Thompson

Well, it ain't a perfect square :-)

In fact, it isn't even rectangular.

Yes, sure, and I have that data. Well, sort of. I was pressing for that information from the manufacturers and they didn't have it (!). Could not believe it.

Thanks, Tim, that was very kind.

There'd be the chance of a wee problem with the TSA folks. "What's this large copper thang?" ... "It's to shield from evil rays from outer space" ;-)

In theory. But one of the nice things about FEM tools is that they'll automatically generate the mesh.

I don't know (I'm no FEM pro) -- but clearly the FEM tool folks do!

It turns out, in LT Spice, that you can make a rectangular grid of resistors by just touching the ends to one another... no wires. If you leave two sides of an array with dangling ends, you can copy and align them and they stick together to build bigger and bigger grids.

But it's messy.

John

It also doesn't converge to the right answer very fast as N increases.

Cheers

Phil Hobbs

It would take a lot of resistors to, say, make a decent circular contact to the sheet. I *said* it was messy!

I like the ATLC approach: use Paint for input/output. On the input side, colors represent (to ATLC) different dielectrics and conductors. On output, you get a text summary of impedances and such, and a neat colored potential map.

John

Good idea. Just like I (so far) did all my mechanical drawings using my schematic editor because for the life of me I can't get used to the icons and hieropglyphs of my ME CAD.

I sometimes use PADS-Logic just for drawing. But Visio (I'm running the 2002 version, cheap on ebay) is really not a bad drawing program, pretty easy to learn. If you export an emf file into Word, the linework is excellent.

John

What's a "Gound plane?"

;-)

What?!? You are unfamiliar with the work of that noted electromagnetic pioneer, Erasmus Qwincy Gound, illegitimate son of James Clerk Maxwell and Louis Pasteur's unhappily married younger sister?

_What_ is the world coming to?

He meant Gourd Plane. That's a cargo aircraft that's used to ship vegetables.

John