Alumina Substrate

I don't remember shrinkage being that large, but it was 35 years ago, and we ordered the _finished_ boards from Coors... yes, the beer people... who became expert in alumina while developing pressed ceramic beads into beer filters ;-)

...Jim Thompson

Held to 80ohms +/- 5%, IIRC. The "Clark" boards the TCMs mounted to were even more beastly. I can't remember exactly, but there was something on the order of 50 signal planes, with overflows on top and bottom done in twisted pair. The machines that did the overflow wiring were amazing to watch. They put the Gardner Denvers to shame. ;-)

Bog, that sounds like the old "Multiwire" (was it) technology. Does anybody else remember?

We have trouble keeping impedances reasonable on 8-layer boards with Er=4.6.

John

Forgive the dumb question, but what's "Er"?

Thanks, Rich

Dielectric constant, relative to vacuum. FR-4 runs about 4.6, which is about 15 pF per square inch of 0.062 board, side-to-side. That makes a

Make the dielectric thinner, and capacitance goes up, so a 50 ohm trace has to get skinnier. Sandwich the trace inside between planes, and you have stripline, even lower impedance. At 8 or so layers, inner-layer traces start to get too skinny to have reasonable impedances. 100 layers of ceramic, with Er=10 maybe, boggle my tiny mind.

John

Some kind of TV series with doctors falling in love and all that sort of things ...

Wasn't that St. Elsewhere?

Oh, no, wait, St. Elsewhere was 25 years ago when things like AIDS and gay doctors were still a big deal. :-)

St. Elsewhere made doctoring look pretty cool, in the same way that Chips (at the time) made highway patrol work look fun. :-)

Yeah, Doctors and Nurses doing it on the stretchers while others are in the waiting room dying!... If you were ever in one of those rooms waiting and just happen to hear what sounded like some one screaming for life "Oh God I'm coming", you can now rest assure that seeing is believing and hearing is well, questionable! :)

Well, here I have to confess that I am TV deprived. I don't remember any of those series and nowadays I only see the last couple minutes of ER or CSI, and only because it ends just before a newscast. That's what we watch, the news, and then the power button is pushed again.

Thanks. I know what dielectric constant is (I am a real tech, after all); I just didn't know that that was what 'Er' meant. :-)

(Isn't that also the name of the George Clooney show? ;-D )

Thanks! Rich

And after you retire, you can sell swampland in Montana on late-nite teevee. ;-)

Cheers! Rich

It really should be a lower-case epsilon sub r. Er is relative to the E0 (epsilon-sub-zero), the permittivity of free space. With real materials, of course, the dielectric 'constant' will vary with frequency etc.

Best regards, Spehro Pefhany

Plus, typical FR4 has a capacitance tempco of around +900 PPM/K. That's enought to use it as a temperature sensor.

John

It's really supposed to be (Greek) epsilon (permittivity), subscript r (for relative) -- Er is about as close as you get with straight ASCII.

What did you watch growing up there in Germany? Weren't we already shipping over David Hasselhoff on Baywatch or at least Knight Rider back then? :-)

A mail room employee at Fox dies every time you say you don't believe in watching American Idol.

---Joel

I am 100% convinced that in another likfe Sherwood Schwartz would have been a used car salesman.

(More on Sherwood Schwartz) I'm amazed, the man's still alive, at the ripe old age of 91! -->

It was covered under my "etc", so it's not really a "Plus" ;-) but an interesting point in any case. So it really starts to degrade tempco of an NPO cap smaller than a hundred puffs or so. Might even be high enough to affect capacitive RH sensors a bit.

Best regards, Spehro Pefhany