C0G ceramics vs humidity

You can do it with ppm accuracy if you don't insist on sine wave excitation.

John,

You are ABSOLUTELY right John. FR4 is something like a sponge facing harsh climatic environments (humidity, salt spray ...etc.) ! Right you are. All experienced engineers have to *NOT FORGET* this !

On VME's cards embedded on railways rolling stock, that is one of the most stringent requirements, believe me.

Habib.

Properly cleaned FR4 boards don't seem to have leakage current issues, unless they get really dirty. But the dielectric constant can drift with humidity, which will only matter in things like precision oscillators.

The footprint of a pad of an 0603 part is so small that the PCB capacitance seldom matters.

Have you seen FR4 problems in the railroad systems?

John,

May be you have no idea on what climatic conditions are on the railroad ! Or you are sensibilized on this crappy things :(

Many years ago (1992) At Alstom (here in France) after months of operational services, some VME's cards are so dirty that i did not recognize my very own design ! We have decided to coat all the VME chassis cards with polish.

The Railways systems operational conditions is not a mickey world as i mentionned sometime ago on this newsgroup.

Habib.

Yup. Once at an optics conference I had lunch with a guy who made optical inspection systems for railway rails, similar to the old Sperry Cars. He said that it took them forever to figure out how to keep the optics clean, and that the only way that worked was to leave the optical port open and blow compressed air out of it continuously.

Cheers

Phil Hobbs

At this moment i spend a long time (and be paid for that ;-) ) to re-considering some designs and physical implementations made by engineers beginners. I admit they have quick understanding, talented and skilled with large overview but no experience on systems reliability and very harsh environment conditions. These skills are not acquired at school.

"keep it simple and efficient in mind" but not simplistic are hard to acquire.

Habib.

At a previous job, my boss was a mechanical engineer. He mentioned that the design loads for some parts of automobiles, mostly suspension stuff, are based on having the car or truck survive the possible freight-train trip from the factory to the other side of the US.

He also told a story of consulting for a railroad that had some piece of equipment that kept breaking. He looked at it and told them that the stresses were such that it would always break in that same spot, no matter how much steel they welded onto it at that spot. I think he wanted to add structure elsewhere to distribute the forces more evenly, but found that this was a hard sell; they thought that fix would have to (only) involve the spot where it broke.

Matt Roberds

Yikes. You might think they would mount electronics in nice Hoffmann-type sealed boxes with hinged doors and gaskets. You know, to keep out the dirt and soot and water and loose nuts and bolts and EMI and such.

How?

^ ^ ^ The Poles are not well known for keeping things in working order... :^)

Tim

They make good filters.

Using imported Japanese war surplus Zeros, no doubt.

Of course "polish" is not the right word, i ment like "nail polish". It was obviously a specific coating material spray. Sorry.

Habib.

You have to make sure they don't gather on the right side of the vehicle, though.

Tim

Purely a joke -- "Polish" (the people) would be capitalized, "polish" (the coating) is not. We spell them the same (although, I suppose the French have different spellings for these things, so it's okay if you missed it!). English being a bastard language, and all that. :^)

Tim

My English is not so perfect i could understand jokes in English or English American ... someone here has pointed my English syntax is awful and he has politely said "If you don't speak it then write it !" he was right .. but i try my best of it.

Habib.

John Larkin writes:

With a charge pump / charge balancing circuit basically.

E.g. drive one end of the capacitor with a crystal-derived square wave.

Other end is switched alternately to ground or to the virtual ground of an opamp integrator using an analog switch.

Average current = f.C.V.

A precision resistor across the integrator capacitor makes the opamp output voltage accurately proportional to average current. You can then measure the output voltage (which is ratiometric to the Cx drive voltage).

Or you can switch an opposing current into the integrator to keep the output from railing, the averaged duty of that switch is your result.

It is all ratiometric to the excitation driver supply so that the absolute accuracy is determined from "first principles" by the crystal frequency and a precision resistor value.

This shows the basic idea. The real circuit can be improved by delaying the steering switch signal and adding some RCs to smooth out discontinuities.

Oscillator frequency f Feedback resistor R Output duty cycle D

Cx = D/(R.f)

Version 4 SHEET 1 1664 680 WIRE 544 -224 352 -224 WIRE 832 -224 624 -224 WIRE 352 -144 352 -224 WIRE 384 -144 352 -144 WIRE 576 -144 448 -144 WIRE 416 -112 -240 -112 WIRE 496 -112 416 -112 WIRE 496 -96 496 -112 WIRE 416 0 416 -112 WIRE 496 0 496 -16 WIRE -240 16 -240 -112 WIRE -112 16 -160 16 WIRE -64 16 -112 16 WIRE 48 16 0 16 WIRE 144 16 48 16 WIRE 192 16 144 16 WIRE 352 16 352 -144 WIRE 352 16 272 16 WIRE 384 16 352 16 WIRE 832 16 832 -224 WIRE 48 32 48 16 WIRE 576 32 576 -144 WIRE 576 32 448 32 WIRE 656 32 576 32 WIRE 768 32 736 32 WIRE 384 48 352 48 WIRE -224 80 -224 64 WIRE -224 80 -272 80 WIRE -176 80 -176 64 WIRE -144 80 -176 80 WIRE 208 80 208 64 WIRE 208 80 160 80 WIRE 256 80 256 64 WIRE 304 80 256 80 WIRE -272 96 -272 80 WIRE -224 96 -224 80 WIRE -176 96 -176 80 WIRE 208 96 208 80 WIRE 256 96 256 80 WIRE 304 96 304 80 WIRE 352 96 352 48 WIRE 48 112 48 96 WIRE -240 144 -272 144 WIRE -112 144 -112 16 WIRE -112 144 -160 144 WIRE 144 144 144 16 WIRE 192 144 144 144 WIRE 304 144 272 144 WIRE -272 160 -272 144 WIRE -144 160 -144 80 WIRE -16 160 -144 160 WIRE 160 160 160 80 WIRE 160 160 -16 160 WIRE -16 176 -16 160 FLAG 496 0 0 FLAG 416 64 0 FLAG -272 96 0 FLAG -16 256 0 FLAG 304 96 0 FLAG -272 160 0 FLAG 48 112 0 FLAG 304 144 0 FLAG 352 96 0 SYMBOL voltage 496 -112 R0 SYMATTR InstName V1 SYMATTR Value 3.3 SYMBOL Digital\\schmitt 768 -32 R0 WINDOW 3 -266 129 Left 2 SYMATTR Value Vhigh=3.3 Vt=1.65 Vh=1 td=1n SYMATTR InstName A1 SYMBOL res 752 16 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R4 SYMATTR Value 100 SYMBOL cap 448 -160 R90 WINDOW 0 -29 -12 VBottom 2 WINDOW 3 -31 -15 VTop 2 SYMATTR InstName C2 SYMATTR Value 100n SYMBOL res 528 -208 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 0 56 VBottom 2 SYMATTR InstName R3 SYMATTR Value 20k SYMBOL Misc\\signal -16 160 R0 WINDOW 3 -29 124 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR Value PULSE(-1 1 5u 10n 10n 5u 10u) SYMATTR InstName V5 SYMBOL cap 0 0 R90 WINDOW 0 -10 30 VBottom 2 WINDOW 3 35 30 VTop 2 SYMATTR InstName C1 SYMATTR Value {Cx} SYMBOL sw -144 144 R90 WINDOW 0 78 74 VRight 2 WINDOW 3 52 12 VRight 2 SYMATTR InstName S1 SYMATTR Value FSA4157 SYMBOL sw -256 16 R270 WINDOW 0 25 92 VRight 2 WINDOW 3 56 93 VRight 2 SYMATTR InstName S2 SYMATTR Value FSA4157 SYMBOL sw 288 144 R90 WINDOW 0 48 95 VLeft 2 WINDOW 3 75 95 VLeft 2 SYMATTR InstName S3 SYMATTR Value NLAS4599 SYMBOL sw 176 16 R270 WINDOW 0 19 126 VLeft 2 WINDOW 3 56 103 VLeft 2 SYMATTR InstName S4 SYMATTR Value NLAS4599 SYMBOL cap 32 32 R0 SYMATTR InstName C3 SYMATTR Value 10n SYMATTR SpiceLine Rser=.01 SYMBOL Opamps\\UniversalOpamp2 416 32 R0 SYMATTR InstName U1 TEXT 448 264 Left 2 !.tran 0 .020 .010 startup TEXT -200 -200 Left 2 ;PR030 MCU Controlled RH Sensor Interface TEXT 448 144 Left 2 !.model NLAS4599 SW(Ron=50 Roff=1000Meg Vt=0)\n.model FSA4157 SW(Ron=1 Roff=1000Meg Vt=0) TEXT 448 200 Left 2 !.PARAM Cx 200p

The real problem, at PPM levels, is going to be replicating those ideal switches. And the DA of C3. And TCs everywhere.

An AC bridge cancels out almost everything but the humidity effect that we want to characterize.

The English language has hundreds of thousands of words. As if that weren't bad enough, each word can have many different, unrelated meanings. And we keep inventing more words!

The ambiguity of English is great for jokes and puns. Puns are probably impossible in Vulcan or Klingon.

"Shakespeare is good in English, but you need to read him in the original Klingon ..."