Quartz crystal failiure?

Hello! I have a weird-ish question here
What causes metal encased quartz crystals to fail?
My problem is this:
I'm working on a microcontroller devboard. It's not mine, someone lent it to me. When I got it, it was kinda' old and dusty, like it stood for maybe a year or 6 months on a shelf or something. When I actually started playing around with it I noticed that I couldn't make either the built in 2x16LCD or the serial port work. I kept getting garbled characters and all sorts of weird things like that. After pulling my hair out and blaming the tracks and switches on the board itself, someone helped me figure that the quartz oscillator was to blame. One 10MHz crystal later, all problems seemed solved! That was one week ago. For two days now, symptoms similar to those caused by the bad, old crystal are begining to surface: characters missing from serial strings, slightly erratic LCD, and other random occurences of similar nature.
What's going on? Could my shiny, new 10Mhz xtal be failing? If so, why? Can crystal failiure be induced by electrical causes (and I'm not talking about sticking it into the wall socket)?

Crystals especially which are built in metal case, rarely fail if they are new. Only cause may be rusted legs. If you say your crystal is a new one, check the connections to the LCD which may habe some dry solder joints or may be some where else too. Dry solder is like to expect as the board is pretty old one.
Cheers

Thanks for the quick reply!!!

I agree. I found it hard to believe that my brand new, shiny 10meg xtal would fail after a couple of days.
However, the entire system (xtal+board+uC) worked fine between the changing of the old busted crystal and a couple of days ago. Why?
If, indeed, there were corroded or dry joints on the board, caused by the age of the board, the symptoms should have been constant. But they're not. The number of non-mangled characters that get returned from the board is getting smaller as we speak.
There's something dying on the board and I have no clue what it is...
Frankly this is kindof a nightmare...

The entire board (xtal, board, uC) hsan't been exposed to thermal or mechanical extremes or shocks. Yesterday power went out in the entire building, but that's about it...

Pav

Could it be the oscillator chip (presumably the micro)? Try the crystals with a simple oscillator circuit.

Keith

It had a powered oscillator and you have now fitted a crystal?

I would try another oscillator as the drive level may be incorrect for the crystal you have fitted. Excessive drive can fracture crystals. Correct load capacitance?

It is easier to match an oscillator to a micro than a crystal.

Problem is I don't have a replacement...

About the driving/oscillator problem mentioned above

It looks like a HC-49 type package, 10MHz, parallel mounted with 2 22pF caps. It replaced a similarly packaged but a lot more battered, presumably 8MHz crystal (the writing on the case wasn't that readable but probability seems to indicate it was a 8MHz one). As far as I understand (both crystals and oscillators are kinda' mysterious to me at the moment) there's no powered oscillator to talk about. The dev-board I'm using is a EasyPIC4 by Mikroelektronika.


Furthermore there's quite some way from the crystal to the osc1 osc2 pins, and there's 2 jumpers in between. Could it be the work of stray capacitances across the board, from under the board, or EM interference (working in an office of an electronics/automation/IT/C firm, who knows what kind of noise I'm getting)?

Added after 1 hours 2 minutes:

I've been working on the board this morning for an hour and a half... The problem hasn't resurfaced... Serial comms work smooth, and the LCD displays propperly...

You know, I kinda' wish it DID resurface, 'cause this is just kinda' too annoying...
It's not that I'd believe in some supernatural explanation, and that would make the whole occurence mysterious or something... It's just that the darned thing got me started on trying to "fix" it, or finding out what causes it and now it just... left... The nerve...

A bit obvious, but worth checking: are the jumpers in the right position? Are you sure the configuration bits are set correctly for the clock?

Keith.

Yep... They're (jumpers and all) ok...

I haven't touched the code since yesteday when it DIDN'T work as expected (severe mangling of strings), and the jumpers for weeks.

Yesterday when I left I put the board back in its box and this morning I took it out and put it on my desk. I can only figure that the movement dislodged some particle of foreign material (the board is rather old, and rather dusty... also a bit... sticky... in some parts) that caused... some sort of... problems... with the tracks... and the crystal... and stray capacitances... Yeah, that's what it was...
I have no clue what happend since yesterday, but if similar problems arise ever again, I'll just rebox and then unbox it

Corroded jumpers? If they are tinned rather than gold plated then they could be making a bad connection? Even if the jumpers are gold the shorting link may not be.

Keith.

u have failure only when serial communication and display it at lcd rite??
try to check the max232 chip if the board use that chip..

just suggest..

lockman_akim said:

u have failure only when serial communication and display it at lcd rite??
try to check the max232 chip if the board use that chip..

just suggest..

Click to expand...

Don't have a problem anymore... it just went away... somehow...

keith1200rs said:

Corroded jumpers? If they are tinned rather than gold plated then they could be making a bad connection? Even if the jumpers are gold the shorting link may not be.

Keith.

Click to expand...

The jumpers don't look bad... But might be... I doubt I'll actually ever know the REAL cause...
I should also probably trim the crystal's terminals... It's not only placed in a socket (another probable source of trouble), but hovering about 1cm above the board because, in my haste to test the (then newly working) serial port, I simply neglected to do that...

Edit:
You know, I was thinking of getting an EasyPIC6 devboard for myself, because to me they seemed hugely configurable and packed with features, but seeing what happend to this one, I'm beginning to have doubts about the whole concept. I mean it seems to me that all that configurability comes at the price of reliability via all those sockets, jumpers and dip switches.