"Campbell, Todd" <[EMAIL PROTECTED]> wrote:
> Does anyone have any experience or information about the upper
> limits of the Palm's (operating) temperature range? PC Computing
> did an article where one was succesfully tested to 130 degrees
> (and 180 degrees while off), but I'd like to know if anyone's
> actually discovered where the "on" limit really is.
This is something that absolutely cannot be "discovered". True, you can
take your own device and run it through various tests to determine its
ambient temperature limits of operation. But that is just one device,
from one production run, from one assembly line.
The "real" operating temperature limit is based on analysis of the
circuit, using specifications provided by the manufacturers of the
electronic components used in the circuit. Manufacturers specify
absolute limits of operating temperature. However, the device does not
have the same behavior across that entire operating temperature range.
The operating characteristics of electronic devices vary with
temperature. They also vary with manufacturing lot. A logic devices'
performance specs are given as ranges, not as a single value (e.g.
propagation delay < 5 nsec, NOT propagation delay = 4 nsec). Capacitors
are specified as having a nominal value to a certain accuracy,
specified at a given temperature, along with a temperature coefficient.
Likewise resistors.
The design engineer performs timing analyses on his circuits, using
these spec ranges, and comes up with the allowable range of operating
temperature. (Or, someone higher up specifies an operating temperature
range, and the engineer designs the circuit to meet that requirement.)
Any particular unit may (and probably does) perform acceptably at
temperatures outside the specs. However, no unit is *guaranteed* to
operate outside the specified temperature range.
The bottom line is this: regardless of the number of individual units
you test and find to work reliably outside the specified limits, you
can't be sure that the next one will.
Just this engineer's $0.02.
--
Roger Chaplin
<[EMAIL PROTECTED]>
