[time-nuts] Temperature sensors and bridge amps

Bill Hawkins bill at iaxs.net
Fri Nov 12 16:26:03 UTC 2010


As I understand it, the problem to be solved is stability to a
millidegree or some such. That kind of accuracy is not required
because the flat spot of crystal tempco is not narrow. +/- 1%
would be accurate accuracy.

For stability, you must remove all sources of variation. Self-
heating is not a stability problem, it is an accuracy problem.
Operating at the bridge null reduces the effect of bridge power
variations.

The challenge is to get enough gain from the low offset error
amplifier to maintain the required error range without having
so much gain that the loop is unstable. This usually means
taking physical design steps to eliminate dead-time or lags
in the heater control loop. Of course, you can't get to a
millidegree from ambient with just one oven. And you can't
eliminate time lags if you have any thermal mass.

Note that you have the same loop stability problems if you use
a crystal as the sensor and a counter as the detector.

I worked for Rosemount (part of Emerson since 1976), a maker of
industrial sensors including 100 ohm platinum. The four wire
Kelvin connection offers the most accuracy, but frugal industry
finds that a three wire connection is adequate for very long runs
of cable. This connection puts a lead wire in both legs of the
bridge.

Bill Hawkins

-----Original Message-----
From: Bob Camp
Sent: Friday, November 12, 2010 6:28 AM

The 100 ohm standard for RTD's dates way back. The assumption was that you
had it on a *long* run of cable (2 pair / sense leads of course). The
insulation leakage was a bigger issue than anything else in the equation. 

On Nov 12, 2010, at 5:36 AM, Poul-Henning Kamp wrote:

> This was one of the things that I wondered about:  How large currents
> are used ?
> 
> Can't be too much because that would lead to self-heating...
> 




More information about the time-nuts mailing list