NTC thermistor sensors

[Bernd Felsche]

Brian Dessent tapped away at the keyboard with:

> Does anyone know the usual way in which ECMs measure the
> resistance of the NTC thermistors?  The most straightforward
> methods I can think of would be to apply a constant current [or
> voltage] and measure the resulting voltage [or current].  It seems
> that the constant-current method would expand the high end of the
> range and compress the low end, and the opposite for
> constant-voltage.  The usual thermistor varies from about 50 ohms
> to 50k ohms exponentially, right?  I guess it depends on whether
> you are interested in the low-end range of temps or the high-end.

Thermistors vary. The Digifant ECU for example, measures the NTC
resistance using a constant voltage applied across a resistor in
series with the thermistor. Open- and closed-circuit conditons are
then easy to detect - used to set CO, idle speed and basic timing in
Digifant servicing.

> Anyway, does anyone have any experience on this issue from
> reverse-engineering ECMs?  Also, is there a "standard thermistor"
> curve such as with thermocouples?  What I would like to do is

Yep... there are standard curves.
  R = A e ^ (B/T)

 A and B are the thermistor-specific coefficients.
 T is temperature in Kelvin
 R is resistance in Ohms.

If they're not documented (e.g. in a service manual), you can
determine A and B from 3 or more measurements using a curve fit.

> piggy-back some of the sensors currently in use by the stock ECM
> -- make their signals available for datalogging or dash-display
> while not affecting the readings that the computer sees.  If the
> ECM supplies a test current and measures resulting voltage, then
> this should be rather easy with an op-amp buffer.  If the ECM
> applies a test voltage and measures current, then a current mirror
> would seem appropriate.  However, if the test current varies (to
> aid linearity or dynamic range) then it would seem that a full
> "resistance mirror" op-amp circuit might be necessary.  Any
> advice?

With Digifant (and probably most other Bosch setups) you'll only
have to buffer the voltage to make use of the thermistor for your
own purposes. The voltage will of course disappear when the engine
is switched off, so that method is unsuitable for controling
under-bonnet temperatures after shutdown.
----------------------------------------------------------------------------
To unsubscribe from diy_efi, send "unsubscribe diy_efi" (without the quotes)
in the body of a message (not the subject) to majordomo at lists.diy-efi.org