Of course the simple answer of how to take care of ALL the other
uncertainties, is to just make a 6.2 volt reference such as something like
the attached.
It need have NO other critical parts except the Zener that limit it's
performance.
AND when using a good linear 8 digit voltmeter, a 6.2 volt ref can be as
accurate as a 10 volt standard as a calibration point.
ws
Hi Warren,
That does seem a very simple way to get the diode at a steady temp, but then
you need to keep the rest steady to stop all the other uncertainties
swamping the nice steady diode you have just created.
From: warrensjmail-one at yahoo.com
It is well known that by changing the nominal operating current of a 1N825
type reference diode, that its voltage turn over temperature point can be
moved.
Usually the turn over temperature can be set to the nominal operating
temperature, there by giving a Zero TC reference voltage over a limited
temperature range.
For the parts I'm testing, the voltage changes less than 0.1 PPM with a 5
deg F temperature variation, and 1PPM over a 20 deg F range when the zero
TC current is set for room temperature.
Not bad, but being an extreme Volt Nut, I always want to do better.
Two ways to make it better, Add an X^2 (second order) temperature
compensator or put the reference in an oven.
What I did was make a very simple Mini-Oven using an 1N825, a glass
thermister & a 1/4 resistor,
all heat shrink together in a small package about the size of a 1 watt
resistor.
Now by Appling about 1/8 W nominal power to the resistor from a simple
temperature controller, the thermistor can be held at a constant
Temperature.
This allows setting it up so that normal temperature changes have true
ZERO effect on the zero TCed zener voltage.
Next thing I want to make is a simple times 8/5 digital gain amp so that
the gain resistors will have no effect when amplifying the 6.25x volt
zener up to 10.000000 Volts.
(or times 10/7 for use with other types of voltage references)
ws
Of course the simple answer of how to take care of ALL the other
uncertainties, is to just make a 6.2 volt reference such as something like
the attached.
It need have NO other critical parts except the Zener that limit it's
performance.
AND when using a good linear 8 digit voltmeter, a 6.2 volt ref can be as
accurate as a 10 volt standard as a calibration point.
ws
*******
Hi Warren,
That does seem a very simple way to get the diode at a steady temp, but then
you need to keep the rest steady to stop all the other uncertainties
swamping the nice steady diode you have just created.
*********************
> From: warrensjmail-one at yahoo.com
>
>
> It is well known that by changing the nominal operating current of a 1N825
> type reference diode, that its voltage turn over temperature point can be
> moved.
> Usually the turn over temperature can be set to the nominal operating
> temperature, there by giving a Zero TC reference voltage over a limited
> temperature range.
> For the parts I'm testing, the voltage changes less than 0.1 PPM with a 5
> deg F temperature variation, and 1PPM over a 20 deg F range when the zero
> TC current is set for room temperature.
> Not bad, but being an extreme Volt Nut, I always want to do better.
>
> Two ways to make it better, Add an X^2 (second order) temperature
> compensator or put the reference in an oven.
>
> What I did was make a very simple Mini-Oven using an 1N825, a glass
> thermister & a 1/4 resistor,
> all heat shrink together in a small package about the size of a 1 watt
> resistor.
> Now by Appling about 1/8 W nominal power to the resistor from a simple
> temperature controller, the thermistor can be held at a constant
> Temperature.
> This allows setting it up so that normal temperature changes have true
> ZERO effect on the zero TCed zener voltage.
> Next thing I want to make is a simple times 8/5 digital gain amp so that
> the gain resistors will have no effect when amplifying the 6.25x volt
> zener up to 10.000000 Volts.
> (or times 10/7 for use with other types of voltage references)
>
> ws
