LM285-2.5: LM285BXMX-2.5/NOPB Vref shifted after aging

Hi Chin,

I agree with your observation that this can be aging but at the same time this can also be thermal hysteresis.

In most precision voltage references we have specified the long term drift and the thermal hysteresis.

Some of the older devices do not have all the specifications. In this case the LM285 has long term drift but it does not have thermal hysteresis.

Long term drift:

Long term drift affects all semiconductor devices such as voltage references.

On the device datasheet we specify the long term drift for 1000hr at 25C because from our testing this is where most of the changes occur. In our higher precision voltage references we show 1000hr and 2000hr and the difference of the 1000->2000hr is usually 1/5 or better of the original. One way to eliminate this is to use the device for 1000hr at 25C but this is unreasonable. A common technique to eliminate most of the effect is to bake the devices. What I often see is that before the device is in use the device undergoes 85C+ temperatures for several hours to eliminate the majority of the long term drift effect. This is because there is a Acceleration Factor Coefficient that says that 1000Hr at 25C is equivalent to less hours at higher temperature. The exact formula is in the blog below.

In this case you can look at the long term stability line on the EC table. All devices on average shifted 20ppm. This can be 20ppm up or down from 2.5V per 1000hr at 25C.

To know more about long term drift or how to calculate the change through EoL, you can check out the blog below.

IC long-term stability: The only constant is change

Some of the most frequently asked questions we get in the Precision Amplifiers forum here in the E2E Community™ have to do with an IC’s long-term stability of various parameters. Nothing in nature is static…

By Marek Lis over 11 years ago in Archives > Precision Hub

Thermal Hysteresis:

After initial calibration if the device goes under thermal excursions such as 25C -> 85C -> 25C this will affect the output voltage. But just like long term drift, after the first cycle the 2nd and 3rd will be minor.

The way I would see deviations on VREF is below.

Device from factory error: initial accuracy

Post solder: initial accuracy + solder shift

System testing with temperature: initial accuracy + solder shift + long term drift + thermal hysteresis.

Depending where you calibrate will determine what errors will remain.

-Marcoo