How to age a MAX6350 Voltage Reference IC?

[t1d]

I have mounted the chip on a breakout board. It has capacitor, trim pot, etc.

1) Does it need to be aged?
2) If so, how do I age it? It is in series with the DMM, so it is not on and it does not see a load unless the meters is attached. Fake the meter with a resistor across the output pins? How many Ohms/watts?

Data sheet attached. Thank you for your help.

[RandallMcRee]

As far as I know, aging simply refers to normal operation on manufacturer's datasheets.

So just keep it on 24x7 and (ideally) monitor the output for shifts.  If you see abrupt shifts that is bad and may indicate popcorn noise which typically is inherent to the device and cannot be aged out. These are on the microvolt level.

A gradual trend is the normal, expected aging.  "Hours" is the typical measurement unit for this and months of accumulated time is typical. The datasheet sometimes gives the expected shift over time, but regardless they are typical numbers and there are no guarantees from the manufacturer.

Hope that helps,
Randall

[Andreas]

I have mounted the chip on a breakout board.

Which package?
CERDIP8, DIP8 or SO-8?

After my experience at least the plastic DIP8 package needs many years to stabilize.
Attached the first 1500 days of 2 samples.

Between the lines of a Maxim datasheet: the stability is directly affected by the package.
(the more stiff and the larger the package the larger the stability).

On a AD586LQ reference I had good experiences with cyclically loading the 5V output with 3*1K resistors (15 mA) over night.
After some months the output stabilized.

So I would probably build a cirquit which draws 15 mA for 1.5 hours and 0 mA for 0.5 hours cyclically from the reference output.

with best regards

Andreas

[dietert1]

It all depends on the intended use. I mean Andreas' references seem to drift in a predictable way and could well be used for ppm level measurements if one includes a numerical prediction of drift. Except the red one exhibited regular behaviour only after one year...

Regards, Dieter

[ignilux]

Andreas-

Having had a look at the datasheet, what clues you in to the package-dependent long term stability? I see the different tempcos given, but they have the same long term drift spec and none of the figures jump out at me.

[Andreas]

Hello,

it was not in the MAX6350 data sheet but in a different one.
See e.g. MAX6126.

In the MAX6350 datasheet the hysteresis is only given for the PDIP package.
For the long term stability they simply have "forgotten" to give the package. (maybe only valid for CERDIP package?)

with best regards

Andreas

[Gyro]

It's a well known manufacturers trick. The specify the drift for the hermetic package, even when said package is long since discontinued. They don't bother to characterise the plastic ones (unless there hasn't been a hermetic one). The drift figures are "typical" anyway - prove them wrong.