OPA128: Photosensitivity?

Hi Justin,

Many precision analog integrated circuits exhibit light sensitivity and it is a factor that we must deal with when testing the die while in wafer form. It is semiconductor junctions that make up the analog integrated circuits behaving as photodiodes and phototransistors. When the junctions are exposed to light it disturbs the bias conditions and cause changes in electrical parameters such as voltage offset and bias current. Passive devices and circuits on the die are generally immune to light because they do not incorporate semiconductor junctions.

The sensitivity to light varies with the semiconductor processing, intensity and the wavelength of the light so it can be quite different for different types of analog products. As I recall, like many precision operational amplifiers, the OPA128 does exhibit sensitivity to both ambient and laser light. We use specially developed measure, trim, measure... sequences to make sure the laser is off and the OPA128 output has settled before each measurement is made during the trim process.

All of our internal characterization of light sensitivity is done in conjunction with the manufacturing processes used to process the devices. We don't do any formal light sensitivity characterization of packaged devices because it is not normally an issue.

Regards, Thomas
Precision Amplifiers Applications Engineering

Hi justin,

I remember an application where we had an unexplainable high offset voltage. Everything was tested and checked but we couldn't find the mistake. Then, suddenly the offset voltage dissapeared. What has happend? A sheet of paper was a covering a part of the circuit and prevented ambient light from illuminating a low leakage glass diode which was mounted at the input of an OPAmp for protection purpose. It turned out that the ambient light was enough to considerably increase the leakage current of this low leakage diode due to the photo effect. Yes, the low leakage diode with the transparent glass body suddenly behaved like a photo diode...

Another effect which can drastically increase the offset voltage of an OPAmp is mechanical stress. In the early days of analog electronics a widely used method for assembling expensive OPAmps were to stick them into sockets. It was the final step of the assembly and it was very carefully done as if you were breathing life into the circuit. For a DC coupled phono amplifier I once selected a handful of OP27 for ultra low offset voltage. After sticking them into the sockets I was surprised to see them producing huge offset voltages. After a while I found out that the amount of offset voltage depended on how exactly I did stick the OPAmps into the sockets. Yes, bending the package (and by this the die!) of an OPAmp can cause the generation of a considerable offset voltage! Even SMD OPAmps can show this effect: Press with a screw driver onto the top of an OPAmp in a SO8 package and you will notice that the offset voltage is changing by some hundreds of µV.

Kai