Help With Photodiode and Transimpedance Amplifier

[RoGeorge]

So I fitted up a new sensor using your suggestions.  I powered it with a 9 volt battery and still got .6 volts with my DMM.  Even walked it out in the middle of a field and got the same result.

I ran across someone else online who mentioned getting .6 volts without any light shinning on it.  I am not sure what my next steps are going to be; I can easily offset the dark output in programming.  I just don't like it when things don't behave the way the should.  Makes it hard for me to trust.

Apparently OPT101 was a Burr-Brown.
https://pdf1.alldatasheet.com/datasheet-pdf/view/56811/BURR-BROWN/OPT101.html

TI acquired BB and their portfolio of products, thus the OPT101 is branded TI, now.  Maybe TI updated/changed the fabrication process for OPT101 (for example the SIP capsule is not any more in the TI datasheet), and the new OPT has the dark voltage 0.6V, but they didn't update the datasheet accordingly.    That's only a speculation, no hard evidence.

I would try out as a curiosity to power the OPT with differential voltage (the 9V battery for the positive foltage, and yet another AA battery for the negative voltage, like in Fig.2 from the BB datasheet, then measure the dark voltage between Common (GND) and the output pin.  You can put all in a metallic case (an Altoids box, or a metallic box for cokkies, a coffee can with a lid, etc. and only route to outside the 2 wires for the measuring DMM) in case you suspect high levels of mains hum might affect the measurements.
- if the Common to Out still measures 0.6V with dual voltage, then something happened with the photodiode
- if the common to out voltage comes to normal (less than 10mV), maybe they changed the opamp with a model that is no longer rail-to-rail output

Also make sure the PCB or socket used for OPT101 are clean.  Dirt and fingerprints (skin oils and dry salts can make invisible films, and this might be enough to affect very small currents like the ones in a TIA).  Same for the soldering flux, some brands will leave conductive debris on the PCB and on the pins.  For example, Bob Pease was recommending a wash for contaminated chips (though the application there was more fussy than the one here in OPT101):
https://www.electronicdesign.com/technologies/test-measurement/article/21766108/whats-all-this-femtoampere-stuff-anyhow

Guard rings around the opamp inputs might help too, in case the OPT101 is soldered on a PCB.  If you want to investigate all these, add a photo of how the sensor is wired and it's fixture.


Anyway, for casual applications there is no need to insist on the cause of the 0.6V dark output.  Eventually the offset can be compensated in software, or the ADC can be used in differential mode relative to a 0.6V local potentiometer (so to not loose the resolution bits corresponding to the 0 to 0.6V interval).

[RoGeorge]

I've just noticed there is a simulation model for OPT101 (can be downloaded from the TI webpage, a SPICE model that was prepared by Burr-Brown).  Gave it a try out of curiosity, and the simulated dark voltage is 7.5mV, so not 600mV.

Maybe I'm missing something, but I don't know why in reality the output measures 0.6V.  My best guess is those chips are something else than in the datasheet. 

[nfmax]

One thing you might try to determine the cause of the 0.6V dark output voltage is to feed a small, adjustable positive current into pin 2 from a potentiometer across a battery via a large value resistor. As you increase the current, the output should go more negative. If it doesn't, the TIA output is somehow saturating at an abnormally high voltage. But if the output does go more negative, then it's a problem of leakage current into pin 2 - either photodiode dark current, or circuit leakage.

If you have an oscilloscope handy, it's also worthwhile looking at the TIA output voltage.

• Is it oscillating in the dark? If so, supply decoupling needs to be improved
• If you apply a pulsed light source, say an LED driven at ~400Hz, do you see a corresponding signal?

Following on from this, you would be better off (mechanically) chopping your light source and synchronously detecting the TIA output. That way you don't care about DC offsets, so long as they don't cause the TIA to saturate

[voltsandjolts]

Is the OP certain that the OPT101 sensor is actually in complete darkness? The black electrical tape will not be completely opaque, and perhaps also the orange coloured holder (homemade adapter?).

[Kleinstein]

With 1 Mohm the 0.6 V of extra offset is already some 600 nA. It needs quite a bit of light to produce this much offset. It is worth checking if it is really dark, but the offset looks rather large to be explained by a little residual light.  It is still worth a check if the offset changes, e.g. with the room light turned off.  Getting it truely dark can be a challange, but this is more at 1/10000 the offset seen here.

[TimFox]

Comment about darkness:
I found during my job that plastics were not trustable to keep a diode in the dark.
Regular aluminum foil (properly insulated electrically) is reliable, if you avoid gaps.