photodiode sensitivity and DMM diode test

[jcarmody]

I have questions about the operation of a photodiode and the relationship between sensitivity and the diode voltage I mesaure with my DMM.

One of my products uses a photodiode to detect the magnitude of incoming light.  Recently there has been a dramatic step-change increase in the uncalibrated/default value of this measurement.  The technicians trying to salvage these units noticed that they could use their DMM's diode test to discriminate between "good" and "bad" photodiodes.  Every unit that failed calibration had a diode measuring 0.6V in this test; old production units contained photodiodes measuring 0.5V.  Replacing the 0.6V photodiode with one measuring 0.5V makes everything all right.

• Is the DMM diode test result related to the photodiode's forward voltage?
• Is the sensitivity of a photodiode related to the parameter measured by this test?
• What do I even ask the manufacturer if I want what I used to get?

Thank you,

[AnalogKid2DigitalMan]

Jim:

Yes, the DMM will measure the the forward voltage drop. Probably best measured in a dark environment.

For responsitivity, set the DMM to mADC and connect the photodiode to the DMM and compare what currents are produced with the same illumination.

Better yet, reverse bias it with a few volts with a curreent limiting resistor in series and measure the voltage generated across the resistor.

The data sheet should spec the response curve (mA/mW) across the range of wavelengths.

Are you using them in voltage or current mode in the end application?

-AK2DM

[Henrik_Volkers]

I guess the tolerances in the dark current leads to this effect. Hard to say without schematic

Options: 

-- Part selection, grouping of diode&resistor  😞

---Hardware redesign (current mode, reverse bias , ...)

-More flexible software calibration...    (I bet that it the first thing you have to check 😉 )

[LV_Pro]

I would also talk to your vendor to determine if there has been a recent change in their supplier. We have had similar issues, in our case with LED's color values changing, failing our tests (sort of the inverse of your problem).

[jcarmody]

---

@LV_Pro wrote:

I would also talk to your vendor to determine if there has been a recent change in their supplier. We have had similar issues, in our case with LED's color values changing, failing our tests (sort of the inverse of your problem).

---

Our LED vendor proudly informed us that their product was improved and would henceforth be brighter.  Do not want!  We bought a lifetime supply of the "old" ones...

Your comment hits our dilemma, though.  Something changed and the only way we can discriminate between the old and the new is with the diode test on a DMM.  I need to know how to ask our vendor for product that matches the old ones.  We don't believe that the new photodiodes are out-of-spec, we just can't use them.  We'd like to approach our vendor with a requirement that will meet our needs, but we don't know what to ask for.  Do we just ask for product with the lower voltage?

[johnsold]

Jim,

I think you are on the right track in trying to ask the right questions. Unfortunately it may not be easy to determine what those questions are.

What is the wavelength or spectrum of your light source? What is the intensity at the photodiode? Do you have stray light or other noise issues? Is the illumination uniform over an area much larger than the photodiode or are you measuring a small spot? How fast are things changing? Has anything else changed in the product or the test setup? 

In response to one of your original questions: Every handheld DMM I have used applies a fixed current (often 100 uA or 1 mA) with a compliance voltage ranging from 1 to 4 V and then measures the voltage across the diode.

Any junction diode, including LEDs, is a photodiode if light can get to the junction.  The forward voltage of a diode depends on the materials including the doping levels, the size, and any internal resistance including lead and connection resistances. The sensitivity depends on those things as well as the optical configuration, exposed area, and probably other things. I doubt there is any simple correlation between forward voltage drop and sensitivity. You probably got lucky on this one.

How closely does the performance in your product correlate with any of the specifications on the photodiode datasheets?

Lynn

[JÞB]

Jim,

Yeah,  That type of "Improvement" is what gives "Component engineers" their dinstinctive grey hair color (and the notoriously unusual bald patches)

so set <Sea story = "True">

in the late 90's our supplier for a 10W RF pre-amp changed the mounting hardware from 6 #6 Captive Screws to 6 #8 captive screws.  these held the amplifier to a heat sink.  (C-Band amplifiers are not especially effeceint and well, spectral purity was a bit more important anyhow so the waste heat really needed to go somewhere).  I read our "Spec" which said " 6 #6 OR LARGER captive screws"  the amps met spec with #8 screws. I could not reject the parts as non-conforming.  I wound up changing the spec and returning existing stock for upgrade to the new mounting hardware on our dime.  Since I had systems deployed globally, and no way to track which had what screws, 

The app eng from the vendor called me for explaination.  It took much longer than you would expect to convince that guy that I could allways drill and tap a #6 hole LARGER for an existing system that needed repair but drilling an existing #8 hole SMALLER would be inconvienient at best.

<End Sea Story>

Summary: Component engineers earn their hairdos