Do Opto-isolators degrade with time?

They do, but if you design your circuit correctly (considering temperature derating for the emitter, and one half of the minimum CTR for the detector) they will reach their specified MTBF without problems.

Is the degredation due to the LED or the clear material turning opaque?

The LEDs themselves degrade over time. Lower current makes for longer life.

If there is any degradation of the clear epoxy, high LED current will cause local heating and make that worse.

So, keep the LED current low.

but high enough that as the CTR drops with age yo still have enough output current

Yup.

put power (Pc) and forward current (If), the current gain (hfe) of the phototransistor, and the ambient tempera- ture. The CTR of devices within a given product family can vary considerably."

Thanks. No mention of optics degredation, only LED output power with time.

That's what I thought.

I'm not sure, but I would guess that if the LED is driven correctly, the degradation of the junction is by far the major cause. If you over-drive the LED is is possible that the heat will affect the plastics too.

The operating life for most LEDs is specified when the light output falls at 50% or 70% of its nominal (initial) value. That's why if you consider a half of the minimum specified CTR you should be safe.

Yup, they do. The LED's output drops with time (especially if you drive it hard to overcome the putrid current transfer ratio), and the goop inside the package turns yellow from heat or age or both.

I expect the cheap'n'nasty Chinese ones are probably worse, but I don't know for sure.

Cheers

Phil Hobbs

Fig 6. But that might vary a lot between different types.

I am wondering when I should replace my bicycle headlights. They run the LED flat-out, I ride with daytime lights, and I ride a lot. I guess since it's so gradual I won't notice.

Yes, if the plastics in there would turn brown they should fire their chief chemist :-)

I think the actual failure mechanism is just a decrease of the LED's output. If you run the LED at very high current (over 10 mA) then it degrades faster.

Jon

Yes, I think so also.

I did a survey of optocouplers. The ones that talked about degredation referred only to the LED, and none mentioned degredation in the optical coupling.

There are a number of different constructions. Ones where the die lay side- by-side use a silicone lens using total internal reflection to couple the light. It is not clear if silicone degrades with time.

Some use clear epoxy. I would suspect this may degrade with time, but it could take a long time.

Other constructions have the die facing each other. These include a barrier sheet between them for high voltage isolation. The barrier could be several different materials, including mylar, polyamide, glass and so on. There doesn't seem to be a degredation of optical coupling with time.

The problem with optical coupling degredation is presumably the degredation would continue even when the LED was turned off. Nobody seems to refer to this effect.

So the main effect seems to be LED degredation with current, time, and temperature.

Not until you wake up in hospital.......

Well, I try never to out-run my lights. This is why I have these powerful lights so I can keep a reasonable speed at night. Sometimes I compare the road bike light with the mountain bike light of same type. On the road bike it is used a lot more because it is always on when I ride, day or night.

I assume that's some sort of thermal thing, If you could keep your led's cool....

(We need to have artist's design, cool, cheap heat sinks/light shades, for our led bulbs... something 'prettier' than the James A. beer can. :^)

George H . (a bit of a shift to consumer led's.)

The LEDs of decades ago could dim 30% in a year: nowadays, with high current densities in illumination type LEDs, there MUST have been bunches of treatments/solutions to the aging woes. Alas, data on aging of LEDs comes mainly from old assemblies, so the longevity of a modern buy (off a modern assembly line, with possibly a modern redesign) is probably good, but not easy to quantify.

If an IR LED in an isolator uses the same optical coatings as a white LED, it won't yellow nearly as much ( blue photons can create more kinds of photodegradation of polymers than IR can).

A good light meter, and a monthly wipe-the-lens-and-aim-at-a-white-wall exercise would tell you a lot. I did some A/B in the old days, and was pleasantly surprised with a variety of incandescents did almost exactly the lighting they promised (so halogen won, having the best aging curve).

I was also surprised to find a bad monitor (backlight uneven from top to bottom) was almost impossible to see by eye, even though a light meter showed me

Some of the "LED flat-out" circuits I've seen use current limiters with thermal limiting ( AMC7135) which will complicate the measurement, BUT they control peak current which might be a saving grace. Assemblies available here:

... or pick up a JANSJO at Ikea, those lamps use similar modules (~4V 700 mA power).