Shorted Optos

[seten]

Building a flying pan clone, was working fine and then the phaser pooped. Think the problem is in the optos so I lifted them out. Turns out the positive leg is shorted to the negative leg - and same thing on the other opto. What exactly couldve happened here?

Guessing I'll just have to get new optos but I'd like to know how this couldve happened. The optos are sealed in a way that I do not see how anything couldve gotten jumbled around and shorted internally - especially not on both of the optos. First thought was both LED's got burnt out somehow while I was fiddling but then the LED should be a break, not a short? I went ahead and burnt out an LED to make sure I remembered correctly.

TLDR; under what circumstances could an LED become a short?

Schem:
http://byocelectronics.com/skilletschematic.pdf

[seten]

For the purpose of including all available info I guess theyre not perfect shorts, one opto is 15ohms the other is 5ohms.

[ElectricDruid]

It sounds like maybe the series resistor isn't suitable for the opto's internal LED? An LED is actually pretty easy to fry if you power it up without a series resistor, and that goes for optos too. Even 3V will do it if the LED is 2V forward voltage.

So either the voltage across the opto was too big or the series resistor was too small. That's my first guess.

[duck_arse]

I've only ever seen leds fail open, as in crack and split to allow the gasses out. I think the bonding wires are meant to work as fuses in extreme cases, no? that would result open, too.

interestingly, in the past couple of weeks, I've allowed a few leds on the breadboard to cop full supply by accident, no clr - they go pretty bright, but I caught them in time and they still work fine [with clr].

[seten]

It sounds like maybe the series resistor isn't suitable for the opto's internal LED? An LED is actually pretty easy to fry if you power it up without a series resistor, and that goes for optos too. Even 3V will do it if the LED is 2V forward voltage.

So either the voltage across the opto was too big or the series resistor was too small. That's my first guess.

if they were fried though shouldnt they be open, not shorted?

[PRR]

That schematic could throw a half-Amp onto the LED. I think there should be 50 or 100 Ohms in series with the LEDs.

I don't trust an ohm-meter too far. Put 100-220 Ohms in series with a 9V battery and measure the LED voltage drop, forward and reverse. We expect ~~1.7V one way and 8.99V the other way. If both the same, and the LED not claimed to be bidirectional, yes it is probably melted.

Open? Closed? In decades of burning transistors I have seen both ways. A really BIG hit will blow the Silicon wide open. But a fatal yet not huge hit may melt the P and N back into undifferentiated Silicon, which would be tens of Ohms.

[Eb7+9]

That schematic could throw a half-Amp onto the LED. I think there should be 50 or 100 Ohms in series with the LEDs

The LEDs are driven by a current source, sticking xtra resistance in the loading zone won't do anything to "scale" down current ... rather, R36 should be brought up to at least 100 ohms ... I suspect that drive circuit was originally designed to run on incandescent bulbs - otherwise, why use a Darlington'd source ?!

[PRR]

> do anything to "scale" down current ...

Not trying to "scale down". Normally it won't have excess current because everything is trimmed just so. Yet the builder report does smell of burning. Hundred ohms of dumb resistance has no effect for normal 10mA-20mA currents but will prevent absurd currents due to workbench troubles.

[ElectricDruid]

It sounds like maybe the series resistor isn't suitable for the opto's internal LED? An LED is actually pretty easy to fry if you power it up without a series resistor, and that goes for optos too. Even 3V will do it if the LED is 2V forward voltage.

So either the voltage across the opto was too big or the series resistor was too small. That's my first guess.

if they were fried though shouldnt they be open, not shorted?

Maybe, but that's not the sort of thing they put in the datasheet If you fry it, all bets are off, and that includes how it behaves afterwards.

[willienillie]

Maybe...maybe "fried" also includes "charred"...and "charred" means "burnt black"...and the black part is carbon.  Carbon is conductive.  Enough of it and you might see a low-ohm short.  Maybe this is more likely to happen in an optocoupler than a naked LED, because of whatever the potting material is made of.  Maybe.

[Eb7+9]

Yet the builder report does smell of burning. Hundred ohms of dumb resistance has no effect for normal 10mA-20mA currents but will prevent absurd currents due to workbench troubles.

umm, no mention of burning smells by the OP - rather, the thing is working for a while (held in some delicate balance) and then poof ... that's what I'm addressing

two ways of limiting current here: clamp or scale ...

if you just clamp the current source (exact same technique as Bob Moog does in the Maestro S/H using 22k resistor to save the OTA from going over its 2mA drive limit), then you're not addressing the potential for sudden shift towards high current and then, if it did (suddenly again), the effect will have a clamped response which would necessitate recurring adjustment ...

if it happened once it could happen again - maybe with temperature

by scaling you bring everything down to levels appropriate to the cell LED's (datasheet) and then the risk associated with the extra fine balancing act set on the edge of precipice is completely eliminated ...

I would replace R36 by a 500 ohm trimpot, once adjusted to a sane value the problem will never come back ... notice, the original adjustment scheme relies on controlling Base current going into the Darlington pair ... a very iffy proposition at best

[ElectricDruid]

A different thought:

If the OP's meter uses a "reasonable" voltage to test continuity, could it be enough to switch an LED on? Anywhere above a couple of volts would do it in most cases. I know my meter doesn't manage it, but others might.

Then you'd test it, it'd beep for continuity, and you'd think it was shorted, since that's what shorts do, right?

[PRR]

...umm, no mention of burning smells...

The report smells. Not the LED. Not the OP.

We distinguish between "normal" operation and "oops!" operation. My house normally sucks 2 Amps to 65 Amps from the power line. But if I "OOOPS!" at the meter pole, current is limited to <1,000 Amps by stray line resistance.

[seten]

Thanks yall - one is 0.5v both ways and the other is 0.8v when reverse biased and 0.5v the other way. So get some new optos and throw 100ohms in series with em to make sure it doesnt happen again - sounds like a plan