http://web.ncf.ca/ag136/dashLEDs.html

This is valid, however you must make sure you don't forward bias the diode inadvertently, or you'll let the magic smoke out. In other words, the end with the grey line on it must go to the switched 12v feed, and the other end to ground.
Another way would be to put a forward biased diode in-line with the 12v, but that would mean breaking the wire.

The problem with sticking another diode in series forward-biased is knowing how high a Peak Inverse Voltage rating you need.

Sooo, yeah.....when you figure out if this cures the problem, let us know (in English) how we can install the diode. 

Extra points for photos!

The problem with sticking another diode in series forward-biased is knowing how high a Peak Inverse Voltage rating you need.

That's hardly a problem. Components are out here for 600V that would work perfectly well.

Check out this page

http://www.diodes.com/datasheets/ds28002.pdf

A 1N4002 would be perfect for this job. They've been around for many years. There are probably a dozen alternatives, but this is a tried and tested device and you'll most likely find it at your local Radio Shack.

I haven't been following this issue too closely, but if somebody will post the part number for this device I'll take a look at it and find out why it's blowing. For a LED with internal current limited there's no excuse for that. I find it hard to believe that with that battery in the circuit you're getting reverse bias pulses. I wouldn't have thought that there was enough inductance in the wiring to allow it. I could be wrong, of course - I often am - but I'm suspicious of the diagnosis.

(I'm not the world's greatest mechanic, but this is one area where  I can presume to offer reasonable advice).

Well, the ones that have been dying in batches look like this and have a series resistor of 460 ohms, limiting the forward current to somewhere around 25mA. The reverse breakdown voltage of the LED is higher than 12V because you can put them in backwards and it doesn't damage them.

When they fail, the little resistor is still good and still soldered to the diode terminal. The failed diode is an open circuit in both directions.

Got them on eBay
http://www.ebay.com/itm/10x-T10-Car-Auto-192-194-168-W5W-White-LED-Light-Lamp-Wedge-Inverted-Side-Bulb-/141018226515?pt=Motors_Car_Truck_Parts_Accessories&hash=item20d5575f53&vxp=mtr
but no idea of the part number.

A while ago, I worked as a discrete semiconductor component engineer for good old Nortel Networks so I have some knowledge of this stuff, but unfortunately no current access to instrumentation beyond a digital voltmeter.

A while ago, I worked as a discrete semiconductor component engineer for good old Nortel Networks so I have some knowledge of this stuff, but unfortunately no current access to instrumentation beyond a digital voltmeter.

Good to have someone to talk to about this stuff with then, George :-)

From what you say about their specs it is a mystery. The forward drop is, what, about 2V or so? So that's only 60mW dissipation - not enough to cook them, even totally enclosed. Vibration would be another unlikely cause: if the mounting is soft enough for a filament bulb it's certainly soft enough for an LED.

I wonder if there is high frequency, high voltage spiking from the alternator getting through? Perhaps we should look at something like a TVS or maybe even a simple capacitor to smooth out any such spikes?

My interest in this has spiked a little (forgive the pun!); I'll look into it a little further and see if I can come up with a simple solution. I do have access to a 'scope, so I can look for spikes on my own bike and try out a couple of ideas. I will also talk to my department's electrical technician: he is a wizard and is much more up to date on modern components than I am.

Meanwhile, the 49 cent 1N4004 is paralleled with the LEDs reverse biased. If it does something, yippee, if not, well I was in there anyway to replace the damn LEDs.

    A few years ago I installed LED bulbs in my C10 dash and after about 6 months they began to flicker and then quit working. I replaced them with a different brand and the same thing happened. Checking voltage with an analog tester read 14.5 but when I tried a digital tester, the voltage was all over the place with spikes higher than 20. I traced the problem to the battery ground cable where it attaches to the engine case. You know, down under the airbox where it is almost impossible to reach. After cleaning that connector with a wire brush, the voltage checked out at 14.5 with the digital gauge. The third set of LEDs are still working fine.

thanks for sharing that info. I will check that while my bike is apart as I just installed leds in my dash.

thanks for sharing that info. I will check that while my bike is apart as I just installed leds in my dash.

Do install the diode too: discussing this with a technician friend we figured out that reverse bias spike CAN be generated when the ignition switch is switched on. It doesn't take much of a reverse bias to take out an LED.

Well, even with the protective diode, after about a year, 2 of my 4 dash leds are starting to be intermittent. I gave up and went back to incandescent 194s.

I did take out the diode and make sure it was still healthy.