LED bulbs

Hi I have had several LED bulbs fail on both my Land Rover and car all different types / sizes fitted in different locations (dash, side lights, indicators etc.) The Land Rover did at one point have a full set of LED bulbs but about half of them have failed after a few months, I normally send them back and get replacements ok but they never last long and tonight 2 more aren't working so I've changed them all back the traditional bulbs again. Going to change the car back to bulbs tomorrow because another one on there has failed. Since it's happening on two different vehicles I am assuming that its the LEDs that are faulty, I did however temporarily fit a meter that records peak voltage to the Land Rover for about 2 weeks to see if there is a problem with over voltage burning them out but that only recorded a peak of

14.67 volts. All the LEDs I've been getting are from u l t r a l e d s . c o . u k is there anywhere that may sell more reliable LEDs out there?

Thanks Liam

Reply to
Liam
Loading thread data ...

On or around Mon, 31 Jul 2006 21:06:52 GMT, "Liam" enlightened us thusly:

not had any notable problems with the ones here. actually, tell a lie, one was coming apart, but a bit of superglue fixed that. maybe that's what's happening to yours? Mine are the "35" ones, and the top bit with the axial LEDS on had come loose on one.

superbright leds in america are cheap, but the construction is simialr.

Have you sent any back to ultra, or got onto them about it? they might want to know...

I keep thinking about LED arrays to put in lamp units, which would be a better solution, especially for discos and such.

Reply to
Austin Shackles

Given that the life is of an LED is typically 50,000 hours, something is screwed somewhere....

Are they getting too hot ?

Steve

Reply to
Steve Taylor

I have emailed UltraLEDs in the past but got no reply I have emailed them again, if I don't get a reply I'll phone them instead. I haven't sent any back to them but I think I will as I seem to be building a collection of them now. I've had a good look at the bulbs and all appear to be ok physically. I've normally had good service from LEDs that I've used on other applications at work on machinery control panels etc. for years, don't think I've had one fail yet. Suppose I'll have to wait and see what UltraLEDs have to say.

Liam

Reply to
Liam

A few thoughts..

My background is in electronic/software design (mostly telecomms), so I've had some experience of LEDs in general, although not much with the sort of high power arrays now used in car headlamps

LEds themselves are normally *very* reliable. However, there used to be (maybe still is) a problem with mechanical shock. If you cut the lead of the LED which forms the base for the LED substrate with ordinary sidecutters, then the shock wave as the cutters snipped the lead travelled up the lead and could shatter the actual LED itself. I'm sure this has been improved over the years, but it may be that even now some LEDs are more prone to vibration damage than others. Having said that, they shouldn't be more vulnerable than incandescent bulbs!!

Another thought is whether the failure is due to the LEDs themselves or to the associated elecronics, wiring and mounting around them.

Higher end automotive LED arrays that I've come across don't use the car battery voltage directly to drive the LEDs but use switched mode power conversion from the input variable battery voltage to the required constant LED voltage and current. Looking at the ultraleds website, however, the design doesn't look that sophisticated.

Spikes on the supply to the unit are more likely to be a problem than "slow" changes in voltage and a peak voltage detector may well miss these if they're short.

I would therefore guess that the problem may vary significantly from one manufacturer to another and would guess at the possible failure modes as follows (in decreasing order of likelihood):

a) For a simple bulb (with no internal electroinics), the LED itself is vulnerable to temperature (the ultraleds website hints that this is a significant problem) or voltage spikes. Perhaps the two together are more than it can stand! b) For the more sophisticated bulbs, power convertor design is marginally toleranced with regard to temperature or voltage spikes. b) The elecronics (if present), wiring and physical mounting are vulnerable to shock or vibration. c) The LEDs themselves are vulnerable to shock.

Have a look at the locations of the bulbs and see if any of them get hot (especially this weather!). Is there any correlation between the bulbs failing and really hot days? How much is vibration reduced (e.g. are there rubber gaskets or mounting around the bulbs to cushion vibration)?

If you find out any more, I'd be interested to know.

Reply to
ArthurC

This is the reply I've got back from UltraLEDs:

Liam the bulbs only really fail due to over voltage - we don't many returned back to us so maybe the problem is a power spike that can also kill the leds -

Kind Regards Robert Bennett (Director)

So they think that a power spike is causing the failures. Trouble is on the Land Rover I do have some LEDs that have been on for years (in fact I'd forgotten about them!) and not failed, I have 4 extra rear marker lights, 2 in the roof and 2 low down on the body also the switch panel I built on the dashboard is lit by 4 LED bulbs. Why haven't these failed? They are all on the same circuit. I did get them from a different supplier about 3 years ago. I could understand power spikes on the old Land Rover but not on the 2 year old Jag, but isn't power spikes normal for a battery and alternator setup? I would have thought that if these spikes are a problem for the LEDs then they shouldn't be used on cars or the LED should be protected from them in some way. Think I'll just forget about them in the Jag for now but is there something that I can fit to the electrics on the Land Rover to protect the LEDs from power spikes?

Thanks

Liam

Reply to
Liam

A bloody big capacitor will do it - one of those silly 10 Farad jobs should do nicely. I'd stick an oscilloscope on your supply if you have one and just see what happens when you are running - maybe you have a regulator fault too.

Like you say though, "normal" automotive electronics has to withstand horrible transients - as the alternator dumps load for example, when the output can sometimes spike up to 100 volts, and down to -45 volts, levels which are used in actual functional testing for the automotive industry.

Steve

Reply to
Steve Taylor

Liam it couldn't be that the alternator is bleeding AC could it as a DC meter would not read the surge- maybe an early warning that the alternator Diode pack is on its way out Derek

Reply to
Derek

LEDs really hate reverse biasing - and load dumps can do that too. Check your alternator.

Steve

Reply to
steve

On or around Wed, 02 Aug 2006 21:04:01 +0100, Steve Taylor enlightened us thusly:

fackinell. you could dispense with the battery and the engine...

10 Farad is enough energy to power a small bomb.

Faraday was working with lightning, ISTR - so his units are about a million times bigger than the ought to be.

Reply to
Austin Shackles

On or around Wed, 02 Aug 2006 23:09:32 +0100, steve enlightened us thusly:

they ought to be protected against reverse voltage - and in fact, now I think of it, they are - I've put Ultras 35-LED units into wrong-polarity holders and not blown 'em up.

Reply to
Austin Shackles

I don't think Faraday invented the unit - it is named after him, not by him and the size is the logical result of the rather large size of metres, kilograms and seconds as basic units. JD

Reply to
JD

I didn't slip up. You can buy 2500 FARAD capacitors, albeit at only

2.6v. 12V 10F caps are /used to be available from Maplin, for sub woofer in car freaks.

If you WANT, you can buy rackmount caps, with capacity of 1.6kW/minute.

Steve

Reply to
Steve Taylor

A 1 Farad sphere would be 1 million miles across.

Steve

Reply to
Steve Taylor

Derivation of the unit has nothing to do with a sphere - it is the capacitance that will store one coulomb at one volt - no mention of how you construct the capacitor. JD

Reply to
JD

Course not, I was making an illustration of how big unit Farad can be as a plain sphere. Clearly when you can pack 500 of 'em in a shoebox, construction is not from plain isolated spheres.

Steve

Reply to
Steve Taylor

On or around Thu, 03 Aug 2006 19:50:37 +1000, JD enlightened us thusly:

I still reckon it's got to do with the original leccy lot playing with lightning. You start using 100KV plus, you're going to think big.

Reply to
Austin Shackles

On or around Thu, 03 Aug 2006 09:49:02 +0100, Steve Taylor enlightened us thusly:

it's still a lot of energy. does the amount of energy depend on voltage? I don't think it does.

Reply to
Austin Shackles

All the LED bulbs I have look like they have diodes inside them I presume they are up to the job. Just noticed that the other LEDs I have on the Land Rover that have worked for years have what looks like the 'E' marking on most of them, where as the ones that are failing do not have any markings on them. Could the 'E' marking mean that they are built to a more robust standard that is more suitable for vehicle applications? So fitting a capacitor may help, I may experiment with that, how exactly would that be wired in? Could there be any other side effects fitting one? What size capacitor would it be worth fitting?

Thanks for your help Liam

Reply to
Liam

Power is Volts times Amperes.

Energy is power times time (seconds).

1 Ampere is 1 Coulumb per second.

So Power is Volts times Coulombs divided by seconds.

And Energy is Volts times Coulombs. (time cancels out)

Since one Farad is one Coulomb at one Volt, the energy in a capacitor is proportional to the square of the voltage. Something like 10 Farads at

15 Volts will run a 1-bar electric fire for just over 2 seconds, which, while pretty fierce in terms of current, is only 66 Amperes.

Of course, if you had a capacitor rated at 1500 Volts (which is the voltage at which nasty things will happen inside the capacitor) you could stuff 22 megaJoules into it, and that is a things-go-boom event when you dead short the terminals (or the insulation inside the capacitor fails).

So don't do that.

Reply to
David G. Bell

MotorsForum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.