Help pls - Alternator voltage too high, blowing LEDs?

[Jeremy]

Grateful for advice folks.
I fitted Marek's combined reversing/centre brake light LED unit and the white LEDs have burned out. On Marek's advice I checked the alternator output and it hits a max of 14.68V at about 3000rpm (using a Fluke multimeter that I completely trust). It's a Fosseway Lucas 11AC-Style 70A unit (new) with integral regulator. I got it cos it looks more original than the baby Denso unit. Battery is new Halfords. Voltage at the reversing light end of the wire (open circuit) maxes out at 14.60V. Marek kindly replaced the LED light unit, but says I must reduce the alternator voltage. Otherwise I will damage not only the new LED unit, but other parts of the car too
I have no idea how to do this!
Or is the alternator defective? Should I complain to Simon at Fosseway and demand a replacement?
Thanks in advance for help,

[Gfhug]

Jeremy, have Fosseway given you no information about the alternator? Access to a potentiometer to adjust the voltage as you would on a TR?

Geoff

[malcolm]

Is 14.6V too high? I thought 14.5 V across battery terminal was normal under charge, with a non-charge level at the battery of 13.7V? So 14.6V is maybe a tiny meter error? Or an insignificant extra voltage?

[Heuer]

14.3v is standard but I doubt 14.6v is going to make any difference to anything in the car including LED's. Under normal operation the voltage at the supply lines ranges between 9 V and 16 V (12-V system) whilst Garages and emergency road services have been known to utilize 24-V sources for emergency starts. High voltages such as these are applied for up to 5 minutes and sometimes with reverse polarity. My alternator regulator failed in the wrong way and it was outputting 21v for about 15 minutes before I twigged something was wrong. Only two casualties - the battery and the electronic tach board I was using (no Zenner diode overload protection).

Marek's LED's require at least 11v to operate I seem to recall so will probably not illuminate unless the engine is running. Certain exterior safety signals may be expected to produce light down to supply voltages of 6v, sometimes for as long as 2 minutes and can potentially create a rear-collision hazard when not lit. This type of voltage drop occurs when starting and cold ambient conditions aggravate the voltage drop. This means switching power converters are needed for driving signal and lighting LED's to cope with under and over voltage.

Just sticking some LED bulbs in the standard E-Type lighting circuit simply will not do unless they include the correct power supply drivers. You could find yourself on the side of the road in the dark with no lights if the battery is low or engine not running! Hence I never fit LED's as they are a nightmare unless you understand what you are doing, but no one listens

If you have tail/stop/hazard/indicator LED's fitted you need to check they have suitable drivers included in the circuit and can output constant current and supply voltages of between 6v and 26v.

This data sheet from a suitable LED driver explains things: https://people.xiph.org/~xiphmont/thinkpad/PT4115E.pdf.

You must buy car LED lights that incorporate such a driver unfortunately not many suppliers detail if they are included.

[64etype]

I'm listening! I used LEDs in the instrument lighting circuit only. I missed the caution regarding installation of a device like the PT4115. So far I've had no "bulb" failures in the LED instrument lighting circuit, but I guess it's just a matter of time. Failure of the strip lights in the tach and speedo would be a nightmare. Is it sufficient to put just one of the PT4115 items in the instrument power supply wire ? Also, I installed one of those simple "floating" rheostats (white strip with three buttons) in the tach and speedo wiring. I wonder if the PT4115 device will impact the operation of that rheostat? Is there a 'noise" issue regarding radio reception, for example? (I am a complete amateur regarding electronic circuits). Any advice greatly appreciated.

[malcolm]

Hmmm. I have instrument-only LEDs. They're great, same green colour, not too bright, but can see clearly at night. They were a proper kit, but no drivers etc. Just replacement bulbs/fittings. Never had any problems over last 2 years. Just 12v DC LEDs, so I don't quite get why there should be a problem, but there again I'm not great at this stuff! The battery supply is never going to be less than 12v though unless it's dead, and the DC LEDs I use in the house simply get dim when the batteries start to go flat. So in the range 12V to 13.6V, I'm not sure why it should be an issue?

[cactusman]

14.6 volts is well within normal range for a 12 volt system. The LEDs will be fine. If it gets much above 15.5 volts I would start to get a bit concerned and anything above 16 and the alternator regulator has probably failed. The LED bulbs will probably be fine but the battery may not fair so well...

[Heuer]

LED's have a 'strike' voltage above which they illuminate and vendors usually quote as 12v but is typically 10.8v although it varies greatly. Anything below that voltage and the LED will not illuminate and a battery left for a few days will probably read below 11v. When you attempt to start the car it could drop to 6v. If your alternator fails you could still get home with what remains in the battery however your LED instrument lights plus any LED signal bulbs will switch off at some point. Filament bulbs will continue to glow of course. Over driving the LED's will mean they draw more current and will quickly die so the upper voltage has to be clamped. Bear in mind that switching off the ignition in a running car can produce a 40v spike in the supply!

Properly designed and purposed LED bulbs will have the power driver built in which will allow them to work at any voltage between 6v and 30v. Just like the power brick for your tablet or PC, which you can feed with 110v or 240v, they will output 12v regardless. Modern cars have the LED driver supply built in to the harness so many LED's do not need or indeed have a driver board. For older vehicles the individual LED bulbs can have the driver circuit embedded in the housing. They look like this:


Trying to get the information about how the bulb operates and its rating is very difficult as most do not have any technical data included. You need to seek expert advice or find a supplier who knows his stuff. For your safety the LED bulbs must work on any voltage between 6v and 30v DC.

Eric: The strip LED's certainly do not have a driver, they just have a resistor in circuit with each one so it may be worth considering putting an in line driver into the circuit. They only cost about $6. Having said that they should be OK with up to 16v; it is the current that kills them not the voltage but the two are inter linked so a complex subject.

Jeremy: There is nothing you can do about the alternator output which is well within specification.

[64etype]

David, Several years ago and in the same timeframe that I installed the strip LEDs in the tach and speedometer, I built several LED strip light arrays for the boat. At the time, properly stabilized individual bulbs were extremely expensive. My strip light arrays for the boat were all burned out within a few months. I'm sure this is the reason. Since the instrument lighting in my car is all very small bulbs and strip lights, I'm sure none of them is stabilized. An in-line driver makes sense for me. That said, I don't have a clue as to how to wire it into the circuit and set it up...

[Jeremy]

David and Julian: many thanks for this information, and the reassurance that my alternator appears to be working correctly.
Marek's other recommendation is that I put a small diode in line with the reversing light. Might this represent the missing inline driver you're referring to...?

[cactusman]

Too complex to go into here but light emitting diodes should be powered from a constant current supply unlike filament bulbs which are powered from a constant voltage supply. Any good quality LED bulb will incorporate a current source and will usually operate from about 10 volts to 24 volts or more with no significant change in current through the light emitting diode elements. Putting a standard silicon diode in series with your reverse lamp will have no effect at all.

[64etype]

I can't figure out how to wire the PT4115 driver described in the link attached to Post #4 above, but I'm guessing it or something like it is at the heart of the installation ready devices described in this link.

http://www.luxdrive.com/content/2008B-PowerPuck.pdf

The input wiring to this driver/buck devices looks obvious enough, but it appears that the LED grounds on the output side are independent/isolated from the chassis/battery? If so this would seem almost impossible to achieve in an auto installation. Can someone help me understand how this would be wired into an instrument panel lighting circuit? Thanks

[cactusman]

The simple answer is they are not really suitable for use in a vehicle. BOTH sides of the LED bulb here to be isolated from ground. They seem to limit at 350/700/1000 mA which will be for high power leds that will be far too bright for instrument lighting. Buy well designed (not cheapo) led bulbs designed for automotive use. They all have drivers within them.

[Heuer]

You could use a driver on the strip LED's because they are 'floating' and do not have one side to earth. Just a matter of finding one that will work with the strips. Maybe something like this http://tinyurl.com/ycw7nboe
without needing to use the remote? Not a recommendation though!

Hopefully the LED lighting Ray Livingston is developing will have full protection and the correct drivers.

[64etype]

The simple answer is they are not really suitable for use in a vehicle. BOTH sides of the LED bulb here to be isolated from ground. They seem to limit at 350/700/1000 mA which will be for high power leds that will be far too bright for instrument lighting. Buy well designed (not cheapo) led bulbs designed for automotive use. They all have drivers within them.

I'll look around for bulbs with drivers, but doubt the little LED peanut lights that fit in the strip for the switch labels can be had with drivers. Hope I'm wrong. That said, I'm actually far less concerned about any of the lights on the center fold out panel, than I am about the strip lights inside the tach and speedo. The former are easy to replace...the latter would be a major project. If I can protect those strip lights I'll be happy.

David, I wonder if that little white push button dimmer that was discussed a few years back (which I installed in the speedometer and tach circuit) serves the function of one of these limiting drivers? At it's optimum setting It does cut back the brightness of the lighting significantly from "full on".

[cactusman]

If you want to build your own driver it is not difficult providing the forward voltage drop of your LED string is less than the 12 volt supply voltage. The one in your image has a buck converter that steps the output up to 32 volts to drive long strings of LED's and seems to need the output completely isolated from ground.

If your string draws less than about 100 mA then the simple way is to include a resistor in the positive. Take the nominal 12 volt supply, subtract the voltage drop across the led. this gives the voltage drop across the resistor. Then use ohms law to find the resistance where R=V across across divided by led current in amps. So if your string dropped 10 volts at 100 mA the voltage across the resistor would be 2 volts. R=2/0.1 = 20 ohms. Power dissipation in the resistor would be 2 volts x 0.1 amps = 0.2 watts. I would use a 22 or 27 ohm 1 watt resistor. It wouldn't be a constant current as the current will rise a bit when the engine is running but at.currents up to 100mA I would not worry. Beyond 100mA you might be better with a proper semiconductor current source. Not difficult to build but a bit too in depth to describe here....

[64etype]

Thanks, Julian. The dimmer device in the photo is presently installed in the tach and speedo circuit, which is indeed "floating" in my installation. It works well in reducing the light intensity. I have no idea if it also serves the function of a limiting driver. (I do know the selected/reduced light level stays constant as voltage from the alternator varies). There are less than 30 small strip LEDs in the tach and speedo combined.

I'll look at adding a resister, but wonder if adding another properly sized floating device like the ones described in the post #12 link would perform the protection function?

[cactusman]

Id guess your dimmer device is a variable pulse width modulator type. It very rapidly turns the leds on and off. By varying the relative off and On times the dimming is achieved. If it works then fine.

[MarekH]

For people who have contributed to this thread, I'd be interested to know what voltage they measure on a running car (i.e. alternator powered) at the reversing light connection. Obviously this would have to be done with the aid of an assistant. It'll help make sure this happen to anyone else.

Jeremy measured 14.6volts and the LED unit failed because it was supplied with too high a voltage - it's operating window is between 12.8 to 14.4volts.

kind regards
Marek

[64etype]

Id guess your dimmer device is a variable pulse width modulator type. It very rapidly turns the leds on and off. By varying the relative off and On times the dimming is achieved. If it works then fine.

Could you (or perhaps Marek or Ray) speculate as to whether a PWM type dimming device (I'm sure you're right) would also provide protection for the over voltage scenario? It would save me and others the additional step of adding another part in the circuit. Judging from the discussion in early strings about LED instrument lighting, there are potentially quite a number of unprotected strip lighting installations "floating" (or not) around. (Thanks for bearing with me....I'm obviously electronics deficient).