LED tail light problem and my solution

[WZ JUNK]

I have recently been working on a problem with LED taillights on a  AV8 coupe.  The lights are in a spreader bar that is made of stainless and has openings machined in the bar for the rectangular light assembly.  The turn/brake lights worked okay until you turned the headlights on and then the signals and brake lights were then to faint to see them working.  They worked better when the engine was running but still not good enough.  I thought it was ironic that just when you need them most, you let off the gas and hit the brake or turn on the turn signal, the engine slows down and the lights do not work.   I took some voltage readings and noticed a significant voltage drop when we turned on the headlights.  The LED seems to require full voltage.  I tried some different things but the solution was to place two relays in the trunk, that are triggered by the existing turn signal/brake light wiring.  I ran a direct fused wire to the positive cable connection at the starter.  This eliminates the drop I was getting through the wire and switches.   The lights work fine now.

[Beck]

This sounds like the tail lights I bought for use in my '32. I have not installed the spreader yet. Do you have any idea what caused the voltage drop. I shouldn't have been the wire size to the tail lights or the brake/blinker switches. The lights would have been dim all of the time if they were the problems. Do you think your wire size to the light switch is undersized? Are you using high current headlights that really pulled the voltage down? I have read that using the relays for the headlights can really brighten them up.
Beck

[WZ JUNK]

This sounds like the tail lights I bought for use in my '32. I have not installed the spreader yet. Do you have any idea what caused the voltage drop. I shouldn't have been the wire size to the tail lights or the brake/blinker switches. The lights would have been dim all of the time if they were the problems. Do you think your wire size to the light switch is undersized? Are you using high current headlights that really pulled the voltage down? I have read that using the relays for the headlights can really brighten them up.
Beck

Prior to trying the relays, I replaced all the tail light wiring from the column back with heavier gauge.  When the car was originally built it had a similiar spreader bar but bulb lights and everything worked fine.  

I like the LED's when they are working correct but this is not the first problem I have had with them.

[HOTRODSRJ]

There may be two problems at work here even tho you have solved the voltage drop from the fuseblock to the leds by circumventing the box in a backhanded way really...but it worked. This may be a sign of other issues tho?

I suppose or I am assuming you are pulling full headlight feeds thru the headlight switch from the fuse block....... right (in other words you do not have relays for the headlights)?  I have seen this problem before and it's not the size of the tail light fed bus, but the voltage drop that is taking place at the fuse block due to the current hogging of the headlights.  This indicates that either you need to increase the feeder DC bus size from the battery/alternator to the fuseblock to decrease that drop and/or relay the headlights and get the current coming directly from the battery/alt circuit to the headlights. This will make the headlights brighter as well.

Compound that with a alternator that looks to me by your description like its dropping below the designed turn-on shaft speed and really drops the voltage when idling at slow speeds. This may be able to be remedied with either "overdrive" pulleys and/or a "high output" (at idle) alternator which are getting plentiful these days.  Most alternators have to keep 1750rpms (shaft speed) to keep the output up. So, do the arithmetic on the pulley ratios to the crank rpms and see if your engine idle rpms are enough to keep it "on".

And remember the now OEM "gold" standard to relay anything and everything that is over 10amps.

[Mr Cool]

Its possibly a problem with the headlight switch, dirty/worn/bad contacts.
When you turn on the tail lights only it uses a different contact to the position where the headlights also operate.
My guess without actually seeing it is this is where the problem will lie.
I do alot of fault finding with my work and I see this as a common thing.
Not likely to be a problem with the wire guage, if it were the manufacturers would have had problems earlier in their design.
Another possibility is a bad/corroded connection somewhere, maybe in your fusebox. Ive seen this happen also especially is damp/humid areas, such as basements, coastal etc. The connection carries enough power to operate the tail lamps (and with LEDs its very low power draw) and when you turn on the headlights the extra power required to operate the headlight relays is enough to make a bad connection much worse and creating voltage drop. Relays will operate with as little as 4-6 volts, depending on brand/quality.
The only real way to fix the problem is find where the voltage drop is happening, you'll only end up with the relays not working sooner or later too as it gets worse.

[WZ JUNK]

Its possibly a problem with the headlight switch, dirty/worn/bad contacts.
When you turn on the tail lights only it uses a different contact to the position where the headlights also operate.
My guess without actually seeing it is this is where the problem will lie.
I do alot of fault finding with my work and I see this as a common thing.
Not likely to be a problem with the wire guage, if it were the manufacturers would have had problems earlier in their design.
Another possibility is a bad/corroded connection somewhere, maybe in your fusebox. Ive seen this happen also especially is damp/humid areas, such as basements, coastal etc. The connection carries enough power to operate the tail lamps (and with LEDs its very low power draw) and when you turn on the headlights the extra power required to operate the headlight relays is enough to make a bad connection much worse and creating voltage drop. Relays will operate with as little as 4-6 volts, depending on brand/quality.
The only real way to fix the problem is find where the voltage drop is happening, you'll only end up with the relays not working sooner or later too as it gets worse.

I checked the voltage drop at the light switch and the high/low beam switch and there was not enough drop to indicate a problem.  If I had an extra headlight switch yesterday, I would have changed switches just to see if there was a difference.  Next winter we are going to take it all apart again and look at the left tail light.  I really think it is the problem.  It is not as bright as the right light and if you look real close there is a one inch section that is brighter than the rest of the light.  I suspect that a short is connecting the two sides of the light ( the brake/turn and the tail light ).

[Mr Cool]

It could also be a problem with the earth connection in that case.
Check them and make sure they are connected to CLEAN metal.

[PeterR]

I took some voltage readings and noticed a significant voltage drop when we turned on the headlights.  The LED seems to require full voltage. ..... I ran a direct fused wire to the positive cable connection at the starter.  This eliminates the drop I was getting through the wire and switches.   The lights work fine now.

The overwhelming cause of interaction between brake, tail and turn lamps on vehicles from golf carts to road trains is poor earthing, and on reading the first few lines of the original post that was my simplistic explanation.  However further along the text indicates the problem was cured completely by supplying the lamps with a higher voltage, and no mention of the earthing being altered.

Others have made sensible suggestions such as fitting relays in the headlight circuit and increasing the bus size, --well worth doing, but these do not address the mystery of why the LED lamps appeared so dim when fed by a sagging supply but the old conventional lamps were OK. There would seem to be something peculiar about the LED lamps themselves.

An operating LED has a voltage drop around 2.5 volt and as there is no internal current regulation a ballast resister performs this task.  I had always assumed there is one resistor per LED and if fed from a typical 12.5V supply, 2.5V would be across the LED and 10V across the resistor.  This means if the supply drops to a worst case of say 10V, there is 7.5V across the resistor, the current falls to 75% of the nominal value and the light output a similar amount.  Our eyes have approximately a logarithmic response to intensity so this would be noticeable but not "too faint to see them". (Any photographers out there will remember one stop is a 50% reduction and that is not the difference between bright and faint.)

After some detective work on a couple of LED supplier web sites, I have been able to enlarge the postage stamp sized image of their circuit board and can see that my original assumption is wrong.

In the configuration described above there is a 10V drop across the resistor and 2.5V across the LED, which means 80% of the power is dissipated in the resistor and 20% LED.  To further increase the efficiency of the LED lamp, three or even four LEDs in series are controlled by a single ballast resistor. In the case of a string of three LEDS, 7.5V will appear across the LEDs and 5V across the resistor. Now 60% of the power is utilised in the LEDs which is a great improvement in efficiency, but what happens if the supply voltage drops.

Let's say it goes to 10V again, the voltage across the ballast falls from 5V to 2.5V, causing the light to halve.  In the case of a four-string lamp, if the voltage drops to 10V, the LEDs will barely illuminate.

The lesson from this is that lamps with a series string of LEDs are very efficient but their light output is highly sensitive to supply voltage.  Generally this will be of little consequence because the current drain from the LED lamp is so low it does not cause significant voltage drop even at the end of a long skinny wire.  My guess -and it is very much a guess is that voltage at the light switch is being pulled down by the headlamps, and while this did not appreciably affect the old filament tail lights it is causing havoc with the LEDs.

For a quickie test:
Pull the three pin connectors off the back of the headlamps.
Connect the LED lamps directly to the feed wires.
Switch on the headlights (they will not illuminate)
Check the rear light functions.
If they function correctly this confirms the source of the problem is the sagging switch voltage.  You can remove the relays and use them to operate the headlights where they will provide improved lighting and switch longevity.

If anyone has first hand knowledge of the construction of LED lamps I would welcome their comments.

[HOTRODSRJ]

The lesson from this is that lamps with a series string of LEDs are very efficient but their light output is highly sensitive to supply voltage.  Generally this will be of little consequence because the current drain from the LED lamp is so low it does not cause significant voltage drop even at the end of a long skinny wire.  My guess -and it is very much a guess is that voltage at the light switch is being pulled down by the headlamps, and while this did not appreciably affect the old filament tail lights it is causing havoc with the LEDs.

Bingo!...... I experienced this same phenom on my 57 before relaying my headlights (high output ones at that).  By relaying the headlights and circumventing the fuseblock, this allowed the LEDs on the "V" third brake light to really be bright even tho the brake lamps at that time which were incadescent showed no appreciable effects from lower bus voltages.

[jeffa]

I adopt the same rule as the manufacturers: anything over ten amps should have a relay driving it. Saves heavy switch contacts, long, heavy wires into the cabin and works out a lot cheaper when wiring up a car.
My Daughter's car is a 1974 Gemini. When I found the headlights trying to work at 9.5 volts (with the engine running!) I asked an auto electrician if this was normal. Common, but not normal was his response. I installed a relay into the headlight wiring and one into the starter solenoid wiring (the starter solenoid was also trying to operate off 9.5 volts) and made a world of difference to the headlight light output and to how well the car starts.
By the way, using a relay for the headlights will extend their useful life. A headlight (particularly halogen filled bulbs) operating at it's designed voltage will burn off carbon deposits from inside the envelope. If operated at a reduced voltage, they won't reach sufficient temperature to burn these deposits off, so they eventually go black, resulting in loss of output.

[WZ JUNK]

Thank you Peter.  A very good explanation.  In fact if I had put the relays in the headlight circuit, which is more accessible, it might have cured the problem.  I was chasing a symptom and not the cause I guess.  I am going to rewire the headlight circuit of my truck soon and when I do, I will add a relays to operate the headlights.  It has an additional problem in that my front light circuit plugs in to allow me to remove the whole front fiberglass clip when I need to work on the truck.  I suspect that the plug that I am currently using is causing some voltage loss.

I supplied my son and daughter with a Dodge Daytona each to drive through high school and college.  I bought 4 to keep 2 running and I spent a lot of time messing with them over the years.  Daytonnas have banks of relays.  I have never worked on a car that had so many.  I suppose the designers thought it was the right thing to do.  The problem is once you find a relay problem, you have to find the correct relay.

[PeterR]

By the way, using a relay for the headlights will extend their useful life.

The cold resistance of a lamp filament is a fraction of that at operating temperature and this means the initial current can be up to five times the operating value.  During the fraction of a second until the lamp stabilises it is exposed to a thermal shock which is the main cause of lamp failure, and you will have noticed how often household globes "blow" at switch on.

The maximum filament current at switch on is influenced very much by the resistance of the supply circuit. The lower the resistance of the wiring the greater the inrush current, the worse the thermal shock and the more likely the lamp will fail.

So an unfortunate down side of relays and heavy lighting wiring is reduced lamp life.
 
Though rarely acknowledged, inrush current can also be a serious problem with modern lighting switches.  Old vehicles had chunky dash mounted on-off switches and robust foot operated hi-lo switches, but these have given way to flimsy column mounted switches with peewee contacts.  If column mounted switches are used without a relay, failure is just about a certainty. The heavy inrush current burns the switch contacts, the resulting high contact resistance then causes heating and cooks the surrounding plastic components eventually destroying the switch.

Overall, the reduced lamp life is a small price to pay for extended switch life and improved lighting efficiency.

[Mr Cool]

By the way, using a relay for the headlights will extend their useful life.

The cold resistance of a lamp filament is a fraction of that at operating temperature and this means the initial current can be up to five times the operating value.  During the fraction of a second until the lamp stabilises it is exposed to a thermal shock which is the main cause of lamp failure, and you will have noticed how often household globes "blow" at switch on.

The maximum filament current at switch on is influenced very much by the resistance of the supply circuit. The lower the resistance of the wiring the greater the inrush current, the worse the thermal shock and the more likely the lamp will fail.

So an unfortunate down side of relays and heavy lighting wiring is reduced lamp life.

Actually jeffa clearly made his post in reference to HALOGEN bulbs, which have a totally different operating characteristic than incandescent bulbs which you are talking about. Halogen needs the high operating temperatures to burn of excess carbon as already stated. Incandescent bulbs eventually burn out due to the heating/cooling of the element, stressing them to failure.
As for LEDs, as the voltage decreases marginally, their light output decreases ALOT. So if you decrease the voltage drop across the LED by 10%, the light output is actually around 50%. They do not decrease light output on a linear scale like a bulb.
Many of the better LED lighting circuits now use voltage regulators to keep the voltage at a constant regardless of input, and can operate at anything from about 8-20 volts, with NO effect on light output.
have you seen the latest in LED lighting? A single LED can blind you they are so bright! And they require special driver circuitry to ensure they have constant voltage and current regulation.
One of the items we manufacture from scratch (and I mean we even build some of the electronic circuit boards) is the flashing arrow boards as used by roads contractors. Each of the 15 LED lamps uses a series of up to 60 individual LEDs, controlled by about 4 voltage regulators for redundant circuitry (they will still work if one section goes faulty). We decrease the light output by PWM control rather than decreasing voltage input. ie: by very rapidly pulsing the input voltage to reduce the on-time. the eye will see this as reduced light output, and uses about half the power, very critical when using solar energy and batteries.

So in summary, if using standard non-regulated LED lighting, a small voltage drop will also mean a large drop in light output. This is what you've encountered and have addressed the problem with relays, but the question still remains, where is the voltage drop occurring?

[HOTRODSRJ]

This is what you've encountered and have addressed the problem with relays, but the question still remains, where is the voltage drop occurring?

In the instances where I have investigated....I will stand on my original diagnosis of DC feeder bus issues/size/length.  This is because usually the wire that feeds all the total electrical system from the battery/alternator is not big enough...or has problematic connections.....and/or especially with larger loads being seen... and....on top of that....many put batteries at greater distances from the alternator/fuseblock and the total length of the feeder may increase as well, presenting even further issues.  This is primarily where the large voltage drops occur.

For all practical purposes, in today's automotive electrical environment and depending on load of course, I would put a minimum size of a # 8 from the battery/alternator to feed all the systems electrical need for a moderately equipped vehicle (max 75 amps continuous load @ 6' leader length typical.  For heavily equipped systems of over 65 amps or real long lengths of leader feed, a #6 will do up to 100amps continuous.  For you lighter loaded guys and gals, the least you would ever want is a #10.  This is good for up to 40amps continuous. All of these recommendations are within the OEM automotive guidelines of less than 2% voltage drops.  

Now I will tell you that I am willing to bet most of youse have only #12 in your current systems. This was/is common practice still today.  Don't fret...just add another #12 to the system paralleling the feed. This should suffice in most cases.

[1FATGMC]

just add another #12 to the system paralleling the feed. This should suffice in most cases.

I'm wondering how will those little electrons know which wire to take?  They might get confused and just stay at the junction  

c ya, Sum