Led light turn on at specific Voltage

[djmixalot]

Hi,

Setup is very simple, updating led lights on a vehicle, the older 3" round led lights have been replaced with new ones.

This vehicle put out a constant ~10.2v on the lights wiring. (turn signals, brake lights, 3rd brake light) when the ECU drives the circuit to ground the voltage goes from ~10v to 12V and the lights light up accordingly.

The older lights, lets use turn signals as an example, would turn on when you activate the turn signals. When the turns signals were off they remain off despite the 10v.

These new round led lights have about a dozen led in this 3" round light. Due to the fact there is ~10v on the turn signal wiring, it lights up about 4 of the 12 led at all times. When you actually activate the turn signal, it turns on the other 8 led on and off as expected but these 4 remain on all the time.

The goal is to make the new lights work like the old lights. So a modification is needed for each of these lights. I'm looking for a clean solution since this needs to be done on each of the lights.

1. Running resistors in serious dimmed the 4 led which were on all the time but than none of the lights work when the turn signal is turned on.

2. Added a pair of diodes schottky but it made no difference.

Was thinking a comparator would be ideal however its all IC mounted and not ideal. I'm thinking maybe Zener diode would be the best option.

Thoughts?

 

[rjenkinsgb]

Try a resistor across the light wires, without the LED to start with?
eg. 100 Ohms, and measure the voltage across that with ignition on but lights off??

I'm guessing the 10V is a low current feed intended to detect continuity through filament lamps, for lamp failure warnings; if so, the voltage shoudl drop right down when loaded.

If that's the case, just adding an appropriate resistor in parallel with the new lamp should extinguish it.
If 100 Ohms works to drop the voltage to ~2V or less, you could try different values to find something near the highest that would work reliably.

(The lower the value, the more power it will dissipate while the light is on - for a 100 Ohm, it should be 3W rating at a minimum and fitted so it has airflow, as it will get quite hot. A lower value would need to be higher power rated).

 

[danadak]

Something like this :

Replace the JFET with AL5809 from Diodes Inc, a constant current diode. I could
not get its model to work right in simulator. So used a generic JFET to show rough
sim. So when ramp hits Vzener + Vth JFET the constant current diode turns on.
So worst case the AL5089 and Vz choice/tolerance to get behaviour you want.

Linear LED Drivers

Our linear LED drivers are constant current LED drivers for a broad range of applications, with a low BOM cost, and no inductor required.

www.diodes.com

https://www.diodes.com/assets/Datasheets/AL5809.pdf

One issue with new LEDs is uA of current will make them glow, so pick Vz to have
low leakage.


Regards, Dana.

 

[djmixalot]

Try a resistor across the light wires, without the LED to start with?
eg. 100 Ohms, and measure the voltage across that with ignition on but lights off??

I'm guessing the 10V is a low current feed intended to detect continuity through filament lamps, for lamp failure warnings; if so, the voltage shoudl drop right down when loaded.

If that's the case, just adding an appropriate resistor in parallel with the new lamp should extinguish it.
If 100 Ohms works to drop the voltage to ~2V or less, you could try different values to find something near the highest that would work reliably.

(The lower the value, the more power it will dissipate while the light is on - for a 100 Ohm, it should be 3W rating at a minimum and fitted so it has airflow, as it will get quite hot. A lower value would need to be higher power rated).

The EV came with LED lights to begin with so its not looking for filament lamp failure

adding resistors just causes the new bubs to not even turn on.

I'm thinking a simple 12v zener diode should work. Thoughts?

 

[djmixalot]

Something like this :

View attachment 143874

Replace the JFET with AL5809 from Diodes Inc, a constant current diode. I could
not get its model to work right in simulator. So used a generic JFET to show rough
sim. So when ramp hits Vzener + Vth JFET the constant current diode turns on.
So worst case the AL5089 and Vz choice/tolerance to get behaviour you want.

Linear LED Drivers

Our linear LED drivers are constant current LED drivers for a broad range of applications, with a low BOM cost, and no inductor required.

www.diodes.com

https://www.diodes.com/assets/Datasheets/AL5809.pdf

One issue with new LEDs is uA of current will make them glow, so pick Vz to have
low leakage.


Regards, Dana.

Hi Dana,

thanks for the info, curious why the need for the AL5809 since the ECU knows there is an LED on the other end already? Wouldn't just a zener by itself work?

 

[rjenkinsgb]

The EV came with LED lights to begin with so its not looking for filament lamp failure

A lot of LED lamps intended for vehicles have an integral parallel resistor to provide a path for the lamp monitoring system - just because it came with LEDs does not mean the failure detection does not exist.

Have you actually tried a parallel (not series) low value resistor??

(Another simple test is to measure the voltage across one of the original lamps; if that is low when a new one shows 10V, it has a bypass resistor or similar).


I believe Dana is showing a circuit to drive a "bare" LED, not a 12V LED lamp assembly??

 

[danadak]

I believe Dana is showing a circuit to drive a "bare" LED, not a 12V LED lamp assembly??

Thats correct.

thanks for the info, curious why the need for the AL5809 since the ECU knows there is an LED on the other end already? Wouldn't just a zener by itself work?

Thats just to "fix" the current at a known current value when on. Note there is a slope
based on V when system V is in transition, because a Zener has a > 0 ohm impedance. A
better solution is a comparator based one where the high G "switches" the LED on. Or
a small micro with A/D would do it :


What is your required LED current ?

A quick and dirty way doing it with 8 pin micro :

Note I could have easily added PWM control over brightness to minimize energy
needed as well. Lots of variations easy to add. The above is mBlock, easy to use
GUI block programmer. Just drag and drop blocks, set their values if needed, and
mBlock converts to C/C++ for you, right hand window. Then you use Arduino board
to program chip.

Something like this, diode to protect from load dump in vehicle :

Note I did not correct for V divider in code to compute 12V, leave that up to user.
So that correct V is computed in code. Just alter the sysV computed V by divide
factor. Also ATTINY85 needs 5V supply, a regulator rated for load dump or a R and
Zener would suffice.

Regards, Dana.

 

[Tony Stewart]

Schottky Diodes may not be enough to make the difference. Try Silicon diodes in series. 1N400x type.
If lamps have a series R to regulate the current, then brightness may be affected as they assumed a different standard,, but not as bad as a series R. as 1A diodes are only ~1 ohm or more at lower currents.
So I am assume the lamps are <= 10W

 

[djmixalot]

A lot of LED lamps intended for vehicles have an integral parallel resistor to provide a path for the lamp monitoring system - just because it came with LEDs does not mean the failure detection does not exist.

Have you actually tried a parallel (not series) low value resistor??

(Another simple test is to measure the voltage across one of the original lamps; if that is low when a new one shows 10V, it has a bypass resistor or similar).


I believe Dana is showing a circuit to drive a "bare" LED, not a 12V LED lamp assembly??

Yes I tried the resistor in parallel and the end result was pretty much the same, if the resistance was high enough to dim the lit led's it would make it inoperable when 12v was applied none of the led would light. Very similar to series expect it took more resistance.

That is probably true on the can bus monitor

 

[djmixalot]

Thats correct.



Thats just to "fix" the current at a known current value when on. Note there is a slope
based on V when system V is in transition, because a Zener has a > 0 ohm impedance. A
better solution is a comparator based one where the high G "switches" the LED on. Or
a small micro with A/D would do it :


What is your required LED current ?

A quick and dirty way doing it with 8 pin micro :

View attachment 143877

Note I could have easily added PWM control over brightness to minimize energy
needed as well. Lots of variations easy to add. The above is mBlock, easy to use
GUI block programmer. Just drag and drop blocks, set their values if needed, and
mBlock converts to C/C++ for you, right hand window. Then you use Arduino board
to program chip.

Something like this, diode to protect from load dump in vehicle :

View attachment 143878

Note I did not correct for V divider in code to compute 12V, leave that up to user.
So that correct V is computed in code. Just alter the sysV computed V by divide
factor. Also ATTINY85 needs 5V supply, a regulator rated for load dump or a R and
Zener would suffice.

Regards, Dana.

Hi Dana,

Appreciate the thorough and detailed response. What i'm looking for a small dummy proof solution which will last and be reliable. I'm trying to avoid any circuit boards simply because I have to make at least 4 of these and trying to keep it simple using the existing wires at each light to add some small modificaiton and call it a day if that makes sense.

My initial though of using a zener was incorrect b/c I assumed a 12v zener would do what I need but thats not how it works as it was pointed out in another forum. A 12v zener breakdown would pass 0v at 12v so my idea to use a reverse bias was miss guided. The suggestion was to use a 11v zener and transistor so i'm trying to understand how that works.


Keeping it simple and small is ideal if that makes sense

 

[djmixalot]

Schottky Diodes may not be enough to make the difference. Try Silicon diodes in series. 1N400x type.
If lamps have a series R to regulate the current, then brightness may be affected as they assumed a different standard,, but not as bad as a series R. as 1A diodes are only ~1 ohm or more at lower currents.
So I am assume the lamps are <= 10W

Hi Tony, these leds draw 65mA at 12v. I'm not sure the diodes will do much here as i was hoping by just testing the resistors in series and parallel. Really need a simple comparator but I have not found anything that I think would work in a small footprint that doesn't require a ciruit board.

Even a relay that was triggered off a specific voltage would probably work but that constant 10v is an issue

 

[Diver300]

Do the original lamps only have two wires for each lamp? Most lamp clusters have a common ground wire for each lamp, but some lamp clusters now have signals, such as a LIN bus going to the lamps.

You haven't mentioned make or model of the car. The type of car alone is not a good indication of what the original lamps are.

Most devices intended to run in cars are rated from 9 - 16 V. I would be very surprised if the lights of a car were supposed to be off with 10 V applied. However it's very possible that the 10 V comes from a lamp detection circuit, or LIN or something.

Car manufacturers have a really bad habit making the electronics look for faults even when lamps are off and then inhibiting the lights even though there is no indication at all that inhibiting the lights could be helpful. On a lot of cars, incandescent lights are expected, and if an LED lamp is fitted, the electronics will fail to see the LEDs when they are turned off, and then refuse to run them.

On one car, the standard lamps had two LEDs in series. A monitoring circuit detected the difference between the voltages across the two LEDs, and would inhibit the lamps if there was a big difference. That monitoring continued when the lamps were turned off, and if some other light source, like the sun or other car lights, put too much light on one of the LEDs, a voltage would be generated due the photocell effect, there would be a voltage imbalance and the lamp would be inhibited when it should have been turned on.

I suggest that you test both the old and the new lamps away from the car. Measure current for voltages from 0 V up to either 14 V or when you think that the lamp is as bright as original. Also record what if the lamps are illuminating, and whether some or all of the light sources are lit.

 

[rjenkinsgb]

if the resistance was high enough to dim the lit led's it would make it inoperable when 12v was applied none of the led would light.

If in parallel, higher resistance would have less effect, not dim them??

 

[djmixalot]

If in parallel, higher resistance would have less effect, not dim them??

with enough resistance it dimmed the led but that net affect was the same with resistor in series or parallel with different values it didn't solve the problem

 

[djmixalot]

Do the original lamps only have two wires for each lamp? Most lamp clusters have a common ground wire for each lamp, but some lamp clusters now have signals, such as a LIN bus going to the lamps.

You haven't mentioned make or model of the car. The type of car alone is not a good indication of what the original lamps are.

Most devices intended to run in cars are rated from 9 - 16 V. I would be very surprised if the lights of a car were supposed to be off with 10 V applied. However it's very possible that the 10 V comes from a lamp detection circuit, or LIN or something.

Car manufacturers have a really bad habit making the electronics look for faults even when lamps are off and then inhibiting the lights even though there is no indication at all that inhibiting the lights could be helpful. On a lot of cars, incandescent lights are expected, and if an LED lamp is fitted, the electronics will fail to see the LEDs when they are turned off, and then refuse to run them.

On one car, the standard lamps had two LEDs in series. A monitoring circuit detected the difference between the voltages across the two LEDs, and would inhibit the lamps if there was a big difference. That monitoring continued when the lamps were turned off, and if some other light source, like the sun or other car lights, put too much light on one of the LEDs, a voltage would be generated due the photocell effect, there would be a voltage imbalance and the lamp would be inhibited when it should have been turned on.

I suggest that you test both the old and the new lamps away from the car. Measure current for voltages from 0 V up to either 14 V or when you think that the lamp is as bright as original. Also record what if the lamps are illuminating, and whether some or all of the light sources are lit.

Yes 2 wires.

It is a custom EV not a traditional car brand.

The 10V constant is from the vehicle computer which is detecting for faults ie Can Bus

Bottom line is this is how the vehicle behaves and I'm looking for a solution. I've already taken all the measurements valid to know what is happening.

I'm looking for a simple solution to fix this problem.

thanks

 

[rjenkinsgb]

with enough resistance it dimmed the led but that net affect was the same with resistor in series or parallel with different values it didn't solve the problem

That still makes no sense?? What value resistor did you try, and did you try it across the feed and measure the voltage without the lamp connected, as I requested back in post #2??

 

[djmixalot]

I have a resistor box so I tried different resistors all the way up to couple hundred mega ohms. I know you keep harping on this but it doesn't matter because it is not a working solution.

I tried with the lamp connected because that's the only thing that matters is how it behaves with the lamp connected.

If you have an alternative solution that's simple I would love to hear it

 

[rjenkinsgb]

I know you keep harping on this but it doesn't matter because it is not a working solution.

It's a diagnostic method to figure out WHY the 10V is present.

100 Ohms, WITHOUT the lamp - what's the voltage then??

 

[Diver300]

The 10V constant is from the vehicle computer which is detecting for faults

If the vehicle is detecting faults with 10 V and the lamps off, what current is flowing? Is there any digital signalling going on?

What is the voltage with and without the lamps connected?

 

[djmixalot]

It's a diagnostic method to figure out WHY the 10V is present.

100 Ohms, WITHOUT the lamp - what's the voltage then??

The 10V is the computers can bus, it does not change the fact it is using 10v for its bus.

I still need a viable simple solution.

Shame there isn't a relay that triggers at a specific voltage that would be easy enough to wire in.