Re: Russ's 2.1 SED Auto
Posted: Thu Jan 23, 2014 3:38 pm
Dimming LEDs
Saw this (enormous) thread today, and was aware there had been a good bit of discussion about LEDs as dash-illumination, and as "long life" substitutes for small tungsten lamps.
Several things to note about LEDs:
1. To the circuit in which they are placed, they look just like a Diode (i.e. they will conduct in one direction only).
2. Their tolerance to reverse current (i.e. being connected the wrong way round) is limited - they will blow.
3. They will emit their maximum designed light output at a specified forward current (If, or Ifwd), which is often around 10-30mA for single 'indicator' LEDs.
4. It is this forward current which is critical. If the LED is spec'd as needing 20mA, that is the maximum that should flow through it, and a series resistor should be provided which will allow just this current to flow. So, for a 12v supply, the series resistor would be (by Ohm's Law), 12/0.020 = 600ohm.
5. The forward resistance of an LED is effectively Zero, so it should never be directly connected to a supply without a series resistor. If you do directly connect it, even to just 5v, it will blow. (Five divided by Zero = Infinity!!)
6. The supply voltage for an LED doesn't matter - it can be 5v, 12v, 15v, 24v, 500v... What does matter is the forward current. Get this right.
7. LEDs can be connected together in series without a problem, like old-style christmas tree lights. Their total "chain" resistance is still almost Zero, so the same rule applies: you need ONE series resistor for the lot, at the correct value. (Of course, if one LED fails, they all go out.)
8. You can wire up LEDs in parallel, but it is usually recommended to limit the number to about five, with one current-limiting resistance for the group.
Dimming:
The output (brightness) of an LED across its designed output range is pretty linear, but LEDs which are to be dimmed require very tightly controlled (stable) circuitry and supply voltages for them to work acceptably - which is why cheap dimming units may prove unsuccessful.
Two approaches:
- Analogue (variable, but stable, current supply)
- PWM (variable pulse-width modulation at a frequency the eye will not see as flicker)
More info here:
http://www.digikey.com/us/en/techzone/l ... n-led.html
For dash-illumination purposes (maybe 20+ LEDs) it is possible to dim the LEDs simply by reducing the current flowing through them, but
(a) it won't be possible to do this with the existing rheostat (for controlling a much higher current tungsten array) - one of dedicated value, or a variable-output DC supply pcb will be needed (together with series resistors for each of/groups of the LEDs), and
(b) although the light output can be brought down from maximum to something lower by this means, the behaviour of the individual LEDs will become a bit erratic as their light output becomes very low. They won't dim to black smoothly; for this you need PWM, where the LED is operated at 'design' light output, but only for very brief periods, which the eye sees as 'dim'.
(c) There is obviously an issue about the direction of light output from an LED. Most of them project light forward (180deg cone), while dash illumination often requires omni-directional/side-output lighting (eg behind a display panel, or between a pair of instruments). Choice of LED design will be important, or additional LEDs may be necessary. An out-of-car in-the-dark bench test will be essential!
Sorry if you knew all this before!
Chris
See also:
http://led.linear1.org/category/led-basics/ - What's possible
http://led.linear1.org/led.wiz - LED arrays and circuits
http://www.rapidonline.com/Electronic-C ... cd-55-2228 - Example of a "warm white" LED; plenty more
Saw this (enormous) thread today, and was aware there had been a good bit of discussion about LEDs as dash-illumination, and as "long life" substitutes for small tungsten lamps.
Several things to note about LEDs:
1. To the circuit in which they are placed, they look just like a Diode (i.e. they will conduct in one direction only).
2. Their tolerance to reverse current (i.e. being connected the wrong way round) is limited - they will blow.
3. They will emit their maximum designed light output at a specified forward current (If, or Ifwd), which is often around 10-30mA for single 'indicator' LEDs.
4. It is this forward current which is critical. If the LED is spec'd as needing 20mA, that is the maximum that should flow through it, and a series resistor should be provided which will allow just this current to flow. So, for a 12v supply, the series resistor would be (by Ohm's Law), 12/0.020 = 600ohm.
5. The forward resistance of an LED is effectively Zero, so it should never be directly connected to a supply without a series resistor. If you do directly connect it, even to just 5v, it will blow. (Five divided by Zero = Infinity!!)
6. The supply voltage for an LED doesn't matter - it can be 5v, 12v, 15v, 24v, 500v... What does matter is the forward current. Get this right.
7. LEDs can be connected together in series without a problem, like old-style christmas tree lights. Their total "chain" resistance is still almost Zero, so the same rule applies: you need ONE series resistor for the lot, at the correct value. (Of course, if one LED fails, they all go out.)
8. You can wire up LEDs in parallel, but it is usually recommended to limit the number to about five, with one current-limiting resistance for the group.
Dimming:
The output (brightness) of an LED across its designed output range is pretty linear, but LEDs which are to be dimmed require very tightly controlled (stable) circuitry and supply voltages for them to work acceptably - which is why cheap dimming units may prove unsuccessful.
Two approaches:
- Analogue (variable, but stable, current supply)
- PWM (variable pulse-width modulation at a frequency the eye will not see as flicker)
More info here:
http://www.digikey.com/us/en/techzone/l ... n-led.html
For dash-illumination purposes (maybe 20+ LEDs) it is possible to dim the LEDs simply by reducing the current flowing through them, but
(a) it won't be possible to do this with the existing rheostat (for controlling a much higher current tungsten array) - one of dedicated value, or a variable-output DC supply pcb will be needed (together with series resistors for each of/groups of the LEDs), and
(b) although the light output can be brought down from maximum to something lower by this means, the behaviour of the individual LEDs will become a bit erratic as their light output becomes very low. They won't dim to black smoothly; for this you need PWM, where the LED is operated at 'design' light output, but only for very brief periods, which the eye sees as 'dim'.
(c) There is obviously an issue about the direction of light output from an LED. Most of them project light forward (180deg cone), while dash illumination often requires omni-directional/side-output lighting (eg behind a display panel, or between a pair of instruments). Choice of LED design will be important, or additional LEDs may be necessary. An out-of-car in-the-dark bench test will be essential!
Sorry if you knew all this before!
Chris
See also:
http://led.linear1.org/category/led-basics/ - What's possible
http://led.linear1.org/led.wiz - LED arrays and circuits
http://www.rapidonline.com/Electronic-C ... cd-55-2228 - Example of a "warm white" LED; plenty more


