Moderated Discussion Areas
ContinuousWave: Small Boat Electrical
LED Sensitivity to Voltage
|Author||Topic: LED Sensitivity to Voltage|
posted 07-26-2013 01:27 PM ET (US)
Four years ago I installed a couple of 3" courtesy lights on the side of the RPS on my 170 Montauk as an aid to boading/docking at night. Each had 3 white LEDs in it. Two of the three in each fixture are now burned out.
One of the reasons I got LED lights was supposed longevity, they are not replaceable and this basically means my fixtures are junk.
I notice in newer models of these fixtures that the LEDs are flat in profile, whereas those in the old fixtures were more conventionally bulb-shaped. So it appears the bulb technology has changed.
I also observed that the old units listed input voltage as 12VDC - not a min-max range. Since the DC system at steady state when the alternator is running is in the high 13's, I am wondering if they were being overdriven. That seems like poor design when the bulbs are not replaceable.
Do I have a better chance with the newer units with the flat LEDs? They too list "nomimal 12V input", and I'm hesitant to spend time and money if they too are going to crump.
posted 07-26-2013 01:48 PM ET (US)
The difference in the appearance of the LED lamps is probably due to the use of surface mount packaging of the newer LED lamps.
An LED is a diode. The light output produced is proportional to the current flow through the diode junction of the LED lamp. There is a limit to the amount of current flow through the diode. If that limit is exceeded, the diode can be destroyed.
The voltage needed to operate an LED is only the forward-voltage drop of the diode, which is usually in the range of 1.0 to 2.0-Volts. An LED diode cannot be connected directly to 12-Volts; too much current will flow, destroying the diode in a very short time, although there may be some brilliant light emitted during that time.
The general circuit for powering an LED must provide for protection against too much current. One way to accomplish this is to include a resistor in series. The value of the resistor is chosen to limit the current to a safe value. I will explain with an example.
Let us say we have a light-emitting diode that has a forward voltage drop of 1.5-Volts. The maximum current permitted is 0.5-Ampere. If we wish to power the LED from 12-Volts, we must limit the current with an appropriate resistor. The voltage drop across the resistor will be 12.0 - 1.5 = 10.5-Volts.
In a circuit where 10.5-Volts occurs with a flow of 0.5-Ampere, the resitance must, by Ohms Law, be
R = E/I
R = 21-Ohms
Putting a 21-Ohm resistor in series will limit the current in this curcuit to about 0.5-Amperes.
The above is not a very robust design. There are several problems. First, the source voltage is likely to be variable, and, in the case of a battery being charged, could rise to as high as 15-Volts. Also, there can be some variation in the LED's for forward voltage drop. We might get one with more or less voltage drop, which would affect the current.
Also, like any diode, there is a peak reverse-voltage that can be tolerated, For LED's this value is often very low. It might be prudent to protect the LED against reverse-voltage transients.
A more sophisticated circuit might employ a few additional components to create a precision current limiting circuit. Such a circuit would limit the current through the LED to a specific value, no matter how the input voltage might varying over a reasonable range, say from 10-Volts to 18-Volts.
The experience cited above with very short life of an LED is probably common in automotive after-market lighting. There are probably a lot of poorly designed LED replacement lamps on the market. The lamps probably did not have very good protection against over-current or inverse-voltage. Also, the life span of the LED is very proportional to the current. If the LED is being run with a lot of current--needed to get the bright light output that is often the attraction of the replacement lamp--the life span is going to much shorter than the same LED run at lower current and less light output.
posted 07-26-2013 04:37 PM ET (US)
Thanks for the complete explanation.
Without reverse engineering the circuit board, it is easy to see there are only a couple of resistors on it.
These were Attwood Marine Lights. Obviously saying "Marine" does not automatically make it so.
posted 07-26-2013 05:21 PM ET (US)
One other unstated but huge disadvantage to using a resistor divider rather than a voltage regulator is power. In the example of the 21 ohm resistor, it is dissipating some 5 watts to limit the current to the diode. This negates the low power advantage enjoyed by LEDs.
Quality LED lighting uses active voltage regulation. You'll pay more for it initially, but will realize the cost benefits in the long run.
posted 07-26-2013 07:26 PM ET (US)
I have not really investigated LED control at all, but I see there are now some specialized integrated circuit devices engineered just for the purpose of controlling LED lamps. For example, see
I am sure these devices provide much better power efficiency compared to a series resistor dropping the voltage.
The other approach to LED wiring is to put several in series. If the forward-voltage drop were 1.5-Volts (as our hypothetical LED has), then eight of them in series would produce 12-Volts of drop. In theory you could connect the string directly to 12-Volts.
posted 07-27-2013 11:44 AM ET (US)
Assuming these so-called marine lights have a simple resistor divider, how sensitive would the LEDs be to overvoltage?
Most consumer devices with LEDs in them are nonreplacable, hence I assume they must have some tolerance to a range of voltages.
It is also possible that voltage regulators are so inexpensive that all LED applications now use them. This technology, like all, is fast-moving - I remember when it was a big deal to get white LEDs - not so long ago.
posted 07-28-2013 12:02 AM ET (US)
Yes, most consumer LED devices have a current-limiting chip today.
I've peeked inside a couple.
posted 07-30-2013 10:03 PM ET (US)
I have been browsing, looking for a reasonably simple circuit for driving an LED. It looks like using a three-terminal linear voltage regulator in a constant-current mode is probably the simplest. The drawback of this approach is the power dissipated in the regulator will again be higher than the power used by the LED. A string of three or four LED's in series could be run from one regulator.
If I come up with a workable circuit, I will pass it along.
posted 08-01-2013 02:23 PM ET (US)
Jim, you want something that does pulse width modulation.
No power dissipation. On or off.
Purchase our Licensed Version- which adds many more features!
© Infopop Corporation (formerly Madrona Park, Inc.), 1998 - 2000.