Can a LED replace a voltage regulator?

[Lebe]

I need to drop from 5 to 3 volts in my circuit. I originally thought of just using 5/3.3 voltage regulator, but the thought just came into my head: couldn't I just use an LED with a voltage drop of 2V instead? It should work the same right or am I missing something?

[Bill_Marsden]

Seems reasonable to me, though they aren't predictable, you'd have to have some sort of adjusting or compensating mechanism. I've never really seen a graph of voltage vs. current, but I'm under the impression it is pretty constant once their turned on.

[blocco a spirale]

As long as the current doesn't exceed the rating of the LED. 20-30mA.

[DickCappels]

Just be careful peak current. If your power supply has a fast rise time and there is a large decoupling capacitor on the other end of the LED, then the peak current through the LED could be pretty high.

And assure that there is a minimum load so your load can ever drop to "near zero", causing the voltage could rise (if, for example, your load is a microcontroller that goes to sleep).

[chrissyp]

The led will not act as a voltage regulator, it will only reduce the output voltage by a fixed amount. If the input voltage rises so will the output voltage , a true voltage regulator will maintain the output even with fluctuations in input voltage

[Bill_Marsden]

That fixed amount is exactly what a zener does, only reverse biased. You can even use a regular voltage drop in much the same way from a regular diode. Modern LEDs drop as much as 3.5 volts, which puts them comfortably in the bottom of a zener diodes range, and from what I understand a zener that low isn't too stable, so an LED might be OK.

[beenthere]

For more current, a rectifier diode will drop a lower voltage but handle much more current.

[hgmjr]

Originally Posted by Lebe

I need to drop from 5 to 3 volts in my circuit. I originally thought of just using 5/3.3 voltage regulator, but the thought just came into my head: couldn't I just use an LED with a voltage drop of 2V instead? It should work the same right or am I missing something?

While you can use the led as a regulator, I would caution you that the regulation will not be very tight.

hgmjr

[SgtWookie]

Unless your load is very constant I suggest against it, because the Vf of the LED will change quite a bit depending upon the current flowing through it.

It will be fairly good if you have a load that will always be between 10mA to 20mA, but you'll have several mV variance even in that range. If your load is reactive (ie: capacitive or inductive) the transient loads would likely cause the LED to fail.

Standard silicon rectifiers such as a 1N4001-1N4007 would be much more robust, but even those will have a significant variation in Vf with load current.

Months back, I graphed a few diodes for Vf over a range of currents. I've attached a graph for a 1N4002 diode. What this graph doesn't show is how the Vf rapidly increases as the diode reaches it's maximum rating - because the test only went to 100mA. A similar test of a 1N4148 diode showed a marked increase in Vf above 50mA.

You could perform your own test using an LED, some resistors, a power supply and a couple of meters. If you are careful and record your results accurately, you may learn quite a bit on how LEDs perform in a circuit.

If you exceed the maximum current rating of the LED, it will rapidly burn out. But, LEDs are cheap nowadays

[Audioguru]

Simply look at the poor voltage regulation on the datasheet of an LED.
My MV8191 LEDs have a max continuous current rating of 40mA.
Their voltage changes a lot when their current changes.

The actual voltage could be anything from maybe 1.3V to 2.4V at 20ma.