# How to light an LED with very low voltage and low current?

The goal is to demonstrate, visually (with a light of some sort) that power is being generated, albeit very, very small amounts of power. What kind of set up could I use to light up an LED (or any visual indicator).

My power supply is incredibly small; about 0.1 volts and I'm not sure of the current, but consider it low.

I've heard of things such as a "joule thief" that could maybe allow me to light up an LED with very low voltage, but I believe that works by sucking current, which I don't believe I have enough of, to do that with this very small power supply.

Another option-- if my power supply is constant, but very, very small, can I somehow use a capacitor to store the small amount of energy trickling out of my power supply to then make an LED or other indicator light up (even just flash) once every few seconds or every few minutes?

• And I assume you can't just use a voltmeter for some reason? Feb 28, 2013 at 20:13
• Might be able to use a boost topology, to get a higher voltage, charge a capacitor over a few minutes and flash the LED using the energy stored on the cap every few minutes... Feb 28, 2013 at 20:27
• "Low" is a useless spec for current. Feb 28, 2013 at 22:16
• I was able to light an LED via a joule thief powered by lemons, potatoes, etc. with copper and zinc wires in them. Just put more of them in series ;) Mar 1, 2013 at 1:17
• You may struggle, this is about the best I've seen: dangerousprototypes.com/2013/02/02/… Mar 1, 2013 at 14:08

This is going to be very difficult at best. Transistor circuits don't work with just 100 mV power supply voltage. If you really need to run something from this low voltage, the circuit will need some initial external power to get going. This could charge up a cap to a few volts, which could then run a boost converter which would boost the 100 mV up to the few volts to keep itself going. This is of course assuming there is enough power available to not only run the circuit at the higher voltage after conversion losses, but to have enough left over to light up a LED to a noticable level.

Let's say you get the best efficiency LED you can find and that it is bright enough a 500 µA and drops 2 V in the process. That's 1 mW. Without conversion losses and power needed by the circuit, that would mean the 100 mV supply would need to source 10 mA. If your 100 mV supply can source a few 10s of mA, maybe you have a chance, but you'd still need some initial external energy to get the bootstrap process going.

• Heh, you just wrote, "power voltage" Mar 1, 2013 at 1:08
• @Gbarry: Actually I said 100 mV power voltage. This makes it clear I am talking about a voltage, so "power" in this case is used to describe which voltage. How would you have said it? I suppose "power supply voltage" would be more complete, but was what I said really ambiguous? In any case, I added the extra "supply". Is that good enough for you? Mar 1, 2013 at 13:52
• Your technical insight is VERY appreciated, Olin. This gives me much to think about. My background is not in electricity, so forgive my ignorance with this question: If a power supply trickles out a constant amount of energy, will a capacitor eventually store up enough energy to then release it at a given amount, if only for a very small amount of time? I'm still learning, a lot.
– Zach
Mar 1, 2013 at 20:47
• @Zach: Yes, a capacitor can store energy slowly and then release it quickly. However, the circuit that will occasionally pump some energy from the 100 mV to the 2 V level still has to be powered at least at its sleep current level. If there isn't enough average power to go around, this scheme isn't going to work. Mar 1, 2013 at 21:02

If you can squeeze just a little more voltage out of your plant-based power source, the TPS61201 boost regulator can step 0.3v up to 3.3v, which would be enough to light an LED.

As other have mentioned, you may want to charge up a supercap and discharge it into the LED every few minutes. You can get 555 timers that work at 3.3v and draw 30 µa.

• You can get a microcontroller that works at 3.3 V and draws far less than 30 uA. A 555 timer probably blows the whole power budget on its own. Mar 4, 2013 at 21:28
• @OlinLathrop, good point. More options for control with the micro. Mar 4, 2013 at 21:38

I think your only way out is going to be with a galvanometer. Perhaps you can "demonstrate visually" by blocking or reflecting a light source when the galvanometer moves.

• I had not thought of a galvanometer in this way, and I will certainly consider it. Thanks for your thoughts!
– Zach
Mar 1, 2013 at 20:53

One very interesting answer to this question is in the area of Energy Harvesting, which many makers/hobbyists/beginners don't know about.

Since there seems to be general knowledge on this site about Energy Harvesting, I don't know why nobody has mentioned this before.

Energy Harvesting is taking advantage of some kind of (often low-level) energy present in the environment in which the circuit will operate, to power (or partially power) the circuit itself, with popular sources being vibration=>piezoelectric=>buck converter, or heat-and-or-cold=>thermoelectric generator=>boost converter, or small solar cell=>boost converter. These are special converters, having to work with especially low currents and/or voltages, and having to be especially efficient themselves, and implement an especially efficient power conversion. Because it is fairly challenging to do this, it is usually implemented using a circuit highly tuned to capture a specific frequency / low voltage range / low current range of energy.

There is a lot of information on the web for this, if you google "Energy Harvesting" (with the quotes).

Here are two links more closely targeting your application (but still being somewhat general):

1. Energy Harvesting at DigiKey

2. Energy Harvesting at Analog Devices (looking at LTC*):
#1 -- Parametric search,
#2 -- Google of "Energy Harvesting" at analog.com

Now specifically related to your particular application...

The LTC3109 integrated circuit, an "Auto-Polarity, Ultralow Voltage Step-Up Converter and Power Manager", has been designed for use with TEG's -- thermoelectric generators -- to be able to power sensors and data collection in challenging environments where it's impractical to use conventional power sources (like in space, at the top of a mountain, at the bottom of the ocean floor, or Antarctica -- anywhere where it would be expensive or inconvenient to replace batteries).

If you look at the LTC3109 datasheet, this chip is a boost converter that uses a transformer with some capacitance to form a resonant tank circuit, and the transformer helps boost the voltage from there. MOSFETs inside the chip probably perform synchronous switching to maximize efficiency at what are still fairly low voltages.

It will actually work down to an astounding 30mV! That's 0.030 of a volt! This circuit fits your application very nicely, since your circuit produces 100mV.

Then you will have to pair this boost converter chip with an ultralow power microcontroller, such as MSP430* from Texas Instruments (there are many others you could use like the excellent examples from Microchip.)

You'll have to program the chip to sleep often in order to conserve energy while the Energy Harvesting chip charges up a capacitor or supercapacitor, and then, when sufficient energy is reached, you can very briefly flash an LED. If you do this, you will learn a lot!

The transformer that is required must be very specific to the energy being harvested. From the datasheet it says (emphasis mine):

For input voltages as low as 30mV, transformers with a turns ratio of about 1:100 is recommended. For operation from higher input voltages, this ratio can be lower. See the Applications Information section for more information on selecting the transformers.

From the datasheet, here is a table that lists different suggested transformers, along with their turns ratios and the corresponding minimum voltage they should operate with:

You should be able to use the 1:20 transformer, but you could try the 1:50 and the 1:100 and see how they work out -- you may find out that one works better than another for your application.

Here is the block diagram of the chip from the datasheet (linked above) so that you can have some idea of what it is, and what it's doing:

Full disclosure: I have no relationship with Linear Technologies (LTC) or Analog Devices, and so have no conflict of interest here. I just like these technologies, and this chip in particular (though I've never used it).

I know this post is 10 years old, but I thought that this answer was still worth posting. Enjoy!

LED will be hard; instead you can use the lcd off of a small calculator. very small amount of power required to "light" them. Remove the batteries (if any) and remove the solar cell. Connect your supply to where the solar cell connects.