Well, here then, is another Big Clive video, describing a circuit called the "Joule Thief", which runs an LED (forward voltage ~2.5v) on a nearly completely depleted AA cell (0.5v measured):

It works by using flyback voltage from a tiny transformer hand wound on a ferrite bead, and uses only 1 common NPN transistor and a 1K resistor...surely you can make a single LED run for a long time on a rechargeable battery with this circuit. But you can't defeat the laws of nature; the total energy capacity of the battery must be enough to drive the LED at your desired current for the period you want ...this circuit just squeezes the last bit of energy from a battery, and makes it useful, even if the voltage is too low to be used directly.

Hope you find this useful. For more info, a Google search for 'joule thief' might be instructive...

A Joule Thief is also unrelated to running an LED continuously as 20mA.

Please stop posting answers until you are sure that they actually relate to the question.

 

I have seen no discussion here of the required brightness of this LED. Everyone here is assuming 20mA. But that is ridiculously high for a modern LED that just needs to be visible in normal indoor lighting. Modern high-efficiency LEDs are quite adequate at as low as 1mA, allowing a battery 1/20 the capacity being discussed. At 1mA, two AA cells could run this for 80 days.

There are rechargeable Li Ion AA “batteries” with a buck converter built in and a standard charging port, two of these would easily light your LED at even 5 mA for 15 days.

On the other hand, if it needs to be viewable over a wide angle in full sunlight, the 20mA LED will not do it.

So @PFAB, how bright do you really need this? The environmental lighting level is important, even the angle over which it must be viewable is important. Let’s do some engineering here instead of sticking our wetted finger up in the wind.

 

I am still wondering about the application of such, other than as a waypoint marker for cave exploration. Orpossibly tunnel network traveling, maybe to warn about traps set in a tunnel? Caves and tunnels are the places that do get inky dark.This relates very much with the previous post. The application does matter a whole lot.

 

MisterBill2 BobTPH schluderjupp Ya’akov schwartzenheimer


First-off I’d like to profusely thank everyone who has been chiming-in to my query about this circuit I’m typing to achieve.
For complete transparency, I am not an electrical engineer of any sort, and feel like a total newbie here. Many people think I’m a pretty clever guy in some areas, this definitely ain’t one of them.

Based on a lot of your input so far, I’ve had to rethink not only my approach, but also my needs. I also want to be cognisant and sensitive to everyone’s time (and not waste it!), especially given that you’ve all been incredibly generous so far.

I now realize that I should have better thought-out both in my overall approach and my needs. Speaking of the latter, I had originally mentioned in my original post that the LED would have to be on for a “continuous” period, or “as long as possible”, be it anywhere from 10 to 15 days. What I was (unsuccessfully) trying to convey is that I’m looking for a way to not have to recharge this circuit often.

SO… as Mr. BobTPH but it so eloquently, “Let’s do some engineering here instead of sticking our wetted finger up in the wind.”

A FEW CLARIFICATIONS:

• I’m building a device that will feature a single orange (or amber) LED. 3V seems ideal. I’m using 20mA as a general load.
• Some applications will require an always-on (steady) LED, some will require a blinking (on-off) LED, and some will require a “Breathing” LED (Slow pulse). I’m rating all of these generally at 3V 20mA as per the LED manufacturer’s specs.
• This LED will be powered by a battery (more on this later)
• The LED is activated by a small push-button switch (most likely a momentary switch, not a toggle)
• The LED will stay lit for 3 hours, then automatically turn-off
• Someone can repeat this once a day (but is not obligated to), and will be able to do this for about 6 months (approximately 180 days) before having to fully recharge the battery. The key here is not to have to recharge this daily, or weekly or even monthly.
• A direct connection to a power supply to power the device is out of the question. The only direct connection is to recharge it.
• The device has to be safe.
• The battery (or battery configuration) is rechargeable, therefore must in all likelihood have a BMS.
• The ideal physical size of the battery should lean towards thin. No more than 1/2" thickness would be ideal.
• The connection, or connector, to recharge the battery is also important. I’m going back and forth on this. It could of course be as simple as a USB-C plug and a 5V charger. That’s what I have now, but while it’s mainstream, it’s not my favorite. Even though the plug is reversible, it still requires you to line things up in order to make the connection. If not that, then a simple barrel connector might work even better than a USB-C approach. In an ideal world, an induction approach would be best: think the magnetic charger for an Apple Watch (magsafe).
• The device will be shipped to consumers, and therefore MUST be safe in regards to transportation.
• The battery cannot be recharged by Solar or Wind (The end product lives indoors).
• The LED bulb should be as bright as possible, within reason, without taxing the battery excessively. I realize that this is not a mathematical explanation. The bulb should be clearly visible as lit in a room during daylight hours. The diameter of what I am trying to illuminate is about 5mm wide.
  

Circuit Design

Say i’m using a simple circuit design, such as 3V, 15mA LED and a 3V battery with a capacity of 31.6Wh, the key components will be:

• Battery: I would use a 3V battery with a capacity of 31.6Wh. Since the battery’s voltage matches the LED’s requirement, I don’t need any voltage regulation.
• LED: The LED I am using is a 3V, 15mA one.
• Current Limiting Resistor: Normally, I’d probably need to use a current-limiting resistor to protect the LED. But here, in my simple scenario, since the LED and the battery have the same voltage, and the battery might have internal resistance, I might choose not to include it. If I decide to include it, I would calculate its value using Ohm's law, R = V / I, where V is the voltage difference between the battery and the LED, and I is the current (0.015A). In this case, the resistance value would be very low.
• Timer Circuit: I need a simple timer circuit that can turn the LED on for 3 hours. I could use a 555 timer in monostable mode, or a basic microcontroller like an Arduino or a simpler MCU that can be programmed for a 3-hour delay.
• Switch: A manual switch, like a push-button or toggle switch, is necessary to activate the timer.
• Optional Protective Components: Depending on safety concerns, I might include a fuse or a polyfuse for overcurrent protection.
• Circuit Operation: When I press the switch, it activates the timer circuit, which powers the LED for 3 hours. After the duration, the timer automatically turns off the LED.
• Circuit Diagram: In my diagram, the battery would be connected to the switch, which in turn is connected to the timer circuit, and the timer is connected to the LED (and a resistor, if used). The ground (negative side) of the battery connects back to the ground of the LED and timer circuit, completing the circuit.
  

I think that this design would be straightforward and fulfill the requirement of keeping the LED on for 3 hours once activated. For the timer, I could choose a 555 timer for simplicity or a microcontroller for more versatility. I would imagine that a 555 timer would also be more efficient as far as battery life is concerned?

This setup illustrates a basic circuit where the switch controls the activation of the timer, which in turn controls the LED.

Battery Comparisons:

With so many options out there (thank you for those suggestions), I ran a quick comparison of Lithium-Ion (Li-Ion), Lithium Iron Phosphate (LiFePO4), and various Sodium batteries, including Sodium-Ion (NIBs), Sodium-Sulfur (NASBs), and Sodium-Metal (NMBs). Here's my simplified comparison:

Feature/Type	Lithium-Ion (Li-Ion)	Lithium Iron Phosphate (LiFePO4)	Sodium-Ion (NIBs)	Sodium-Sulfur (NASBs)	Sodium-Metal (NMBs)
Chemistry	Lithium cobalt oxide (LiCoO2)	Lithium iron phosphate (LiFePO4)	Sodium and carbon compounds	Sodium and sulfur	Sodium and metal compounds
Safety and Stability	Lower (risk of overheating)	Higher (more thermally stable)	Moderate	Moderate to high (high temperatures)	Moderate
Lifecycle	Moderate (500-1,500 cycles)	High (2,000-3,000+ cycles)	Moderate to high	High	Moderate to high
Nominal Voltage	3.6-3.7 V	3.2-3.3 V	~3.6 V	~2.0 V	~3.0 V
Energy Density	High	Moderate	Low to moderate	Moderate to high	Moderate to high
Cost	Moderate	Higher (initially)	Lower	Moderate (infrastructure cost)	Moderate to high
Environmental Impact	Moderate (use of cobalt)	Lower (more eco-friendly)	Lower (abundant materials)	Moderate	Lower (abundant materials)
Operating Temperature Range	Standard	Wider range	Standard	High (operates at high temperatures)	Standard

For powering a single LED 3V diode (orange color), the best choice would depend on my specific needs:

• If compact size and high energy density are priorities, a Lithium-Ion battery might be suitable.
• For enhanced safety, longevity, and if size isn't a primary concern, a LiFePO4 battery is a good option.
• If I’m looking for an environmentally friendly and potentially cost-effective solution, and if the slightly bulkier size isn't an issue, Sodium-Ion (NIBs) would be worth considering.

Considering the typical requirements for a small, single LED diode, where high energy density isn't as critical, and assuming a desire for safety and longevity, LiFePO4 might be the best fit. It provides a good balance of safety, lifespan, and has a suitable voltage range for a 3V LED. However, if cost and environmental impact are my main concerns, then Sodium-Ion batteries could also be a viable choice.


Battery Capacity Calculation:
I will calculate the battery capacity required to run a 3V LED at different current ratings for 3 hours daily over a period of 6 months. I will consider scenarios where the LED runs at 20mA, 15mA, and 10mA.

• LED Power Consumption:
  • At 20mA: P = IV = 0.020 A x 3 V = 0.06 W
  • At 15mA: P = IV = 0.015 A x 3 V = 0.045 W
  • At 10mA: P = IV = 0.010 A x 3 V = 0.03 W
    
• Daily Energy Consumption:
  • At 20mA: 0.06 W x 3 hours = 0.18 Wh/day
  • At 15mA: 0.045 W x 3 hours = 0.135 Wh/day
  • At 10mA: 0.03 W x 3 hours = 0.09 Wh/day
    
• Total Energy Consumption Over 6 Months (180 Days):
  • At 20mA: 0.18 Wh/day x 180 days = 32.4 Wh
  • At 15mA: 0.135 Wh/day x 180 days = 24.3 Wh
  • At 10mA: 0.09 Wh/day x 180 days = 16.2 Wh
    
• Battery Capacity Required (including a 30% buffer for inefficiencies):
  • At 20mA: 32.4 Wh x 1.3 ≈ 42.1 Wh
  • At 15mA: 24.3 Wh x 1.3 ≈ 31.6 Wh
  • At 10mA: 16.2 Wh x 1.3 ≈ 21.1 Wh
    

So, I would need approximately :

42.1 Wh for running the LED at 20mA

31.6 Wh at 15mA
and
21.1 Wh at 10mA, to last for 6 months with daily use. (3 hours per day)

Does that generally make sense? Am I on the right track ? Y/N ?

-------------------------------------------------

You all have all contributed amazing knowledge so far.

For example:
- I was unaware about the use of different chemistries that I might consider…
- I was unaware about the fact that a used/depleted battery might still have quite of actual capacity left in it, and that if you mix Vodka and Pepsi you can create a Joule Thief circuit that allows you to capitalize on this. (LOL)

As for the final application (I’m looking at you MisterBill 2) there are 4 people involved in this project so far, me, myself and I… and my boss (the wife). 3 of those would gladly sink a few Vodkas & Pepsi concoctions and share the goods, but the only one that matters is the boss right now and she's sworn me to secrecy until the time is right, hope you can understand my position.

Again, all of this is SOooooo out of my sphere, and this is why I’m looking for guidance from real professionals or gurus such as yourselves.
Have a great weekend and wishing you all a Happy Easter to you and your families.
Cheers,
PFAB

 

Thank you, now we are getting somewhere.

My first comment is about your choice of 20mA. You said:

I’m rating all of these generally at 3V 20mA as per the LED manufacturer’s specs.

20mA is the max current the manufacturer says you can use continuously, not the recommended current. The LED can be run at any current below that, and, is in fact more reliable and efficient if you run it at less. And two different LEDs run at 20mA may produce different amounts of light and over different angles.

Second is about the 3V. A 20mA orange LED will actually be about 2 to 2.2V when run at 20mA. At 5 mA, it might be 1.8V. If you see a 3V 20 mA LED, it is likely an LED with a resistor. You don’t want this because it will limit you.

Now lets talk a little more about what you are trying to achieve, because it will profoundly affect how much light you need.

At one point you say:

The bulb should be clearly visible as lit in a room during daylight hours. The diameter of what I am trying to illuminate is about 5mm wide

So are you viewing the LED directly, or are you viewing an area illuminated by the LED? This could have a dramatic effect on how much light is needed. If viewing via reflected light, how reflective is the area being illuminated?

And from what range of angles are you viewing it? If The LED is viewed directly and always from the same position, it might take 1/1000 of the power that it would to illuminate a 5mm area that is not very reflective visible from every direction.

Each LED spreads it light in a cone with a particular angle, which might be 120° (wide) or 5° (narrow). This is part of the manufacturer’s specification.

This is important whether you are viewing directly or illuminating an area, and something you have to consider to get the most efficient design.

The kind of considerations I am talking about here are what distinguish an engineer from a hacker.

If all of the above come out to the best case, i.e. directing viewing over a narrow angle, a coin cell could be all you need. Simply designing for 20 mA continuous might be okay for the worst case, or might not. We still don’t know enough.

 

Thank you, now we are getting somewhere.

My first comment is about your choice of 20mA. You said:

20mA is the max current the manufacturer says you can use continuously, not the recommended current. The LED can be run at any current below that, and, is in fact more reliable and efficient if you run it at less. And two different LEDs run at 20mA may produce different amounts of light and over different angles.

Second is about the 3V. A 20mA orange LED will actually be about 2 to 2.2V when run at 20mA. At 5 mA, it might be 1.8V. If you see a 3V 20 mA LED, it is likely an LED with a resistor. You don’t want this because it will limit you.

Now lets talk a little more about what you are trying to achieve, because it will profoundly affect how much light you need.

At one point you say:



So are you viewing the LED directly, or are you viewing an area illuminated by the LED? This could have a dramatic effect on how much light is needed. If viewing via reflected light, how reflective is the area being illuminated?

And from what range of angles are you viewing it? If The LED is viewed directly and always from the same position, it might take 1/1000 of the power that it would to illuminate a 5mm area that is not very reflective visible from every direction.

Each LED spreads it light in a cone with a particular angle, which might be 120° (wide) or 5° (narrow). This is part of the manufacturer’s specification.

This is important whether you are viewing directly or illuminating an area, and something you have to consider to get the most efficient design.

The kind of considerations I am talking about here are what distinguish an engineer from a hacker.

If all of the above come out to the best case, i.e. directing viewing over a narrow angle, a coin cell could be all you need. Simply designing for 20 mA continuous might be okay for the worst case, or might not. We still don’t know enough.

Hi Bob,
Thanks again for your insights and commentary, much appreciated.
regarding:

Second is about the 3V. A 20mA orange LED will actually be about 2 to 2.2V when run at 20mA. At 5 mA, it might be 1.8V. If you see a 3V 20 mA LED, it is likely an LED with a resistor. You don’t want this because it will limit you.

You are again correct. 20mA seems to be the high end of it all, but these could be run at less.
And indeed again, the color choice of the LED yields different ratings.

Regarding :

So are you viewing the LED directly, or are you viewing an area illuminated by the LED? This could have a dramatic effect on how much light is needed. If viewing via reflected light, how reflective is the area being illuminated?

The LED will be placed inside of a 3/16 inch thick foamcore board though a small hole in the board. The LED illuminate a thick paper that is glued to the foamcore board. So what the viewer will see is about a 1/4" illuminated circle/glow on the paper. Hence, the bulb should be as bright as possible dso that it can illuminate sufficiently though the thick paper. Also, this will be viewed at all angles. Not sure if that makes sense.

 

So what the viewer will see is about a 1/4" illuminated circle/glow on the paper. Hence, the bulb should be as bright as possible dso that it can illuminate sufficiently though the thick paper.

Since energy use is a concern, I would experimentally determine the minimum current that will adequately illuminate the spot on the paper.

 

The LED will be placed inside of a 3/16 inch thick foamcore board though a small hole in the board. The LED illuminate a thick paper that is glued to the foamcore board. So what the viewer will see is about a 1/4" illuminated circle/glow on the paper.

Good information.

To get the most light on the 1/4 inch circle, you will want an LED with a clear lens and a fairly narrow angle because you want all the light concentrated in the forward direction. Ideally the beam from the LED would just fill the circle of paper you want to illuminate. If you use a diffused LED one with wide angle, most of its output is wasted.

Don’t worry about the viewing angle, the paper acts like a diffuser and turns your narrow angle LED into a dot of light viewable over a 180° angle.

With an efficient LED with the correct angle dispersion, I think you will need way less than 20mA.

Check out this about beam angles:

LED beam angles

 

Good information.

To get the most light on the 1/4 inch circle, you will want an LED with a clear lens and a fairly narrow angle because you want all the light concentrated in the forward direction. Ideally the beam from the LED would just fill the circle of paper you want to illuminate. If you use a diffused LED one with wide angle, most of its output is wasted.

Don’t worry about the viewing angle, the paper acts like a diffuser and turns your narrow angle LED into a dot of light viewable over a 180° angle.

With an efficient LED with the correct angle dispersion, I think you will need way less than 20mA.

Check out this about beam angles:

LED beam angles

Hi Bob, thanks again. I actually just received some micro LEDs that I had on order to try them out. One of them is the perfect candidate. I've also returned back to using two AA batteries in series to achieve close to 3 V. I tried to power this new LED with this power supply, and it's absolutely wonderful. I still don't really know how to go about creating a mechanism to turn off the LED after three or four hours. I looked at a few videos online, many using a CMOS-type 555 timer. This starts to get a big complex for me even though I'm sure it's incredibly simple for others. Ideally, I'm probably more at the stage right now where I would love to engage the services of an electronics hobbyist or engineer to help me achieve my goals. Could you recommend a person or a website as a resource?

 

I have seen published a circuit for a long time delay using a simple series FET biased on by the charge stored in a high value capacitor. Very simple, but I am not certain that it is a solution, but it might be what the application requires.