Hi Everyone!


I'm a bit of a super capacitor fanatic, and I've made dozens of circuits that employ them.  This circuit is a prototype that I'm turning into a DIY kit.  Itis relatively simple, and is pretty darn efficient.  There is also a lot of room for customization!  When I get my custom PCBs made, I'll be throwing this device into an old flash light housing.  For the time being, I'll be talking about the circuit as it is.

This is my entry for the MAD SCIENCE FAIR contest, as well as the MAKE IT GLOW contest, so if you liked this instructable, I'd sincerely appreciate your vote or a rating =)  I've done my best to be AS THOROUGH AS POSSIBLE!


What The Circuit Does:
Unfortunately, super capacitors can only be charged to lower voltages; typically around 2.5v or 2.7v as a standard.  If you place some super capacitors in series, you can charge to higher voltages, but you lose a tremendous amount of capacitance.  When you plug this device into a wall transformer ((I designed this device around a 9v@1A transformer),  the on-board microprocessor turns on a relay that connects power to the capacitor bank.  The series super capacitors then charge to 5.2v through the relay contacts.   The capacitors an be interchanged to use higher or lower values, depending on how much you want to spend.   The voltage on the capacitor bank is constantly being sampled by an ADC (Analog to Digital Converter) that is embedded in the microprocessor.  When the voltage exceeds a value of roughly 5.2v, a flag trips in software, and the MCU turns off the charging relay, at which point the green LED indicator will start and continue to blink as an indicator to show the user that the caps are charged.  You can leave this device plugged in for as long as you want, and the caps will be very much safe and sound.

When the caps are charged, the user can flip a switch that connects power from the capacitors into a DC-DC voltage booster.  The voltage booster takes the 5.2v from the capacitor bank and boosts it to a calibrated 8v.  The output voltage from the booster can be boosted anywhere from 3.4v to 34v, and is easily calibrated by an on-board 10-turn variable resistor.

Since I've calibrated the booster to output 8v, as soon as you flip the switch, the output of the booster will provide a constant 8v to the LED bank that acts to emit light.  The LED bank is meant for 12v, but works great at 8v, and consumes MUCH less current. However, the LED bank is much brighter at 12v.  The booster will continue to source power to the LED bank until the capacitors drain down to 3.4v, at which point the circuit shuts down.  At this point, if you can plug it in again, to charge back up to 5.2v.  

When the booster is tuned to output 12v, the circuit consumes quite a bit more current, but the light output is much greater.   If you're going to consider maximum brightness, you're going to want to use 2x 400f 2.7v caps in parallel with one another.  I also took the liberty of hooking up a $1 LED flash light head that I purchased from the dollar store directly to the capacitors as opposed to the booster, and it lasts MUCH longer.  See the video.


ITS CHALLENGING TO ME................

Well, yes. I have yet to meet a cat that could hold a soldering iron, let alone phathom ohms law.

I'm really intrigued by the possibility of powering things from capacitors instead of batteries.


* Super Caps have a longer life: can be charged/discharged far more times than any battery -- on the order of millions, compared to 500 to 1000 times for a secondary battery.
* No "memory effect".
* Super Caps can be discharged all the way to zero volts with no damage.
* Because of their incredibly low internal resistance, Super Caps can deliver tremendously high current at very little loss. Pound per pound, they pretty much kick the butt of any battery chemistry in terms of current delivery.
* Super Caps have a far better Power Density than any battery chemistry.
* Super Caps have a slightly better working temperature range than any battery chemistry.
* Super Caps tend to be safer to handle and far more "forgiving" to abuse [e.g. Lithium battery fires, Lead Acid hydrogen explosions, the tendency of batteries to leak corrosive chemicals]. Of course, a short across a fully charged Super Cap can be pretty hazardous -- shower of sparks, hot molten metal, etc. Also, a Super Cap cares not about it's orientation -- no concern for chemical leakage like with some batteries.


* Batteries, in general, have a much greater Energy Density than Super Caps, mainly because the energy that a chemical battery can store per unit of weight, is far greater than that of a Super Cap.
* A Super Cap's linear discharge voltage requires, either a shortened discharge time, or some sort of active compensation. Most battery chemistries are capable of keeping the discharge voltage relatively stable across the discharge curve (when drained at a nominal current rate for that battery).
* A Super Cap has a higher self discharge rate then most battery chemistries.
* Batteries/Cells can easily be ganged in series for higher voltages. Super Caps can also be ganged in series, but require some sort of charge balancing.
* Super Caps have a much higher cost/watt.

ReverseEMF7 months ago

Super Capacitors are difficult to charge efficiently. They present such a low impedance, they behave, essentially, like a hefty short, especially when they are, or nearly are, fully discharged. So, to charge them efficiently, the charger needs to have an extremely low output impedance [highest energy transfer occurs when the input and output impedances match].

The best way to do that is with a switch mode buck converter. Program the output voltage of the converter to the maximum charge voltage [which is really, the "over charge protection" feature]. The converter will repeatedly charge an inductor, and then dump the inductors energy, through a switching diode, into the SuperCap, until the capacitor's voltage reaches the converter's output voltage set point. If you use a converter that is capable of "burst mode", it will, then, only transfer energy to the super cap, if the cap's voltage sags -- which is likely, since SuperCaps, inherently, have relatively high leakage current [the leakage becomes greater, the higher the voltage across the cap].

Also, there is a reverse proportionality between voltage across the capacitor [i.e. the charge voltage] and the life of the capacitor. Holding a SuperCap that is rated at 2.7V, AT 2.7V, will significantly reduce the life of the capacitor over charging it to only 2.5V. This from the Maxwell Technologies BOOSTCAP Ultracapacitors Product Guide – Doc. No. 1014627.1:

30% reduction in rated capacitance may occur for an ultracapacitor held at 2.7 V after
5,500 hrs @ 65 oC
11,000 hrs @ 55 oC
22,000 hrs @ 45 oC
44,000 hrs @ 35 oC
88,000 hrs @ 25 oC
15% reduction in rated capacitance may occur for an ultracapacitor held at 2.5 V after
5,500 hrs @ 65 oC
11,000 hrs @ 55 oC
22,000 hrs @ 45 oC
44,000 hrs @ 35 oC
88,000 hrs @ 25 oC

Something to consider, when making the claim that the thing will last "forever". 88,000 hrs IS 10 years, but that's still less than 'forever'.

Solar panels, BTW, are excellent at charging SuperCaps. Since there are, essentially, current sources, they can deliver peak current, even into a short. The trick is to use a solar array with an open circuit [OC] voltage equal to or less than the highest voltage you want the capacitor to charge to. So, say you want your SuperCap to charge no higher than 2.5V, and lets say the OC voltage of a set of solar cells is 0.56V peak [at a 90 degree angle to full summer sunlight] and a blocking diode with a forward voltage of 0.4V at maximum current. (2.5 + 0.4)/0.56 = 5.17, thus, the panel will have no more than 5 cells in series and output a max voltage of 5 * 0.56 = 2.8V and with the blocking diode loss, the final highest voltage applied to the SuperCap will be: 2.8 - 0.4 = 2.4V.

Thus, the solar panel will efficiently charge the SuperCap to 2.4V and stop [and, actually, it will go a little higher, because as it approaches 2.4V, the charging current will drop off, thus lowering the forward voltage of the diode. The system will reach equilibrium when the charging current is equal to the SuperCap leakage current. So, a wise engineer will determine, via experimentation, what the forward voltage on that diode is, at the equilibrium current, across the projected range of temperatures the Cap+charger will be subjected to in the field [or at least the range specified in the instruction manual ;) ]

piddy05048 months ago

Hi great project thanks for sharing,just wondering can i use PIC12F683 chip instead

seanmft1 year ago

Pretty cool but I just want to point out that while you lose capacitance by putting capacitors in series, you don't lose energy. J=C/2(V^2). So with two 2.7 V, 50 F caps in series that's 25F/2(5.4V^2) = 364.5 J = 100F/2(2.7V^2). Whether you put them in series or parallel, their energy storage potential remains additive. You are however losing energy through the power converter.

Also, why so many LEDs? You could improve the optics greatly just by arranging the LED bank on a round board, and positioning it centered at the very very bottom of the reflector. Try it with a light meter, you'd be surprised. I'll bet it'll be brighter with half the LEDs running at full power

EngineeringShock (author)  acmefixer3 years ago
Power resistors and a diode to eliminate back powering. Currently, I'm charing my caps at 500ma (In this video). I've got an LM338 based charger at home that can charge up to 2.5A at 12.5v if you have the right resistor. To charge at 12.5v at 2.5A, you need:

R=E/I (12.5v/2.5A) = 5 Ohms
Power = E x I = 12.5v x 2.5A = 31.25 Watts
I'd use a 5 Ohm 40W resistor

I'd use a 16v 4A wall transformer, and a well heat sinked LM338 based variable power supply to tune to my charge voltage.
EngineeringShock (author)  acmefixer3 years ago
The LEDs work at 8v. Tried and tested. They are much dimmer than at 12v, which is nominal. 10v is good because you get good brightness, and it takes much less power. 12v is blindingly bright. However, 8v works too. Much, much less power. No bosting on the board. Parallel strings of 3x LEDs series LEDs with one current limiting resistor. No boost.
EngineeringShock (author)  acmefixer3 years ago
Hi there
i thought I mentioned it? Sorry if not. The booster board takes any voltage between 3.4Min-34vMax at the input, can can boost up to 34VDC.
wooo, jonny 5, 2nd movie is the best.
rimar20003 years ago
Awesome project, but too difficult for me...

Anyway, thanks for sharing it.
EngineeringShock (author)  rimar20003 years ago
I just made a very simple version of this flashlight, and an instructable if you're interested. No software, and no complicated circuitry =) Check out my channel if you're interested!
Oh, thanks. I will see that.
EngineeringShock (author)  rimar20003 years ago
Agreed. It is a bit complicated. I'll make an easier one next week, and cut down on the theory and such =) Thanks for having a look!