Introduction: How to Test Super Capacitors
Did you get a great buy on Ebay? Or was that "super deal" a "super dud". Have you duplicated one-to-many of the sketchy YouTube super capacitor videos? Or, maybe you have a capacitor bank that has been in service for a while and just want to know how it is doing. Using limited test equipment and some simple tests, I will help you take some of the mystery out of that super capacitor.
Ok, I will admit it. I traded the cow for some "magic beans" again. This instructable was created because I was being cheap. I had not played with super capacitors for a while, so when I stumbled on a great eBay deal I bit, and they reeled me in, hook, line, and sinker. Sure the price was less than half of what the big guys were selling for, and included free shipping from China. I was sure they were just passing on a really great deal that they had scored. What the heck I was thinking. I should know by now, there really is no such thing as a free lunch.
Do you ever get that sinking feeling? I sure did as I started to test my super duper deal capacitors. The first one I grabbed would not even draw any current when I hooked it up to the bench power supply. Hmmm, ok, maybe it's a dud, it happens. The next one I grabbled seemed to charge but when I removed it from the supply it immediately dropped 1.2 Volts. This is not looking good. And then I wondered, how the heck do I test these. Commercial super capacitor testers were certainly out of the question, with price tags in the "if you have to ask, you cannot afford it” range. Well let's see what we can do with some common test equipment.
We are going to be examining three parameters of super capacitors. Capacitance Value, Equivalent Series Resistance(ESR) and Leakage Current / Self Discharge. I will show you several ways to test each parameter and let you know what worked well for me.
Step 1: Equivalent Series Resistance
Why Is ESR Important?
When super capacitors first started to become available ESR ratings were in the low Ohm range. This does not seem like a big deal until you start to figure out voltage drops in higher current circuits. Say that you needed to supply a circuit for 6 seconds with 12 V at 10A and your circuit will work down to a voltage of 6 V. You do your calculations and determine that a series string of 6, 50 Farad capacitors will provide the power you need. That is until you calculate the voltage drop from the series resistance. Lets say that these early generation capacitors have an ESR of 0.5 Ohms, 6 in series yields a total ESR of 3 Ohms. E = IR, so 10 x 3 = 30 Volts of drop. Not going to work. You would have to parallel a whole bunch of smaller value caps to get the ESR down to a useable level. Now if we wanted to run a 100 mA load for a much longer time that would be doable. 0.1 x 3 = 0.3V, 12V-0.3 V = 11.7 V. Low ESR becomes extremely important for high current applications like the audio peak power boost caps that have become so common, or for starter battery replacement packs and energy harvesting from regenerative braking systems. Not so important for memory back up and low power applications. Luckily for us, modern super capacitors have ESR's in the low mOhm range which makes high current projects very possible.
Constant Current, Voltage Measurement
Take a look at the simplified super capacitor schematic above. All capacitors simplify down to the three elements shown. ESR is simply the series resistive element of our capacitor and as such, lends itself well to analysis using Ohm’s law. For this test we need a constant current source. Either a current limited bench supply or a constant current load will work. What we are going to do is apply a 1 Ampere current source or sink and measure the instantaneous Voltage change. Think of the capacitor part of our circuit as being a perfect current source/sink. The trick here is to measure in the middle of the voltage range (0.0 V - 2.7 V for a single cap) so you are sure to have a true 1 A current flow. The classic instrument for this measurement is the Digital Storage Oscilloscope (DSO) Using the measurement cursors you can resolve down to around 10 mV on the typical hobby scope. Don't fret if you do not have a scope, you can still get a relative idea of your capacitors ESR using a good quality Digital Multimeter (DMM). And if you are lucky enough to have a Mooshimeter it will work almost as well as a DSO.
I will assume you have a variable current limiting bench supply. If you don’t, I highly recommend it be next on your purchase list. With decent 0-30V 0-5A supplies being available for 50 to 60 USD there really is no reason not to have one. That being said, if you have a power supply that does not have adjustable current limiting, you can build the super easy constant current load that I have included for less than 10 USD. You could also use one of the low cost DC to DC boost/buck power supply modules available on eBay for around 5 to 10 USD.
Prepare your capacitor by discharging it through the appropriate current limiting resistor and then a direct short for about 15 minutes. This should get you close to 0.00 V if not you probably have a high ESR issue. It is a good idea to check the calibration of your power supplies current meter. On the cheaper units these tend to be a bit off. Short the leads of your supply and set your current limit to 1.0 A. Set your DMM to measure 2 A or more and verify the power supply calibration. Set your power supply voltage less than or equal to the capacitor rating and while you are at it measure that the displayed voltage is the same as what your meter reads(don’t forget to move your leads, we all have done it)!
Connect your DMM or scope leads to the capacitor under test. Connect your power supply leads to the capacitor under test, but do not connect to the supply yet. Set up the measurement ranges on your monitoring devices. With a 1A test current the ESR in Ohms will be equal to the instantaneous voltage differential that occurs upon applying or removing the 1 Amp test current. ( I.E. 50 mVolt = 50 mOhm) On a new cap you should see under 10 mV change. ( you probably will not be able to measure less than 10 or 20 mV ) On an older or bad cap, you may see as high as 1 V. Measure the voltage on the cap and record. Connect the power supply and immediately hit the hold button on your DMM. The stored value minus the capacitor start voltage will be your ESR. Using a DMM for this is more of an art than a science. Depending on your meter, the best resolution you might achieve is pass/fail. Hopefully, you have a known good cap to measure that will give you a baseline measurement to compare to. If you're running tests on a DSO or Mooshimeter, depending on your settings, you can stop now or let it charge to its completion point. Magnify and measure the start up voltage spike before the capacitor begins to charge. This voltage change value will be your ESR.
If you want to measure the ESR using a constant current load, set the load to 1A. Charge your capacitor to its rated voltage and remove from the supply. Measure the voltage on the capacitor. Apply the load and immediately hit the hold button on your DMM. The voltage drop will equal the ESR (Volts = Ohms at 1A charge/discharge current)
Capacitor ESR Meter
Yes, you can use a capacitor ESR meter to measure your super caps ESR. One thing that you must do before measuring is to discharge the capacitor to as close to zero volts as you can get it. Any appreciable DC voltage stored in your capacitor will damage the meter. These meters use DC clamping circuits to make sure that the device under test is discharged. They are not designed to handle the large amounts of energy stored in super caps. Another factor to consider is that the meter probably uses a 100 kHz sine wave instead of the 1 kHz most capacitor manufacturers specify for testing. This can result in slightly different readings than a 1 kHz ESR bridge and both will be different than the DC value.
Battery ESR Meter
This is hands down the easiest and most accurate way to measure your super capacitors. The 50 USD unit that I have is able to measure down to 0.001 ohm and can be used on a fully charged capacitor. The cheap unit I have never zeros completely so my measurements have a plus 2 mOhm error. Even so, the ease and consistent measurements obtained makes this a great investment.
Step 2: Capacitance Value
This is the Big Kahuna of our measurement trifecta. It’s all about the power baby! Well, that and the money, but that’s a different instructable. I don't know about you, but if I order a beer, and the glass is only a little more than 3/4 full, we have a problem. But that is exactly what some capacitor manufactures can do to you. Before you plunk down your hard earned cash on a super cap, go to the manufactures web site and check the data sheet. Find out what the allowable deviation for capacitance value is and use the lowest value for your design calculations. It might be wiser to spend a couple extra bucks to buy a cap rated -5%,+10% than one rated +-20%. A +-20% 500 Farad cap could actually be a 400 Farad cap even before it starts to degrade from age and use.
So we know the range of acceptable values for our capacitor. In my case I have six 100 Farad capacitors rated at +-20%. This gives me an acceptable value range of 80 to 120 Farad. But how do we measure it. Your capacitor meter will burst out in laughter if you try to measure anything anywhere near this large. I guess we are stuck with an indirect capacitance value based on the energy we can push into or extract from our capacitor.
Ok, I hate to do this, but I need to break out a little math. You know I am a hit-it-with-a-rock, poke-it-with-a-stick kind of guy so this hurts me more than you.
Average current in a capacitor i = C ( dV / dt ) with C = Farads, V = Volts, t = seconds
Solving for C = i ( dt / dV )
Since we are going to use 1A of current and measure the time the voltage changes 1 Volt, this makes our final equation
C = dt
So in other words, to determine the capacitance value of my capacitor I simply have to measure the time interval in seconds while charging or discharging at 1A between 1 Volt and 2 Volts. Or 1.5 Volts and 2.5 Volts, any 1 volt differential should yield the same result.(As long as my charge current is constant) So my 100 Farad +-20% capacitor should take between 80 to 120 seconds to go from 1 Volt to 2 Volts. Here again a DSO is the preferred test instrument but a plain old DMM and a stop watch will work almost as well. Piece of cake!
So what if you are really lazy like me, or just don't believe the math and want to see the total energy stored in the capacitor. Then you my friend need a Re:load Pro from Arachnid Labs. This open source, constant current load only runs 125 USD and has a tremendous number of features. One of them is the ability to calculate total power dissipated.
We know that the potential energy stored in a capacitor is
Wh = 0.5 ( C x V^2 ) / 3600
So for our 100 Farad cap,
Wh = 0.5 ( 100 x 2.7^2 ) / 3600
Wh = 0.101
To determine the capacitance value from the measured dissipated power simply rearrange to solve for C
C = 2 ( 3600 x Wh / 2.7^2 )
Remember that your capacitor must be charged to 2.7 V and discharged to 0.0 V for an accurate measurement. What could be easier? A handy thing to remember is that for a 2.7 V capacitor 1 Farad is roughly equal to 0.001 Wh or 1 mWh of energy storage.
Step 3: Leakage Current
I thought that this would be the easiest topic to cover. Leakage currents are specified on all manufacturer data sheets. The only trick is the value is usually specified after 72 hours of charging. But like most things in life, nothing is ever as easy as it seems. I have had several capacitors that met their 72 hour leakage specification but when removed from charging began to discharge at a rate far above what the measured leakage current would justify. Now it turns out that our simplified capacitor schematic is not really up to handling leakage current in the real world. It gets very complicated with the actual model looking like multiple capacitors in parallel with voltage variable resistors in series and parallel. It is called a transmission line model and we are not going there. ( Nope, no way, no how, better chance of teaching monkeys to fly the Space Shuttle than me ever figuring it out.) Dielectric absorption and charge redistribution are two factors that can affect self discharge and can have a significant effect on Voltage if a short charging cycle is used, but as far as I can tell, should be minimal after about 4 hours of charging. So why were some of my caps acting like they did? Not a clue.
Self discharge may not be an issue depending on your application. If your circuit is a constantly powered UPS or only needs to supply boost current, probably not an issue. If you are building a solar charged battery bank that has to hold a charge over several cloudy days, big issue. Let’s try a more real world approach to dealing with self discharge or leakage current. These are my “ballpark” tests, your milage may differ.
If your application is self discharge tolerant and your capacitor passes ESR and value tests; charge at rated voltage for 8 hours. leakage current should be 2 mA or less.(Up to around 600 Farads or so, higher values may have more leakage)
If your application will have a short duration charge cycle and/or you may have long periods of no charge. Determine what is the shortest likely charge duration. Charge your capacitor for this period of time. Disconnect from the supply and measure capacitor voltage. Start a 1 hour timer and measure the capacitor voltage at this point.
i leakage = C ( dV / dt )
This will give you a closer approximation to real world leakage than a 72 hour test.
So lets run through this. Let's say we have 100 mV change in 1 hour with a 100 Farad cap.
i = 100 x ( 0.1 / 3600)
i = 0.003 A (Not great)
For equivalent shunt resistance
R = 2.65 / 0.003
R = 883 Ohms
Lets say you are running a load that is manually turned on and off so it does not consume any power until switched on. Its drop dead, no work voltage is 1 V.
t = -C x R x ln (V1 / V2) where V1 is the discharge Voltage and V2 is the charged Voltage
t = -100 x 883 x ln (1 / 2.7)
t = 87,704 seconds or about 24 hours.
The good news is most of the super capacitors being manufactured today have leakage ratings in the micro Amp range.
Step 4: Build a Constant Current Load
It's a quick and fun project that can be built on a scrap of perf board. Make sure you use a heat sink. The heat sink I show in the picture is only good for around 5 Watts continuous, but it will handle 10 to 15 Watts for short durations. There are a million designs out there, but I like this one. It uses an LM10C op-amp which contains its own 200 mV reference. This allows it to operate with varying supply voltages and still maintain excellent current regulation. The IRF510 mosfet is rated for 43 watts of power dissipation but even with a very good heat sink I would not push it past about 25 Watts. If you need more power an IRFP150 could be substituted.
Step 5: Everything Else
Take a look at my capacitor test results. Of the seven units obtained on eBay (Drum roll please) not a single one met the manufacturers specifications. They were so bad (How bad were they?) that I have to write cheesy jokes. I cannot imagine that these were actually new units as the vendor advertised. Counterfeit, manufacturer rejects rescued from the scrap heap, or maybe OEM rejects? Who knows. I do know that it wasted far too much of my time, and from now on, this is a product that I will be buying from known vendors. The Eaton brand capacitors that I purchased from DigiKey passed all tests with flying colors. I have purchased 8 additional ones that test like they are clones of the first two. The uniformity is very reassuring.
Samwha or Shamwha?
Not a lot of love on the net for the Korean Company that may or may not be part of Samsung. Ever since the so called capacitor plague of the early 2000s where Taiwanese low ESR electrolytic capacitors were failing at an alarming rate. Samwha has received quite a bit of negative press. Probably much of it not deserved. I have had good results with Samwha products and you can be sure that with global market competition, bad product manufacturers do not last. I would love to compare some Samwha super caps from a real supplier. My guess is that it is a fine product and what I got stuck with was no more than a shady deal.
eBay And The One Time Access Code
eBay is a great company. They provide a service that makes it possible to buy just about anything and their customer service is first rate. They even got my money refunded from crapacitor gate. But I am starting to worry. They now issue you a one time code for access to a human. How many times do I get before the soul less electronic gate keeper switches my access to an endless Mobius strip of voice prompts and canned recordings without end. I hope I get an email when I get close.
Cyber-stalked by a Chinese merchant?
“hello big customer i am so sad not make you very happy
pls change the feedback
have a nice day"
And four more like it. It is very simple. Stop selling defective products and forcing your customers to open cases with eBay to get a refund.