What's all this supercapacitor stuff anyhow?

[CWatters]

I calculated that your source should charge a 3.3F cap from 0V to 3V in about 160 seconds ignoring losses. In the charged state it would hold about 14 Joules of energy.

If your load works from 3V down to 0.9 V the capacitor could deliver 60mA for about 115 seconds.

Useful theory...
Q=CV
dQ/dt=CdV/dt
dQ/DT=I
So...
I=CdV/dt

You can rearrange that last equation to calculate a lot of things such as the charge or discharge time given the current, capacitance and allowed change in voltage.

 

[CWatters]

That's the rub. I can generate up to about six volts at 1 to 4 mA. Takes a long time to charge a supercap like that. I can charge a very large STD cap much quicker but it drains out of juice in seconds at 60 mA and that won't work for this remote sensor application.

Ah I used your earlier figures or perhaps missunderstand post #14.

 

[dnyberg2]

Then somethings wrong somewhere. With a load of 60mA I get about 4 minutes of power down to about 1.5 volts. With a charge current of 10mA, it takes 24 minutes to charge the cap back to 3 volts...

 

[dnyberg2]

I find it interesting that the low current charger I have can be six volts @ no load but the that supercap pulls that down until it gets up near full voltage. Then it seems that the "load" of the cap itself can hold the low current CHG CKT down around 3 volts. I was surprised to see that behavior but I've never really worked so closely with supercaps before.

 

[dnyberg2]

I'm going to to have to do a side by side experiment. CW DC at as much CHG current as I can get versus storing small amounts of it faster in a CKT that bursts a higher current pulse into the supercap...

 

[CWatters]

Then somethings wrong somewhere. With a load of 60mA I get about 4 minutes of power down to about 1.5 volts. With a charge current of 10mA, it takes 24 minutes to charge the cap back to 3 volts...

Could be the capacitance of the actual capacitor is much greater than specified. Some have a wide tolerance of -0% to +80%.

 

[dnyberg2]

Oh yea, I see that now. Crazy wide tolerance. I wonder if that moves around very much over the life of the part?

 

[CWatters]

Then somethings wrong somewhere. With a load of 60mA I get about 4 minutes of power down to about 1.5 volts. With a charge current of 10mA, it takes 24 minutes to charge the cap back to 3 volts...

Both charge and discharge figures suggest the actual capacitance is 9.6F.

 

[dnyberg2]

Thats a MUCH bigger # than the 3.3F on the wrapper!

 

[dnyberg2]

I wonder why the tolerance band is so crappy for caps unless you spend $100 a piece for them?

 

[dnyberg2]

That brings up another important question. Is there such a thing as an accurate capacitance meter that can measure tens of farads? If I were to market this thing, I'd need a way to quality control the actual capacitor value.

 

[CWatters]

I find it interesting that the low current charger I have can be six volts @ no load but the that supercap pulls that down until it gets up near full voltage. Then it seems that the "load" of the cap itself can hold the low current CHG CKT down around 3 volts. I was surprised to see that behavior but I've never really worked so closely with supercaps before.

C=eA/d where d is the plate separation. So for max C you want min d but small d reduces the breakdown voltage. I think it's more complicated with supercaps but my guess is that over their rated voltage they are either breaking down or the leakage is going way up dragging down the source voltage.