On 8/11/2014 12:44 AM, Robert Baer wrote:
rickman wrote:
On 8/9/2014 6:09 PM, Robert Baer wrote:
Jon Elson wrote:
EngineeringGuy wrote:

When connecting "super capacitors" in series to increase the working
voltage above the level of the individual capacitors, does anyone
have an
active voltage clamp that would not discharge the capacitors when the
power supply is disconnected? Most "super" capacitors have a working
voltage of 2.5 or 2.7 VDC... I would like to connect 10 capacitors in
series for a 25
VDC capacitor stack. Most suggestions call for connecting equalizing
resistors in parallel with each capacitor, but I would rather have a
circuit that would not discharge the capacitors when the power is
off...
anyone have any ideas?
Well, a couple diodes across each cap would not leak a whole lot
when forward biased below the forward conduction voltage.
With plain Si diodes, that might be 5 diodes in series. they would
start leaking somewhere below 3 V, and conduct pretty strongly
at 3.5 V.

You could also look a Schottky diodes and see what values of Vf you
see. If they start to conduct at .4 V, then 7 in series would
give you 2.8 V turn-on. You can get surface mount SOT23 dual diodes
wired in series, so that would only take 4 parts.

Jon
Sounds silly to me; why use the diodes in the forward bias mode,when
the reverse bias mode seems far better?
That way, the capacitors are always reverse polarity protected.

Because the reverse breakdown voltage is not well specified. It is
guaranteed to be above some value, but where exactly will that be? How
many diodes do you know of that will protect a 2.7 volt cap when the
diode is reverse biased?

LEDs might be a good choice for forward biased protection. Some of them
work at 2.7 volts or so. Add a single Si diode and you will get a very
small amount of current flow in the cap working voltage range with the
knee in the I/V curve somewhere above 2.7 volts depending on the color
used.

Err..any (reasonable) silicon diode reverse polarity connected WRT
the cap will conduct in the region of 620mV; one certainly does not need
to have more in series. And the diode reversed biased (capacitor forward
biased) will will have essentially zero current.
Now,if 620mV is a bit much for these soup-er-boop-a-doop capacitors,
then instead of sand power, go for flower power (Ge) at about 320mV.
Unfortunately, they are not exactly the bees' knees and are a bit
leaky of the electronic honey.

I don't follow your reasoning at all. The OP wants to work at increased
voltages by using multiple super caps in series. Why would you think
620 mV would be a high enough voltage?

The way resistors work to equalize the voltage on the caps is to provide
a small leakage current around the cap, higher when the voltage is
higher reducing the charge on that cap relative to the others. The
difference with the diodes is that they conduct much less current when
the cap is in its normal working range. Only when it exceeds the
working range by some margin does the diode start to conduct
significantly bypassing a significant portion of the charging current.

There has to be some headroom between the working voltage and the damage
voltage since the diodes won't have a highly stable knee voltage. But
given enough margin this should work well. If there is not enough
margin then something with a better regulated knee would need to be used
like a voltage regulator.

--

Rick