Ultra Capacitors

[DanG]

All I'm saying is you've gone a little overboard in sizing the caps. That thing at 28.8V will throw 1700 watts power off for a minute.

http://www.electronics2000.co.uk/calc/capacitor-charge-calculator.php
http://www.mhi-inc.com/Converter/watt_calculator.htm

No doubt that is a valid application - if/when several devices are enabled simultaneously down stream the 100ms~ inverter lag/sag and inrush is wear and tear on the entire system...  and the 'boost' may help stretch the battery charge life/levels via Peukert's law, and buffer deep draw shock from accelerating battery plate aging.... It may be by a minute degree but over 5 or 7 years service I'm pretty sure it will pay off, as long as you overlook the liabilities of damaging the capacitors from a thousand different oversights, oops or acts of God. 

I have VRLA telecom batteries that have a 0.18V maximum DC float ripple specification to help get the 10-year warrantied service life. I'm counting on my supercap to help absorb the wild AC ripple from wind turbines etc. AND provide boost for heavy loading (AC motor - compressor starting etc.)

don't/won't dump their charge all at once like a true capacitor

BZZT. http://www.youtube.com/watch?v=EoWMF3VkI6U <-- the fireworks here after the 5 minute mark shows the discharge being limited by the connections and wires themselves. If its wired in heavy enough to be useful in a 3000w inverter circuit then its wired in heavy enough to really dump its potential. If you want car starting battery strength you'll need car starting battery connections. You have enough energy there it can/will be dangerous.

Just reminding you the quote you offered from the Maxwell site speaks of elegant tuned systems, not duct taping 1300 'D' Cells together

[mike90045]

One thing that's getting lost in all this, is that only a small portion of the stored energy in the cap is being used.  Say - + .5V to take out the big 120Hz surges.  The entire charge is not being dumped, just a bit off the top, hopefully enough to :
a) improve inverter performance
b) reduce wear on the battery bank, increase it's life, reduce it's heating.

So the cables to the cap bank, need to be less inductive than the battery cables, in order to be effective.

The capacitor needs to have a longer service life at 120Hz, than the batteries do. Ripple currents = heat, some ripple caps are meant to be heatsinked

The capacitor fuse has to a) be less resistive  b) blow fast enough to be useful


http://www.newark.com/pdfs/techarticles/cornell/selectinvcap.pdf
or http://www.cde.com/tech/selectinvcap.pdf   2 links, same article,
Cornell Dubilier - Abstract— Aluminum electrolytic capacitors for inverters

http://www.cde.com/catalog/inverter/   speaks for itself.

http://www.ampbooks.com/home/amplifier-calculators/LC-ripple-filter/
LC Ripple Filter calculator - to it's credit, it did not choke on the insane values I plugged in


http://www.murata.com/articles/ta06d2.pdf   Murata high power ceramic caps

http://www.maxwell.com/  Maxwell super caps
( "With a cycle life of 1,000,000 cycles"  = 8,333.3 seconds = < 1 day which has  86,400 seconds in it !)  But, we'd not talking "full cycles" but maybe 0.1 cycle