Inverter input ripple capacitor(s)

mike95490 · November 2011

Ok, starting this here, because it's been mentioned in a couple other threads, and I am keenly interested, because my new Nife bank, is likely higher internal resistance than my old Pb-H2So4 bank.

Some were saying Farads needed, some "supercaps" and I'm attaching some links from some tech bulitens I found over the weekend, while fighting a cold!

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

Ultracapacitors with Dr. Andy Burke (you tube)

Anyway, I hope this will start some discussion on the subject

* fusing the capacitor bank
* Precharge & bleeder resistors

john p · November 2011

Re: Inverter input ripple capacitor(s)

the batteries you are going to be using . How are they configured? one string? voltage ?....multiple strings parallelled ??voltage ?

Super caps dont have as long a life as people think if opperated near their voltage and capacity.. Have used banks of them for experiments at work.

mike95490 · November 2011

Re: Inverter input ripple capacitor(s)

john p wrote: »

the batteries you are going to be using . How are they configured? one string? voltage ?....multiple strings parallelled ??voltage ?

Super caps dont have as long a life as people think if opperated near their voltage and capacity.. Have used banks of them for experiments at work.

Battery = single string, NiFe, 48V 825ah (photo album in my .sig in first post)
I'm thinking super caps won't do 120hz very well, and am now starting to research the links I've dug up.

john p · November 2011

Re: Inverter input ripple capacitor(s)

It was hard for me to tell how they were interconnected in the photo.
I have been thinking about what you want to do and had a short talk to 2 engineers here at work.. They say your batteries may have more problem charging the super caps than feeding enough current to the inverter? They have a very low impedance ,that makes them so easy to charge vey quickly with a suitable charger.
We have a 500,000 farad bank and 12v . but that banks charging is only limited by the 400a charger we have. They are amazing in some ways .. but useless in so many ways. and at the moment grossly overpriced.
I am having 1 engineer look at your problem . so will get back to you if he comes up with good? idea. but he is an engineer so dont expect result today..

RCinFLA · November 2011

Re: Inverter input ripple capacitor(s)

The problem with super caps are their limited voltage rating. The guy in the You Tube video claiming he tested the cap to 400 volts, I think he meant to say he tested the cap to 400 amps of discharge.

Voltage rating are between 2.3 v to 2.7 vdc per cap so you have to put a lot of them in series to accomplish a worse case of 65 vdc for a 48 v inverter system. (I am not sure what max voltage you are looking at for NiFe battery array).

Putting 24 in series to accomplish the needed voltage will also reduce the capacitance by 24 so you take 5 F @ 2.7vdc cap to 208,000 uF @ 65v max wvdc. You might as well use computer grade electrolytics and 0.2 Farad won't do much for ripple reduction so you need about 20 series strings of caps to get up to 4 Farads. (up to 24 x 20 = 480 caps)

For super caps, a significant issue is balance and making sure that no cap sees more then 2.7 vdc during its charge and discharge cycle. Like the issue with LiFePO4 battery charging, you are going to need a charge (and discharge) balancing system to avoid overvoltaging any of the caps. The caps may explode if overvoltaged.

If the caps are carefully matched you can get by with just bleeder resistors. Then you have to worry about caps diverging in matching as they age. You also have the continous load loss of the bleeder resistors.

Forget Supercaps for 48 vdc battery system. Find youself a deal on a boat load of computer grade 75 wvdc electrolytics.

(Picture is largest cap I happened to have and it is only 0.12 Farad at 40 wvdc. Size about same as quart of liquid)

waynefromnscanada · November 2011

Re: Inverter input ripple capacitor(s)

Great info "RCinFLA". Thanks for sharing. I for one had no idea they were limited to such low voltages, or that their lives were so relatively short, especially considering the way I've heard them promoted for years as being the up and coming answer to the world's renewable energy problems. Quite the let down, but at least now I know.

RCinFLA · November 2011

Re: Inverter input ripple capacitor(s)

I think you have more then a short term Rs issue with NiFe batteries. A moderate current draw for several minutes will cause the voltage to fizzle. They need time to recover (redistribute) their charge.

You could put a small AH lead-acid or LiFePO4 in parallel to take the short term high discharge current. This is probably a better solution they any parallel cap approach.

Back in one of my prior lifes we put zinc-air battery in a wrist watch pager. The zinc-air batteries have a small hole in the back to draw in air for the reaction. Problem is they can't stand high current. It could run the pager for a long time but the battery could not even power the alert transducer. We put a small button rechargable LiIon battery charged from the zinc-air to get the needed peak current.

john p · November 2011

Re: Inverter input ripple capacitor(s)

While its true that single consumer cells are about 2.7v for ultra caps. You can buy high voltage modules that can be rated at anything from 12v to 24v to 48v to 75v to about 200v if I remember correctly. The bank of caps we have at work are 12v modules connected in parallel. they will happilly take 16 volts.
We use no ballancing between the modules and no problems.

Never have found a good use for them as storage yet. as you can get much the same result just paralleling lead acid batteries.
But they make a great toy to destroy things with.. 500,000 farads at 12v dumped in less than 3 seconds is actually quite spectacular.:cool:
But you cant do that at home unless you have a huge budget to buy them.:D

I see the greatest problem in using them is if they get close to fully discharged it takes a lot of current availability to get them charged again as they appear to the charger as close to a dead short. So if between the batteries and the inverter they may create more problems than they solve.

Fullpower · November 2011

Re: Inverter input ripple capacitor(s)

Can you give more detail on the nickle iron batteries?
What nominal voltage are you operating, and what inverter are you using?
Is the battery bank in operation at this time?
What limitations on discharge current are you seeing?
Are you getting voltage sag under load, and having inverter shut down?
I have been wondering what the discharge curve looks like with the NiFe cells, and what sort of working load versus Amp-Hour rating ratio is usable.
Are the batteries rated at a 20 hour discharge rate, or some longer interval?
Thank you for the information.

mike95490 · April 2012

Re: Inverter input ripple capacitor(s)

Fullpower wrote: »

Can you give more detail on the nickle iron batteries?
What nominal voltage are you operating, and what inverter are you using?
Is the battery bank in operation at this time?
What limitations on discharge current are you seeing?
Are you getting voltage sag under load, and having inverter shut down?
I have been wondering what the discharge curve looks like with the NiFe cells, and what sort of working load versus Amp-Hour rating ratio is usable.
Are the batteries rated at a 20 hour discharge rate, or some longer interval?
Thank you for the information.

Sorry, I ddn't spot this till now !

42, cells in series, 800ah cells. gives about 52V nominal, but needs 71V for quick charge. Supposed to be safe to discharge down to 43V
I'm using an XW6048 inverter, it's good up to 70V DC input - i've limited the battery charge to 68V
It's been running since novenber 2011, and has not had any issues, but it's not been stressed. It passes the 5 days of clouds, while running a fridge.

westbranch · April 2012

Re: Inverter input ripple capacitor(s)

Mike, what made you go 42 cell/52V config?

mike95490 · April 2012

Re: Inverter input ripple capacitor(s)

westbranch wrote: »

Mike, what made you go 42 cell/52V config?

I'd wanted a couple spare cells, just in case, since replacements take 3 months to ship, so I can always drop a dud out of the system. Apparently, just keeping a "spare" on hot standby, was not as good of an idea, as I'd thought months earlier, so I decided to use them all.

Inverter input ripple capacitor(s)

Comments

Categories