Replace your car battery with capacitors!

[redpoint5]

Rated voltage is why I purchase Maxwell. They are rated at 2.7v, so a 6 series string has no risk of overvolting.

The BCAP3400 is rated to 2.85v, so I put together a 5 series string. That's cutting it close, but the literature allows a max of 3v.

LEDs are fine for balancing because caps that are already closely matched will not go that out of balance to begin with. The point isn't to bleed off massive overvoltage from an alternator, but to bleed off slight imbalances between the caps.

[cts_casemod]

Originally Posted by freddyzdead

There will be a 25% reduction in capacitance, which is already suffering from series string disease. I don't know that it really hurts them if they occasionally nudge a little over 2.5 volts. The weak link is the dielectric layer, which in this case is very thin. Excessive voltage will cause a breakdown, which may or may not be self-healing. One day if/when I have a spare one that I don't need, I will run some tests. It certainly would be useful to know what is the failure mode of these things.

Connecting a series string of zeners directly across the capacitor/battery is a bad idea. A transient spike, perhaps from the starter motor, could zap one or more of them, resulting in disaster. The proper way to do it is to have a transistor across each capacitor, turned on when the zener starts conducting. Probably cheaper and safer to just buy a balancing board. The LED idea is appealing because of its simplicity, but it doesn't really do the job, and can't dissipate enough power to drain away the excess charge quickly enough.

It is assumed the pass element (zenner/balancing mosfet, etc) already has a series limiting resistor, same as a LED (which too, is a form of zener) or any mosfet in a battery pack balancing system. Balancing is limited to a few mA and not meant to limit fast rising transients. The latter are dealt by the very low ESR of the cap.

I don't know of any LED that conducts at 2.5V. IRdA is about 900mV, standard types at 1.75V and white/blue/HB GREEN at 2.8V. I don't see this as an optimum solution.

Of course this is a free forum, this is just my opinion - Why not use the ubiquitous TL431 with a 100Ohm series resistor for the same effort/cost, and much better regulation?

Transient spikes are handled by the cap low ESR. This supersedes the much higher impedance of the battery in lieu of which these systems are designed, and expected, to operate.

Originally Posted by redpoint5

LEDs are fine for balancing because caps that are already closely matched will not go that out of balance to begin with. The point isn't to bleed off massive overvoltage from an alternator, but to bleed off slight imbalances between the caps.

I agree 100%. And a LED for such terminal voltages makes sense too, perhaps that's how this method first came to be?

[freddyzdead]

Ok, I don't have the benefit of actually having tried any of this, as my capacitors still haven't arrived. But the reason red LEDs aren't much good is a typical red one will start conducting at about 1.7 volts and less than 1 milliamp, which is negligible for this purpose. At 2.1 volts it's around 10-15 mA. You could pad it with a Schottky diode, but the forward drop is load-dependent and it would be hard to know what you were getting. A white/blue LED is on about 50 microamps at 2.7 volts, so that's no good. Whether or not the red one does the job, it will deplete the charge over time, or not work at all. I can't know any more until the capacitors arrive.

The 7.5Ah LiFePo4 battery has arrived and I have installed it in the car.

I had to pack it with a big chunk of wood and some foam pieces to keep it in place. It weighs almost nothing, makes you wonder how it can work.

I left the original battery terminals on, just in case I want to revert. I'm not planning on it.

The first thing I noticed is the starter cranks over noticeably faster than it did with the old battery. Since I only bought the car 8 months ago, I have no idea what is the condition of the battery. A good test is to put the roof up and down without the engine running. The hydraulics puts a pretty heavy drain on the battery, and it had begun intermittently failing partway through the cycle.

With the new battery, the roof zips up and down no trouble at all. So, I'm impressed. This is just a motorcycle battery, after all, not meant to be putting in cars. Looks like I don't really need the capacitors, but I will put them in anyway. It has to make life easier for the battery, I think. I will update when they arrive.

[redpoint5]

Originally Posted by cts_casemod

Why not use the ubiquitous TL431 with a 100Ohm series resistor for the same effort/cost, and much better regulation?

I wasn't familiar with this low cost solution as I rarely tinker with electronics. Seems a superior solution to me. No satisfying light though! I've had LEDs stay lit for months on a fully charged cap bank.

What package TL431 would you recommend? If I wanted the activation voltage at 2.9v, what would I need to do?

Originally Posted by freddyzdead

...the reason red LEDs aren't much good is a typical red one will start conducting at about 1.7 volts and less than 1 milliamp, which is negligible for this purpose...

Looks like I don't really need the capacitors, but I will put them in anyway. It has to make life easier for the battery, I think. I will update when they arrive.

I used LEDs because I already had a bunch on hand, along with the diodes necessary to solder in series to get the activation voltage close to where it needs to be. This is a proven effective solution, and common due to how cheap and easy it is to implement.

You probably don't need the caps, but what is the burst (short duration) C rating for your battery? I've thought about wiring in a low ohm power resistor to limit battery output. It would put the demands of starting on the caps, but allow the battery to keep them topped up when the vehicle is off.

Perhaps not an issue in Australia, but LiFePO4 is not rated to accept a charge below 0 C. The LiFePO4 in my Prius is inside the cabin where it is significantly protected from freezing.

[freddyzdead]

TL431 would work better than a LED, but maximum current is 100mA, it would take a long time for it to drain anything significant off a 500F capacitor. As I said before, I don't know how much over-voltage it takes to permanently damage one of these, or what is the failure mode if it happens. Such a big capacitor is outside most peoples' experience, anyway. Most are astonished that capacitors that size are available, when I tell them what I'm about to do. When I went to electronics school many years ago, we were told that a one-farad capacitor would be as big as a house.

The battery is rated at 380CCA. And PCA (I'm not sure what that is) is 620A. It doesn't get that cold where I am, so I don't think there's any temperature issues. Weather here is pretty much like Southern California. That's why I live here, largely.

If the cycle life of this kind of battery is typically 3000 or so, then it seems to me it should last at least 10 years; with capacitor bank to help it, maybe a lot longer than that, if you don't try to run things much without the engine running.

[cts_casemod]

Balancing is totally optional and meant to deal with small currents to top up the individual elements equally, not to deal with prolonged transients.

Good quality elements keep themselves fairly balanced whenever batteries or capacitors.

[freddyzdead]

How do you know this? The tolerance on these capacitors is +/- 20%. While it is true that components from the same manufacturing batch should be fairly well matched, there is no guarantee of this. So if you had one that erred on the high side and one on the low, you could have a serious overvoltage on one of them. The ebay seller advises to charge them to 16 volts, discharge to 2 volts, repeat 6 times. I can only suppose this is to get them balanced. He doesn't mention that this applies to 6 capacitors in series, which might confuse some people. Anyway, he warns against just connecting them straight across the battery.

I am still waiting for the damn things to arrive so I can assess their behaviour for myself. But I think you'd be mad to not provide some kind of protection against overvoltage.

[redpoint5]

I tend to agree with cts. If you have a large enough voltage headroom, balancing is likely unnecessary. The caps have a sort of self balancing in that they leak relatively large amounts of current as they approach rated voltage, and almost nothing just a few tenths of a volt below.

With my 5-series BCAP3400 2.85v rated caps, there is no headroom, so balancing is more important. With my 6-series BCAP350 rated at 2.75v, I've got lots of headroom. Even so, I still put in balance LEDs because I didn't know better at the time.

[teoman]

They have a 3v 3000f version now if i do not remember wrong.

[teoman]

Enki, why do you need to have the pump and fans running?