Replace your car battery with capacitors!

[Enki]

Originally Posted by redpoint5

Excellent post!

Thanks! I try. I actually messed up and originally posted this in the already linked battery thread. Lol.

Originally Posted by redpoint5

I've been charting the leakage of my 350F series of capacitors over the course of the past few days and will put it into an Excel sheet soon. Mine are balanced by the LED/diode method, so most of the initial "leakage" is just driving those lights. Do you find balancing to be unnecessary? I've wondered if higher charged cells would also have a higher leak rate, effectively balancing themselves. This would be a worthwhile experiment since eliminating the balance circuit simplifies things and lowers the parasitic drain.

My final product (in the project car) will be balanced. For 6 cells (ideally 7), balancing *may* not be required if all the cells are relatively close; the cells that charge faster probably have a higher failure voltage threshold as well, so I feel that for generic use like my jump pack, a non-balanced setup should be fine. I forget if the "fatter" charging cells have a higher leak rate (my earlier posted results may bear this out...?) but I'd wager that if balancing is not required after the initial charge (all cells at a safe voltage, starting from 0v) then it shouldn't be required at all (potentially, of course; this would require repeated, analytical testing which I'm too lazy to do).

Originally Posted by redpoint5

Your parked parasitic draw of 250mA is extreme. I'd only expect that draw on a cheap aftermarket alarm system install, or a failing electrical component. Have you measured the drain after letting the car sit for a longtime without opening the door or messing with the door locks? I ask because my TSX has a higher draw just after parking, and whenever the door has recently been opened. It eventually settles to 40mA.

The numbers were pretty sobering, for sure. What I did was close everything up, key off, then I put my multimeter in line with the positive and the battery to find the draw. Then, I set up the cap bank, withe the multimeter set up to measure voltage only (lest the meter catch fire), started the car up until voltage stabilized, then shut it down and watched the draw. That data looks like this:

Time (minutes) - Volts
0 - 13.77
5 - 13.62
15 - 13.23
30 - 13.05

As a side note, calculating usage via farad method, my Mazda draws about 550-570 amps when starting (cold or hot).

Originally Posted by redpoint5

The 14v-10v 560mA figure in your spreadsheet is very helpful because I found conflicting methods of how to calculate it (please PM me your spreadsheet or the formula). It looks like my truck, with a 20mA parasitic drain, can only sit for 28hrs before your 3000F caps would fail to start it. I believe I can get the drain down much further after disabling a thermometer with constant LED back-light. I'll try that tomorrow.

Not a problem. None of this is secret; it just took me a while to figure out how to do it, but I'm still fairly new to this as it is.

Bank_Amp_Hours = (Max_Volts - Min_Volts)*Farad_Rating/3600

Hours_of_Runtime = Amp_Hours/(Multiplier*Draw)

Originally Posted by redpoint5

Can you explain the V+W -> A and the V+A -> W table a little further? I know how they are calculated, but I'm just wondering what the figures have to do with replacing a car battery with capacitors. I'm also curious what you are planning with those resistor calculations.

It's just wattage/amperage/voltage translation. With any two of the three, the third can be calculated:

Watts / Volts = Amps
Amps * Volts = Watts
Watts / Amps = Volts

Edit: This math also helps figure out charging rates when using lightbulbs as charging ballast.

As for the resistor calculations, those are for charging ballast for the capacitor bank. More on this in the next sectioned quote/reply.

Originally Posted by redpoint5

I bought 2 amp loggers to measure the amps going into, and out of a capacitor or battery so that I can determine the charge/discharge efficiency between capacitors, LiFePO4, and lead-acid. When I start getting more than 1 day off per week, I'll start experimenting and share my methodology and results.

I can tell you right now that the efficiency/speed of charging for those three energy storage methods will be in this order:

1. Capacitors
2. LiFePO4 (a very distant second)
3. Lead Acid (a fairly distant third)


Originally, I hooked up my bank to a "smart" battery charger. It was initially supplying 6 volts to the bank, unknown amps. Thing sounded like it was going to burst into flames at any moment, so I hooked a car headlight up to it as a charging ballast to limit current flow (dead capacitors act like a short, after all; that should tell you what you need to know about charging rates). This is where the resistor equation comes in, though I've since decided on using small, switched halogen lights in the 12 volt range (20-50 watt) for charging the bank, as they are a great visual aide for charging/discharging the bank as well as potentially providing a working light.


As for discharging:
Here's a video going over the discharge rate of an older, 2600 farad 2.5 volt capacitor and bank:

http://www.youtube.com/watch?v=EoWMF3VkI6U

Some warnings:

Do *NOT* charge a dead capacitor bank off your alternator unless the farad rating is *LOW* or your vehicle is set up for SPL competition (car audio loudness) otherwise you could hurt your electrical system/melt wires.

[redpoint5]

Originally Posted by Enki

I forget if the "fatter" charging cells have a higher leak rate (my earlier posted results may bear this out...?) but I'd wager that if balancing is not required after the initial charge (all cells at a safe voltage, starting from 0v) then it shouldn't be required at all (potentially, of course; this would require repeated, analytical testing which I'm too lazy to do).

I'll add this one to my to do list, although it'll have to wait until I buy another set of capacitors since all of mine are already soldered up with balance circuits.

The numbers were pretty sobering, for sure. What I did was close everything up, key off, then I put my multimeter in line with the positive and the battery to find the draw. Then, I set up the cap bank, withe the multimeter set up to measure voltage only (lest the meter catch fire), started the car up until voltage stabilized, then shut it down and watched the draw.

Try getting a reading hours after not touching the car, or the next day, taking care not to disconnect power to the car when hooking up the meter (tricky, but doable). Cycling power to the car computer will put it back in high draw mode.

You can also start pulling fuses while keeping your ammeter connected to see what circuit is the one responsible for the 250aH draw. That will narrow it down, then you have to locate the individual part that is drawing all that power. One guy on Youtube found some circuit near his seat belt was responsible for high draw.

Supercap battery replacement aside, it would be important to reduce parasitic load even with a lead-acid battery to prolong the life. My lead-acid motorcycle battery is still going strong 10 years later just because I've been very good at keeping it topped off.

I can tell you right now that the efficiency/speed of charging for those three energy storage methods will be in this order:

1. Capacitors
2. LiFePO4 (a very distant second)
3. Lead Acid (a fairly distant third)

Whoa, whoa, don't say that too loudly here or this thread might go to the unicorn corral, because mostly proven facts based on sound reasoning doesn't go over well with the establishment. YOU haven't proven that it's more efficient to charge and discharge capacitors than lead-acid batteries, ergo you are talking about unicorns.

I'm feigning offense, and don't really care what goes to the unicorn corral. I just find it odd when legit concepts are blown off.

[Enki]

As far as "unicorns" go, I'll add this to the thread:

The weight savings is very real. These caps weigh almost nothing, despite their size (510 grams, 1.12 pounds each according to datasheet; this is for the huge 3k Farad caps). They weigh more than the LiFePO4 cells I've had my eye on, but still, the savings over a regular sized battery is massive.

They aren't as efficient for overall energy storage (amp-hours), but everyone already knew that. If you need to run the electronics, etc in the car without the engine running, then Lithium Iron batteries will be your best bet; a 10 amp hour setup should be plenty, but it will cost for the batteries, charging controller, etc; something that small won't last as long as a cap bank will (as far as charge/discharge cycles goes).

I dug up the spec sheet on my caps as well as the lithium cells I've had my eye on; that info is as follows:


Cap cells:

Cost:
$20 (used on Woot) x 6 = $120

Cycles:
>1,000,000

Operating temps:
-40c to 65c (-49f to 149f)

Weight:
510g/each x 6 = g (6.7 lbs)

Max surge discharge rate (< 30 seconds):
210 amps (sustained until empty; peak < 1 second = 2200 amps)



LiFePO4 cells:

Cost:
$18 x 4 = $72 + charge controller/balancer (another $30 minimum; both from batteryspace.com)

Cycles:
>1,000

Operating temps:
Charging: 0c to 45c (32f - 113f)
Discharging: -10c to 60c (14f to 140f)

Weight:
340g/each x 4 = 1360g (3 lbs)

Max surge discharge rate (< 30 seconds):
100 amps

[redpoint5]

Originally Posted by Enki

LiFePO4 cells:

Cost:
$18 x 4 = $72 + charge controller/balancer (another $30 minimum; both from batteryspace.com)

Cycles:
>1,000

Operating temps:
Charging: 0c to 45c (32f - 113f)
Discharging: -10c to 60c (14f to 140f)

Weight:
340g/each x 4 = 1360g (3 lbs)

Max surge discharge rate (< 30 seconds):
100 amps

I'm going to have to thank you again for finding these batteries. Much better than the Hobbyking pack I bought at 4.2Ah for $40. The Hobbyking does have a rated surge of 168A, but 100 should be good enough when paired with a Supercap.

[serialk11r]

Enki have you considered pairing the motorcycle battery with a small cap bank? As demonstrated by the video in the original post, you can start the car with a surprisingly small cap bank thanks to the high power density. Your motorcycle battery sounds like it's still good to go.

My own plan is to go for a Deka ETX12 battery next time around (so a little less than half the capacity of stock), lead acid since I feel like there's no reason to put that big of an investment in at this point, and 6 400F ish capacitors. If it can start a V8 it can surely start my wimpy 1.8L 4 banger. The smaller capacitors will also bleed less current which will help.

[Enki]

Yes I have thought about that, but I'm short on room and I doubt that the cap bank would make things significantly better, as the set of 650 caps I have will have a total capacitance of 108.3 farads, or a voltage drop of 5.17@ 560 amps of draw. Even if it does help with battery lifetime, I already have a spare, and I've found that I can double the life expectancy of each battery by running it through a de-sulphating process on the charger. As a side note, the 650s have a bank leakage of about 9 mA (as opposed to the 3k farad caps total leakage of 31.2 mA), so adding the cap bank would probably make my drawdown issue worse.

Instead, I'm going to use the 650 bank as a big, fast charging LED flashlight with one or more buck converters powering parallel arrays of LEDs (making near full use of the bank). Charging such a thing off the alternator shouldn't be as big of a deal, as to go from 0-~14 volts should take less than 20 seconds, and with 10x25mA LEDs, last 86 minutes. Not a bad trade, considering.

Also the spec sheet for all the variants of these caps is here: http://www.maxwell.com/products/ultr...s_10153704.pdf

[Enki]

Originally Posted by redpoint5

I'm going to have to thank you again for finding these batteries. Much better than the Hobbyking pack I bought at 4.2Ah for about $70. The Hobbyking does have a rated surge of 168A, but 100 should be good enough when paired with a Supercap.

Does the Hobbyking pack have an integrated charge/balancing controller? If not, and if the $18 cells remain relatively balanced, a set of four of those with a bank of 650s would probably be downright amazing.

[serialk11r]

Originally Posted by Enki

Yes I have thought about that, but I'm short on room and I doubt that the cap bank would make things significantly better, as the set of 650 caps I have will have a total capacitance of 108.3 farads, or a voltage drop of 5.17@ 560 amps of draw.

Okay I know you're looking at the numbers and it doesn't look like it adds up. I did that too, and didn't think this would work. If you watch the videos though you'll see that the wimpy 58.3F pack is capable of starting a 1NZ-FE engine starting with less than 11V. The Maxwell caps don't have an extraordinarily low ESR either. What I would conclude from that is that a 108.3F bank should be able to start your 2.3L low compression ratio engine, and a small battery would keep it topped up at a sufficient voltage to do so for at least a month.

I think this probably has to do with the fact that the starter usually only has something like 7V to work with so tiny capacitors giving it a 8V jolt is a massive improvement.

[redpoint5]

See this thread for updates on my battery replacement project.

I worked on my truck yesterday and found that the parasitic drain was 45mA, not 20mA like I had thought. The backlit thermometer that was wired directly to the battery was responsible for 39mA of that!

After removing the thermometer, the truck had a parasitic drain of only 5.89mA. This drain is likely entirely due to the aftermarket radio since the truck doesn't have any power features. I removed the faceplate of the radio to see if drain further reduced, which it did by a measly 0.3mA. Interestingly, I noticed periodic spikes to 11mA after removing the faceplate, and found this was due to an anti-theft LED that blinks with the faceplate removed.

Using 3000F capacitors, I could let the truck sit for 4 days before needing a boost. Combined with my 4.2Ah LiFePO4 pack, I could go for 3 weeks. If I were to use your 10Ah batteries instead, I could let the truck sit for almost 2 months!

I found that my 2.5W solar battery maintainer is totally inappropriate for maintaining the charge of my 350F (58F series) capacitors. With overcast sky, the panel put out 40mA of charge, and in direct sun late in the evening (6pm) I managed 140mA. It could probably peak at 200mA in direct noon sun. When I checked on the state of charge, my bank of 6 series 350F capacitors were charged to 16.02v and rising. The balance LEDs were beaming brightly trying to dissipate the extra charge, but were being over-run by the solar panel.

A 1W panel like this is probably more appropriate. At 85mA peak output, it would likely not over-run the balance LEDs on the capacitor bank.

Originally Posted by Enki

Does the Hobbyking pack have an integrated charge/balancing controller? If not, and if the $18 cells remain relatively balanced, a set of four of those with a bank of 650s would probably be downright amazing.

My mistake on the price of the Hobbyking pack. It's $40, but shipping was $27! It doesn't have integrated balancing, but it does have balance leads, and I purchased a balance charger for $27 and a quick connect voltmeter for $2 that gives the voltage of each cell.

The 8.4 Ah pack has a burst discharge rating of 336A!

Originally Posted by serialk11r

Enki have you considered pairing the motorcycle battery with a small cap bank?

That would be likely be unnecessary. I'm running just a set of 6x 400F capacitors on my motorcycle, but could easily replace them with just a 4.2Ah LiFePO4 battery. My bike only takes 70A to start, which is easily supplied by only a LiFePO4 battery. Besides, the battery compartment in motorcycles is extremely small. I couldn't fit both a capacitor bank and a small battery in my bike.

My recommendation for all motorcyclists is to run a single 4.2Ah LiFePO4 battery. My bike has a parasitic drain of 1.25mA, which means that a 4.2Ah battery has enough reserve to sit for over 100 days without being charged, and still start the bike with no damage to the battery. That means it could sit all winter without any maintenance and would still be in top shape in the spring.

At $40 for the LiFePO4 battery, nobody should be buying a lead-acid battery for their bike.

My own plan is to go for a Deka ETX12 battery next time around (so a little less than half the capacity of stock), lead acid since I feel like there's no reason to put that big of an investment in at this point, and 6 400F ish capacitors. If it can start a V8 it can surely start my wimpy 1.8L 4 banger. The smaller capacitors will also bleed less current which will help.

Have you measured the parasitic drain on your car? The Deka ETX12 only has 10Ah, of which probably only 4Ah is usable before you cause damage to it. It also weighs 12 lbs and costs a minimum of $40. I'd go LiFePO4 instead because it's much lighter, costs the same, and you can use most of the capacity without causing damage. It should last longer too.

[Enki]

I had a thought.

At 14.4v peak, a balancing circuit isn't needed, and here's why I say this:

With a maximum safe voltage per cell of 2.7 volts, equilibrium *should* ideally be around 2.4 volts, or 88% of maximum charge (with listed max safe actually being 2.85 per cell or a hair over 17v). Let's say that one cell is charged to only 2 volts while the other hits 2.8..that would be a variance of 29%, which is huge.

Bearing that in mind, what happens to charge rate as the capacitor bank comes up on fully charged? The charging rate drops off.

Thus, my argument is that a cap bank of identical condition/make/model capacitors will always self balance to a degree; what that degree is, I'm not sure, but the variance on my bank was only 5.186% from most charged to least charged. I'd wager that over time, that number would likely reduce slightly, and that the most charged capacitor(s) could likely hold more voltage before having an issue anyway.

I am hoping that moving to copper interconnects/connectors will help reduce some of this variance as well, as I know aluminum has a higher innate resistance and may be forcing more into the capacitors closer to the negative side of things. I may wind up testing this out after a while, but don't hold your breath.

Thoughts? Concerns?