Supercapacitors are really cool in their ability to charge and discharge very high currents without breaking a sweat. Unfortunately they dont have the energy density to get you much further than flat on your ass when your board shoots out from under you and promptly dies, but I want to know if they have a place in esk8!
There are several EV’s that are using a battery system that is a hybrid of traditional Lithium Ion battery tech (for its energy density purposes) and supercapacitors for their current capabilities. Why arent we doing the same?
The current load in esk8 is highly variable, as anyone who has looked at a VESC log file will know. When you are accelerating from a dead stop, you get a big current spike to the motors which very quickly drops off as you get moving, since it takes less energy to keep you moving now that your inertia has been overcome (oversimplification, I know). Then when you go to brake, especially from high speeds, all that energy gets dumped into your battery at whatever level you set to be safe for your cells. So our braking force is limited by the cells and their current capabilities.
Seems to me like some supercapacitors would be perfect to provide the punch of amps to get you moving, and to take the big dump of current to slow you down, without stressing our Liion batteries. So I’m calling all the folks who can pick apart all my hand-waving explanations and talk about what an actual application of supercapacitors (any maybe even ultracapacitors, if they become cost effective) could be for esk8.
Paging @DerelictRobot cus I found this when I was digging through the archives:
Also paging @Battery_Mooch because he seems to always have the answers about battery related questions, and also cus I like talking to him.
Please tag any other big brains who might know about battery systems.
Dude, me and my friend where talking about this but for an electric kart, (as esk8 doesn’t have a lot of room to spare)
the only thing is that I’m nowhere near knowledgeable enough to design a “discharge controller” that would have this incorporated.
I’m even having problems to figure out how to control a 10w 12v common anode rgb led with an arduino for a custom lightsaber. Im learning electronics now.
Yeah I’m wondering about this as well – is it good enough to just put supercapacitors in parallel with the battery, or do we need some special circuitry to make best use of them?
Not an expert in any way, but I’m always looking at battery research, but not much on SuperCapacitors specifically
A few problems with this mixed system, they have the energy stored across all the voltage range, that means to get access to all of it, you have to discharge it to 0 V, this doesn’t go well with our batteries, so just having them in parallel with a battery does almost nothing. Let’s say you have a capacitor bank that have 42V fully charged to match the battery, and whatever battery you are using has a peak voltage sag of 1 V, this means that this is the range the capacitor can work, the 100 Wh/kg magical energy density turns into less than 5 Wh/kg
You could have a buck/boost converter that takes whatever the energy stored into it and convert to the present state battery voltage, but to have something powerful enough to actually take advantage of the capacitor power output, it won’t be small
In summary, almost all the applications I’ve seen comparing supercapacitors, you would be better off just taking whatever space and weight this system would have and stuffing more batteries in
As others have mentioned, the incredibly low energy density of ultracaps typically makes them a good choice only where their low internal resistance is needed for very short periods of time…milliseconds to hundreds of milliseconds. Unless you can afford to have a huge and terribly expensive pack, that is.
The low voltage rating for ultracaps means you would need about 20 of them in series for a 12S li-ion pack (running at 2.5V each for decent cycle life). Then add on the balancing circuitry. For that size and price those caps need to clearly outperform what additional li-ion cells could do.
The photo below (some stuff blurred out) shows a 6S1P ultracap pack (using 100F caps) that straddles a 4S1P A123 26650 pack for IC engine starting. It works great but you can only get about 20A for one second from the ultracaps alone. That’s terrible energy density. However, it also means you can get around 400A for a couple dozen milliseconds and that’s enough to unstick an engine. Then the A123 cells are enough to spin things long enough for the engine to start. This is where ultracaps really shine, huge amounts of current for very short periods of time. Way beyond what li-ion cells can do unless you have a huge pack.
We have look at what problems we want to solve using ultracaps. Just doing some creative cooling and/or adding some additional cells in parallel can solve a lot of the issues at a much lower cost. Not nearly as cool but saving hundreds of dollars is also pretty cool.
The bursts of current needed during esk8 starting and the high regen current levels we would want are both things that happen for at least a couple of seconds, if not much longer. Now we’re talking about ultracaps of hundreds of farads each (see the second photo, 350F each), twenty of them, and their balancing circuitry and additional wiring. Or we can just add one or two more cells in parallel and the pack will run cooler, be stressed less, and cost much less than adding ultracaps.
I would LOVE to see a li-ion/ultracap hybrid pack be effectively used for esk8 but IMO it’s just not a good application for the low internal resistance and very high power density that ultracaps can give us. Having said this, I have been considering the use of a hybrid pack for my first board for a few weeks now. Not because it solves any real issues but just because…well…just because.
what if the caps were just used to dump braking current? Assuming they’re in parallel with the battery, they would have to be on some kind of mosfet that connects them just as you hit the brakes, and only when the braking current is greater than whatever the charge current at their current voltage would be so that they don’t suck power out of the battery. Then you’d have to dump their power somehow - I guess once they get above your pack voltage you could dump the power into the pack but that brings you back to the original problem. Plus, how would you discharge them to zero usefully?
One other thing - I don’t know much about supercaps, but google is telling me they have less energy density than a li-ion cell, something like 100wh/kg. A 70kg rider moving at about 22mph has something like 1Wh of kinetic energy, so maybe you don’t actually need a ton of energy density to get some usable performance out of them. I dunno.
The 100F ultracaps in my first photo above have about 0.05Wh to 0.07Wh of usable energy each when in a series string paralleled with li-ion cells. The ultracap is about the same volume as an 18650 but has about 1/100th to 1/140th the energy density of a decent power 18650
But with a carefully designed system that perhaps only “absorbed” the energy that is created at above the pack voltage an ultracap setup would be perfect. Better than dumping all that energy into resistors, that’s for sure.
The electronics for balancing, monitoring, and switching the caps in/out would be a big job though.
This is something I was also thinking about. But how would that be implemented? What would the circuitry look like to make the supercaps an energy dump without having them suck current from the batteries to keep themselves full? Some beefy diodes?
I’ve done too much reading on this topic hoping to make it worth it somehow. The allure of the idea still makes me want to try even if it’s not really worth it for any practical reason.
The problem I found, mentioned earlier, was that most caps are rated for a relatively low voltage. Higher voltage high farad caps tend to be super bulky and very expensive. To go with cheaper lower voltage caps so you’re not spending $500+ on massive caps, you would need to add them in series.
The problem with caps in series is your total Farads drop exponentially … (Is it exponential? It’s a lot, see:)
Any more than a couple high F caps in series and the effective capacitance becomes basically useless for our application.
Look for single caps we can parallel rated for the voltage of our packs and the selection gets super tight.
Those big F pre-made cap banks look sick, but dropping 10s of Ks of $ on a battery … … Look up the caps they use to make them with and hundreds of $ per cap …
Why? Your guess is as good as mine… it’s a bummer, that’s why I started looking into higher voltage rated caps rated for greater than my pack that I could parallel with each other and the pack…
Going off the top of my head (+/- an order of magnitude probably since I haven’t calculated it recently)… It’s roughly the cost of a high end pack itself for a relatively bulky cap bank that will discharge ~80-100A for just a few seconds to soak up the voltage drop of a liion pack, say 3V drop (caps doing their thing from say 47V down to 44V).
Really it’s not out of the question…
Maybe for some exceptionally energy dense cells, low discharge rated, 5P config, paired with a big ol cap bank… It would have almost no voltage sag for those few seconds, then start to load the pack… At that point I’d just opt for a bigger pack though, or lipos which already have minimal sag.
I feel like if they had a place, it’d be for welding on the bench with 18650 cells that can’t support the types of current that welding requires. Like a 3S1P 30Q cell welder with a supercap and probably some regular caps as well.
… Look up the caps they use to make them with and hundreds of $ per cap … 
Damn is that true? Why is that? Capacitance as a unit of measurement goes straight over my head. I was only just wrapping my head around V/A/W haha.