supercapacitors + battery

[jimmyhackers]

wonderfully strange subject here. as always from me

ive seen a guy on youtube combining supercapacitors with lipos as a lightweight replacement for a car battery. with some very promising results

this gave me an idea on how i could possibly boost my SLAs for around £20.

at the moment my slas dont have the puff in them to provide full amps/torque under acceleration. past 15mph the amp loads reduces and they can. most of the time im going faster than 15mph and it take me all of 5 -7 seconds to get to 15mph.

so my idea is attaching in parrallel either one supercapictor across 48vs of batteries (4xslas) or 4 individual capacitors (one across each sla). with the intention that the slas charge the capacitors when not in use. then when under load the supercapacitors add their minimal farads/amphours the the batteries for the 5 or ten seconds i need, reducing the load on the batteries.

few questions....

anyone sure how farads calculate out into amp hours, the internet is vague and im wondering what size ill need?
should i go for one across all four in parrallel or one per battery?
i know capacitors are sometimes dangerous. i plan on having them attached/wired semi perminantly to the batteries, no switching or control etc. could they in any way mess up my batteries?

thanks in advance for the help
jim

 

[Deleted member 4366]

The number of Whs in a supercapacitor that you can afford is so low that it's negligible. Lipos would be better.

 

[jimmyhackers]

hopefully i wouldnt need that many whs. lets say im currently at 1000w. thats 1000watts per hour. an hour has 3600 seconds in it. so per second im using 0.27777watts. i only need a max of ten seconds so 2.7watts.

i know the higher C discharge rate from my slas really effect the actual ah you get from it. for me this is at its most from standstill. with the multimeter on the voltage dips are only really noticeable from standstill for a second or so. so even 0.2777watts might help.

ive just dug out somthing i bought a long time ago (7years), its a car voltage stabilizer, its probly whats given me the idea.

basically the same concept minus the supercapictors, just regular caps in this

im going to do some testing on a single battery and see if it makes a difference.

 

[trex]

you've got your units mixed up. Watt is a unit for power, WH is a unit for energy, a 40W lightbulb consumes 40WH in an hour. 1WH = 3600 Joules.
Your idea of using supercapacitor is a non starter. You need 2 Farrads at 1V to store 1 Joule in perfect condition.
Let's say you want to give to your motor 1000W boost for 15 seconds. You will need 15,000J for that. The formula is E = 0.5 * capacitance * voltage square. The voltage is 40V, the capacitance required for 15 second boost is 15000(J)/800(V square) =18.7F
You can find on ebay 100F @ 2.7V caps, using them, you need to build an array of at least 18 caps in series.
This guy sells them for £12.47 for 10,
http://www.ebay.co.uk/itm/500-Farad-2-7-Volt-Super-Capacitor-Green-Cap-EDLC-Unit-Quantity-2-/181691282233

PS: even if you want to spend the money, you'll need to build a PSU to convert supercaps to stable 40V supply rated at 1000W to to make it work. The discharge profile of supercaps are not suitable to power motors.

 

[Deleted member 4366]

Another thing, they won't go below about 10v per SLA because the SLA will hold up the voltage. They will only discharge the voltage that the SLAs sag, so about 30% at most, which is just as well because they short out if you fully discharge them.

What have you got against lipos? They would do exactly the same thing except that you get more charge and they're cheaper.

 

[jimmyhackers]

thanks for the info. not quite sure if agree trex. or maybe you think im after some other effects.

technically my batteries can provide the 1000w. they just sag in voltage when doing so. i want to use the capacitor to lessen the sag (not much else). so the actually power available to the motor will be the 1000w from the batteries plus the 5-10 seconds of what ever wattage that equates to from the caps. 2.7watts

and so far the practical looks to prove so

welll.......ive been out today with just one voltage stabilizer (old car battery one i had ) on one of my batteries with the most sag.

max dip without it was 10.2v, max dip with it was 11.6v. thats a 1.4volt improvement!!!!

i then tried it on what i think relativeley new battery max dip normally 11.3 without and with it its max dip was 12.1. thats a 0.8volt improovement.

with just the one on my worst battery my led meter stayed full, and power delivery seems faster and smoother. top speed was the same however.

sorry to be going against what ive just been told guys but ive had a look on ebay. about 12 quid i pop and they now come with a numbered led volt meter built in.

im buying some more

 

[jimmyhackers]

another way to look at it is the amps side. 48v at 20amps makes the 1000w give or take. thats 20amps per hour so divde 3600 (seconds in an hour) by 20 and you get 0.005263 amps per second

so for ten seconds its 0.05263amps. also as i said in above post the battery can still provide this per second anyway. adding 0.025amps will still give me a 50% available amp increase (10ah boost for the ten seconds)

 

[Deleted member 4366]

Your units are all over the place. Amps are amps. If you have 20 amps, its 20 amps per second or per minute or however long you have it.

 

[4bound]

D8veh is right. Amps are a unit of current. If current flows for a certain time, say you use 20Amps for an hour that's 20 Amps. If you have it for ten seconds its still 20 Amps. Adding 0.025 amps will make diddly squat difference to a 20 Amp current.

Time comes into it when we start talking about power storage - so if you can take 20 Amps for an hour from your battery, or 10 Amps for 2 hrs, or 40 Amps for 30 mins that's a 20 Amp hour battery.

 

[Alan Quay]

another way to look at it is the amps side. 48v at 20amps makes the 1000w give or take. thats 20amps per hour so divde 3600 (seconds in an hour) by 20 and you get 0.005263 amps per second

so for ten seconds its 0.05263amps. also as i said in above post the battery can still provide this per second anyway. adding 0.025amps will still give me a 50% available amp increase (10ah boost for the ten seconds)

Jimmy,

Just to back up what others are saying, you seem to have Amps and Amp Hours confused.

In theory, you 20ah battery can supply 20a for 1 hour. During every second of that hour, it will be supplying 20a.

Hope that helps.

 

[selrahc1992]

The number of Whs in a supercapacitor that you can afford is so low that it's negligible. Lipos would be better.

I admire your originality Jimmy (I too like building frankenbikes and dream ofsome very weird creations), but in the real world 16 ah hobbyking lipos go for 80 quid and give 20 C, comfortably enough for a 2.5 kW BPM job i built,though the jury is out on the number of cycles I'll get

 

[trex]

Jimmy, I take it that you watched this video:

The guy on this video replaced the car battery with the lipo, like d8veh suggested. His supercaps did not help much.
His supercap array consists of 6 x 350F Maxwell D cells connected in series. You'd need 3 of those arrays. You will get 15 seconds boost but I think the boost won't be strong enough to make your spend worthwhile. Boost your battery with lipo packs but lipos are dangerous. Supercaps are 20 times bulkier than lipo for the same job.

 

[jimmyhackers]

sorry i should of been clearer between amps and amp hours in my ramblings.

i think reason i havnt is because im using 20ah batteries and my motor roughly takes 20 amps (at full tilt). meaning i havnt had to really factor in the calculations that require much differenciation between the two.... sorry my bad

so factor that mistake i made out between amps and ah and my maths still makes some sence, and in the real world ive seen it do as i expected which is a benificial effect. little bit of performance gain and less strain on the batteries.

you get where im coming from though.....?....right?
i only need the littlest amount of energy at certain times to fill in the gaps (flaws) that are inherant in the design of an sla, the peukert effect is it?

the less peukerting my battery does the further itll take me "in theory"

which brings me onto lipos, and why i havnt fully gone there yet

i have run it on 4x 3300mah 30c rating lipos. making 48v. which cost me around 80quid for the 4. the cells being so small i still got an equivilant if not worse battery voltage sag and i got all of possibly 5 minutes worth of riding.
not really too impressed

my charger tells me exactly how many mah it puts into a battery on charge and despite the 30c rating my lipo batteries are still suseptable to the peukert effect at my bikes hubs discharge rate. probably due to their tiny ah size. if i ran them constantly like this im sure id tire them out pretty quick.

to minimizer this pukert effect i need to have a lipo battery rated at 20ah or above so im getting less than or equal to a 1C discharge rate. i dont have 400 plus quid for 20ah lipo, and i dont want to spend around half for a
10ah one becuase it "will do" or another 80 quid to turn my 3300mah bodge lipo battery into a 6600mah battery.

the weight difference when i went to lipo.....honestly i couldnt notice a thing when riding, it went, stopped an turned as well as with heavier slas. which should of been the main benefit when goign to lipo i thought.

sorry if youve heard this from me before but i have at least 8 good slas that i use (4 at a time) and if i can boost them to a comparable lipo standard (in everything but weight) for 40 quid..... compared to 400quid it makes sense

so.....if all goes well i have at least another years worth of use from my current batteries. when they do go caput there might be an even better battery tech out there

 

[jimmyhackers]

lol you posted while i was typing all that rubbish above....

yep you got the video

the supercaps in that instance are used as i am using my regular caps (in my battery stabilizer).... to protect the battery with a buffer zone.

lipos probably need it too (in the car application)... the CCA cold crank amps of a car battery are in the hundreds and a lipo by itself may turn over a car, but to protect it hes using supercaps.

the only thing he need to protect is the starter motor draw which as i said is in the hundreads so even an 5ah lipo at 30c can produce available 150amps. and 150amps may not be enough for some cars to start...

 

[trex]

I am very interested in your experiment. The cap in the controller that is used to stabilize the average 36V Lithium ion battery is only 0.47F. If you go with 350F Maxwell D cells, the array will be equivalent to 350/18 = 18.4F, 41 times the capacitance of the one in the controller.
Please come back and post the result here.
PS: I still don't think it's worth the money but it's only money.

 

[mrgeoff]

I suggest you get hold of a 1 farad capacitor, put it across a 55w 12v bulb, and see exactly just how long it lights it for.

You will be VERY disappointed...

I have a 33 farad supercapacitor that I have toyed with, still pathetic for its size compared to the same size volume of Li ion :-(

 

[jimmyhackers]

i have done many a playing with caps in the past. ive made gauss rifles, guitar foot pedals and amps, strobe light etc etc. i get the theory behind basic electrical concepts and components.

as much as you say it shouldnt work.... it is. so something in the numbers is wrong or missing.

ive thought of another analogy for what the caps do (helps if you know computers)

if you view the caps as ram in a computer. electrical power as information, a slow hdd is my lead acid batteries and the processor is my motor.

the rams high input and output speeds in a computer are used to bridge the slow data transfer speed the hdd can provide. giving overall better performance.

and for those in the know i could further the anaolgy that an ssd could be a lipo. also awsome but also flippin expensive like lipos in comparable storage size/price ratio

it could be that....the nature of a brushless motor and pwm control. is that it pulses. off and on all the time. on a graph it may look like loads of large spikes of energy draw with valleys of no draw inbetween.

the caps may be smoothing out these spikes (or at least taking the top peaks worth of load off of the battery) meaning my battery isnt as stressed as its peak power draw is reduced? less stress means less peukerting, especially in lead acid batteries which are very susceptable to it. which could be why im seeing an improvement

i have 4 new ones on the way (deizin volt stabilizer). when they arrive ill open my old one and see whats what. its a clear case and so far i can tell its 4 largeish capacitors a few other smaller ones and not a lot else. its also got a board mounted 1.5amp fuse meaning max input or output was intended at lower than 1.5amps for however many (less than seconds).

1.5 amps or less seems to be doing it though. i have......a fair few caps lying about.....i know ive got 4 units on the way but.....i could make my own version.

if someone can confirm my power spiking graph due to pwm..... and if what im doing could be lessening the peukerting effect. it could be capacitor hybrid batteries arnt that far fetched of an idea??? especially in brushless motor application or anything thats pwm?

thanks for your input again guys

jim

 

[trex]

This is the voltage graph of the three phase wires of your motor:

The frequency of the alternating voltage at your three phase wires is around 20kHz to 30KHz when your bike run at 15mph.

Think of the capacitor as a lift, going up and down carrying people.
Its usefulness is measured by the number of passengers.
A fast lift that can take only 1 person but going up/down 20,000 times a seconds is a lot more efficient than a slow one that carries 50 but can only go up and down once every 15 seconds. The fast lift is the 0.47F capacitor that is built into your controller, the large and slow lift is the battery stabilizer supercaps that you want to add to your SLAs.
Of course you are parallelling the supercaps and the existing voltage stabilizer, so you'll get something for your money but a 0.47F is good enough for this job, you get a very diminished return for adding more caps. You can calculate the array's impedance to evaluate the effect of adding supercaps. If your battery drops from 14V to 11V when you draw 20A, the ESR of the battery is 3V/20A = 0.15 Ohm. The impedance of your supercap array is Z=1/(omega * C), omega is the frequency, as you use the array only 15 seconds, omega = 0.066, Z=1/(0.066 * (350/18)) = 0.77 Ohm. Z is zero to start with then increases rapidly in the next few seconds. The effect is there, but only small - about 15%.That effect has an exponential shape, you should feel it strongly at the beginning two three seconds then it dies quickly, see the discharge profile of supercap below.
You can read more about electrical impedance to make sense why a capacitor is sometime inefficient as voltage stabilizer.
http://en.wikipedia.org/wiki/Electrical_impedance

Comparision of discharge profiles, supercapacitor versus battery:

http://en.wikipedia.org/wiki/Supercapacitor

discharge profle of supercaps:

choosing the right power source:

 

[jimmyhackers]

thats some interesting stuff........ been out on the road testing and experimenting for a week or so and i thought id update....... its working well so there has to be an explanation to it all. i think ive found it

not quite sure where the 0.47farad cap in my controller is :s ..... the one im seeing is more like a 63v 470 "micro" farad capacitor.
so really its 0.00047 farads??? not to sure about my number of 0's there

thats quite a lot smaller than what im adding which is in the order of 0.048 farads or 48,000 microfarads

ive yet to recieve my chinese parts so i made my own 16v 12,000 microfard banks. consisting of 2x 16v 5600microfarad caps in parrallel. one bank per battery seems to mimic my lessened voltage sag from the proper raizin battery stabilizer. generally a volt per battery

been out for a ride nearly every day this week in the lovely weather.....apart from a popped tyre its been seemless. the extra runtime coupled with more poke just makes it that little bit better.

im getting another 20 -30 minutes runtime with these. making a grand total of an hour and a halfish untill my battery meter finally goes to the lowest of 3 settings. torque is defintely up and the bike feels a lot more pokey

ive noticed looking through all my capacitors.... that there seems to be an exponential increase in ratio between voltage and farads in capacitors of similar sizes.

i.e in my hand right now i have two similar sized electrolitic capacitos. a 1000microfard 25v cap and a 22microfarad 250v cap.so i get 40 microfards per volt in the 25v cap and 0.088 microfards per volt in the 250volt cap.

my caps are rated at 16v and 5600microfards.... two in parrallel over a 12v battery could make around 1000 fards per volt instead of my controllers measly 7ish farads per volt.

being extra silly im going to go out on a limb here and say that these banks ive made are about 100 times larger (in farads) than the controller cap alone. this is enough to take the stress of the high/max amp load off of my batterys and place/share it onto the caps. giving me more ah in the battery due to a lessening of the peukert effect.

feel free to have a bash at my math/theory, its more than likely some of it isnt quite right but im getting a noticeable/quantifyable 50% increase in runtime.

 

[trex]

the voltage stabilizer capacitor is the largest capacitor you have in your controller. If that is a 470 microfarad then adding extra capacitance will help because that's not enough to support a large DD motor.
Still, you get dimishing return as you keep adding. I reckon 0.1F to 1F is the range, most controllers' voltage stabilizers don't go beyond 0.1F.
afer that, it's not worth the money.