DIY battery + Capacitors, useless or ok?

[spuzzete]

Hi everyone,

I searched the forum but I could not find any thread where I can find an answer.

I have built a 36v 24ah (10s12p) li-ion battery for my ebike using 18650 cells from laptop batteries. The typical power usage is between 300-500w when travelling at full throttle and not pedalling. But during stop & go the current draw is up to 32A/1200w. This causes some of the cells to go under 3.2v, which I want to avoid.
I tried smoothing the throttle adding a variable resistor to the ground wire of the throttle and things have improved a bit but sometimes I still draw too many Amps during stop & go.

I would like to avoid that "stress" to the battery adding some capacitors in parallel to the battery. I read a discussion here and people have different opinions. Moreover I cannot find any real data about the use of the capacitors in parallel with the battery.

I want to make it clear the I do not intend to use capacitors instead of the battery, I already know it's not worth it, I am just wondering if they would avoid some stress to the battery during stop & go.

I was thinking to use some capacitors of pc's atx power supply as I could have access to those for free. Since I don't have to pay for them maybe it's worth a try?

I appreciate any opinion on this subject.

[cal3thousand]

Sounds like a good idea to me. But I'm no expert.

Not sure if the discharge would be handled properly though. I'm interested

[Hillhater]

do you really think 2.7A /cell ( 1.3C) for a few seconds, is "stressing them ?.
You could always reduce the max amp draw from your controller.

[CamLight]

I was thinking to use some capacitors of pc's atx power supply as I could have access to those for free. Since I don't have to pay for them maybe it's worth a try?

I appreciate any opinion on this subject.

Capacitors could definitely supply enough current for a second or two (whatever's needed) to take the load off the battery but the caps from ATX power supplies aren't nearly enough. You'd probably need hundreds of them as capacitance in the range of hundreds of Farads are required.

[spuzzete]

Capacitors could definitely supply enough current for a second or two (whatever's needed) to take the load off the battery but the caps from ATX power supplies aren't nearly enough. You'd probably need hundreds of them as capacitance in the range of hundreds of Farads are required.

So I basically need some supercapacitors just to supply enough current for a second or two.

Maybe I can try with a few capacitors just to see if I have any changes (at this point I doubt it), but I always like to try just for the fun of it. As there are different capacitors in the Atx power supply should is it best to choose the 50v/63v ones or the 200v ones? the 200v ones are 680uF, the others several values of capacity.

[SamTexas]

But during stop & go the current draw is up to 32A/1200w. This causes some of the cells to go under 3.2v, which I want to avoid.

It's perfectly ok for 18650 laptop cells to go under 3.2V under heavy discharge. That's still less than 2C (for your 888wh pack). And you can safely do 2C for short periods (30 secs give or take). If you look at Panasonic datasheet you will see that they discharge their cell down to 2.50V under load. I personally have cells down to 2.70V regularly when I am using my smaller pack (600Wh) during accelerations from stop. The pack is about 18 month old now and shows no sign of degradation.

That said, go ahead and try your experiment. It sounds like a fun project. Do come back and give a full report, I'm interested.

[dogman]

I'm just confused how you get A123's in 24ah size to sag that much under 1200w. Sure they are real? Sounds worse than a ping pack's sag under the same load, and those are 2c stuff.

Bottom line, you have a battery problem fixable only by weeding out your weak cells. That is, if they aren't all weak cells.

Maybe I just misunderstand. They go to 3.2v. From what starting voltage, 3.65? Or are they just sagging at the end of your discharge, when they are nearly there at the edge of the cliff.

[CamLight]

So I basically need some supercapacitors just to supply enough current for a second or two.

Maybe I can try with a few capacitors just to see if I have any changes (at this point I doubt it), but I always like to try just for the fun of it. As there are different capacitors in the Atx power supply should is it best to choose the 50v/63v ones or the 200v ones? the 200v ones are 680uF, the others several values of capacity.

The 200V caps would be the safest to play with (once discharged completely in ATX supply!!!) but what are the highest capacitance values for the 50V/63V caps?

If you had twenty of those 680uF caps, you'd get 10A of current for only about 16 milliseconds.
If you wanted to get 20A out of them, you'd only get it for about 8 milliseconds.

Here's the equation: Time = (Capacitance in Farads) x (Start Voltage - End Voltage)) / Current in Amps

Not sure if it's worth experimenting with, but give it a try and let us know how it goes!

[megacycle]

Could try Maxwell's site
Edit and online shop,
and look thru the supercaps and do the math and work out what you need,they need to be low esr and obviously not for back up supplies seriesing is an issue as you could need to balance them.

[SamTexas]

I'm just confused how you get A123's ...

He doesn't use A123. He built his pack with 18650 laptop recycled cells. You haven't had your morning coffee, right?

[spuzzete]

Thank you guys for all the replies!

I will try to make a capacitor bank big enough (20-30-40 of those 200v big caps).
So far I tried with only 4 of those 680uF caps and of course there is no big difference.
Can I safely mix caps of different voltages as long as I stay above 50v (battery fully charged is 41v)?

I will keep you updated

[Tommy L]

I want to make it clear the I do not intend to use capacitors instead of the battery, I already know it's not worth it, I am just wondering if they would avoid some stress to the battery during stop & go.

I appreciate any opinion on this subject.

I have a friend that does a lot of things with capacitors. To build something worthy will be bulky and add weight. Probably better to add more cells.
But from what I'm reading here, a closer look at your pack might be in order

Is your pack 24ah or 2.4ah? Somethings not quite right.... \\m//

[spuzzete]

Is your pack 24ah or 2.4ah? Somethings not quite right.... \\m//

My pack is 24Ah (twenty-four). As I wrote before it's made out of 120 Li-Ion (li-co) 18650 cells recovered from "dead" laptop batteries.

What's wrong?

[CamLight]

Thank you guys for all the replies!

I will try to make a capacitor bank big enough (20-30-40 of those 200v big caps).
So far I tried with only 4 of those 680uF caps and of course there is no big difference.
Can I safely mix caps of different voltages as long as I stay above 50v (battery fully charged is 41v)?

I will keep you updated

No problem mixing different voltage-rated caps as long as you don't match or exceed the rating for any of the caps.

A little math is probably worth the time here. Using the formula: Time = (Capacitance in Farads) x (Start Voltage - End Voltage)) / Current in Amps,
and assuming 0.5 second and 20A to be the absolute minimum that might make any difference at all, and a 30V end-of-pulse voltage, then....

0.5 = (Cap) x (41 - 30)) / 20
Capacitance = 909,091uF

That's 1,336 of those 680uF caps.
Ideally, you'd like a couple seconds of current from the caps and that would require 5,348 caps.

[spuzzete]

A little math is probably worth the time here. Using the formula: Time = (Capacitance in Farads) x (Start Voltage - End Voltage)) / Current in Amps,
and assuming 0.5 second and 20A to be the absolute minimum that might make any difference at all, and a 30V end-of-pulse voltage, then....

0.5 = (Cap) x (41 - 30)) / 20
Capacitance = 909,091uF

That's 1,336 of those 680uF caps.
Ideally, you'd like a couple seconds of current from the caps and that would require 5,348 caps.

Hi,

thank you for posting the formula! So based on your calculation I would need about 1F for 0.5 seconds with the data you used.

So I should expect to "see/feel" some changes using a 5F supercapacitor.

I am thinking to use the cell-log8 to have a log and see the behaviour of the bat with and without capacitors. For now I will not use supercaps, I will just try with all the free stuff I can get.

Any suggestion about accurate logger(s)?

[Tommy L]

Is your pack 24ah or 2.4ah? Somethings not quite right.... \\m//

My pack is 24Ah (twenty-four). As I wrote before it's made out of 120 Li-Ion (li-co) 18650 cells recovered from "dead" laptop batteries.

What's wrong?

Without knowing the exact specs of the 18650 cells, I'd only be guessing at this point. The guess would be that the pack you built
is not capable of sustaining 30amps. 33amps x 36v is 1200 watts. If you have built a 24ah pack, then it's not capable of sustaining a
1.375C. Possible weak or dead cells, bad connections(cold solder joints). I would think that a healthy 24ah/36v pack should/would be able to
sustain 1.375C without an issue. When you have a solid pack, you will not be looking at Caps.

Tommy L sends... \\m//

[spuzzete]

Without knowing the exact specs of the 18650 cells, I'd only be guessing at this point. The guess would be that the pack you built
is not capable of sustaining 30amps. 33amps x 36v is 1200 watts. If you have built a 24ah pack, then it's not capable of sustaining a
1.375C. Possible weak or dead cells, bad connections(cold solder joints). I would think that a healthy 24ah/36v pack should/would be able to
sustain 1.375C without an issue. When you have a solid pack, you will not be looking at Caps.

Tommy L sends... \\m//

From what I gather from people that use the same cells it's pretty common for this pack to have "voltage sag" this way (discussion here viewtopic.php?f=3&t=26383) . These are not high discharge cells. Thing is when I pull 30 amps and the motor is already moving the voltage doesn't sag so much (3.2v alarm is not triggered).
The longest trip I had with this pack is 27km and the voltage under load looked still good (above 35v). Tomorrow if the weather is good I will go for a 40km+ trip, let's see.

[ProDigit]

I would just buy any 48V battery pack, preferably the regular Li Ion batteries, and plug that one in parallel with your current pack.
That way not the battery alone, but also the extension battery will provide the amps.
I personally would choose the cheapest Li Ion battery, because too much risk in destroying, stealing, breaking, accidents etc.
But even if not, 48V@10A provides you with an additional 480W sustained power, and it extends your range somewhat!

Although a capacitor might be really good to catch voltage fluxuations and spikes, it won't be able to catch the cranking amps the motor demands in the beginning.
Large capacitors are expensive, but perhaps a capacitor meant for industrial purposes (like the ones rated for 3000V, bout the size of a hot dog) would do the job.

[CamLight]

Although a capacitor might be really good to catch voltage fluxuations and spikes, it won't be able to catch the cranking amps the motor demands in the beginning.

Actually, ultra-capacitors (super-capacitors) are incredibly good at supplying hundreds to thousands of amps for even several seconds without problems. I have a 6S ultra-cap string (using 400F caps) here that can easily start a 1000CC motorbike.

Yup, these things are big.
Yup, they're expensive too.

[CamLight]

Hi,

thank you for posting the formula! So based on your calculation I would need about 1F for 0.5 seconds with the data you used.

So I should expect to "see/feel" some changes using a 5F supercapacitor.

I am thinking to use the cell-log8 to have a log and see the behaviour of the bat with and without capacitors. For now I will not use supercaps, I will just try with all the free stuff I can get.

Any suggestion about accurate logger(s)?

A supercap string with an effective capacitance of 1F would be the absolute minimum I'd recommend trying out. A 5F (effective) string would be better.

Assuming 2.7V supercaps running at 2.5V max, you'd need 17 caps in series to cover a fully charged 10S pack.
And since the capacitance is reduced when you connect caps in series, you'll need a string of seventeen 20F supercaps to get a 1F string that will handle 42V without problems. Better to have a string of seventeen approx. 100F supercaps though (effectively 5F).

Caps in series formula:
Ceff = 1 / (1/C + 1/C + 1/C ....etc.)
Where Ceff= effective capacitance of the cap string and C is the capacitance of each cap.

Be sure that the ESR (internal resistance) of each cap is very low because putting seventeen in series will severely limit the cap string's effectiveness otherwise. A few passes at some Ohm's Law equations will let you know how much the voltage will drop for any particular caps you find.

[Edit] Almost forgot....
You'll need to balance the string of supercaps since they typically operate quite close to their maximum rated voltage. You can do that with a 250 ohm resistor across each supercap, allowing 10mA of balancing current to flow. Unless you charge each cap individually before assembling the string, it will take a few days on the charger/power supply to balance the caps. And, there will be a continuous 10mA flow from the battery pack through these resistors as long as the cap string is connected to the pack. You can disconnect the cap string and allow it to discharge but it will draw a huge surge from your pack when you connect it again.

Just a few of the reasons why supercaps are only good for a very few, specific applications.
They can be a lot of fun to play with but a new battery pack would be a lot less expensive.

[megacycle]

3 of these @ 8F X 48v would give that bike a serious kick in the ass. I'm thinking 150A or so they could deliver,
not much sag then, own built in balancing too
Shame they're $135 each pretty small @ about 100 x 50 x 50mm, 500+ group buy give them a price $5 each .
http://www.tecategroup.com/store/index. ... ts_id=1225.

Supercaps are coming down in price, they're be one hanging off every heavy duty lipo set up soon.
Thet will have to do until they bring out these in a year or two.
http://www.technologyreview.com/article/40220/

[ProDigit]

Just one question,

We all know capacitors can blow.
What happens if they do? Will it short circuit, or will it disconnect? (usually when they blow,is when one of the anode plates touches a cathode plate, and it blows.

If you ask me, the higher the voltage handling, the better the cap, as it will probably last longer on lower voltages.
Instead of finding on that operates on 36/48/96V or even 110V, wouldn't it be better to have one that operates at 1kV or so?

Too high values aren't good neither, as their leakage may be larger (eg: 5V leakage on a 5kv capacitor, may be insignificant for those 5kV operations, but if it's operating at only 20V, 5V is a lot).

[ProDigit]

3 of these @ 8F X 48v would give that bike a serious kick in the ass. I'm thinking 150A or so they could deliver,
not much sag then, own built in balancing too
Shame they're $135 each pretty small @ about 100 x 50 x 50mm, 500+ group buy give them a price $5 each .
http://www.tecategroup.com/store/index. ... ts_id=1225.

Supercaps are coming down in price, they're be one hanging off every heavy duty lipo set up soon.
Thet will have to do until they bring out these in a year or two.
http://www.technologyreview.com/article/40220/

Why not simply solder 2 lower powered ones in parallel, to get the same rating if that's cheaper?

Also,the cap you linked to has an operating voltage of 16V, you must go higher than your max voltage (that is, in case the battery is fully charged, and the motor is regenerative breaking (think 60V +)


I think audio capacitors might be really nice!
Capacitors used in audio amplifiers (especially for high power car amplifiers), are reasonably priced.
They usually handle upto about 60V, though I can't tell from the amazon specs.

This one would look quite cool to mount under the bike, especially because it has a blue led built in (10Farad, ~$65):
http://www.amazon.com/Boss-CAP10-10-Far ... =de_a_smtd

[Tommy L]

I would just buy any 48V battery pack, preferably the regular Li Ion batteries,

I had to kick in here..... you say: "Regular Li Ion Batteries".

Read this..
A battery is a transducer that converts chemical energy into electrical energy and vice versa. It
contains an anode, a cathode, and an electrolyte. The anode, in the case of a lithium battery, is the source
of lithium ions. The cathode is the sink for the lithium ions and is chosen to optimize a number of
parameters, discussed below. The electrolyte provides for the separation of ionic transport and electronic
transport, and in a perfect battery the lithium ion transport number will be unity in the electrolyte. The
cell potential is determined by the difference between the chemical potential of the lithium in the anode and
cathode, ¢G ) -EF.

My question is: what is the regular cathode material to be a regular Li Ion cell?

CoO2 (cobalt oxide)
Mn2O4 (manganese oxide)
NiO2 (nickel oxide)
FePO4 (ferrous phosphate, also know as Iron phosphate)

So the combined chemistry gives the cell a certain potential.
What concerns me here is that we are all trying to help each other, so I'm trying to point out that there is
no such thing as a Regular Li Ion Cell or Battery Pack for someone to purchase. Each chemistry has
pro's and con's and care/maintenance to have longevity and safety.

What one to pick and how much you need is almost always based on what you are trying to achieve and being able
to understand it's limitations and safety requirements once you choose a Li chemistry. \\m//

Tommy L sends.....\\m//

[ProDigit]

with a 'regular' Li battery I mean Li-Ion, which is currently the standard of Li batteries (soon to be replaced by lifepo4); but also one of the cheapest batteries, as lifepo4 can be quite expensive.

They are used in electric bikes, are extremely lightweight (more so than liFePo4), and handle good current flow.
Whatever power generated by the motor through regenerative breaking (if applicable) can be shared and absorbed over both batteries.

any battery has the same properties as a capacitor, in that it can charge and discharge like a capacitor, only a little slower, but handles much more capacity.

I know there are lots of Li-type of batteries out there, the best would be LiCd, because of it's superior power handling, and peak draw, but it's considered unstable, and can catch fire. But just any battery you can get your hands on, like the ones they sell in BMSbattery.com

For the purpose of just having more torq when accelerating, that depends on the motor. The motor can only handle a certain amount of power, and there's no use going above that. However, if your battery only delivers 15 sustained amps, and 30 peak (@48V that is 1440W); but you're equipped with a 700W motor that can handle a peak draw of 2kW, you will benefit more from installing a second battery than a capacitor.

For instance:
An electric bike does not have much storage space,
This battery is only 66mm X 168mm X 158mm, just about as large as a 2 large capacitors.
http://www.bmsbattery.com/48v/166-48v-1 ... -pack.html

It handles power fluxuations well too
It gives more than enough power to the motor to max out the motor's powerhandling, when plugged in parallel with the stock battery
Any voltage spike will be shared over the two batteries, meaning spikes will be halved, current draw (cca) will be significantly reduced
On top of that, it gives you longer mileage,something a capacitor does not do.