[Solarsail]

If you are looking for 96V average, i.e. the controller can work with higher voltages but expects the average voltage to be 96V as the cells discharge, then you need to divide that voltage by 3.6 and not 4.2. The cells reach their average midpoint capacity at about 3.6V (actually 3.63V). 4.2V is the maximum voltage at full capacity. My understanding of motor/controller manufacturers is that 96V means the average voltage and not the maximum voltage.

In this case, you would need 96/3.6 = 26 cells in series. This is commonly written as "26s30p". So there are groups of 30 cells in parallel, and then 26 of these groups in series to make the pack.

With only 23 cells in series, you will be starving the controller half the time, and it may shut down prematurely to save itself. And you may be draining too much current from the battery pack when its voltage drops below 23*3.6 = 83V which may cause the battery pack to shut down or blow a fuse. If the system shuts down at 83V you still have half the battery capacity left, which will be unusable.

 

[Solarsail]

My suggestion is that instead of building one huge 26s30p pack, which will weigh about 43 kg (95 lbs), to build four smaller modules, each 13s15p. So you get two modules and tie them in parallel, and then the other two in parallel, and then tie these pairs in series to get the 26s30p or 96V 100Ah.

There are several advantages to smaller modules. The weight is now down to 11 kg, and the size is a quarter, and you can distribute the four modules throughout the vehicle. The charger can also be 1/4 the rating of the full blown charger with half the voltage and half the current rating, and that will be a lot easier to find. You can now use 4 chargers if you are in a rush, and if you have the power supply to charge all at once.

Furthermore, it would be a lot easier to find a balancer and protection board for 13s15p compared to a 26s30p. All e-bike power packs have a balancer and a protection board. I really don't know how you want to get away without them.

I am building a 13s10p module, and have started a thread for that here. You may want to take a look at it. Cheers!

 

[Solarsail]

For each of my 13s10p modules I will be using one 900W CCCV charger, such as:

http://www.ebay.com/itm/DC-DC-BST90...894694?hash=item1eb1da54e6:g:fzcAAOSwGjpXTOlP

This can supply about 16A at 54V, so it can charge the module in 2 hours (or 3 hours for a 13s15p module). Cost $20. It needs a 36V or 48V power supply, and there are many to choose from on eBay anywhere from $30. The power supply should be shared between the modules, and its wattage depends on how much power you can get from the plug. If you want lots of charging power, then get four 1000W supplies instead of one 4000W supply.

Then there is one balancer-protection board per module, such as:

http://www.ebay.com/itm/13S-Li-ion-...735711?hash=item361877c25f:g:x-cAAOSwX9FZG-Ki

Note the rating is for 20A which would probably mean that one should not draw more than 17A or 0.5C. Depending on how much power you want to draw from the pack, you need to rate this board to allow for the power. Cost is $10.

 

[Solarsail]

Very good questions Jayson. Happy to say none a problem.

2) Your original layout was 23s30p 3.4Ah cells. That gives you 23*3.6 = 83V x 102 Ah = 8.45 kWh. The range depends on V*Ah and not just on Ah. So when you go to 26s you have more V, and you can reduce the Ah by reducing the number of parallel cells so that the range remains the same, and the number of cells are approximately the same. 8450/(3.6*3.4*26) = 26. So you can have 26 cells in parallel instead of 30 cells in parallel, and you will have the same power and the same range. (Power is volt * amp, and is not Ah). Therefore you can go to 4 modules 13s13p = 676 cells instead of 23s30p = 690 cells. Or you can go to 4x 13s14p = 728 cells.

3) You do not need to reconnect the modules because you will never separate them. They are always connected. Even if you have to occasionally separate and reconnect, it is not a problem because the two modules will have exactly same voltage when you connect in parallel. If they are not equal, just put a 1 ohm resistor between them, and let them equalize for a while before connecting. This said, it is not necessary to disconnect two parallel modules for charging. Thus let the charger charge both of them while in parallel. They will balance out. I also believe you can put two chargers in parallel as long as they are adjusted both to 13*4.2 = 54.6V - check the charger specs. Thus if one 900W charger is too slow, put two 900W chargers in parallel - both must be set at 54.6V exactly.

 

[Solarsail]

1) If 17A is not sufficient, then use a bigger BMS. Below is a link to a 45A balancer. So now your pack can produce 2*45A or 90*3.6*13s= 4200W. I think your title says 3000W. So what you need is two parallel modules to produce 3000/(3.6*13) = 64A, or each module to produce 32A. So you could use a 35A balancer. Note that Ah is unrelated to power (W=V*A) and should not be confused. To get 3000W, you need 64A total or 32A per module.

The BMS is so cheap and easy to install, but it does many important protection jobs that you must use it. It saves your cells from going below 2.8V each, which will partially destroy them, and is a safety issue. If you leave your lights on one night, and the cells go down to 2V or 1V, you have lost $3,000. It saves the cells from going over 4.2V each, which will shorten their lifespan and is dangerous if the charger goes bad and keeps on raising the voltage. The BMS will also balance the cells to make sure they are all equal to 4.2V when fully charged. Also in case of short circuit, it will cut the current. If too much current is requested from the pack by the controller, the BMS will cut the current. And if too much current is forced into the battery during charging, it will cut that too. Finally, the BMS detects the module temperature and cuts off the current if it becomes too hot during a runaway event.

http://www.ebay.com/itm/For-48V-13S...642990?hash=item36166f8bae:g:Gv8AAOSw3ZRY9Utg

 

[Solarsail]

1) So for charging them, do you make a charging port for every module or just do 2 modules per charging port and then do parallel charging? which is better?

Yes, each module can be charged independently. In case you remove one for testing, etc. So each module has a charging port. When you install the module and put it in parallel with the other, then the two charging ports should go in parallel, and you can connect one charger to charge both, or connect two chargers in parallel, one on each port. Or you can have a separate charging port in a convenient location that connects to both charging ports in parallel.

The other module pair that is in series, must have its own port, and you cannot charge two serial modules in parallel (unless you install hefty switches or contactors to rewire for charging, and not forget to switch it). However, you can have one big charger 2x13x4.2V = 109.2V and charge the whole system with a single charger. I still prefer at least two chargers, each 54.6V, in series, one for the upper pair and the other for the lower pair. If you lose a charger in a trip, you still have the other one that can charge any pair.

I will also put an ammeter with LED display on each module, as they can point to problems. Also one LED voltmeter that can measure the module voltage, or using switches measure the voltage for each group of cells. So the switches will allow 14 readings by the voltmeter. One for each group and one for total. Occasionally upon charge and discharge one must check all 13 groups in the module. If there is one that is different, then it probably means one cell has gone bad. You need to replace that cell in the group as not only it reduces total energy storage disproportionally, but could result in future safety issues like a fire.

If you plan to go 26s30p, then I would suggest you split that into 6 modules: 2 series x 3 parallel x 13s10p. The modules become smaller and may be easier to install. Also, if you lose a module (due to a bad cell), you can disconnect two modules, and still have 4 working which gives you 2/3 range. On the other hand if you had 4 modules total, and one goes bad, your range drops to 1/2. Also if you have one or two bad cells in a group, or you are unsure, you will have to open all cells in the group to find which one it is. In a 13s15p module, you will have to open 15 cells. In 13s10p, you open 10 cells.

 

[Solarsail]

Whoa ... I thought by e-bike you were building an electric motorcycle. But it seems you are building an electric bicycle. If so, why so much capacity, and how do you intend to carry the pack?

A 26s30p pack will be about 10 kWh. That is a range of almost 1,000 kilometers! Do you need so much range?

A bicycle travels anywhere from 70 km to 125 km /kWh. A moped goes about 37 km/kWh. A motorcycle about 25 km/kWh.

High capacity 18650 batteries are about 4 kg/kWh. With 20% overhead for the enclosure, cabling, electronics, your pack will weigh 50 kg! Can the bike even carry this? And the size of the pack can be 25 to 30 liters.

Do you really plan to travel 100 kph on a bicycle?

 

[Solarsail]

Those are very good batteries. They are much cheaper in China (about $3 a piece) but add $1.1 for air freight. See my project:

http://www.diyelectriccar.com/forums/showthread.php/18650-13s10p-project-48v-x-34ah-188618.html

 

[Solarsail]

At 120 kph, the unit range of km/kWh will drop drastically due to aerodynamic drag, and motor inefficiency, and friction etc. Even if it drops to 30 km/kWh, you are still at nominal:

96V * 3.5Ah * 30p * 30(km/kWh) / 1000 = 300 km range.

But you can't drive the whole way at that speed. Let's say you drive 1/3 at 20 kph, 1/3 at 60 kph, and 1/3 at 100 kph. The respective ranges are let's say 80, 65, 30 km/kWh for an average of 58 km/kWh.

So your range will be 585 km - still too high. I would suggest a range of 150 km, to make the pack as light and small as possible. So the p becomes:

150 * 1000 / (96 * 3.5 * 58) = 8 in parallel. Since the 3.5 is really a 3.3, and the 96 is really a 93.6, and in practice the 58 becomes a 50, and add a 10% margin, you need:

1.1 * 150km * 1000 / (26 * 3.6V * 3.3Ah * 50km/kWh) = 11 cells in parallel

I.e. two modules of 13s11p. You are down from 50 kg (110 lbs) to 17 kg. Much more manageable.

Note that the 150 km range includes high speed runs. If you were to travel 40 kph uniformly, your range will be about 225 km.

 

[Solarsail]

4200W is an awful lot of power. In steady state riding on level ground at 100 kph, you will never use this kind of power. The 4200 peak power may be reached during abrupt acceleration or steep hill climbing. An automobile running at that power will probably do 45 - 50 kph.

When the manufacturer says 65A peak, it means that is what the controller and motor can handle. It does not mean it will handle it at 96V. As the power requirement approaches 4200W maximum power, then add the motor and controller inefficiencies (10% at such high power and 10% respectively), and you need 5,100W. But what if your battery are almost empty at 3.0V each? Then the amount of current you must draw becomes

5,100W / (26 * 3.0V) = 65A. So your controller and motor can handle that at a lower voltage. But when hill climbing on full batteries, it is using only 5100/(26*4) = 49A.

It does not mean that at 120 kph you will need all the 4200W. Just ask the seller what the power requirement is for 100 kph on level ground no acceleration. I have assumed 30 km/kWh at this speed. So output power is 100km/h / 30km/kWh = 3,300W or 1.21*3300/(26*3.6) = 43A input current. Even this is quite conservative, and I think it will be much lower, depending on aerodynamics. Normal bicyclist output on level ground at let's say 30 kph is 100W. So using the square law, it should be (100/30)^2 * 100 = 1100W output power at 100 kph, or 1,350W input.

Maybe the motor or controller is not very efficient. Which means at those powers, it will be generating so much heat that it will probably burn up.

 

[Solarsail]

I am about to write up the charger section for my project which I gave you the link. Keep on checking on it, till I make the post.

You need 1) power supply, 2) charger, 3) PMC (protection-balance). Also optionally switches, fuses, ammeter, voltmeter, plugs and jacks.

I use a 800W 48Vx16.7A supply, for a module. This will be set to 46 to 51V.

http://www.ebay.com/itm/110V-220V-t...hash=item212b6dd93c:m:mkXyMOfLDgX-IuNj3le0sCw

Charger I use is 1200W, 20A, but will be running at 800W. It is CCCV and is isolated. This will be set to 54.0V exactly (4.15V per cell for longevity). Charging will be set at 1.5A per cell. May increase it later. Li-ion chargers must be CCCV - no exception.

http://www.ebay.com/itm/1200W-20A-D...968318?hash=item5d6a280d7e:g:0tUAAOSwn-tZLpgy

PMC I have already posted link.

The charger module you linked is not suitable because it is not CCCV. And its power is no good. Doesn't even say what is the power. My guess is it is not more than 4A at 54V, which means it will take 12 hours to charge one module. It is meant for Lead Acid. I already posted a link to a 900W CCCV charger with digital display.

Yes, AC goes to power supply, which then goes to the charger, which then goes to the - of group 1 and + of group 13. - of group 1 should obviously NOT be connected to chassis. The - of group 1 also goes to the PMC which can cut it off, and from PMC comes out of the module to be the module - (not to be grounded to chassis). This is then connected to the - of the parallel module. Then if the pair is the upper half, the minus goes to the + of the lower modules. If it is the lower half, then it goes to the controller.

Note: each cell should also have its own little fuse (a thin wire). I believe e-bike pack makers omit this.

 

[Solarsail]

get a charger that outputs somewhere around 120V, watch pack voltage high/low and balance charge it. Good to go!

Thanks, but the link is to a LiFePO4 BMS. Will not work with Lico.