DIY “Home Made” Balancer for Li-ion / Li-Mn /Li-po Cells

 

Hi there,

Long time since the last post. I have been working on a lot of projects but I haven’t posted much on the blog.

Here is the project I am working on at the moment. It’s cell balancer for Li-ion / Li-Mn /Li-po cells. Basically it detects if a cell goes above a pre-set voltage during charging and “activates” the mosfet and power resistors in the circuit, draining/bleeding the charging current allowing the cell to stay at or below the pre-set voltage. Depending on the resistor (and mosfet) used it’s possible to set the maximum bleeding/draining current. In the video above the maximum draining currents turned out to be 1.3-1.5 A.

It’s very important for all the lithium chemistries that the single cell don’t go above a certain voltage to avoid damage. In case of Li-ion / Li-Mn /Li-po cells it is recommended not to go over 4.2v per cell. I set my balancer to activate at 4.15v per cell.

Some of you would say: “why don’t you just use a BMS?”. Well, BMS, especially for bicycles and motorbikes have limitations. In particular the charging and discharging current, and most importantly the balancing current, which is normally set to 0.3 A or less. Therefore using a standard BMS can take several hours or days to balance an unbalanced battery pack. The circuit above is designed to speed up the process when bulk charging.

I will have new pcb ready in a few weeks as the one I am currently using has a design error. The current boards are 5s.

Please comment if you have any question or contact me if you are interested in having one of the boards to test it on your setup.

 

Advertisements

E-Bike improvements

IMAG0210

It’s finally warm here in Canada and I have started to use my e-bike to go to work. I introduced some improvements to keep things tidy and easy to use.

Starting with my e-bike computer I used an old printer cable (with many cores) to re-wire the connection between the LCD screen and the main board. I also added a toggle switch that powers off the unit while I am at work or when the bike is not in use, avoiding battery draining.

IMAG0189

 

Everything looks so much tidier now!

I also re-wired the main board with 10AWG cables.

IMAG0211

 

The bike now uses the Hybrid battery (see previous post), 54V nominal and about 9.5Ah (with 7.5Ah of usable capacity to preserve cycle life). I was able to shrink wrap the battery in black PVC tubing, making it safer and more professional looking.The battery is secured to the frame with “backpack” straps.

IMAG0212

 

I am still working to improve the bike.

Next things to do:

  • Use the main battery to power up the e-bike computer (now is powered with an external battery)
  • Introduce a plug for recharging outside the wooden battery frame, weather insulated, and protected ( with a fuse).
  • Reduce the amount of wires.

 

I am also working on a “cell balancer”, I have done a few tests and I should be able to have a pcb ready for more testing next month..I will keep you posted!

 

 

Hybrid battery

IMAG0123

It has been a long winter here in Canada and while my electric bike was parked and unused during those cold months, I started to test laptop cells to “expand” my current battery.

The battery I have been using is a 20s4p, 74v 6Ah nominal made from makita cells. Those cells are great and can be discharged up to 10C, and can be discharged at 3C continuous giving good performance. Laptop cells instead are capable of only 1C continuous maximum.

I don’t need 10C on my electric bike, but I would like more range, so I decided to increase the capacity of the battery using laptop cells.

I tested a lot of laptop cells (all 2.2Ah nominal) using the CBA 4 Battery Analyzer, and selected only the ones that had more than 75% of the original capacity left.

All the cells tested were “coupled” 1s2p:

laptop cells -pass

As you can see from the graph, there is a lot of variance in capacity between cells. To reduce this gap in capacity I separated the cells and “re-coupled” them to obtain a smaller gap in capacity:

Excel-cells

It’s still not perfect but much better than having 15% or more difference in capacity between cells.

Then, after this operation I soldered the laptop cells to the existing battery the battery:

IMAG0121

On the left the battery before the addition of the laptop cells.

Then I did a discharge test on the “new” battery on 3 cells connected in series at 1C (expecting a capacity of 9.6Ah) to confirm the capacity and to see the behaviour of the cells and make sure their gap in capacity is not too wide during discharge.

Hybrid_3S_1C

The capacity is slightly less than expected but the discharging curve is still good. The temperature of the cells at the end of the test was about 35-37°C for the Makitas and 53-56°C for the laptop cells.

Here the voltage log during discharge:

Cell_voltage_3s_1C

The discharge graph denotes a pretty uniform discharge until the very end of the test. This graph suggests that the LVC should be set at 3.0v or higher for the cells.

All the test were performed at an environment temperature of 21-24°C.

Choosing the right battery for your e-bike. The “C” factor.

CBA 4

Because I have been a very good boy, Santa brought me a great present for Xmas, a professional computerized  battery discharger: West Mountain Radio – CBA 4.

I made some test on my cells, and here below there are some graphs that explain how batteries are affected by heavy loads.

These are real tests and not some data taken from a data sheet.

All the test in this page have been performed at an average ambient temperature of 21°C.

In the next two graphs the same battery was discharged a 1C, then recharged and discharged at 2C. These cells are rated for discharge at 1C continuous with occasional 2C short burst.

The battery is a Li-Co 18650 Panasonic (used in laptop batteries) 1S2P 4400mAh.

The vertical red line represent 75% of the nominal capacity. The cut-off was set to 2.8v to avoid any damage to the cells.

Graph Voltage vs Ah

LiCo_1C_vs_2C_Ah

Graph Voltage vs Time

LiCo_1C_vs_2C_Min

In the next two graphs the same battery was discharged a 1C, then recharged and discharged at 3CThese cells can support a 3-4C discharge continuous with occasional 10C short burst.

Those are Sony Konion cells,  LiMn cells, 18650 format 3.7v 1500mAh for each cell.

The cells are connected in parallel, so the battery is a 1s2p 3.7v 3000mAh nominal.

Graph Voltage vs Ah

LiMn_1C_vs_3C_Ah

Graph Voltage vs Time

LiMn_1C_vs_3C_Min

As you can see from the graphs the Sony Konion cells are more suitable for an electric vehicle. For the same size (cell format 18650) they store less energy compared to the laptop cells (1500mAh vs 2200mAh) but can deliver more current.

In other words LiCo laptop cells can still be used for electric bikes but have to be discharged at less than 1C to avoid damage  and shortening dramatically the cycle life of the battery pack.This means building a bigger and heavier pack compared to cells with high discharge rate.

Batteries with a high continuous discharge rate can be used to build small capacity packs that can deliver high current.

It’s always best to choose the battery according to the specifications of the vehicle (Amps continuous and peak) and needs of the user (range for example).

Cell-log LVC “BMS” project.

I decided to make a Battery Managemet System using the cell-log 8s/m.

Those little cell/battery voltmeters have an alarm port that can be triggered at a selected voltage (battery pack voltage or cell voltage). I made a simple circuit exploiting this feature.

In the video you can see my prototype circuit to cut the load using the cell-log alarm port. When the threshold voltage is reached the alarm port closes the circuit and the mosfet opens the circuit.

I used a 12v 50W light bulb as load.

The battery is a 5s4p 18.5v 6Ah li-mn.

The alarm is set to go off if any cell goes under 3.79v. You can see the light bulb flashing as the alarm start going off. This is due to the fact that the voltage of cells goes up once the load is cut.
The alarm(s) can be set to go off to any value within the cell-log specs.

The next stage of the project is to cut the load permanently (no more flashing) as the fist alarm goes off. Then I will add the possibility to connect more cell-logs to manage a bigger battery.

The beauty of this system compared to a standard ebike bms is that  there can be no current limit (depending on how/where the load is “cut”) so it’s suitable for high power setups.

A big “Thank you!” to my friend Adriano for helping me out with the circuit.

Happy new year!! It was a great 2012 for Jacopo.tk!

The WordPress.com stats helper monkeys prepared a 2012 annual report for this blog.

Here’s an excerpt:

600 people reached the top of Mt. Everest in 2012. This blog got about 6,000 views in 2012. If every person who reached the top of Mt. Everest viewed this blog, it would have taken 10 years to get that many views.

Click here to see the complete report.

New PCB is working!! Yay!! :)

Finally I had time to solder the components onto the PCB I ordered from China and got a bit of time to test it. Unfortunately when I designed the circuit I forgot a pull-up resistor and that made things harder because I had to do some debugging as nothing seemed to work at first. But not everything works great!

The values on the LCD are random because the sensors are not connected 🙂

I am looking forward to get some real data once I’ll test the PCB on the e-bike!

Another pic: