Lithium cell tests

[terrorhurtz]

Well we ran them for the first time at a competition at Guildford. They performed very well, with no problems. They were barely warm after the fights. However, we only had about three fight, and two of those were against a slot machine...

We will be running them again at Exeter, so that should be a better test.

John

[terrorhurtz]

Again, no problems at Exeter. Worked well.

The only question now is how to balance them easily and well. I suspect that some of the commercial balancers out there dont do a particularly good job of it.

At the moment I am balancing them manually, with varying degrees of success, using a 16 channel analogue-to digital card in a PC to view the cell voltages.

John

[typhoon_driver]

i thought that one of the main pluses about these cells was that they balanced themselves in a similar way to nicads?

[terrorhurtz]

No, this is one of the differences between these and NiCads - like LiPos, they need to be balanced.

If building packs, then balancing wires need to be built in.

If a cell in a pack is repeatedly over-charged then it will be damaged. There may be some small self-balancing effect (the over-charged cells will get rid of the excess energy as heat). This should not be relied upon.

Another difference from NiCads is that M1 Lithiums do not like getting hot. Damage will start to occur if the cells get over 60 deg C. Fortunately them seem to stay cooler than NiCads under discharge.

The standard DeWalt drill packs have connections to the end of each cell built in and the DeWalt charger balances the cells at the end of each charge (if it deems that they are out of balance).

(Message edited by terrorhurtz on July 14, 2007)

(Message edited by terrorhurtz on July 14, 2007)

[mute]

From what Ive read from the about the M1 cells they only need balancing every 10 or so cycles. No sure how this pans out in the real world though.

Ive been looking at balancers that can be slaved together via a data link to work with upto a 12 cell pack. Ive no idea how good they are but from memory the data sheets say they balance within 0.01v... may have been 0.05v though? I Might be a bit off though cos all my info is at the office.

What is the imbalance tolernace like on the M1 cells?

[terrorhurtz]

I am still experimenting with ways to balance these cells, but already its clear that they can only be effectively balanced right at the end of charge, or maybe even after the first cell has peaked.

See http://www.terrorhurtz.com/a123/images/DeWaltBalance2.pnghttp://www.terrorhurtz.com/a123/imag...ltBalance2.png

If this pack had been balanced at half-charge, it may have seemed that the pale blue cell was high. But in the last few minutes of charge, it turns out that it is low.

This graph is of the DeWalt charger doing its thing. Notice that all the balancing is done after the first cell has peaked. The first cell is used as a reference to which the others are then matched, which is rather more sophisticated than the other balancers that I have seen. The balancing is done at around 0.15 Amps.

Regarding voltage difference between cells - at peak I would say 0.05V is sufficient. But just before peak, the voltage of the cells is changing rapidly, so achieving less that 0.05V difference at that point is not straightforward.

This graph is a rather extreme example - they would not normally be that far out.

(Message edited by terrorhurtz on July 15, 2007)

[woody]

John,
I know very little about electrickery ... but ... and this may well be a daft idea ...would it be posible to use a diode ,zener and resistor combo across each cell to attain balance?

[terrorhurtz]

Hi Woody,

It seems that none of the easy/obvious balancing schemes work, or at least not well.

e.g. effectively getting the high cells to charge the low ones - the lower ones need to be charged to a higher voltage initially and then they drop back.

or charging all the cells individually to the same voltage doesnt seem to work either - they seem to remember that they didnt get there in the same time, even after resting for an hour.

[leorcc]

Looking for a balancer solution
This is interesting:

quote:

. High temperature performance

It is detrimental to have a LiCoO2 battery working at elevated temperature, such as 60C.. However, a LiFePO4 battery runs better at elevated temperature, offering 10% more capacity, due to higher lithium ionic conductivity.

Source: http://mysite.verizon.net/adaptron/batteries/Lithium/Iron/Phosphate/index.htmhttp://mysite.verizon.net/adaptron/b...hate/index.htm

Point number 4 is interesting too.

[big_nipper]

A123s website is not very informative either but I guess the ones in the dewalts are the ANR26650 cells. The only thing I can find on how to charge them says you need to charge using a Constant current, then constant voltage upto 3.6 Volts per cell starting at 3 Amps, but can be fast charged at 10 Amps. Interesting to note they also said that the low voltage cuttoff is 2V per cell at 25 degrees C and 0.5V below freezing, but this is not what the dewalt chargers do after looking at Johns graphs.

I was just wondering if you charged them like this if you would still get the 1000 cycles life they claim. If so I think it could be done relativly easy but you would really need one charger per cell, maybe a balancer could be used and one charger per pack but we wont really know until someone tries it a 1000 times without any problems.