Trying to understand this LiFePO4 BMS system

[jerryk]

 I have battery (LiFePO4 4S2P) that failed so I cut it open to see if I could understand the cause of the failure and learn about the BMS system.  I found that both of cell #3 (BT1-3, BT2-3) in the two packs were at zero volts.  All other cells in the two packs were at 3.3V and balanced.  The mosfet (Q7) that balances both cells #3 had some surface swelling and a small discharge on top of the chip.  Using my DCA75 it tests fine.  I do suspect that this component was either the cause of the failure or it's visual deformity the result of it.

I have attached a schematic I drew of the Battery and it's BMS and am curious if there are any guesses at to why just cell #3 in both packs discharged to 0V.

Also I'm trying to understand how this BMS does all the things that the manufacture claims.

Exclusive Features:

–Over discharge protection (#1 killer of all batteries and other brands do not protect against this and void the warranty)

–Over charge protection (Other brands do not protect against this and void the warranty)

–Built in cell balancing technology (Other brands must use a special balancing charger in order to level the charge of each cell. Your vehicle is what charges the battery most of the time, if you do not have built in cell balancing technology, how is your battery balanced while using it?)

– Short circuit protection up to 1,000 amps (Other brands void the warranty if you short the battery)

–Excessive cranking protection  (Lithium technology is different than a lead acid battery and needs to be protected from heat)


In general terms I can see cell balancing in the schematic but am confused as to exactly how this BMS does all these things claimed.  I have also included a photo of the pnp transistor and the zener diode marking if anyone has an idea of their specific identification.

Any thoughts or references that will help my understanding this BMS is very appreciated.

Jerry

[Jim-0000]

I am unable to help you with Battery Management Systems. I would very much suspect that you LiFe cell failure was a result of the BMS malfunction; despite the reason for it's inclusion in the device you are powering was to protect them.
However, I have had lots of experience with LiFe chemistry batteries, particularly when used in model aircraft.
My observations might be offtrack for you, but I will list them.

In the model aircraft application, common management practices include the following:

1. Charge them only with a charge designed for LiFe. (The charger can also charge other types, but it must be set on the LiFe setting).
2. Balance charge at 1C or less.
3. Limit discharge to a minimum, around 40% capacity remaining.
4. Discharge rate to suit the cells, usually around 5C or so Max.

If at all possible (it might not be!), I would discard the BMS and buy or make a battery charge level indicator and a good smart charger.

LiFe cells are considered to be the Lithium chemistry cells that have the greatest safety and lowest maintenance requirements.

If this information is applicable, I can link you to threads on rcgroups that explain and discuss this far more extensively that I have, or can.

Jim.

[jerryk]

Thanks for the reply.  I'm desperate to learn anything about this stuff and every bit helps.  The battery is marketed as a drop in replacement Snowmobile and Motorcycle battery.  It's subject to the charging systems of these types of vehicles.  It has also been adopted by the light experimental aircraft market and is used to start engines up to 180 hp or so with a 14.2V regulated charging scheme. 

I was real surprised to see this BMS circuit given the claims listed in my first post.  It either shows how challenged I am at understanding this stuff or the claims are just not true.

Take the over discharge protection claim.  I thought in an ODP scheme there would be a way disconnect the cells from the load.  I can't for life of me see that in this circuit.

Jerry

[floobydust]

That BMS board is a piece of junk. It does not do all the claims, many parts are not populated and instead jumpered out. Toss it in the garbage.
It is four identical shunt regulators. Each will activate when cell overvoltage occurs and shunt the cell's excess charging current into heat. It can't take much heat before the mosfet cooks and dies.
So it can keep the cells equalized, in a sloppy way as it does not have precise voltage control and the BMS turns any overcharging into a few watts of heat.

Something seems to be special about cell #3, I think it's because that mosfet has worst cooling due to other ones being right next to it. Semiconductors do not develop "surface swelling" and still work, so I'm not sure what you are seeing.

I would get a replacement LiPo BMS off eBay, they are much better.

[jerryk]

That BMS board is a piece of junk. It does not do all the claims, many parts are not populated and instead jumpered out. Toss it in the garbage.
It is four identical shunt regulators. Each will activate when cell overvoltage occurs and shunt the cell's excess charging current into heat. It can't take much heat before the mosfet cooks and dies.
So it can keep the cells equalized, in a sloppy way as it does not have precise voltage control and the BMS turns any overcharging into a few watts of heat.

Something seems to be special about cell #3, I think it's because that mosfet has worst cooling due to other ones being right next to it. Semiconductors do not develop "surface swelling" and still work, so I'm not sure what you are seeing.

I would get a replacement LiPo BMS off eBay, they are much better.

Thanks for that information.  I have attached a photo of the swelling that can be seen under my microscope.  There two small cracks.  One is through the bottom of the first number 1 and the other above the same number.  They are difficult to capture because of their size.  There were two small discharges on the surface at each crack.  Connected to my DCA75 it shows it as an N-Channel Mosfet.  I have not tested it under any load yet.

When you say it does not do all the claims can you confirm my assumption that there is no over discharge protection at all in this board?

I also can't understand where they come up with the short circuit protection and excessive cranking claims.

Is it possible to determine at what voltage the balance circuit start to bleed the excess voltage through the mosfet?  Is it at the zener voltage? The unidentified pnp transistor has an hfe of 199 if that helps.

Again thanks for your help here.

Jerry

[floobydust]

I think it's your usual ripoff BMS, notice there are are extra D2PAK mosfet footprints so those need to be there for over-discharge, over-current protection, as well a control IC. Most of the circuitry is missing to meet all the protection claims. It's just a lowest cost build out of china.
What's there only does some over-charge protection and equalization - nothing else.
The over-discharge protection switch (D2PAK mosfet) is not on the board, so it can do nothing to disconnect the battery to protect it from going flat dead.

LiFePO4 battery min. 2.5V, nominal 3.2V, max. 3.65V

The shunt voltage is well below the zener voltage+the VBE of the transistor, which is not intuitive. The zener rating VZ of 4.7V is at high currents, several mA on the datasheet. This shunt reg runs the zener at low currents, well below the knee on the curve, perhaps at 0.05mA because the PNP transistor amplifies things.

I estimate the shunt at 3.87V at 20mA and 3.93V at 200mA (0.8W each mosfet hot). So it can overcharge the cells- unless the zener is a lower voltage part. You can do a LTSpice simulation, I did not look at how it does when hot or have exact part models.
The SOT-223 mosfet can't get rid of much heat, it's tiny so I said 0.8W or 200mA overcharge at most it can deal with.

The mosfet has a cosmetic defect or blemish but notice the laser etching of the part number was over top. So I would say it's just a lower quality part and did not pop or anything.

I can't see why cell #3 is the one to fail in the two packs. Unless it gets hotter than the rest or they put a weak cell on lucky number 3.

[jerryk]

Floobydust - That information helps me a lot.  The D2PAK foot prints were right there in front of me but I did not recognize them as such.  Now that I look at the unpopulated area in that light there are 8 D2PAk foot prints ( 4 on back of board ) and I would imagine populated on other battery models for the ODP missing here.

I am going to try to salvage 4 of good packs on this battery for a learning project and use another BMS that does what this one claims to do.  The BMS that I'm looking at is of the type mentioned in this article http://www.kerrywong.com/2017/09/10/modifying-a-4s-100a-lifepo4-bms-module/comment-page-1/ .

Like all this stuff, reviews are all over the place but the price is right.  If you have any suggestions on a 4s LiFePO4 BMS that works that would be great but otherwise thanks the information you have provided.

Jerry