A Manual BMS Design

[Ypedal]

I think what safe needs, is to understand what " Ballancing " is in the first place.

The whole point of ballancing a Lithium battery " Pack " is to have the option of using the packs capacity safely. ( and i don't mean safely as in BANG. kaboom.. wooosh flames.. tho that is another reason) i mean safely as in not harming the cells potential cycle life.

A pack is considered " Ballanced " when all the cells in the pack are charged to the same SOC *( Sate of charge ) . the voltage displayed on a DVM ( digital Volt Meter ) does not accurately tell you that cells specific soc.. only a rough idea.. and that's not good enough.

If you use one charger per cell for every cell in the pack, then you insure that every cell gets a full 100 % charge.

However. this requries alot of wires and alot of chargers.

Using a BMS, this job can be done without your help or the use of X amount of chargers.

What Bob and Gary have done.. is exactly that..

LVC, each cell is monitored, during discharge, under load, etc.. and as soon as 1 single cell goes below set Voltage.. the WHOLE PACK is stopped by shutting down the controller.

Then, during charge, the circuits will monitor Cell-Level voltage and will not allow any single cell to go above X voltage.

This is a simple, effective, reliable, and CHEAP way to insure long battery life !

Simple.

and safe.

 

[fechter]

Hmmm.... what's wrong with using a separate charger for each cell? If you insist on being manual about things, you could manually switch the charger from one cell to the next instead of buying a charger for each cell.

That way if you get explosive diarrhea while asleep on the couch, it won't ruin your $1000 battery.

The zener diode is a great idea. In fact, that's almost exactly how most pack balancers work. The problem with just a zener diode and lamp is at 3.x volts, zener diodes tend to not behave very ideally and their voltage will change quite a bit with current.

 

[safe]

Hmmmmm...

I used the voltage divider concept to configure a way to simulate a higher battery internal resistance. After all, if you can make the batteries resistance appear to be larger then it will tend to not want to discharge as fast... at least not into useful power. This doesn't look right though... while it's true that the final voltage output could be controlled with the voltage divider I'd think that the resistor would burn up as much energy as if you didn't have it at all (so the net cell drain is not reduced as desired).

When charging this sort of idea might make some sense because you could artifically limit the amount of charge going into the cell.

This seems to circle back to the reason for PWM in the first place... that resistors always suck current and are a waste. It doesn't make much sense to balance your pack and lose anything in the process... :?

Ideally I wanted to use this equalizer concept so that it would work on both the charge and the discharge, but it seems that charging is easier to deal with because you don't care about losses very much. Who cares if the charger works a little harder? But if the battery wastes energy that's a no-no...

 

[safe]

Fundamental Premise Likely Impossible

The fundamental premise was that you could LIMIT the weak cells... I'm not sure if that's possible without resorting to PWM or some other mechanism. Resistors don't work because they waste energy and unless you only focus on the charging side you can't prevent the discharge rate from being large.

What I want to do is:

Reduce the rate that the weak cell absorbs a charge.

Reduce the rate that the weak cell discharges.

...going the "other way" means that you use the energy from the strong cells to prop up the weak ones. That seems to be the only way to go and if that's the case then there's no way to create a manual BMS because you can't know based on the strong who is the weak. You would have to create something like an extra pool of energy that got passed around.

This might be why BMS systems are moving to some really fancy computer technology because there seems to be no good low tech way to pull it off. (at least I'm not seeing it)

This would be just so cool... 8) (but it might be impossible)

 

[Ypedal]

What is your goal exactly ?

man.. i give up.

 

[safe]

What is your goal exactly ?

My goal is to create a manually adjustable battery pack where each cell is controlled with something that would function much like an equalizer. The strong cells would be "full on" and have no limiting. The weak cells would be "partially on" and would tend to discharge more slowly and charge more slowly. The idea would be that you would "equalize" the ability of the cells so that the strong could pull strong and the weak would only be ASKED to pull less.

And that's the problem and why the idea might in fact be impossible to achieve.

The reason they invented PWM was so that you didn't have to use resistors to control the current. All the BMS's allow the cells to run full blast all the time, they rob from the rich and give to the poor and in that way you get psuedo equality. But just like in real life this "theft" doesn't do that great of a job because ideally the poor cells would not be asked to so as much as the rich. Rich people pay lots of taxes... poor people should not have to pay so much taxes.

Anyway... many metaphors later... for the same reasons that PWM was introduced into the EV world we probably can't slow down the weak cells for their benefit.

The answer to this quest might be:

"It's impossible"

 

[rgx]

I'm also not sure what safe is trying to establish, but I can confirm a few of his assumptions and ideas.

1. The internal discharge rate of litium cells is very low, which means that balancing can be done either very slowly, seldomly, or even manually/occasionally. If a pack is balanced (or unbalanced) it will stay that way for a relatively long time - at least several cycles. Balancing by shunting away several amps during charge is not necessarily the only option, it can be done anytime. And it can be done by shunting milliamps, it only takes longer. If done continuously the shunt current only has to be larger than the internal discharge current. (Or actually, larger than the difference in internal discharge between cells...)

2. Yes, it is a clever approach not only to take the end-of-charge voltage into account when balancing. More interesting is the discharge voltage, or perhaps the end-of-discharge.

3. 1. and 2. together means that some kind of monitoring system which remembers the state over several cycles is an advantage. Only balancing during charge (or worse, end-of-charge), without taking into account what has happened previously, is not ideal.

So I guess that a manual approach can work well, but you will need voltage cutoff circuits for protection. You would need to balance the cells every ten cycles or so, which shouldn't be an issue to someone like safe. (The average consumer might complain though.) Only if a cell's internal discharge increases a lot will manual balancing not be enough. This is supposed to happen at the end of cell life, so you might have to replace the cell anyway in that case.

I have some experience from occasional balancing myself, because of a flaw in my litium setup. The way it is wired, balancing only occurs when the charger is connected, and only after charging has finished, and I often unplug the charger directly after or even before complete charge. It can go several cycles between any decent balancing occurs.

 

[safe]

So I guess that a manual approach can work well, but you will need voltage cutoff circuits for protection. You would need to balance the cells every ten cycles or so, which shouldn't be an issue to someone like safe.

The question remains....

How do you limit a weak cell?

It's one thing to take higher power AWAY from the strong cell and give it to the weak. It's something else entirely to limit the weak cell so that it resists doing anything. The idea of manual control would be to know which cell is weak and be able to lower it's setting so that it does less work both on charge and discharge.

You could propose an idea.

Charging seems to be the easier side of this. It's reducing the RATE of discharge that I want. I want the weak cell to take longer to finish itself rather than it finishing early and then getting safely cut off. It's the idea of "proportional usage" that is at the center here... and might be impossible.

 

[Jeremy Harris]

Interesting post, RGX.

So, provided we have a decent LVC system wired up to each cell, to prevent over-discharge, then perhaps we could balance a pack by just having a big resistive divider chain across it, with closely matched resistors. If this were hooked up to a trickle charger, operating at a constant voltage, then the whole pack would (eventually) get in balance.

If this only needs to be done every few charge cycles, then a reasonable fix might be to have a multipole charging connector, that also hooks the resistive divider up to each cell. Leaving the whole pack on a float charge, say every weekend, might keep the pack in balance. Normal overnight charging would be unaffected, although balancing effectiveness would obviously be a lot lower for a short hold at the float constant voltage.

Any thoughts on this as a pragmatic approach?

Jeremy

 

[safe]

So, provided we have a decent LVC system wired up to each cell, to prevent over-discharge, then perhaps we could balance a pack by just having a big resistive divider chain across it, with closely matched resistors.

That's a hybrid. :wink:

I was hoping not to create a design that needed two separate systems, one for charging and another for discharge.

Hopefully everyone has "got" what I imagined as an idea, even if it's impossible to implement it is how things should work. The weak cells should work less all the time, no matter if they are charging or discharging.

Really far out idea... :lol:

What would be cool is if there could be some signal that the battery could take that would ALTER it's internal resistance. That would be a battery chemistry issue, but imagine being able to control the batteries production at it's source. You could get rid of the controller completely. 8)

 

[rgx]

How do you limit a weak cell?

OK, I am slowly beginning to realise the extent of the confusion. I think you are trying to solve a problem that doesn't need a solution. So the simple answer is: don't. In a series connected string of batteries, cells participation in the current flow can not be limited. Some sharing/redistribution of energy can be done using the flying capacitor concept, though. I guess controlling contribution from separate cells can only be done easily in a parallel setup.

Of course it is possible to imagine a complex setup, let's say each cell had its own step-up DC-DC-converter, connected to a common bus bar. Then you could control power from each cell individually. The weight, cost, losses and complexity of it all would be way higher than what you gain, though.

 

[safe]

Like I said... my original idea is impossible... thanks for at least understanding what I had in mind...

In the upside down world of the balancing systems (as they are being done) is there a way to get the same effect using a different mindest so that you can in fact isolate the weak cells?

How could you specifically protect the weak cells and not just try for an average?

 

[rgx]

Interesting post, RGX.

So, provided we have a decent LVC system wired up to each cell, to prevent over-discharge, then perhaps we could balance a pack by just having a big resistive divider chain across it, with closely matched resistors. If this were hooked up to a trickle charger, operating at a constant voltage, then the whole pack would (eventually) get in balance.

If this only needs to be done every few charge cycles, then a reasonable fix might be to have a multipole charging connector, that also hooks the resistive divider up to each cell. Leaving the whole pack on a float charge, say every weekend, might keep the pack in balance. Normal overnight charging would be unaffected, although balancing effectiveness would obviously be a lot lower for a short hold at the float constant voltage.

Any thoughts on this as a pragmatic approach?

Jeremy

Well, yes, it would work and basically this is in a way how simple BMSes are designed but if you read my post again you realize I was trying to argue against balancing during end-of-charge. Or balancing voltage only at a certain state, without taking into account the capacity of each cell.

Litium cells age quickly if you keep them for longer periods at their top voltage (end-of-charge voltage). This is one of the main reasons for avoiding this. The weakest cell (with lowest capacity and/or highest internal resistance) will hit the roof first, and sit at top voltage until all the others catch up. Which means that poor cell will age even quicker. It's not solidarity at its best.

 

[Jeremy Harris]

I think that the bottom line here is that you can probably get away with no form of balancing/LVS system IF you make sure that you never over-discharge the pack AND you are happy to accept a limited useful life.

You may be able to improve useful life by just fitting some form of over-discharge protection. Eventually the pack will get out of balance, but the over-discharge system will cut in to stop that being too damaging to the pack.

The next level of life improvement would be to optimise cell balance. I guess there are many ways of doing this, from individual chargers for each cell, through to something as simple as a big resistor chain connected to each cell and a constant voltage charger.

It may well be that we are worrying too much about absolute cell balance. Maybe it doesn't need to be as super-accurate as we think. Judging from the sort of variability reported here in cell open circuit voltages at end-of-charge, I'm inclined to think that cell balance accuracy only needs to be within a couple of percent or so.

Is it actually true that LiFePO4 cells age prematurely if kept at full charge? I know that LiPo cells do, although keeping them cold seems to limit this.

Jeremy

 

[safe]

Minimal System

Since there's no practical way to balance on the discharge (you can't actually slow down the weak cell) the only place you really have control over is the charging side and the easiest way to do that is to have a charger for each cell.

So the "minimal system" would:

Have an individual charger for each cell.

Some type of low voltage cutoff.

The only thing you need to be sure of is that you have a low voltage cutoff that is realistic. You should probably cut all the cells when the weak one is done. (sad to say ) I hate the idea of all the cells getting cut when it's just one weak one that is holding the rest back, but things only get worse for the weak one if you don't stop.

That's about as good as you are going to get without resorting to some fancy computer systems. Ideally you would want to capture historical data on each cell and adjust your charging to fit that, but short of that you're not going to get much of a difference.

Am I correct about this?

 

[Ypedal]

The " weakest " cell in the pack, should honestly not be that weak.. if it's that bad that it impedes the performance of the whole pack.. then it should be repaired ( the pack ) by replacnig that one cell.

Consider your BMS to be a warning system if you will.

If you get 9.6 reliable ah from a 10ah pack.. i consider it fair, however, if one cell prevents the pack from delivering more than 7ah.... then it's a problem that needs to be corrected.

 

[safe]

If you get 9.6 reliable ah from a 10ah pack.. i consider it fair, however, if one cell prevents the pack from delivering more than 7ah.... then it's a problem that needs to be corrected.

Well then it's back to the argument for rotation. If the cells tend to stay within fairly close parameters then if you have a couple of spares you can rotate the ones that are weaker and wear on the others for a while. When the group creates a new weakest cell you replace it and on and on.

All the while you charge the cells individually and be sure to have a low voltage cutoff that is pretty high.

That ought to cover the most worrisome things...

 

[safe]

Out of the Loop

One idea might be to incorporate the "Design An Individual Cell Voltage Meter" discussed on another thread with the ability to manually disconnect (through the wires that connect the cells together) a cell from the pack. That way when the weak cell drops too low you can take it out of the loop and rely on the other cells for a while. Depending on how far gone the one weak cell is this might be enough to get you home in situations where you've made a mistake and gone too far.

Your controllers low voltage cutoff might not allow this however...

 

[safe]

Further Meditations on a Theme

Once you had figured out which cell was weak you could take advantage of your memory of that fact the next time you went out for a ride and START the ride with the weak cell out of the loop. That way the strong cells can run for a while and the weak one can rest. (this might be thought of as your "warmup laps") After about 10 miles you reconnect the weak cell into the pack. As you got to know your pack better you would learn about the cells strengths and weaknesses.

It would like being a coach... you let the weak kid sit on the bench for the really difficult situations and then let him in when the main players start to get tired out.

For a typical 36 volt controller the cutoff voltage is 30.5 volts which translates into 2.54 volts if you have 12 cells. Remove a cell and the remaining 11 cells will cut out at 2.77 volts. So you could get away with this for a good percentage of the ride since LiFePO4 tends to run pretty flat most of the way through.