LiFePO4... Why Use A BMS?

[safe]

...optocoupler

http://en.wikipedia.org/wiki/Opto-isolator

That's a complete topic unto itself.

 

[safe]

The cell 9(in light purple) is weak and also, the best cell in this pack is the cell 3 (in green)

I do appreciate the graphs and the validation that cells are not all alike and some behave badly. That seems like something that is pretty easy to understand.

Does the fact that the light purple cell goes higher mean that it reaches it's voltage maximum earlier than the rest and what we are really seeing is the effect of forced charging?

Is all that extra voltage damaging the already crippled cell?

 

[safe]

Sacrificial Cell?

When you discharge your pack one cell inevitably "takes the fall" and absorbs the damage. That same cell takes another hit when you charge the pack because it's lowered capacity means it absorbs a higher charge voltage for longer. One cell inevitably becomes the "Sacrificial Cell".

Couldn't you simply watch this as it happens and rotate out the "Sacrificial Cell" after a time (since it's always going to be the same one) with one spare that is of lessor damage?

"Balancing" might be like a "tune up" where after every hundred cycles you rotate out the weakest link. Sort of like rotating your tires.

Why not do this and avoid all the fancy electronics?

(this means that you always need to own at least one extra cell in reserve)

A Little Like A Welfare Program?

The idea is that some cell is always going to be failing while others might be highly productive members of the battery society. If you remove the non-productive cell and give it some rest (like living for free on welfare) then the hard working cells will continue to produce until they reach their retirement age and start to drop off. At some point the relatively unused cell that began life weak can be given a job again because the big performers are now getting older and more tired out. The overall battery GDP goes into decline, but it's a balanced decline.

 

[EMF]

I guess you don't "need" a BMS. I think if you put together your own pack, know how to use a cycle analyst, then get to know your pack, check the cells now and again, balance them once in a while, your ok. In this case, you also don't give a hoot about warranty. Plus, your probably more of a technically minded person, and enjoy fooling around on your own with things.

On the other hand, if you are not into futzing with things and don't like manually keeping tabs on your pack, you don't want to worry about such issues as identifying an unhealthy cell, or troubleshooting a pack, you would be a good candidate for a BMS. OR If you are tyring to market a product to the general masses, that's simple and easy to use (not scary to the uninitiated) and want to offer a warranty, minimise damage/trouble - you should supply a BMS. This way, you can minimise problems for your customer. In this case you just want the end user to see the pack as transparent as possible and just want them to ride and enjoy the purchase. Plus, minimise returns!

I think Gary and his friends/associates are on the right track, with trying to invent a system that is just simple and reliable for the end user. Plus, I bet they are having a lot of fun, designing and re-designing their components!

Frankly, I think the BMS and assorted tools that can help keep a pack healthy will be a big plus for the e-bike and personal transportation category. Not to mention, help many folks get a lot more utility out of their money spent. Anything to take complication out of the mix, will help to lure more people to electric.

Nothing like plugging in and then simply knowing your pack is optimised and protected! Then, just pull the plug and go for a long ride! Then let's hope these folks tell their friends how easy it is to have an electric bike.

 

[safe]

Frankly, I think the BMS and assorted tools that can help keep a pack healthy will be a big plus for the e-bike and personal transportation category. Not to mention, help many folks get a lot more utility out of their money spent. Anything to take complication out of the mix, will help to lure more people to electric.

One day we will not even open up the batteries because they will be sealed at the factory. All this balancing stuff is just to cover the gap between now and where things are going to end up.

I guess my point is that for the economy minded person who wants to buy the most in the way of battery without getting caught up in the whole "pre-built pack" purchase price the ability to know how to make things work without a BMS seems to be an important type of knowledge to have... for now.

LiFePO4 seems to have eliminated the NECESSITY of having a BMS, but it hasn't eliminated the benefits of longer pack life that balancing can give. Self balancing seems possible, though more work.

Five years from now this will all be trivia anyway... :lol:

I know that I'll be doing my best when I get LiFePO4 to simply never run the battery all the way to empty. If you can prevent that first cell from dropping then you can prevent it's damage. So the first thing a user can do is monitor their range and be sure that once they know their range that they keep their rides below the limit. That can involve choosing one bike route over another. I got really good at knowing my range on my old bike and learned not to go too far.

 

[EMF]

Well, you wanted to know why a BMS is ever used. The answer is not everyone is interested in the technical aspects of not ruining a pack - they want automation and care-free use with a warranty, to them, this IS the best value for their money. There's your answer.

If we were all the same, we would all be chasing the same girl.

 

[GGoodrum]

Haven't we agreed that charging behaves like Lead Acid in that when a cell is full it simply becomes inert and allows other cells to absorb more charge?

This is completely false. See below...

There is no need for charge balancing... the only critical thing to worry about is when a cell drops below it's low voltage level while the rest of the pack is still going strong. It's this first cell that drops off that will tend to age faster. So discharge balancing matters while charge balancing does not.

I don't even know what this means. What matters is making sure each cell can get as full as possible when charged, and that no one cell will be allowed to be discharged to too low a level. If you only do those two things, your cells will last the longest amount of time possible.

You guys need to learn how to "teach" better.

Slow down the ideas... try to think as if you are a beginner and you are making the path from no knowledge to the final conclusion. Thrown in some fun along the way to make it easier for a novice to enjoy the journey. [/color]

Okay, fair enough...

What SLA and Lithium-based chemistries have in common are the method by which they are recharged. Looking at a single cell, what it first needs is simply a supply that will limit the current to something the cell can handle. With SLA cells, this is typically only about 1/4C, so for a 12Ah battery, that is a max of about 3A. The LiFeBatt cells can handle up to 3C, so we could hit them with as much as 30A, but that’s really not that practical. Anyway, with a “constant current” (CC) supply, the voltage is allowed to float. What this actually means is the cell controls the voltage. As the cell takes in the current, the voltage gradually rises. This voltage rise happens because as the cell becomes “fuller”, it is harder for it to accept so much current, so the voltage has to rise. Voltage is like water pressure, and current is like the amount, or volume of water. Picture a big water tank that is raised off the ground, with a pipe coming out the bottom of the tank, which is connected to a pump. When the tank is empty, it doesn’t take much water pressure to start pumping water up the pipe, into the tank. Once the tank becomes fuller, though, there is back pressure from the water wanting to go back down the pipe. To counteract this back pressure, the pump needs to supply greater pressure in order to keep pumping water into the tank. The same thing happens with charging a cell. It takes greater voltage (or pressure…) to keep pumping electrons into the cell at the same rate. When the cell gets close to being full (about 85%...), the voltage starts to rise at a much quicker rate. The voltage for a LiFePO4 cell when this rapid rise in rate starts is 3.65V. If you are charging a cell at a 1C rate, it will take an empty cell about 45-50 minutes for the voltage to reach this point. After that, left unchecked, the voltage will rise quite quickly to over 4V, in as little as 30-60 seconds. Although more tolerant of these over-voltage conditions than Lithium-Cobalt – based cells, which tend to blow up when this happens, it is still not good for the cell, long-term. A123Systems says that if you did this on a regular basis (i.e. – letting a cell charge to 4V, or over…), it will cut the cell life expectancy in half, maybe more.

If you just stop charging when the cell hits 3.65V, however, it will only be about 85% full. The trick to get the last 15% is to not let the voltage rise any higher than 3.65V, and let the current start to drop. The cell will control this, and if you watch, you can see the current slowly reduce. When it gets down to about 200 mA, or .2A, the cell is about as full as it can get. So, the supply not only needs a CC mode, to limit the current to some max value, it also needs a “constant voltage” (CV) mode to cap the voltage to 3.65V. With a 12-cell pack, an “optimum” bulk charger would have a CC/CV profile with a “crossover”, or voltage limit of 43.8V. SLA cells work the same way, except their optimum voltage limit is around 2.40V per cell, which for an 18-cell 36V setup would be 43.2V. In a perfect world, where all the cells start out, and remain at exactly the same voltage levels and state of charge (SOC), this “bulk” charging method can be used, where you simply have one big CC/CV charger with a crossover point that is a multiple of the number of cells. In reality, however, this is just not the case. Cells discharge at different rates, due to differences in internal resistance, temperature variances and other factors. They also will differ in how fast they will charge. Over time, these differences get larger.

What makes SLA cells different is that during charging, they can accept more current at the crossover voltage, which is absorbed chemically. This unique feature means that a simple bulk charger can be used, and the cells that either started out at a higher voltage, or just charge faster, can simply “wait” for the rest of the cells to catch up. This way the SLA cells will self-balance, and each will get a full charge. Lithium-based cells don’t have this extra current “absorption” capability, so if the voltage is held at a cutoff voltage, the current will keep dropping all the way down to zero. This is a problem for cells that are connected in series because it means the slower-to-get-a-full-charge cells won’t be allowed to finish. This is because all the current that a charger supplies always goes through all the cells. If you have one cell that “finishes early”, and cuts the current down to nothing, that means the rest of the cells get nothing too. This results in cells that end up with less than a 100% charge. This out-of-balance condition will just get worse over time, if left unchecked. This is how cells get damaged.

So, to make sure each cell gets as full as it can get, in whatever time it needs, you either need to use individual cell chargers, or you need to use a bulk supply with individual shunt voltage regulators that will still let each cell gradually taper off the current it accepts, to get the "last 15%" into the cell, but bypasses the rest of the maximum available current it doesn't use, on to the next cell in series. The net result for either approach is the same, each cell gets to its own max charge level, and does it at its own pace.

-- Gary

 

[safe]

This way the SLA cells will self-balance, and each will get a full charge. Lithium-based cells don’t have this extra current “absorption” capability, so if the voltage is held at a cutoff voltage, the current will keep dropping all the way down to zero. This is a problem for cells that are connected in series because it means the slower-to-get-a-full-charge cells won’t be allowed to finish. This is because all the current that a charger supplies always goes through all the cells. If you have one cell that “finishes early”, and cuts the current down to nothing, that means the rest of the cells get nothing too. This results in cells that end up with less than a 100% charge. This out-of-balance condition will just get worse over time, if left unchecked. This is how cells get damaged.

There seems to be some contradiction when it comes to issues about cycle life. If any cell is given too much charge voltage it increases it's aging rate. So because of this the charger should back off so that the voltage never rises too far and this will drop the current towards the end of the recharging process. When you do this the tradeoff is that the rest of the cells that are stronger tend to get only 85% of their theoretical peak charge.

Wouldn't the cells that are only taken to 85% age very slowly?

Seems to me the "rich get richer and the poor get poorer" when it comes to cell life and capacity.

A BMS corrects this situation, but manual cell rotation of the weakest cell out of the pack should do the same thing. It might be a good idea to have a couple of extra cells that you rotate in from time to time and that way the collective aging process would be very slow. I know that this seems like a lot of work, but for the big cells there would not be that many of them and having to rotate them every once and a while doesn't seem like a big deal. I appreciate the idea of a BMS in that it does simplify balancing but it's clearly not like the days when Li Ion would literally meltdown on you if you overcharged a cell. (so in that sense a simplified no BMS system with an educated techie working on it would work okay) The LifeBatt type cells have bolts on the ends so all it would require is a quick turn of the wrench and the cell would be off to be checked or replaced.

Question:

What's the status of the BMS market? How easy and cheap could it be to buy a BMS that would work for something like the LifeBatts or is building your own circuit the way to go at the moment?

 

[Jeremy Harris]

Neat PCB Gary, I'd be tempted if it wasn't for the fact that I live a few thousand miles away from you, plus have an idea that needs a different board layout. I've made a few PCBs using the iron-on laser printer resist film, which is sort of OK for a one-off. I think I will go with this option.

My plan is to build in a really simple microcontroller and memory, in addition to the straight LVC system, to monitor the 12 optoisolator outputs individually and keep a log file of which goes low first (if any of them ever do). I will either make provision for downloading the log file, or more likely provide some form of simple readout, perhaps like the diagnostic fault memory found in cars with OBD. This would be pretty straightforward to do and might give early indications of a potential cell problem.

I'll definitely opt for single-cell charging, as this looks pretty easy to do. The longer charge time is unlikely to be a problem, as the bike is only used for commuting, so spends hours sitting unused. Rather than go for a bunch of separate chargers, I think I may go for a stack of charging circuits that do the same thing, but in one box, a bit like the DC-DC converter system that Doctorbass has used.

This is a good forum for knocking around ideas and sharing stuff - the total worldwide benefit is greater than the sum of the individual contributions.

Jeremy

 

[safe]

The SuperBowl is just about to get started but I had an idea...

Couldn't you set up some kind of circuit that simply told you that a cell had discharged below the target lower limit and then an LED light goes on?

That way you would know that "enough is enough" and you know you are really done even though the pack is still giving more energy. You could even create a circuit that simply cut the throttle to zero. This might mean that you only get %85 of your packs energy but so what... this protects it from damage.

From what I know about comparator circuits they use very little energy (they basically block most of the time) so this could be achieved at low cost and without any real power losses.

 

[Jeremy Harris]

This is pretty much exactly what the simple circuit that Bob Mcree came up with does. Instead of lighting an LED, it asserts the controller cut-off, but it would be very easy to make it light an LED as well, just by adding a resistor and LED to the cut-off output. In fact it would be dead easy to add individual LEDs to each optoisolator output and bring them all out on to a remote display, so you could see which cell(s) were shutting down first.

Jeremy

 

[GGoodrum]

Jeremy is absolutely right, this is exactly how this works, but I don't think you need the LED. The reason is that when any opto trips, and tells the controller to cut the throttle, as soon as the load is reduced, the circuit automatically resets itself. Usually, the first time one trips is when you've got the throttle maxed out, like going up a hill. You feel it cutout, then it comes back in a second, or two, and if you still have the throttle maxed out, it cuts out again. If you back off the throttle, you can actually go quite a bit farther. I've tested this and just cutting the load to about half (i.e. -- from a max of around 40A, down to about 20A...), I can go about another two miles before the LVC kicks in again.

Anyway, you do have an indication whem the first cell gets low, so you can take appropriate action.

-- Gary

 

[bobmcree]

i am glad to see so much intelligent discussion of this subject; i have spent quite a bit of time over the last few weeks testing some bms prototypes, alternating between bouts of bronchitis, which is why i have been so quiet, among other things. it seems critically important to protect the cells against being discharged below about 2.1v, with the possibility of early cell failure being the price paid for this kind of abuse. While there is plenty of anecdotal evidence that A123 cells can be badly abused and will sometimes recover, this is not always the case, and if we are really trying to extract the highest possible number of Watt hours out of our batteries, i think that a pack mounted low voltage cutoff (LVC) is necessary. The cells are too expensive to take any chances of an early failure caused by discharging a cell too low.

LVC can be done with a simple microprocessor but seems like a good job for analog electronics like the simple circuit i suggested recently. We are taking advantage of a part designed to reset a laptop computer when the battery voltage sags below 2.1v. Because these parts are mass produced for computers they are very inexpensive and use very little power. The TC54 we are using draws only a microamp from the battery, so the pack mounted LVC does not contribute to reduced shelf life caused by steady low power drain. These comparators are accurate to 1% and employ hysteresis on the rising signal, so when a battery drops to 2.1v under a heavy load and the motor is cut off, the sensor will switch again after the voltage rises about 150 mv, making it possible to "limp" home on reduced power after the initial motor cut-out warns of impending battery failure.

There has been quite a bit of debate on whether charge cycle management is required, and my testing shows good evidence that there can be significant benefit to monitoring the level of each cell during the charge cycle. Many of us have observed that the LiFePo4 cells will charge at a very steady voltage until they reach about 3.65v (depending on manufacturer this varies a bit), then the cell voltage will rise rapidly. In a charger that measures the pack voltage, this rise in voltage in one cell, or several, may be enough to trigger the charger cutoff, resulting in the other cells in the string never achieving a full charge, but much worse; some cells will be significantly overcharged. If we are expecting these cells to deliver thousands of cycles it only makes sense that this overcharging must be limited. These batteries are not like the NiMH we are used to, that can just change some extra charge current into heat. They cannot be "trickle charged".

The truth is we just don't know how much benefit can be achieved by utilizing a charger that monitors each cell and assures a full charge on each one every time. It seems like it is worth doing to me, so i have come up with a couple of approaches. I am not talking about a system that operates during pack discharge to limit current or keep the cells balanced; I think a fuse is adequate protection against excess current, especially in the case of the LifeBatt cells which can easily provide 10C or more where overcurrent is not really an issue. My limited testing has not shown that there would be a great short term advantage on ebikes in having an RC type balancer that keeps the cells balanced during discharge by shuttling power from the higher cells to the lower ones. This type of system requires a complex expensive module that adds weight to the pack and can itself cause more reliability problems than it cures. The balancers typically only operate at 100 ma. so they just cannot resolve much of an imbalance in packs that are often 10Ah or more. The same holds true for chargers that switch into a low current "balancing mode" at the end of the charge cycle and require 12 hours to balance the pack. What good is having a battery you can charge in an hour or two if it takes another 6 hours to balance so it will not provide a shortened cycle life?

Most of us would agree that a separate charger on each cell would be the perfect solution. One of my solutions is exactly this, 12 or 16 chargers each providing an adjustable output (nominal 3.65v) with each output isolated from the others. One problem of course is that it takes a lot more wires to connect the charger to the pack. If the proper multi conductor connector is supplied, this can be a very minor inconvenience. The return in improved battery life is something we will just have to learn. It turns out we can borrow from the PC industry again, and take advantage of the myriad of 3.3v switcher modules around, some of which are adjustable and provide the voltge range and current we need. I put together one system using 66w converters that can charge my 12 cell 10 Ah string in 1/2 hour. The converters were $169 each at digi key, but i paid $10 each on ebay. The supplies even have control inputs that can be used to qualify the charge conditions. This is of course not a good solution for mass production, but for the hobbyist looking to get the most cycles and the fastest charge it can be a great way to go. It does help of course to have the LifeBatt cells that can be charged at a much higher rate than much of the competition.

The approach i have worked out with Gary G that he will probably be producing soon is a system of active shunt regulators that clamp the voltage of each cell in a serial string. This way, when a cell reaches full charge level the shunt will bypass any more charge current. This provides two benefits; it prevents overcharge, and it balances the pack on every charge cycle. There is no longer the problem that an overcharged cell will reach a higher voltage and result in the charger terminating early. If the shunt regulators are set to 3.65v and the charger is set to cut off at 43.8v the charger will cut off when the last shunt starts to conduct, signifying that every cell has been fully charged. The shunt regulators need to be connected to each cell, so if they are not mounted on the pack they will need a heavy multi conductor cable. The pack mounted system of course is more convenient, but there is a limit to the power dissipation in the size heat sink we would want to carry all the time. This will determine whether the charge current needs to be reduced during the balancing phase.

We are working with 8A chargers from LifeBatt now that look very promising. We figure that since our cells can be charged at such high rates there is a benefit to having a charger capable of getting them there. The charge efficiency of these cells is very high, so the possibility of just stopping at Starbucks and plugging in for an hour to get a full recharge is just about here. The first units have the shunt regulator mounted on top of the charger and a heavy cable that mates with a connector on the pack mounted LVC board. The shunts are capable of handling the full 8A charge current, but the charger will reduce the charging current towards the end of the cycle, which will reduce the heating in the shunt regulator module.

I am working with Gary right now to provide a turnkey LifeBatt system, where the user can either watch a meter or just ride until the bike starts cutting out then charge it for an hour or two, and go again....for years.... Gary is also planning on continuing to sell the LVC and will likely soon be adding the charge management system to his product line.

I know there are some people here in the group who have been buying supposedly similar batteries that are significantly lower priced. Some of them may actually be from the early production runs of the LifeBatt factory, but it is starting to sound like many of the ones that are being sold were not made by PTI so are not LifeBatt cells. People who buy these cells are likely not getting the 2 year warranty we offer, or any warranty at all, but i certainly don't blame anyone for trying to save a few bucks. Gary will probably sell you the LVC and CMS (charge management system) no matter whose cells you are using.

 

[safe]

Great Response Bob Mcree

I thought it was interesting that you felt that the low powered BMS solution that is being talked about is really meaningless because it doesn't offer enough cross over voltage to really do anything of value. One solution can be ruled, out at least for the bigger cells.

Your shunt based solution sounds interesting and I'd like to hear more about how that might work.

What differences take place when cells are in Series verses in Parallel when charging with LiFePO4? The idea of using one charger per cell makes sense because it eliminates all the balancing issues, but in the case of wanting to be able to charge at StarBucks you wouldn't want to carry 16 chargers around with you all the time. Is there any benefit to simply using 16 PowerPoles or Deans Connectors and then switching from series to parallel for charging purposes? The "switcher modules" you were talking about sound interesting, but the supply of them at a good price sounds suspicious.

 

[safe]

What I've Learned So Far...

Some websites that discuss LiFePO4 talk about it's charging charactoristics as being "just like Lead Acid", but that appears to be just plain wrong. You cannot simply flood the pack with 3.6v and expect that things will self balance like Lead Acid does (at it's required voltage) without incurring unacceptable wear and tear on the cells that have lowered capacity. If you do this repeatedly you will systematically destroy one "weak cell" and eventually the entire pack will take on the charactor of it's weakest link. Premature "apparent" aging takes place, but other than the single "runt" cell the rest are fat and happy and unaffected. You could swap out the "runt" and put in a new one, but in the cruel world of LiFePO4 the pack will find a new "runt" and begin beating up on it next. The only way to restore "civilization" to the pack is to create a balancing mechanism. Time to call in the battery cops.

Lead Acid is "civilized", but LiFePO4 is cruel and unjust and needs policing.

 

[Ypedal]

The net is full of " marketing Wrong's " .. it's a sad state of affairs when even vendors beleive their own BS..

A BMS is a vip in a battery " Pack " ..

Accidents happen, for example, using my single cell chargers, i hooked up 16 cells this weekend, one of the 16 chargers was slightly out of the power bar and did not provide charge to 1 cell in the pack.

This pack was taken down to 80 % dod before the charge cycle and i would have destroyed it at 40 amps if it had not been for pure luck that i noticed that the 2nd block of cells only provided 15 red lights.. 2 minutes of probing with a volt-meter quickly confirmed that 1 cell out of 16 had not been charged..

IF this pack had a BMS.. my ride would have ended short.. but my pack would have lived to go another day !

Mess up a 30 $ SLA brick, ok, lesson learned.. wreck a 1000 $ battery pack and it's a whole other story.

 

[safe]

Capacitor Based BMS Solution?

Here's a thought...

Since there is so much measuring and limiting required of LiFePO4 in order to get it to work perfectly. And since a capacitor can hold a charge temporarily before it needs to be used. I wonder if maybe you could rethink the way a pack is assembled so that you first "buffer" the cell to a capacitor or an array of capacitors in series. In this one-to-one relationship (cell to capacitor) you place the cell level logic. (which is at a lower voltage and current... easier to manage) Then at the level of the capacitor you actually extract the energy that the bike uses. (which might actually be in series) This way you create a "buffer" between the cells and the outside world. Given that we use PWM anyway this "pulsed" battery would make a lot of sense.

Anyone thought about doing this?

Does it even make any sense? It might be useful in regen too, so it's something that could benefit in more ways than one. (killing two birds with one stone)

 

[TylerDurden]

The function of the BMS is to monitor and control at the cell level.

Adding caps probably doesn't diminish that need... but caps are being used to buffer peak current flow; in and out.

http://endless-sphere.com/forums/viewtopic.php?f=14&t=3224&hilit=

 

[safe]

Unless I'm misunderstanding the way a capacitor works I believe that when you place them in series they build a voltage equal to the combined voltage of each of the capacitors. But on the cell level... where you connect the cell to the individual capacitor in the series... it's still a one-to-one relationship so all the lower voltage and current that a single cell deals with applies. This means that rather than dealing with 48 volts and 40 amps (1920 watts) worth of heat all the time in your design (which is nasty) you could instead deal with 3.3v and 40 amps (132 watts) as your worst case scenario... much easier to handle with lower powered circuit components.

It's sort of the "divide and conquer" idea... divide the pack at the cell level and then force all the high voltage and current onto the capacitors which are better suited to deal with such things. Protect the cells behind a "buffer" and never even connect them in series... only the capacitors ever see a series connection.

We also already know that the PWM controller is pulling from the battery in "pulses" so the oscillating nature of the capacitor "pump" is very much in line with what we want to have happen. (there's no need for an analog battery when the controller is essentially digital) All that matters is that when a "pulse" is desired by the controller that the battery can provide that "pulse", it doesn't really matter that on the other side of the "buffer" things are recharging between "pulses".

 

[Ypedal]

My lord man.... you sure love to overcomplicate things don't you lol..

The amount of power required by our controllers ie: 20 to 40 amps can be delivered with these lithium cells without the need for this extra hassle... it would only be adding more things to the mix that can go wrong and require fixing. :wink:

 

[safe]

The amount of power required by our controllers ie: 20 to 40 amps can be delivered with these lithium cells without the need for this extra hassle...

But trying to handle 1920 watts worth of power in the BMS circuit is a real problem. Adding capacitors makes it easier on the BMS circuitry.

The main thing is to break the series configuration of the pack.

Once you make that break you are all of a sudden free to make the BMS only concern itself with a single cell... no "balancing" is needed anymore because it's a simplified circuit that simply limits the upper and lower permissible voltage levels for the cell.

It's a little like in my old days as a programmer when you jumped from traditional programming to object oriented programming styles. Initially you think it's worse, but after a while it starts to make sense because on the cell or object level things are much easier to figure out. It means that you can add more cells without having to think about whether your BMS techniques are still going to work.

It's a thought anyway... add a "buffer" by using capacitors...

It's really not about the peak amps the battery can collectively pull... that is still something that the LiFePO4 is very good at... it's about a possibly better way to manage the cells. Trying to "group manage" LiFePO4 seems like a bad idea, but giving each cell individual management seems to have benefits. This would likely even increase the lifespan of the already excellent LiFePO4 cells because the pack surging would be moved off to the capacitor level. (actual instantious surge might be 10C, but behind the "buffer" it might only be 5C)

Could you imagine 5000 cycles? (capacitors never wear out)

 

[safe]

I have to add that going by what Bob Mcree has said the options for high quality BMS's in the LiFePO4 area (especially the bigger cell sizes) are very limited. There are few options out there and most people are inventing systems that just do their best to help the system hobble along until something better arrives. This is an area that is currently not well addressed by retail products. (it's making the "keep my eyes on things and swap out the weak link" idea look good )

 

[Ypedal]

The amount of power required by our controllers ie: 20 to 40 amps can be delivered with these lithium cells without the need for this extra hassle...

But trying to handle 1920 watts worth of power in the BMS circuit is a real problem. Adding capacitors makes it easier on the BMS circuitry.

The discharge power does NOT go thru the BMS..

The BMS only monitors cell voltage.. then stops your controller from drawing any power via thin E-brake wires. The components requried are already in the controller.

 

[GGoodrum]

This just doesn't make sense to me. The controllers have no problem handling 40A. I've got one that Bob modified that puts out 85A with zero problems. Why you keep trying to make things so complicated is beyond me.

As I've said, what you really need on the pack, at a minimum, in my opinion, is low voltage protction for each cell. For charging it is best to let each cell charge to its own 100% level, at its own rate. Individual cell chargers will do this, as will a bulk charger with shunt regulators on each cell. On big EV packs, the shunts are in the BMS, colocated with the cells. We decided to make the shunts external, so that we didn't have deal with dissipating a lot of heat in the pack. All it requires to do this is to bring out wires from the cell junctions, which you also need to do if you are using individual cell chargers.

-- Gary

 

[TylerDurden]

Safe: You should stop drinking when the game ends. :lol:

Here's an analogy just for you:

You are at a party...
You are drinking a 12pack, from all twelve cans at the same time.
If you let any one of the cans get empty early, someone will put a cigarette-butt in it...
The BMS is like having your nephew check each can, until one is almost empty. Then he tells you to stop and pay attention.
The charging system is like having your nephew even-out all the cans again by pouring more beer into them until they are full.

No IFs, ANDs or BUTTs.