A PWM Controller For Each Cell?

[Randomly]

This isn't going to work if you put those circuits in series. Kirchoff's law says that the sum of all currents into any node must equal zero. Another way to put that is current in must equal current out. That means that the battery current on every cell in the chain must be the same. One cell cannot flow less current than the other cells.

 

[safe]

This isn't going to work if you put those circuits in series. Kirchoff's law says that the sum of all currents into any node must equal zero. Another way to put that is current in must equal current out. That means that the battery current on every cell in the chain must be the same. One cell cannot flow less current than the other cells.

Current will be dependent on the weakest cell. This is true. But that weakest cell will be protected by the fact that it will not be asked to deliver as long of a pulse as the others will be required to do.

The weakest cell will deliver a short pulse.

The stronger cell would deliver a longer pulse.

PWM does some pretty crazy stuff allowing more current than you might expect because the inductance downstream allows it to behave like a capacitor. In this case we are actually placing capacitors into the daisy chain so we are adding inductance to the system.

I'm running strictly on intuition so far... so I'm only guessing that this idea and metaphor might work... if I'm really going to discover whether it might work I'll need to make up a computer simulation to see what happens.

Maybe this winter... (it's a good winter time project)

One might ask how this is significantly different than a regular controller combined with a "switched capacitor" balancing system. The goal is to allow the weaker cell to work less and the stronger to work more. What is really different is the fact that it's built with a more modular design and reduces the redundancies of multiple systems layering on top of each other.

In an unbalanced "raw" battery pack the weakest cell is forced to deliver as much as it can all the time and so it tends to empty itself first. If you can somehow force the stronger cells to deliver more of their energy while holding the weaker one back you would get more full usage of the pack. Even if the entire ride is smoothed out to the limit demanded by the weakest cell it still means they all end at the same time. Better to have a long run at a constant (but reduced) rate than a short one that forces the "runt" cell to become exhausted.

It's an idea worth thinking about...

 

[Randomly]

No, really. This will not work. This topology is functionally equivalent to a string of cells with a single PWM FET on the end. Whatever current flows through a stronger cell MUST also flow through a weaker cell. You cannot reduce the current out of a weaker cell without also reducing the current flow in a stronger cell.
To be able to 'throttle' the current independently on each individual cell you would need to have an inductor or transformer on each individual cell to do current transformation and then sum the outputs. Essentially a switching power supply for each cell with the outputs paralleled. That would work, but it would be bulky, expensive, and the efficiency would only be around 85% of what you would get out of just hooking the cells in series. Not to mention the 15% of power going through the system that you would now need to dissipate as heat. This approach can be useful for inaccessible high reliability uses such as satellites, but it's not cost effective and has efficiency, complexity, and volume issues. Here's an example.

http://www.google.com/patents?id=w2UVAAAAEBAJ&pg=PP1&lpg=PP1&dq=patent+6873134&source=web&ots=eAZislP9He&sig=BHYvoLNDVIZDvPo4hjeVkM9GTSw&hl=en&sa=X&oi=book_result&resnum=7&ct=result#PPA1,M1

 

[TylerDurden]

To be able to 'throttle' the current independently on each individual cell you would need to have an inductor or transformer on each individual cell to do current transformation and then sum the outputs. Essentially a switching power supply for each cell with the outputs paralleled.

Unless you can sufficiently warp the current. :lol:

 

[safe]

Essentially a switching power supply for each cell with the outputs paralleled.

Well that's kind of the idea of having PWM on each cell.

Power extracted from a battery has two components. There is the voltage and the current. On the stronger cells the voltage would be slightly higher than the weaker cell so even with the same current throughput that is uniform for all the cells the change in voltage caused by the stronger cell will still deliver more power than the weaker cell.

PWM is the same as a switching power supply and the capacitor is the inductor... so I think while the current is limited by the weakest cell the power does differ coming from the cells. And that's really all we want, for the strong cells to give more than the weak.

Standard controllers use big MOSFET's that have relatively high resistance. Using many smaller MOSFET's with less resistance would achieve the same efficiency.

 

[Ypedal]

Safe.. pls read:

http://endless-sphere.com/forums/viewtopic.php?f=5&t=5007&start=30

 

[safe]

Safe.. pls read:

We're having a civilized discussion of battery theory.

No problems here.

 

[safe]

A Real World Example

My bike currently has a mix of old and new SLA cells. If I had a self balancing "cell level" PWM system I could get the most from the cells, but as it is now I really don't.

Typically the cells with start at:

13.1 (runt)
13.1 (weak, but stronger than the runt)
13.2 (strong)
13.2 (strong)

...now if I go for a ride and come back I'll get:

11.1 (runt)
12.1 (weak, but stronger than the runt)
12.5 (strong)
12.5 (strong)

Now I would guess that the current extracted is based on the controller limit (of course) so all the cells are being asked to deliver more or less equally towards that goal. So all the cells are being asked to deliver 40 amps at first we get approximately:

13.1 * 40 amps = 524 watts (runt)
13.1 * 40 amps = 524 watts (weak, but stronger than the runt)
13.2 * 40 amps = 528 watts (strong)
13.2 * 40 amps = 528 watts (strong)

2104 watts available in the beginning.

...but that ends with:

11.1 * 40 amps = 444 watts (runt)
12.1 * 40 amps = 484 watts (weak, but stronger than the runt)
12.5 * 40 amps = 500 watts (strong)
12.5 * 40 amps = 500 watts (strong)

1928 watts available in the end.

But that's assuming that the weakest cell can even deliver 40 amps which I think is unlikely which means that all the cells are reduced to less than 40 amps. The power falls accordingly.

Using the Inductance of PWM

PWM can take a voltage and a current of say:

48 volts and 30 amps (1440 watts)

...and turn that into:

36 volts and 40 amps (1440 watts)

...that's the way that low end torque is produced in an electric motor. The idea here is to use PWM to convert a lower voltage to give an EQUALIZED amperage to the next cell. So by this logic we FORCE these voltages based on the PWM "duty cycle" from the very beginning:

11.0 * 40 amps = 440 watts (runt)
11.0 * 40 amps = 440 watts (weak, but stronger than the runt)
13.2 * 40 amps = 528 watts (strong)
13.2 * 40 amps = 528 watts (strong)

1936 watts available all the time. (it will vary as things equalize of course, but you get the idea)

So what's happening is the PWM delivers a lower effective voltage while still being able to deliver the current required. The total voltage of the system is lower than it is in the "raw" state, but it's going to be more stable and it will seek to level the strength of the cells as it goes.

Is it making sense yet?

It's the "lowered effective voltage" that is the key here... a cell might be 12 volts but if it's weaker than the others the system might only ask it to deliver 10 volts, but at the amps that are required.

This completely falls apart if "power" and not "current" must be the same across all cells, but that is not the case to my knowledge... only current needs to be constant.

 

[Randomly]

Essentially a switching power supply for each cell with the outputs paralleled.

Well that's kind of the idea of having PWM on each cell.

Power extracted from a battery has two components. There is the voltage and the current. On the stronger cells the voltage would be slightly higher than the weaker cell so even with the same current throughput that is uniform for all the cells the change in voltage caused by the stronger cell will still deliver more power than the weaker cell.

Yes but if the current out of all the cells is the same then it's exactly the same as just hooking the cells in series without the mosfets and capacitors. Although the power out of a cell depends on it's cell voltage, all the currents are the same. You are not able to individually control the cell voltages nor the current out of an individual cell independent of others. Thus you cannot control the power drawn from an individual cell independent of the rest of the cells.

PWM is the same as a switching power supply and the capacitor is the inductor... so I think while the current is limited by the weakest cell the power does differ coming from the cells. And that's really all we want, for the strong cells to give more than the weak.

Unfortunately it is not the same since you cannot achieve the effect you want using a capacitor.

Let me put it a different way. Lets try to simplify your circuit down for clarity. A battery essentially IS a capacitor, so the job of the capacitor in parallel with the Cell will be done by the battery itself. So let's eliminate that capacitor. A FET in this case is just acting like a switch, so let's replace the FETs with switches. The result ends up looking like this.

If any switch is open, no current flows in the circuit. Current only flows when all switches are closed, so you can simplify this down to just the cells connected in series and a single switch.

 

[safe]

Kirchhoff's Current Law

http://en.wikipedia.org/wiki/Kirchhoff's_circuit_laws

At any point in an electrical circuit where charge density is not changing in time, the sum of currents flowing towards that point is equal to the sum of currents flowing away from that point.

This means that the current needs to be constant, but it does not mean that a cell MUST deliver it's full voltage in it's stage. A weak cell might choose to give only 80% of it's voltage but still achieve 100% of the current demand.

This is what makes PWM so cool... 8)

Randomly: Read my "A Real World Example" posting above...

 

[TylerDurden]

What's going to be the inductor?
What happens to current @ 50% duty in cell-A, combined in series with current @ 60% duty in Cell-B?

 

[safe]

You are not able to individually control the cell voltages nor the current out of an individual cell independent of others.

This is where we differ...

PWM in use on electric motors that have sufficient inductance are able to take a higher voltage and deliver a lower voltage while keeping up a higher current.

That's why we have high torque at the low rpms... the current that the PWM delivers is mutated in the bizarre "current multiplication" process.

This is either going to make you go "wow I get it" or "oh jeez he missed some other basic law of physics not yet mentioned". :lol:

 

[Randomly]

You are not able to individually control the cell voltages nor the current out of an individual cell independent of others.

This is where we differ...

PWM in use on electric motors that have sufficient inductance are able to take a higher voltage and deliver a lower voltage while keeping up a higher current.

That's why we have high torque at the low rpms... the current that the PWM delivers is mutated in the bizarre "current multiplication" process.

The key word here is INDUCTANCE.

You need some type of inductor to do the current transformation and you need one for every cell.

 

[safe]

You need some type of inductor to do the current transformation and you need one for every cell.

http://www.csgnetwork.com/capinductancecalc.html

Do you know typical values for capacitors and inductance?

It might come down to picking the right capacitor... with supercapacitors this would be possible, but I'd prefer the least expensive option.

This is one of those calculations that I'm going to need to investigate deeper to make with any confidence.

I'm going to do stuff now... will pick up again later...

I think you at least see the idea now, so that's good.

 

[Randomly]

You are making some mistakes in electrical theory. Look at the schematic I posted of the simplification of your circuit. No current can flow in the battery stack if any single switch is open. There is no way for one cell to have a different PWM duty cycle from another cell since as soon as the first pwm turns off it stops current flow through all the other PWM stages also.

You want to be able to have one cell with a 60% PWM duty cycle, and another with a 50% duty cycle. Where does the current flow out of the 60% PWM when the 50% PWM has turned off? It has no place to go.

 

[monster]

safe, what if instead of using PWM you just have the circuit for each cell wait till it gets to the LVC voltage then disconnects it from the pack and connects the two cells around it together? you don't have losses from PWM and no communication is needed. you get the same end result with less complexity.

well i don't really understand safe's idea but it sounds good

dirty_d's idea i do understand. i have these temperature switches in my NiMH charger to terminate the charge when pack temp>50*C http://www.maplin.co.uk/Search.aspx?criteria=temperature switch&source=15&SD=Y
i was thinking the other day that you could have these in the individual sub packs, powered by the sub pack. but you could take it further than that and have them on individual cells! they would isolate the cell(s) from the battery when the cell temperature exceeds a pre-set limit but allow the battery to continue without breaking the string. as the pack hots-up its voltage would drop as cells go off line.

it would solve cell balancing problems on charge and discharge in large packs and it might extend cell cycle life considerably by preventing hot spots.

 

[monster]

i once saw a schematic of a smart battery system that had diodes in parallel with each cell? i think the diode overflow limit was matched with the cell LVC or something? -so i think there is something in this smart battery idea.

 

[safe]

You are making some mistakes in electrical theory. Look at the schematic I posted of the simplification of your circuit. No current can flow in the battery stack if any single switch is open. There is no way for one cell to have a different PWM duty cycle from another cell since as soon as the first pwm turns off it stops current flow through all the other PWM stages also.

You want to be able to have one cell with a 60% PWM duty cycle, and another with a 50% duty cycle. Where does the current flow out of the 60% PWM when the 50% PWM has turned off? It has no place to go.

A charge pump takes a certain voltage and repeatedly loads it into a capacitor which then increases it like a sort of turbocharger because it raises the base voltage as it cycles. The idea I had was more like a charge pump than a series connected set of cells.

Charge pumps only need to be built strong enough to handle the difference in voltage in each cycle.

...in this scenario rather than cycling the same capacitor again and again you transfer the energy from one capacitor to the next in the daisy chain with each cells battery adding it's own voltage. So rethink this along the charge pump line of thinking.

As for the "where does it go" part... the next cell in the daisy chain had it's own capacitor and it would accept a pulse or two pulses in a row if need be.

 

[safe]

Deeper Better Thought Out Analogy

I had some time while I was working on the bikes to think of an analogy that is better, closer to a reality we can think about. (once you get the concept right then you can think about the technicals to make it possible)

Let's say you first looked at a large scale dam. It has many streams that feed into it, but the only exit is over the spillway and a radical drop in height from it's surface height to the bottom. This most closely resembles our controllers because they take the combined energy resources of the entire battery pack and are forced to chop off little pulses one at a time and deliver them over a big drop. (I know some dams suck from the bottom, but just stay with the spillway idea)

Now, instead of the large dam, think of a series of locks like you might have at a canal like the Panama Canal. Each lock raises or lowers the water a certain height. One lock might raise or lower 50 feet, while another might raise or lower only 25 feet.

If we have a boat that can travel at a constant speed relative to the land (representing the constant current we desire) then as long as each lock takes the same time to fill we will pass through it at the same speed (the time is the same) but the water required to make this happen could be radically different. Filling or releasing a lock that is 25 feet tall means potentially less water than one that is 50 feet tall... but wait... let's introduce PWM logic... we can vary the shape of the lock so that even though the height is double the volume could be the same or even less if we want.

We could make one lock narrow and the other wide!

(we could make one pulse in a PWM system narrow and the other wide)

...so it's possible to guarantee shipping traffic a constant rate of speed through the locks if you design the shape of your locks to match the water resources you have at hand.

Hopefully this analogy will clear up the idea... while the technical hurdles remain I think the idea is something that exists in real life and so having a parallel should help give the "oh I've got it" sensation. (or at least "I understand the analogy that is the basis of the idea" :wink: )

Just think about canal systems... verses dams...

 

[safe]

...i think there is something in this smart battery idea.

Fully implementing the "Smart Battery" is very complex. The whole reason for even thinking about this change in topology is that the history of batteries seems to show one thing layered on top of the next without any attempt at a new perspective. The "Smart Battery" gave us the "Object Oriented" perspective of battery pack design. (all things are viewed from the batteries perspective and everything else serves the battery)

What I'm trying to do is go back and get to the root of how the flows are controlled in the first place. If you could have control of the cells "at the root" then you can possibly get all the benefits of the "Smart Battery" without needing the computerization.

Again I'll repeat the analogy... if you think of the battery pack as more like a canal than a dam... so that each step is addressed by it's own "locks"... then that gets closer to being able to see things in the right way.

I have a feeling it won't be until winter that I actually get to the math and technical stuff on this one.

 

[Randomly]

Deeper Better Thought Out Analogy

I had some time while I was working on the bikes to think of an analogy that is better, closer to a reality we can think about. (once you get the concept right then you can think about the technicals to make it possible)

Let's say you first looked at a large scale dam. It has many streams that feed into it, but the only exit is over the spillway and a radical drop in height from it's surface height to the bottom. This most closely resembles our controllers because they take the combined energy resources of the entire battery pack and are forced to chop off little pulses one at a time and deliver them over a big drop. (I know some dams suck from the bottom, but just stay with the spillway idea)

Now, instead of the large dam, think of a series of locks like you might have at a canal like the Panama Canal. Each lock raises or lowers the water a certain height. One lock might raise or lower 50 feet, while another might raise or lower only 25 feet.

If we have a boat that can travel at a constant speed relative to the land (representing the constant current we desire) then as long as each lock takes the same time to fill we will pass through it at the same speed (the time is the same) but the water required to make this happen could be radically different. Filling or releasing a lock that is 25 feet tall means potentially less water than one that is 50 feet tall... but wait... let's introduce PWM logic... we can vary the shape of the lock so that even though the height is double the volume could be the same or even less if we want.

We could make one lock narrow and the other wide!

(we could make one pulse in a PWM system narrow and the other wide)

...so it's possible to guarantee shipping traffic a constant rate of speed through the locks if you design the shape of your locks to match the water resources you have at hand.

Hopefully this analogy will clear up the idea... while the technical hurdles remain I think the idea is something that exists in real life and so having a parallel should help give the "oh I've got it" sensation. (or at least "I understand the analogy that is the basis of the idea" :wink: )

Just think about canal systems... verses dams...

I understand what you are proposing. You are making the common mistake of drawing an analogy to something else, and then equating the two. Capacitors are energy storage devices like Inductors, but they are not identical and you cannot interchange them. Dams and water flow have some similarities to the circuit you are proposing, but they are not identical. Similar does not equate to Identical. There are some crucial aspects that are different between your circuit and your concept of what it will do.

I understand charge pumps, switched capacitor systems, inductive and transformer couple power systems and I've designed with all of these things many times. You have picked up a good deal of electrical know how but you are missing some points of the crucial basic theories at the bottom of it all. It's the boring drudge theory stuff that's painfully branded on your brain in the first year of engineering college but everybody would rather just skip over. However it's essential to truly understanding how all this stuff works.

At this point I cannot add anything to what I have already pointed out . I can only reiterate that your system will not work as you have described. I know you still think it is viable so I suggest you build a simple 2 cell version of it and see if you can get it to work in any way at all.

If you can get it to work, I will humbly apologize and nominate you for a Nobel prize.

 

[safe]

If you can get it to work, I will humbly apologize and nominate you for a Nobel prize.

Before you completely give up on the idea let's take a step backwards to where I was a few months ago...

The idea of "Distributed Pulse Width Modulation" is not only possible there was even a guy that filed a patent about it. In that idea you have PWM through MOSFETs on each cell but they are all synchronized. Apparently there is some electrical use for doing this other than batteries and so doing things this way both exists and seems to be of value.

I acknowledge that there are philosophical problems with irregular PWM (not synchronized) because it appears to break up the daisy chain of connections, but I'm also of the idea that with a three way dynamic of capacitor / battery / MOSFET that such a system might be viable.

I'll keep trying to post a diagram that clarifies what is not yet clear in my own mind....

 

[Randomly]

If you can dig up a link or reference to that patent I'd appreciate it, I'm always curious.

The basic concept of individually controlling the power drawn from each cell has some merits. I don't see any way to do that with capacitors effectively other than using a flying capacitor approach but that would be prohibitively complex with a very large number of FET switches and high value capacitors with very high current capacity. System Efficiency would not be too good.

The need for a Cell by Cell pwm approach only arises from the problem of mismatched cells. The pack Amp-hour capacity is determined by the cell with the lowest capacity. I think it far more practical to just use matching cells when assembling the pack and putting up with the minor differences.

The cells will 'age' differently and their capacity will diverge from each other over the lifetime of the pack, but unless there is substantial divergence it's not worth all the complexity of the pwm per cell approach. With high quality cells like the A123 they will degrade in capacity at a very similar rate and stay fairly well matched. It would be much easier to just keep track of individual cell capacity and replace any weak cells in the pack.

Unless you can demonstrate a significant problem of cell capacity diverging from each other substantially over time I don't see a need to make such heroic efforts. If the cell capacities stay matched over time, the cell by cell pwm system will provide no benefit at all.

 

[safe]

Unless you can demonstrate a significant problem of cell capacity diverging from each other substantially over time I don't see a need to make such heroic efforts. If the cell capacities stay matched over time, the cell by cell pwm system will provide no benefit at all.

In a typical system that does not have any balancing in it the "runt" (weakest) cell will always be strained at the end of the ride. Every ride increases the divergence more and more until the "runt" dies and for many people with soldered packs that means the pack dies. All the remaining cells still have plenty of life left in them, but the one "runt" cell dragged the whole pack down prematurely. (ask DocBass about him getting old recycled drill packs and extracting the useful cells out of them... only one cell need fail and the pack gets recycled)

So yes... divergence is not only possible it's expected.

Balancing is the right idea, but it's adding a separate layer on top of the regular PWM system. So in the philosophical sense it's a "hack" done after the main design was completed.

The trick of being able to extract less from the "runts" in terms of voltage and converting that to a higher current afterwards through inductance (PWM style) would allow the weaker cells to be balanced from the start.

Well, I'm going to keep it in the back of my mind and keep thinking about how to do it. I like having things that I push deep into the back of my mind because they seem to get processed in my sleep or while daydreaming. Most of my old software projects were "solved" while sleeping (seriously) and the coding part is all I did at work.

 

[ZapPat]

I think randomly is quite right, and you should google for some good pdf's on PWM basics. I know that it's easy to get carried away by a seemingly great idea, but it's gotta be practical too.

I see your idea as actually being a chain of totaly seperate PWM circuits, one for each cell. This necessarily implies seperate PWM controllers, FET driver(s), FETs, inductors and capacitors for each cell. As others have noted, this would be most likely be less efficient than regular battery PWM, but mostly it would be VERY very complex in implementation and fabrication.

Good luck anyhow if you try it... just read more basic theory to validate your idea!

Pat