Hi,

I am hoping to build a balanced regulated LiFePO4 battery pack. I would like to be able to maintain a regulated voltage of at least 3V per cell throughout the discharge cycle. Further, I want to avoid switching elements in the balancing and regulation. As far as possible I would also like to minimise extra current draw due to the solution, though the other criteria mentioned take precedence. I’m something of a novice, so please bear with my questions if they are stupid.


Can anyone recommend a suitable schematic?


If I regulate each cell individually with a VLDO linear regulator will this provide balancing or could it have the opposite effect? I am thinking that one of the cells may have lower internal resistance and thus provide more of the current than the others, causing the regulator equivalent resistance to decrease in order to maintain constant output voltage, thus increasing relative current flow from this cell. This could cause the voltage of this cell to drop more quickly than in others, again causing the equivalent resistance of the regulator to drop (relative to the others), again causing more current to flow from this cell than otherwise. Or is the cell output voltage a linear function of its internal resistance? Or is equivalent resistance the wrong way to think of a linear regulator?

 

The problem is that balancing requires drawing current from the "higher" cells during discharge to prevent the lower cells from being over discharged. That current has to go through a load of some kind and it dissipates a lot of power. This problem has been well known for decades and is one of the reasons all electric vehicles are not commercially feasible and never will be.

Bottom line: every cell in the pack has to be sensed individually for cell voltage and have some kind of variable "load shunt" connected across it. You need a cpu to monitor all the cells and take action based on cell voltage, which of course also takes software. The shunts burn a lot of power so the "balancing" typically has to be done very slowly and the shunt assemblies have to be large enough to dissipate the power safely. I used to work with somebody designing the power units for EVs... who eventually gave up and got a real job. What you are trying to do is not for the novice.


Li cell balancing refs:

http://www.ti.com/lit/an/slyt322/slyt322.pdf

http://www.killerrc.com/forum/viewtopic.php?f=5&t=34

http://www.all-battery.com/tenergybalancecharger-01267.aspx

 

The problem is that balancing requires drawing current from the "higher" cells during discharge to prevent the lower cells from being over discharged. That current has to go through a load of some kind and it dissipates a lot of power. This problem has been well known for decades and is one of the reasons all electric vehicles are not commercially feasible and never will be.

......................

So you are saying the Nissan Leaf all electric car is a figment of someone's imagination?

 

So are you saying it is better to just regulate the cells individually without worrying about balancing and design it to cut out or give a warning or something when anyone battery voltage drops too low?

And does my last paragraph/question of my initial post make sense?

Thanks for the responses.

 

Are you building this for fun and/or an educational challenge? Or, do you need on that behaves differently from the commercially available $25 balancers available for R/C batteries.

 

Are you building this for fun and/or an educational challenge? Or, do you need on that behaves differently from the commercially available $25 balancers available for R/C batteries.

Mostly for fun (i.e. educational challenge). I also want to be able to regulate the voltage, which I am not seeing on the cheap commerical balancers I have looked at, and to keep the battery as a very "clean" voltage source (also mostly for fun/ educational challenge)

 

So you are saying the Nissan Leaf all electric car is a figment of someone's imagination?

Please read what I said:

electric vehicles are not commercially feasible and never will be.

The Leaf is a dud (see below) and without government subsidies, electric cars would not exist at all. They are not commercially feasible. Just ask Tesla: they "sell" a coupe for $110,000 with a range of 150 miles and a gross weight of over 6000 pounds..... 3000 pounds are battery weight.

Electric vehicles are and always have been an ATAMO technology:

And
Then
A
Miracle
Occurs

For 30 years they have been betting the electric car's future on the discovery of a miracle battery that is cheap, super light, safe and powerful. The laws of physics keep saying NO.

The reason EVs exist at all is because idiots that run states like California "mandated" that a specific number of vehicles sold here by a certain date must be zero emissions. So they will sell ridiculously expensive, impractical cars to who ever will buy them.

As Sales Lag, Nissan Cuts Price on Leaf Electric

http://wheels.blogs.nytimes.com/2013/01/14/as-sales-lag-nissan-cuts-price-on-leaf-electric-car/



as for the idiots running this state:

Car companies fight California mandate on zero-emission vehicles


Lobbying groups petition the EPA to reconsider a program that would require automakers to sell a certain number of ZEVs in the state.



http://editorial.autos.msn.com/blogs/autosblogpost.aspx?post=7f309312-ff10-4cd4-ac05-195b33896549

 

So are you saying it is better to just regulate the cells individually without worrying about balancing and design it to cut out or give a warning or something when anyone battery voltage drops too low?

And does my last paragraph/question of my initial post make sense?

Thanks for the responses.

Li cells have to be charged using a CC/CV charge scheme until the cell voltage reaches a specific voltage with very little charging current. The cell maker gives the specs. You can't safely overcharge them so cells in series have to be monitored individually. When a cell reaches full charge, a shunt must start bypassing that cell while the other cells in the string charge to full.

 

Li cells have to be charged using a CC/CV charge scheme until the cell voltage reaches a specific voltage with very little charging current. The cell maker gives the specs. You can't safely overcharge them so cells in series have to be monitored individually. When a cell reaches full charge, a shunt must start bypassing that cell while the other cells in the string charge to full.

So I need to switch cells in/out individually at 3.2V and 3.6V.

If I am regulating the voltage, I do not immediately see what the advantage is of a "load shunt." If I make sure that no cell is ever over or under charged, what is the advantage of draining extra current from the stronger cells? Or does it reduce the current drawn from the weaker cells?

 

Please read what I said:

The Leaf is a dud (see below) and without government subsidies, electric cars would not exist at all. They are not commercially feasible. Just ask Tesla: they "sell" a coupe for $110,000 with a range of 150 miles and a gross weight of over 6000 pounds..... 3000 pounds are battery weight.

..................

I did read what you said:

That current has to go through a load of some kind and it dissipates a lot of power. This problem has been well known for decades and is one of the reasons all electric vehicles are not commercially feasible and never will be.

I was simply responding to the technical point of your post, not to any political or cost aspects of an all electric car.

 

So I need to switch cells in/out individually at 3.2V and 3.6V.

If I am regulating the voltage, I do not immediately see what the advantage is of a "load shunt." If I make sure that no cell is ever over or under charged, what is the advantage of draining extra current from the stronger cells? Or does it reduce the current drawn from the weaker cells?

A load shunt does not drain other cells. It bypasses the full cell. It switches the full cell out of the stack of charging cells. Thus, you are using a switch, thus the LOL on your original idea of charging a bunch of cells without using any switches.

 

I did read what you said:

I was simply responding to the technical point of your post, not to any political or cost aspects of an all electric car.

Power balancing during charging requires the charge cycle to be very long in time because the power dissipated in the shunts at high charge currents is enormous and would cook the battery pack..... that is one of a number of reasons EV's are not feasible for daily private use for any but very short travel distances.... some of the others are that the battery packs simply can not be recharged quickly enough to facilitate travel and on top of that, charging stations do not exist in enough places. If you commute ten miles to work and don't go anywhere else (and don't mind the ridiculous cost) an EV might work. But you would need a second car to do any actual traveling.

Nobody wants a car that goes less than 150 miles then requires at least four hours (probably 3X that actually) to recharge. That means you can only travel short trips and must stay within range of the recharge stations.

The only place I know where some small electric delivery trucks are used required the government to "own them" and set up stations where the battery packs were dropped out and replaced which took about 20 minutes. Actual recharge takes too long so multiple spare battery packs are required which increases the total cost exponentially.

This is a non feasible technology for any type of general usage.

 

So I need to switch cells in/out individually at 3.2V and 3.6V.

If I am regulating the voltage, I do not immediately see what the advantage is of a "load shunt." If I make sure that no cell is ever over or under charged, what is the advantage of draining extra current from the stronger cells? Or does it reduce the current drawn from the weaker cells?

In a series circuit, the current through the entire battery stack MUST be constant by definition. If you are charging at 1A and one cell signals "I'm full" to the controller, it must turn on a shunt that allows the 1A to pass by the cell without going through it..... note that the shunt element is NOT a switch it is a linear device which will have the full battery cell voltage (about 4.2V) across it and the full charging current through it. THAT DISSIPATES POWER which heats the pack. The shunt elements have to be able to handle it. In reality, when doing a "balance charge" the charge current has to be greatly reduced to prevent cooking the battery pack since many shunt elements will be activating as the cells balance and the total power using a high charge rate would melt the cells.

 

Hi,

I am hoping to build a balanced regulated LiFePO4 battery pack.

Here is an IC designed to do this:

http://www.ti.com/ww/en/analog/bq77910_integrated_battery_protection_solution/index.shtml

data sheet:

http://www.ti.com/lit/ds/symlink/bq77910.pdf

 

A load shunt does not drain other cells. It bypasses the full cell. It switches the full cell out of the stack of charging cells. Thus, you are using a switch, thus the LOL on your original idea of charging a bunch of cells without using any switches.

I was intending switching when charging, it was during the discharge cycle that I wanted to avoid the switching (except for low voltage cut out). Hence also my comment about shunting - I thought this was refering to during the discharge cycle. Is the balancing only important in charging? I think I can do the charger side, I was mostly looking at the discharging aspect (including obviously what is necessary to interface with the charging circuits).

 

In a series circuit, the current through the entire battery stack MUST be constant by definition.

Face to hand moment!

If you are charging at 1A and one cell signals "I'm full" to the controller, it must turn on a shunt that allows the 1A to pass by the cell without going through it..... note that the shunt element is NOT a switch it is a linear device which will have the full battery cell voltage (about 4.2V) across it and the full charging current through it. THAT DISSIPATES POWER which heats the pack. The shunt elements have to be able to handle it. In reality, when doing a "balance charge" the charge current has to be greatly reduced to prevent cooking the battery pack since many shunt elements will be activating as the cells balance and the total power using a high charge rate would melt the cells.

I begin to see. Before my previous posts were all focussing on the discharge cycle (though I never really said this), which is why I missunderstood you.

I am reading 2.5A CC (per 2.5 Ah cell) followed by 3.6V CV for a 60 min charge cycle. At what voltage should each cell be switched from constant current to constant voltage charging?

So I will need to be sinking 9W per cell during the charge cycle.

During dischage is there any advantage to regulating the cells individually rather than en masse at the end of the series?

Thanks for all the help

 

Here is an IC designed to do this:

http://www.ti.com/ww/en/analog/bq77910_integrated_battery_protection_solution/index.shtml

data sheet:

http://www.ti.com/lit/ds/symlink/bq77910.pdf

Thanks for the data sheets. I may resort to something like this. I was hoping to implement as much of it myself as possible though.

 

During dischage is there any advantage to regulating the cells individually rather than en masse at the end of the series?

Regulating? I don't think so. Monitoring? Definitely yes.

Any stack of cells will have weaker and stronger cells. You must stop when the weakest cell stops contributing to the energy you need. After that, the other cells will force it into deep discharge, and Lithium cells do not like that!

 

I am reading 2.5A CC (per 2.5 Ah cell) followed by 3.6V CV for a 60 min charge cycle. At what voltage should each cell be switched from constant current to constant voltage charging?

The battery maker defines that. The Sony Li-Ion cells required 4.200V (+/-0.5%) while sanyo had set their final charge voltage at 4.150V. Li cells are very unforgiving about over charge and the final charge voltage of the charger typically has to be at least 0.5% accuracy.

During dischage is there any advantage to regulating the cells individually rather than en masse at the end of the series?

Thanks for all the help

You can't regulate the cell voltage during discharge, you can monitor it and disconnect the load when any of the cells reaches it's end of life voltage.

 

Is the balancing only important in charging?

Both. If some cells are only partly charged (or are weak from age/damage and have lost cell capacity) the whole battery pack's A-hr capacity is reduced to the min value of the weakest cell in the pack. That means under discharge, the run time of the device is reduced.