"BattEQ" Balancers for LiFePO4 Packs

[PJD]

Some of you may be familiar with SmartSpark energy Systems' "BattEQ" balancers for strings of 12V SLA batteries. They operate similarly to Powercheq's, but come in a single package with simplified numbers of wires and supposedly capable of higher balancing current.

Well, it seems that they also sell BattEQ's configured to balance the cells of LiFePO4 packs. Theoretically, if they work well, this greatly simplify things because they work full-time non-dissipatively balancing the cells, so the charge voltage and low-voltage cutoff limits are simply controlled by the charger and the controller's LVC for the whole pack, respectively.

Four, five, or six-cell balancers are chained together through a common cell to accommodate larger packs.

However, the SmartSpark Rep. did write this:

"After responding to your original email I remembered that our
BattEQ does not perform as well as expected on long (greater than 10)
strings of LiFePO4 batteries. If you are still interested, the prices for the 6 cell 25-75 AHr
units are as follows:
1-9 the price is $219 each
10-29 the price is $160 each
30-99 the price is $135 each
100+ the price is $115 each"

So, I have a request for how they define performance as "not as well as expected."

They are pricey at low quantities, ($617 for my 16-cell pack) but reasonable at large quantities.

So this is possibly another of the very few pack balancing options available right now.

Comments and Ideas?

http://www.smartsparkenergy.com/

 

[safe]

Can the switched-capacitor based BattEQ move sufficient current to be effective?

Yes. BattEQ has already been proven to move sufficient current to equalize very high capacity batteries. SmartSpark has invested heavily in optimizing its patented technique. Any source to the contrary is very unlikely to have properly designed the circuit as to make a valid claim.

http://www.smartsparkenergy.com/products/Frequently_Asked_Questions_(FAQ).aspx

So they have some sort of sophisticated switched-capacitor solution.... nice.

 

[safe]

http://www.smartsparkenergy.com/pdf/batteq1.pdf

This really happens...

My Project #001 bike had three 38 Ah cells and for 3000 miles they worked fine until one cell went completely out... it could not hold a charge. Apparently since I had been only looking at the parallel voltage connected battery as a unit when I was charging I didn't notice all along that one was failing for a long time.

Now I've got two new 18Ah cells to replace the one 38Ah cell and here's where it gets interesting. When I run in 48 volt mode (18, 18, 38, 38) the two 18's drain faster than the 38's even though the 38's are older. So I often end up with a lower voltage level at the end of a 48 volt ride with the 18's. But when I charge I find that since the 18's are newer they can actually hold a higher final charge. The old one's tend to want to settle down into about 12.8 volts if you let them settle down a little, but the new one's will hold at 13.1 for much longer. So I find that if I don't physically disconnect the new one's from the old and separately charge them the last 10% that they would go unfilled because the charger sees the others banging against the high voltage cutoff. (the old one's are full and can't accept any more)

Anyway... this is their chart from their pdf.

You might want to download it because it's got some really good stuff in it. 8)

 

[PJD]

The problem is, that Spartspark themselves admit that the use of their equalizers on more than 10 cells or batteries doesn't perform "as well as expected". I will hopefully get an answer that quantifies the problem. Maybe it will be tolerable and we can use them. Then, I'd like to try to organize a group purchase to get the price down.

 

[PJD]

Safe,

And as far as your SLA's in series, I consider equalizers - powercheq or BattEQ to be almost mandatory. I am currently using powercheqs on my 48/60 volt scooters - But, if I was going to continue with SLA's, I'd get the BattEQ's.

 

[fechter]

The concept is valid.

Since the charge is being moved between adjacent cells, it has a hard time moving charge from one end of the string to the other. The longer the string, the harder it gets to move the charge.

I didn't see where it mentioned how much current it could pass.

$600 for a 16 cell setup is way over my budget.
You could get 16 single cell chargers for much less.

 

[GGoodrum]

I'm still intrigued by this approach, and have decided to do a test balancer, using the LM2663 switched cap chip that was mentioned in Richard's PIC-based BMS thread. I know this is probably not the optimum solution, as it at best moves only 200 mA between cells, but some RC-styled balancers only do 50-100mA. The better ones will do 350-450mA. In any case, what I'm looking to do is an external balancer that would be used occasionally to re-balance a pack. I know there is a "worst-case" issue with having to move current from one end of the string to another, but that just means it will take more time. That's not a problem for me, as I would simply plug it in overnight.

Here's the schematic for one channel:

The LM2663 and the cap do most of the work. the rest of the circuit, which Richard helped me with, drives a bicolor red-green LED. Basically, what happens is that if the upper cell is higher than the lower cell by an amount governed by the feedback loop gain (the ratio between R104/R103...), the LED will be lit red. If the lower cell has a voltage greater than the upper one, the LED will be fully green. If the cells are balanced, both LEDs will be on, giving a yellowish color. This way it should be fairly easy to see what is happening. You just wait for all the LEDs to be yellow and then you know the pack is balanced.

Since this is mainly an excercise to test the concept, I was going to breadboard this, but the LM2663 only comes as an SMT part, so I decided it would be easier to just do PCB. Here's what it looks like:

I got a few boards made, and have all the parts now, so as soon as I get caught up on a few things, I will put one of these together, and start doing some tests.

-- Gary

 

[paultrafalgar]

The problem is, that Spartspark themselves admit that the use of their equalizers on more than 10 cells or batteries doesn't perform "as well as expected". I will hopefully get an answer that quantifies the problem. Maybe it will be tolerable and we can use them. Then, I'd like to try to organize a group purchase to get the price down.

Here's an idiot nubee suggestion:
If you didn't have to solder the cells together but could put them in a "clamped" rack, you could swap them around so that the transfer of charge to lower cells didn't have to traverse the whole string?

 

[PJD]

Gary,

The idea of BattEQ's, or Powercheqs, is that they simply stay connected full time, and balance full time, when charging, discharging and in-between.

The quiescent draw is sufficiently small that a 40AH pack, a charging every month or so when in storage is adequate. The Powercheq claims that it delivers 2A of balancing current, which seems believable since just a second of operation (as indicated by a LED) is enough to bring a pair of 40AH 12V batteries to equal voltage.

If it weren't for the high price in small qtys. I would give the BatEQ's a try. Group purchase anyone?

 

[PJD]

paultrafalgar,

I'm not sure what you are suggesting, the electrical connections are the same regardless of how the cells are physically arranged. My application is an electric motor scooter and the cells are in battery boxes between the frame rails and over the rear wheel, not accessible without a bit of disassembly of the scooter body.

Keeping the physical length of the wires as short as possible would certainly help.

 

[GGoodrum]

Gary,

The idea of BattEQ's, or Powercheqs, is that they simply stay connected full time, and balance full time, when charging, discharging and in-between.

The quiescent draw is sufficiently small that a 40AH pack, a charging every month or so when in storage is adequate. The Powercheq claims that it delivers 2A of balancing current, which seems believable since just a second of operation (as indicated by a LED) is enough to bring a pair of 40AH 12V batteries to equal voltage.

If it weren't for the high price in small qtys. I would give the BatEQ's a try. Group purchase anyone?

I think leaving these connected all the time to cells in a LiFePO4-pack is not wise. Justin, from ebikes.ca, had some packs end up with dead cells after leaving packs sit on the shelf for too long. I really don't care if it takes 2 or 10 hours to balance, but I wouldn't leave them cconnected permanently.

-- Gary

 

[PJD]

"I think leaving these connected all the time to cells in a LiFePO4-pack is not wise. Justin, from ebikes.ca, had some packs end up with dead cells..."

My fairly long experience using them shows that they can be kept connected full time all the time on 40AH or larger packs without problems. The specific sequence that you have to connect and disconnect them makes it inconvenient to not leave them connected full time. Even after a month of sitting idle, I found the amount they discharged to be almost undiscernable from just self discharge. But I guess a damaged battery - one that that is virtually shorted, could cause the battery balancers to drain the others.

 

[GGoodrum]

That's the problem. If you do have a bad cell, you can kill all the rest of the cells in the pack, if you leave such a system connected all the time. Also, with LiFePO4-based packs, you really can't do this at the pack level, like these SLA-based units are designed to do. SLA's can be balanced at the pack level during bulk charging, but Lithium-based cells cannot.

The LM2663 chip has a "sleep mode" input, so I guess you could use these in a BMS design that only enabled them when the charger was connected, for instance. I still like the idea of keeping it external, which is what I will test first.

-- Gary

 

[PJD]

SLA's can be balanced at the pack level during bulk charging, but Lithium-based cells cannot.

This isn't true, and it was a long expensive lesson for me finding out.

Without balancing, when bulk charging, some of my SLA or silicone batteries would go up over 16 volts while others were not much more than 13.5. And unlike flooded lead acid, there isn't replaceable electrolyte you can afford to gas off of the high batteries (and also individual cells) while the pack equalizes. I found that equalizers, or bank charging are mandatory for SLA packs longer than 2 batteries. Electric motorcycle/scooter sized packs can cost &700 for a new pack using good-quality SLA's at current prices - and that's shopping around.

 

[GGoodrum]

I agree, you can't bulk charge SLAs in a string, and have them balance. I was talking about single sealed batteries, which have multiple 2V cells. for multiple strings, where you have 3-4 12V SLAs in series, I can see how using a PowerCheq/BattEQ would work fine, but again, if you have one bad battery, it will kill the others in the series string.

With your 40Ah TS setup, you have 16 individual cells, so you would need 15 of these in order to work. that is a VERY costly option, I think, and I certainly wouldn't consider leaving these connected all the time.

The LM2663-type chips do the same basic thing, which is to move current from the fuller cells to the less full cells, until they are all equal. The LM2663 only does 200mA, not 2A, but if my test system works, it can always be scaled up by paralleling multiple LM2663s. These are surface mount parts, so they are tiny. So are the 10uF multilayer ceramic caps.

-- Gary

 

[AbM]

Hello, new here, first posting. Great forum!

I read this topic with interest (despite that it´s quite old). I was thinking, what will it take such that you can connect Gary Goodrum´s circuit permanently? So I have a few small ideas to minimize current drain when the batteries are balanced.

I have a physics background, i am not an engineer, so I roughly understand circuitry schemes, but i don´t know all the partnumbers etc. Therefore it´s only a rough sketch that I´m posting. The drawing only contains the proposed mod, not the entire scheme. Imagine the rest is still there please.

The ideas are as follows:

- When the LEDs are connected to the mid-point rather than the minus of the lowest batt, then they only will light up, when the batteries are not balanced, one at a time. When balance is perfect, the output of the opamps is equal to the midpoint of the two batteries, so the LEDs are off.
You might leave out the LEDs altogether, to save even more, or in case the board is to be put in a place where you can´t see it.

- When the batteries are balanced, the pumping IC should shut down. Therefore the thing with the two diodes and the transistor. When the batts are out of balance, one of the opamp outputs is high. Through the respective diode, it pulls the base of the transistor up, which in its turn will pull the SD pin down to GND, so the pump will start. When the batteries are equal, the opamps are at GND level, the transistor blocks and the resistor pulls the SD pin high so the pump is off. Input resistance of SD is high, so this part doesn´t need to use a lot of current. The resistor only needs to pull SD up to 2 volt.

- when you take opamps with super low resting current, maybe then this circuit would use so little current when the batteries are balanced that it is safe to leave it mounted all the time.

I also have 2 questions:

- Could this circuitry be proportioned, i.e. are there equivalent components available, such that it can do the same trick with 2 x 12V, i.e. SLA batteries? Then it could be a reasonably priced alternative to PowerCheq/BattEQ?

- in the mean time, does anyone know of a supplier that sells PowerCheq/BattEQ directly to european customers?

--AbM

 

[pgt400]

I'm still intrigued by this approach, and have decided to do a test balancer, using the LM2663 switched cap chip that was mentioned in Richard's PIC-based BMS thread. I know this is probably not the optimum solution, as it at best moves only 200 mA between cells, but some RC-styled balancers only do 50-100mA. The better ones will do 350-450mA. In any case, what I'm looking to do is an external balancer that would be used occasionally to re-balance a pack. I know there is a "worst-case" issue with having to move current from one end of the string to another, but that just means it will take more time. That's not a problem for me, as I would simply plug it in overnight.

Here's the schematic for one channel:

The LM2663 and the cap do most of the work. the rest of the circuit, which Richard helped me with, drives a bicolor red-green LED. Basically, what happens is that if the upper cell is higher than the lower cell by an amount governed by the feedback loop gain (the ratio between R104/R103...), the LED will be lit red. If the lower cell has a voltage greater than the upper one, the LED will be fully green. If the cells are balanced, both LEDs will be on, giving a yellowish color. This way it should be fairly easy to see what is happening. You just wait for all the LEDs to be yellow and then you know the pack is balanced.

Since this is mainly an excercise to test the concept, I was going to breadboard this, but the LM2663 only comes as an SMT part, so I decided it would be easier to just do PCB. Here's what it looks like:

I got a few boards made, and have all the parts now, so as soon as I get caught up on a few things, I will put one of these together, and start doing some tests.

-- Gary

Thats a cool circuit. I tried a similar idea but used a DPDT PC relay to switch a 470 uf, 35v cap between 2, 12v agm batteries at 2 Hz. My plan was to have them switching the entire time the battery was charging and off otherwise. I tried it on my bench and it seems to balance the sla's perfectly.

 

[AbM]

Yeah, that circuit from Gary Goodrum is exactly the one that I was proposing these modifications to...
Cool that your trick with a relay works, too. Makes sense though.

 

[GGoodrum]

I had some boards made for this, and bought the parts, but found it nearly impossible to solder those little SMDs, so this never went too far. I still love the concept, though.

-- Gary

 

[dmwahl]

I'm working on a simpler, smaller version of the LM2663 circuit (minus the blinky lights), eventually I'm going to have some boards spun and I figured if enough people are interested it would bring the cost down a lot. I would also buy the required components in bulk and sell a kit that you can just solder together. My rough estimate for cost is approximately $30-40 for a 12 cell version for the pcb and all components required (LM2663 and low esr caps) minus wires. So all you would have to do is solder everything together, connect the wires to each cell, and balance away. Depending on how the layout goes, I may just make the board support 16 or 20 cell packs too, and you just hook up however many you want. The majority of the cost ends up being the LM2663. I'm not doing this for profit, just for fun and to benefit others, so the price may come down a little depending on how much interest I get. The circuit would also work for any lithium chemistry currently available, so the same board could potentially be used to balance packs for rc cars, bikes, etc.

Let me know if you're interested in one of these so I can gauge if this is even worth doing. My motivation is primarily that the pack I use on my bike didn't come with a balancer, so I'm guessing there are others in the same boat. I have some samples from national that I'm going to build a test circuit with tonight, I'll post back with results.

 

[GGoodrum]

You might wantto breadboard and test this, before you spend any money on doing a bunch of boards. What I found is that in order to get the full amount of current transfer (200mA...) you have to have a pretty big difference in voltage between the cells. It is because of the resistance of 3.5 ohms. You'd need a voltage difference of .7V in order to get 200mA of transfer. If you cells are that far out-of-whack, you've got problems a balancer won't help. About the maximum difference I ever see with a123 cells is about 50mV. Most of the time, the differences are down around 20mV. With a difference of 50mV, you would only get about 14mA of transfer current.

-- Gary

 

[dmwahl]

You might wantto breadboard and test this, before you spend any money on doing a bunch of boards. What I found is that in order to get the full amount of current transfer (200mA...) you have to have a pretty big difference in voltage between the cells. It is because of the resistance of 3.5 ohms. You'd need a voltage difference of .7V in order to get 200mA of transfer. If you cells are that far out-of-whack, you've got problems a balancer won't help. About the maximum difference I ever see with a123 cells is about 50mV. Most of the time, the differences are down around 20mV. With a difference of 50mV, you would only get about 14mA of transfer current.

-- Gary

I couldn't agree more , I spent some time last night putting together a test circuit with the 2663. Unfortunately the only caps I had were either electrolytic or was too low capacitance. Most electrolytics have a pretty high esr at the 150khz that the 2663 switches at, so I was getting basically no transfer. My goal is not so much a fast balancer, but something that is continuously connected and can keep cells in balance. I'm going to talk to the guy at work who does most of our power supply design and see if I he can scrounge up some low esr caps to test out. What caps did you use?

Thanks for the input,

-David

 

[Lock]

BUMP

SmartSpark® Energy Systems Inc., in cooperation with the U.S. Consumer Product Safety Commission, is implementing a voluntary recall of BattEQâ„¢ Battery Equalizers.
http://batteq.com/

June 18, 2009
WASHINGTON, D.C. - The U.S. Consumer Product Safety Commission, in cooperation with the firm named below, today announced a voluntary recall of the following consumer product. Consumers should stop using recalled products immediately unless otherwise instructed.
Name of Product: BattEQ Battery Equalizers
Units: About 800
Manufacturer: SmartSpark Energy Systems Inc., of Champaign, Ill.
Hazard: The recalled equalizers can overheat, posing a fire and burn hazard to consumers.
Incidents/Injuries: The firm has received one report of an equalizer that overheated which resulted in fire that caused damage to the equalizer and batteries. No injuries have been reported.
Description: The recalled battery equalizers are charge balancing devices designed to increase the performance and longevity of rechargeable batteries.
Sold at: Authorized distributors and retailers from July 2006 through March 2009 for about $300.
Manufactured in: United States
Remedy: Consumers should immediately stop using these recalled battery equalizers and contact SmartSpark for a full refund.
Consumer Contact: For additional information, contact SmartSpark at (800) 905-6137 anytime Monday through Friday or visit the firm’s Web site at www.batteq.com

tks
lLlok

 

[abala]

Hi,

there are some experience now?

I would like to built an equalizer with the LM2663. Only the lm2663 and a 10uF/10Volt.
Is it possible to use it parallel to equalize 1Amp?

Is it working with a cell difference of 0.02Volt?

many thanks for response
abala

 

[texaspyro]

Is it possible to use it parallel to equalize 1Amp?

No... the best you can expect is around 1 mA per mV of imbalance.