What do you use for Cell Balancing

[Gazza_au]

What do you use for Cell Balancing?
 

[Hakato]

Are you thinking of putting two LiPo cells in series?

Not much out there for cell balancing. I've been looking for a good solution too. This might do the trick (?):   
https://www.maximintegrated.com/en/products/power/battery-management/DS2778.html

I haven't yet used the part myself but I plan to do so for design that I have in the works. Let me know what you come up with should you decide to use it. :-)

-Toni
Helsinki, Finland

[bazza]

For lead acid (AGM) I use....nothing (small system).
For LiFePO4, still doing the research here. Will jump in this year for sure.

[Codebird]

For LiFePO4, still doing the research here. Will jump in this year for sure.

Check the prices though! LiFePO4 is expensive. I'm going to give it a bit more time than that, and hope that as production ramps up so prices will come down.

(I wonder how you'd say that in conversation? Le-fee-po-four?)

[bazza]

For LiFePO4, still doing the research here. Will jump in this year for sure.

Check the prices though! LiFePO4 is expensive.

Cost is comparable to lead-acid when you take into account they can be charged faster/more efficiently, have a (claimed) longer service life and can be discharged far more with less damage. You can buy a much smaller Ah LiFePO4 battery to match an equivalent 'bigger number' Ah lead-acid to fairly compare pricing.

I've looked at electrodacus.com for cell balancing but don't know what I will do yet.

[retrolefty]

The radio control crowd have been using balanced chargers/dischargers forever. Try looking through:

https://hobbyking.com/en_us/chargers/battery-charger-dc-only.html

https://hobbyking.com/en_us/chargers/battery-charger-ac-dc.html

[Red Squirrel]

For lead acid, nothing.  Don't really need to, but it's a good idea to monitor it.  (mid point voltage).  Even the telcos don't balance charge but do monitor it.  In the building I work at we have 24 3900AH cells in series which creates one 48v string, then put several strings of those in parallel to add capacity.   

For lipo/lithium ion it's a bit more complicated as each cell needs to be balanced.  Or more specifically each cell set, as you can have a bunch of cells in parallel act as 1 cell, at least I think.  I have not played much with them myself.  So if you have a battery pack that is 4 cells in series you need  need a charger with 5 wires, as you  need wires going between each cell too.  I'm not sure how it works for alternate energy applications where you want to power the load at the same time too though.  I don't think you can just hook the load up and just use it, there's probably more to it than that as the charger needs to be aware of the current going into each cell and if there is a load it would skew those numbers.   I'd be curious about this myself as I've been wanting to play around more with lithium ion.  With 18650's becoming so widely available they are really starting to make sense vs lead acid for some applications.   Lead acid is probably still more suitable for stationary applications though as weight/size is less of an issue and they are more tame.  Ex: they won't set your house on fire if you do something slightly wrong.

[bazza]

With 18650's becoming so widely available they are really starting to make sense vs lead acid for some applications.   Lead acid is probably still more suitable for stationary applications though as weight/size is less of an issue and they are more tame.  Ex: they won't set your house on fire if you do something slightly wrong.

LifePO4 should be safer than other lithium chemistries. When I dive in, it will be with a small 40Ah / 12v system to learn the nuts-and-bolts of LiFePO4 before moving onto bigger things.

[vealmike]

If that doesn't help, please consider making the question more specific.

[Codebird]

You can buy a much smaller Ah LiFePO4 battery to match an equivalent 'bigger number' Ah lead-acid to fairly compare pricing.

https://www.amazon.co.uk/Lithium-Standard-IP56-Waterproof-LiFePO4-Phosphate/dp/B01N0SAM9P/ref=sr_1_15?ie=UTF8&qid=1490879000&sr=8-15&keywords=lifepo4&th=1

100Ah, 12V, £1200.

https://www.amazon.co.uk/Exide-Start-Battery-EK-950/dp/B00VQ66ZOM/ref=pd_day0_328_2?_encoding=UTF8&psc=1&refRID=1F3G3W4TNPP8B5ZCV59Z

95Ah, 12V, £135.


LiFePo4 can handle more current, true - but not enough to justify ten times the cost! And their service life is comparable to lead-acids too, as those are a pretty durable chemistry if treated well. Plus you have to deal with a BMS which brings further cost overhead, something that lead-acid (and the old NiMH) does quite happily without. I've high hopes for LiFePO4, but right now it's a young technology and bears the cost penalty expected of that. Give it a few more years and I'm sure the price will come down.

[Kilrah]

Balancing what cells for what application?

BMS with balancing would do for lithium cells...

[bazza]

100Ah, 12V, £1200.
95Ah, 12V, £135.

12V 40Ah
$299 Australian dollars (~182 British pounds).
http://www.evworks.com.au/winston-lp12v40ah-lifepo4-battery-12v-40ah
A 40Ah is basically equivalent to a 100Ah lead-acid, capacity-wise (you can't really discharge lead-acid low often without doing major damage to it). So they are basically on-par cost-wise. Coupled with the fact that charging them is more efficient and they are likely to last longer, we are easily at a point where they are cost-competitive over the service life. Plus at about a third of the weight of equivalent lead-acid. The only niggling issue being the cost of the cell balancing system you choose, in my opinion.

In your cost comparison above:
1) you've chosen an uncharacteristically high-priced LiFePO4 battery (as lots of "Available from these sellers" Amazon listings are. LiFePO4 are nowhere near 10x the cost of lead-acid. If they were, I certainly wouldn't be looking at them.
2) you've assumed both chemistries have similar capacities by simply looking at the 95Ah/100Ah capacity designations. LiFePO4 can be discharged to 70-80% fine without damaging the lifetime of the battery too much. That's at least 2x and up to 3x more than is safe for lead-acid. In other words: when you buy a lead-acid batery, you're really buying only about 30% of the claimed capacity - if you want your battery to last, that is. 50% at best.

You probably want to charge it at a rate of 10%-20% of its rated capacity too, depending on exact chemistry (AGM, flooded, etc.). And you will probably have to live with charging inefficiencies that are higher than LiFePO4. At the end of the day, that's more power wasted with lead-acid.

You can opt to buy individual cells at a similar cost to the battery I've shown above:
http://www.evworks.com.au/winston-battery-wb-lyp40aha-wb-lyp40aha-3.2v-40ah
Which may give far more flexibility. It's probably what I would do at this point. 4 of those ond you have a modest 12V 40Ah battery.
$72 AUD = around 44 British pounds per cell.

We are at a point now where LiFePO4 can easily make inroads into lead-acid territory, in my opinion. So it's hardly 10x the cost when you take into account realistic prices, longer (claimed) service life & more efficient charging and discharging characteristics. We are 'there' now.

[h3po]

From what I've read and seen on youtube, It might be reasonable to not have built-in balancing (just monitoring) at all in a house battery. For EV 18650 packs there's people that bottom-balance, top-balace and not balance... It very much depends on how likely your battery is to get out of balance in the first place.

[Gazza_au]

https://www.eevblog.com/forum/projects/building-circuit-for-testing-one-way-balance-system-with-lifepo4-cells/msg1187723/#msg1187723

[Red Squirrel]

On this subject I've been pondering about how one could make a small solar or UPS type system that uses lithium ion.    What is the best way to balance charge these, while the device can still be used?  Ex: you are charging each cell individually while tracking current, thus adjusting voltage as required for each cell, or cell set (a set can be a bunch of sells in parallel, but you can treat as 1 cell).  But the ends of the serried set is also connected to the application that draws power.  If you do draw power while the balance charger is on, it's going to mess up all the readings as it will be drawing current from the charger too.  So how would this normally be done?  Do you have two battery packs and simply alternate between both?  So charge one while the other is discharged, then switch?  Ex: every 10 minutes.   Or is lithium ion just not suited for such an application?  With lead acid you'd just keep them at float and not really worry about current. That's what the telcos do.

[tatus1969]

I am currently working on a protected 6S2P pack for my kick scooter project (www.kicksurfer.de, sorry not much up there yet - its on the way). It is based on Intersil ISL94202 / 94203. That is a quite impressive chip, it can be used as a full blown standalone BMS / balancer, and it can be fully controlled by a uC via I2C. Here are some first impressions. The first prototype is working well, all functions tested and working. Next are charge/discharge testing.






A picture from testing short-circuit. Yellow = pack voltage, blue=voltage across 1mOhm shunt. The pack shows an impressive 250A short circuit current, and the MOSFETs sately interrupt that after 400us.

[Kilrah]

Interersting, did you try that already? Most available kick scooter motors will be anemic at 6s.

[tatus1969]

Interersting, did you try that already? Most available kick scooter motors will be anemic at 6s.

I am using the pack together with a 350W/24V hub motor, I am very happy with that combination. I am limited to 12 cells, because I wanted to place everything underneath the scooter. The pack uses Sony US18650NC1, and it can deliver up to 15A. The scooter does not have a very high climb rate, but the intention of it is more to give assistance during daily work commute with public transport. It weighs only 8kg and I use it every day instead of walking between home, trains, buses, and work. I once calculated that I had 20 minutes in train, plus another 40 walking every day. That was the start of this project.

[Kilrah]

Nice. I've got an electric scooter too, but a bit more heavy (10.5kg) and powerful (450W, 9S/13A stock) with about 25km range (6.5Ah).
Initially I thought having 25km range was stupid and I'd rather save some weight instead, but in the end I wouldn't want any less, given it's really a cool thing to ride on I use it for much more than just the "utility" trips and end up just riding around for fun and doing >10km pretty much every day I use it.

I've since boosted it to 13S/650W by adding a 4s 6.5Ah in series and it's just a blast. Range is 20-30km depending on how aggressive you are on the throttle, and I've got a way to switch batteries around to a 12S/13Ah configuration for a bit less punch but ~50km range. 

[tatus1969]

that sounds like a real beast  Can it do wheelies? Your pack has roughly 13 * 3.7 * 6.5 = 312Wh. My cells are 2.9Ah each, so they achieve 6*3.7*5.8  = 129Wh. With one charge, it travels about 10-20 km, which is more than enough for daily commute. I have to charge it every 2-3 days.

The working principle is also different. It has no throttle or brake, nor a power switch. It senses that acceleration from the kicks that you give it, and assists in maintaining the speed. Braking works similar. It automatically powers on/off when you start/stop driving the scooter. Kind of British understatement  Also charging is simple, just pull out a cable and connect the charger.

I am working on a DIY kit for sale that includes all components for an electrical conversion, except the scooter itself.

[johansen]

hope that as production ramps up so prices will come down.

Might be waiting a long time.

i seem to recall just last year that the cheapest lithium iron phosphate cells i could find were 3$/amp hour, now they are 2$.

anyhow, for cell balancing, try this:


you can build each switching section separately optimized for the voltage, frequency, mosfets available, or you can make them all identical and chain them together.

you will need to ensure exactly 50% duty cycle...

the advantage of this circuit is you can still get power from the system even when one or more batteries are completely dead or open circuit, according to the current capacity of the bms.

[Kilrah]

that sounds like a real beast  Can it do wheelies?

LOL nope given it's a forward drive 

The working principle is also different. It has no throttle or brake, nor a power switch. It senses that acceleration from the kicks that you give it, and assists in maintaining the speed.

Yeah, I tried one of these but absolutely hated it... oh well, different tastes/purposes. I went from 9->13S mostly so I could climb some pretty steep paths with decent speed and while carrying a rather heavy backpack, does the job well. The 40km/h top speed on flat road is a very enjoyable "side effect" of it

[tatus1969]

I tried one of these

Interesting, can you tell me the model? There is only one that I know of, which is the micro e one. That one is different from mine, in that it stops assisting after every kick. It just boosts your leg power, and it will roll out when you stop kicking. Mine works different, it fells like a regular scooter with all friction removed. One kick and it drives you home.

[Kilrah]

Yup it was the micro.
OK interesting, would have to try your solution.

So basically it's just a fancy way to set the speed instead of using hand controls, how do you differentiate the user braking from a climb? Brake switch?

[ConKbot]

Yup it was the micro.
OK interesting, would have to try your solution.

So basically it's just a fancy way to set the speed instead of using hand controls, how do you differentiate the user braking from a climb? Brake switch?

Just guessing, but a tri-axis accel to keep track of the acceleration vector.  Remove any transverse acceleration from cornering, to get your acceleration on the front/rear/up/down plane. Offset front/rear acceleration until the leftover vector is 1.0G and that offset was the amount of user input.  Apply motor torque accordingly.

Obviously a bit simplified, throw on filtering, a bit of checking to make sure it's not tilted over, and compensating  for g forces rounding the bottom of a hill (rate sensor to look for rotation maybe? )  and you have something that can play nicely.