Cheapest Active Cell Balancer Project

Joined
Feb 24, 2017
Messages
20
Hi all,

I am new here, so, let me know if I am doing anything wrong.

Basically, I think I just successfully designed a non-dissipative active cell balancer using cheap jellybean parts. The cost for all material (excluding labour) is around £3 to balance a 2-cell pack at 1A.

I am actually quite surprised by how simple it is and I wonder why no one is selling it yet.
My design uses closed-loop buck converter topology. It looks almost like this: http://www.flyelectric.ukgateway.net/bal-circuit.htm
But, instead of opamps, it uses half-bridge driver.

The design is analogue based with no microcontroller or any special battery-chip involved. The close-loop is achieved by an LM358.
It is so simple that I am very excited to share it.

At the moment, it is all built on a perfboard and I am currently optimising the quiescent current (quite high actually).
The design is intended to be used while charging (maybe even running?).


The advantages are: low wasted heat, higher current (not tens of milliamps), and, small footprint (almost the same size as 9v battery).



I am planning to open source the design as I think the community will be able to improve it further and benefit from it.

But, before I open source it, I want to sell at least 10 of them (maybe on Tindie?) for around £8 each. The reasons are:

1) To recover my RnD costs (£15)
2) I want to buy a hot air soldering station (£25 on ebay) to assemble it
3) I want to teach myself pcb manufacturing (maybe £15?) and selling hardwares


I am willing to give free three prototype samples to anyone that answer these questions:

1)
Do I need to worry about CE to sell it? I tried to research about it, but, it is difficult for me to understand lawyers' languages.
Breaking this down to simple English would be nice. as IANAL.
I know dealing with battery is dangerous, hence the question.
I will incorporate polyfuses and reverse-polarity protection in 2nd revision for extra safety.

2)
What is the best place for me to sell this as cheap as possible? Is Tindie a good start? Or, is ES is good enough?

3)
As this is my first time manufacturing my own PCB and sell it, any tips? I am planning to use SeeedStudio because of the low price (hopefully lower than £15).
Any suggestions?

4)
Maybe a small review about the prototype.


I am in the UK at the moment, so, the free samples can be posted only to the person that are in the UK (to keep low shipping cost).

I am also interested to see someone else's design (that is <= £3) for me to compare against my design (I can't find any on google search, am I the first?).


Any guidance will be greatly appreciated.

EDIT:
I forgot to mention that, instead of wasting energy in resistors, the design will transfer the charge from strong cell to weak cell.

Thank you. :)
 
AfdhalAtiffTan said:
1)
Do I need to worry about CE to sell it? I tried to research about it, but, it is difficult for me to understand lawyers' languages.
Breaking this down to simple English would be nice. as IANAL.
I know dealing with battery is dangerous, hence the question.

It's highly likely that your product will fall within the scope of at least one EU safety standard and would therefore need to be assessed for CE conformity before it could be placed on the market. However, the chances of it being worth you doing this for the sake of selling a handful of units to a hobby community is slim and the chances of Trading Standards taking noticing your activity and taking action similarly slim.

The main impact in failing to comply with CE marking would be in the unlikely event that your product caused serious damage or injury and the matter reached the courts. It would greatly increase your culpability.
 
Hi Punx0r,

Punx0r said:
It's highly likely that your product will fall within the scope of at least one EU safety standard and would therefore need to be assessed for CE conformity before it could be placed on the market. However, the chances of it being worth you doing this for the sake of selling a handful of units to a hobby community is slim and the chances of Trading Standards taking noticing your activity and taking action similarly slim.

I don't think I will be able to sell more than 20 units, so, I will take that as a green light for me.


Punx0r said:
The main impact in failing to comply with CE marking would be in the unlikely event that your product caused serious damage or injury and the matter reached the courts. It would greatly increase your culpability.

Now that you mentioned it, I will definitely add polyfuses in the final design.



I will be assembling the prototype board to send to you. Give me two weeks to get it done and I will PM you for your address for me to post it.



Thank you so much for your reply!
 
In case anyone is interested, I am posting my progress on https://hackaday.io/project/20825-low-cost-non-dissipative-active-battery-balancer
 
Hi :)

Interesting design ! Do you know what the maximum current will be ? Can it set a max voltage per cell ? I want to make a DIY charging dock for 4x 8S1P 18650 batteries with integrated "bms" for charging. Plug and forget, unplug once full.

I see that you use 14S recycled pack so it must be working right ? What max charging level do you usually set ?

I'd love to test this if it can meet my needs. If it can bear high enough current I know a handful guys who'd be interested too for a "battery integrated" solution on their e-boards.
 
Hi :D


Vanarian said:
Do you know what the maximum current will be ?

The maximum current is set by the chip (around ~3A), but I want to limit it to 1A for safety reasons.


Vanarian said:
Can it set a max voltage per cell ?

This is a simple active balancer, the max voltage is always half of 2S cells (it will auto adjust). For example, if you have 2S cells staying at 9V, the max voltage per cell will be 4.5v.
In other words, the max cell voltage is set by the full pack voltage, e.g. 8.4v for 4.2v or 8.0v for 4.0v, etc.

Vanarian said:
I want to make a DIY charging dock for 4x 8S1P 18650 batteries with integrated "bms" for charging. Plug and forget, unplug once full.

The good thing about my design is that, instead of monitoring each cell's voltage, you can just monitor full pack voltage (in your case 33.6V), and the balancer will force all to be 4.2V.
i.e. no need to monitor 8 voltages, just one is enough.


Vanarian said:
I see that you use 14S recycled pack so it must be working right ? What max charging level do you usually set ?

No, it is an ongoing project. I usually charge at 4.1v, but it is difficult to maintain with a passive balancer. Recycled cells are very capacity-inconsistent, this is why I started this project.

Vanarian said:
I'd love to test this if it can meet my needs. If it can bear high enough current I know a handful guys who'd be interested too for a "battery integrated" solution on their e-boards.

That is awesome! Once I fully optimise it, I will send a free board for you to test and review. Hopefully it will be useful for e-boards too. :)
 
Hi, that's great thank you! ;)

OK so basically the balancer splits equally in 2 based on charger voltage input? That's good feature too!

Two more questions :
- how could we help you?
- how is it plugged? Is it like - charger - > balancer-> batteries or charger - > batteries - > balancer?
 
Hi :)

Vanarian said:
OK so basically the balancer splits equally in 2 based on charger voltage input? That's good feature too!

That is the plan. ;) I designed it that way so that i don't have to measure individual cell voltage, this simplifies monitoring.
Feel free to suggest any features. :mrgreen:


Vanarian said:
Two more questions :
- how could we help you?
- how is it plugged? Is it like - charger - > balancer-> batteries or charger - > batteries - > balancer?

At the moment, I am working on optimising the close-loop stability. I post most of my progress on hackaday if you are interested.

In terms of help:
-I might need some advice on the questions in my first post.
-Contribute by improving the circuit that I posted on hackaday
-Provide reviews for the prototype that I will send for free (and hopefully suggest some improvements)



The final application will look somewhat like this:

[charger] --> [battery] <-- [balancer]

The battery can be charged normally, but with the balancer hooked, it will force all cells to have equal voltage.
Without the balancer, the cell voltage is determined by its capacity, i.e. high-capacity cell will gain voltage slower than the weaker cell.
 
interesting project. I presume your using a isolated buck converter between the cells? Does it have any kind of 'deadband' where if the cells are within x% of each other, it stops trying to equalize them?
Seems as if you could technically charge up just 1 cell in any given series string, and this would distribute the energy among the whole pack... Not the most efficient or quick way of doing it though hahah.
 
Hi,


sn0wchyld said:
interesting project.

Thanks :)

sn0wchyld said:
I presume your using a isolated buck converter between the cells?

No, I use a synchronous buck topology.

sn0wchyld said:
Does it have any kind of 'deadband' where if the cells are within x% of each other, it stops trying to equalize them?

It has no deadband at the moment. It will never stop equalising as long as the total voltage is above 5V.
I plan to add auto shutdown later.

sn0wchyld said:
Seems as if you could technically charge up just 1 cell in any given series string, and this would distribute the energy among the whole pack...

Yes, that is how it will work.

sn0wchyld said:
Not the most efficient or quick way of doing it though hahah.

It is better than using bleeder resistors ;)
 
Hi,

Just checked your Hackaday page, good work keep it up! :wink:

I receive a batch of new cells on Monday, can't really wait to try it out on them.

Edit : just forgot, since you don't want to sell a lot of them I assume that you will sell them not as a company but as a private person; hence your sales may remain in the "personal goods" sales category. In that case selling your units as private won't strain you with any certification duties because you are not a professional selling a commercial product.
You need to check how many goods people are allowed to sell in UK before reaching the "professional activity" category though. Hope this helps a bit :)
 
Hi Vanarian,

Thanks for checking the logs, I appreciate it.

At the moment, I got the balancing tolerance down to lower than 0.3%, that is just under 10mV for 4.2V li-ion cells.

Yes, I don't plan to sell many of them as I am still in a uni, I worry I can't time manage myself.

Your cells arrived on Monday? That will be exciting! :)
I think I am ready to send the prototype version to you (built on a perfboard), I hope I can build it fast enough.

If you want a proper manufactured PCB (with safety features included -- overcurrent/reverse-voltage), I can't make it until the end of next month because I have to wait for my student allowance to finance it.

And yes, thank you for your advice, it surely helps! :D
 
I feel you there it's harsh for us students to support projects!

I'd gladly help out though, since I was planning to offer at least to pay for shipping (I'm in France :p) I send you a PM.
 
Quick question, could you really trust that this is working to balance the cells if you have no feedback on the voltages?
 
This is a nice project you have going. When you are ready, I will take a few balancers off your hands.
 
Offroader said:
Quick question, could you really trust that this is working to balance the cells if you have no feedback on the voltages?

It depends on how you define a feedback.
If you looked at my hackaday log, you will see that it does have a feedback at the output node.
 
Vanarian said:
I feel you there it's harsh for us students to support projects!

I'd gladly help out though since I was planning to offer at least to pay for shipping (I'm in France :p) I send you a PM.

Students in da house woot woot :lol: :lol:

I will check with the local post office, hopefully you don't have to pay for the shipping. :)
 
Hi all,

I finally got the chance to learn KiCad to redraw the final schematic of the project.
I decided to open source it because I can't manufacture it myself as I am in a financially challenged situation.

I also posted one of the boards to @Vanarian for a review, unfortunately, it was lost during shipping and the board doesn't reach to him.

Anyways, here is the final circuit: http://www.afdhalatifftan.com/2017/06/low-cost-active-battery-balancer.html

Hopefully, this will benefit the community. :)
 
Cheer up ! I'm a bit mad after my postal country services, was really looking forward to give you feedback too.

In any case that's a great share and as soon as I can I'll make test units for review ; hope more people here will adopt it :) I'll forward your link to the french DIYer community too.
 
Congratulations for the design! It is quite unique!

I have some questions;

1. If the Vin is too low, MP2307 won't start.

Wouldn't be a good idea to tie this pin several level above? If yes, the last few levels might need a step up converter; the lm358 power supply might need moving as well.

2. if some battery overheats, shouldn't it be a good idea to disable charging based on a thermal sensor? Maybe you can tie a thermal sensor to the EN pin and/or allow the whole balancing system to be disabled/enabled by an MCU / push button.


What do you think?

P.S: I am no expert. If any of the questions/ideas are wrong, please apologize my lack of knowledge. Note that I would appreciate if you would tell me where I am wrong.
 
I dig it, it is so easy to solder up I think I may try it on my old old old battery, gosh its got 1000 cycles on it by now because some days I'd charge 3 times.

One main question I have is
Is it safe?
Can it be made safer?
I dont mind spending a bit extra for fuses or diodes or whatever else. What I have now is basically 5S2P segments to make up 10S and 8 or 10P but I'd love for it to balance, knowing I aint buying some Chinese basement design.




balancer1.jpg




hookup.jpg



Afdhal Atiff Tan wrote 05/24/2018 at 00:09

The following is an example for a 4S pack:
Balancer1: Cell1 and Cell2
Balancer2: Cell2 and Cell3
Balancer3: Cell3 and Cell4
Let me know if this is still confusing.
 
For a 10S5P at a cost of say $3 which is the cost of the MPZ307 alone its going to be $150, judging from Digikey.CA's website, more like $5 for each balancer. Easier for me to just buy a BMS from Besttech. It is a very interesting circuit you made there and I am quite impressed by it actually.
 
markz said:
It is a very interesting circuit you made there and I am quite impressed by it actually.

I know right ? He made a sick job, I'm still angry I never got around the Post office to try his proto. And yea it is kinda expensive for bigger builds, wasn't able to afford a full string back then.

Keep me updated if you go this route please!

Edit : Wait actually if you can parallel plug batteries on modules it should work no ? I mean : You usually use a single BMS for a whole battery pack. Can it be made ?
 
Back
Top