First schematic for auto-balancing LiPo charger with all saftey features - RC Groups
Thread Tools
Mar 30, 2004, 04:09 PM
Registered User

First schematic for auto-balancing LiPo charger with all saftey features

I want to get into LiPo batteries but have not found a charger that suits my needs.

My requirements:
- safe & simple operation
- automatically balancing individual cells
- safe in case I try to charge a cell that is discharged too deep
- no damage to LiPo pack in case the charger circuit breaks down (if it will get old enough it finally will die)
- safe in case a cell dies during charging (short-circuit)
- dedicated to 1 battery type so I can make no mistakes with connectors or settings
- state-of charge monitor so I can decide the cells are charged enough for my need and go flying.

I saw a solution that was charging all cells in parallel, but disliked the following:

- during the flight , the cells where configured in a series pack using a purpose-built connector. This means that for 3 cells the current has to pass 6 connectors instead of 2, resulting in 3 times more ohmic losses and potential for failure

- during charge, the cells where wired in parallel by another purpose-built connector. This means that should a cell be out of balance, it will receive the full charge current from the other cells. And since discharge current is much higher then max. allowable charge current, this abuse will further weaken the weak cell.

- during charge, should a cell in a pack fail (and eventually one of them will fail) by short-circuiting, the other cells will also be in short-circuit, potentially resulting in fireworks.

So therefore my solution below.

Short explanation: I set up a current pump, connect the pack to this, monitor indivudual cell voltage and when a cell tries to go higher then 4.2V, I sink the current through an old-fashioned power transistor (could have tried FET's and power resistor).
As an extra security measure, I have an additional transistor that can break the current completely on two events: the whole pack has a voltage lower then 2.5 V (so we should not charge it to avoid danger) or higher then 12.7 V (so one of the opamps or transistors has died, or possibly an already overcharged pack was attached) I then added a couple of led's to give me an idea of state of charge of individual cells (so I would know when I have a cell that is particulary weaker then others), going brighter when charging is more complete.

I quickly estimated cost and for 25 Euro the finished charger should be doable.

My question:

Does the schematic make sense to you experts out there? It's been more then a few years that I last designed something using analog stuff!

So all feedback / ameliorations / scepticism are welcome...
Sign up now
to remove ads between posts
Apr 11, 2004, 09:35 AM
Registered User
well done!
Would you mind to draw the pcb to me please?
Apr 14, 2004, 03:02 AM
White Knuckle Flier
_AC_'s Avatar
Hi kodel,

I was thinking of a circuit with similar features for charging lipos, but lazyness got the best of me and I'm building a 'dumb' cc/cv charger.

I just have one question about your circuit. After each of the comparators that appear to be monitoring the individual cell voltage it looks like you have transistors... they look like they're connected in parallel w/ each cell.
Are they really transistors?

I'm thinking that if one of those transistors ever turned on you might have a fireball on your hands - either the transistor or the lipo cell is going to blow up, the weaker of the two.

I see a few other errors in the schematic, but it would be difficult for me to go into detail without having references assigned to each part.
Apr 15, 2004, 03:27 PM
Registered User
I drew the schematic before reading up on my rusty analog design experience. There are major errors in here, the opamps with the 4 resistors on them should be configured as a differential amplifier with a gain of 1. The drawing shows incorrect resistor connections.
I'm now redrawing the thing with Spice with correct opamp configurations.

Towards your note on the transistors: they are indeed power NPN transistors, the idea is that they are turned off when the cell voltage is below 4.2 V. When the voltage reaches 4.2 V they gradually start to absorb the excess charge current. So, when the pack is fully charged, they will be absorbing the full charge current (say 1.5 Amps).
The only danger I see is that over time the transistor will break down (e.g. after 10 years of usage of maybe much less when no effective heatsink is provided).
But even after 10 years I would not want a lipo fire, hence the safety circuit drawn at the top.
Apr 15, 2004, 03:33 PM
Registered User
I just wanted to add that this is a low-priority project for me and that I will not be able to produce a proper schematic / pcb layout / component values / test report until I manage to get a few days of vacation.
The idea of posting pre-maturely was just to get feedback to see if anyone was interested or if you all want to stick to the simple CC/CV charger types.
Apr 15, 2004, 10:15 PM
Permanently Banned
Larry3215's Avatar
I'm interested!

Apr 17, 2004, 11:22 PM
Registered User
I am interested also.
Sounds just like what I have been looking for.

Apr 19, 2004, 02:03 AM
Registered User
I wonder why the people producing Tanic cells with the tanic taps have never come up with this tpye of charger, it really brings out the add value of the tanic taps.
Apr 20, 2004, 12:18 AM
Registered User

Tanic Charger

OH we have been working on this charger for a few months now and your right it is a great idea, but it really takes some time for all the designing and board layout etc
I really like your drawing but here is ours


Apr 20, 2004, 01:03 AM
Registered User
Hehe, you're clearly at a further stage of development then I am

Attached is the schematic as I'm now drawing it with PSpice, still not finished (some resistors to be calculated, replace current sources with their equivalent components, etc)

Do you already have a ballpark figure of MSRP ? I think my design could after optimisations and production in China be around 50 dollar, however I'm not familiar with distribution structures in the RC niche.


Last edited by kodel; Apr 20, 2004 at 01:05 AM.
Apr 20, 2004, 09:01 AM
Registered User
the first one will be a couple grand from there if we mass produce they will be very affordable if we can get the volumes

May 07, 2004, 03:40 PM
Registered User

Simpeler alternative possible?

I am thinking of an idea on replacing the 2 opamps per cell + npn shunting transistor with an off-the-shelf linear voltage regulator (like LM317), since the components I'm using are actually inside the regulator and they are preforming the same function.

I couldn't figure out how to hook up the regulator, until I came across my RSComponents catalog and wondered what the difference was between positive linear voltage regulators and negative linear voltage regulators. What surprised me even more was that there are types that qualify both as a positive and a negative regulator.

So, I did a small simulation of a positive type. The input voltage remains constant, but at the output I have set a voltage source that slowly rises from 0V to 20V. For the regulator, this means the load diminuishes until it is zero and when the output voltage rises higher then the voltage the regulator is programmed for, the load actually becomes negative.

This means that the regulator would need to sink current instead of source current. From the simulation graph, it is obvious that a positive-only regulator can not do this.

My goal of replacing 2 opamps + 1 npn + a bunch of resistors with a singe regulator with 2 resistors is not possible with the positive regulator type, since the only behavious I'm looking for is to sink excess current before the cell voltage rises.

But, what is this negative regulator type? On to the next post...
May 07, 2004, 03:46 PM
Registered User

A negative regulator should do the trick...

Edited on 16/08: this idea is a no/go, forget about it and read the next post

So, in my quest for a negative regulator I had the bad luck that in the spice emulator the manufacturer did not provide a negative model.

But looking at a schematic from the competition reveals that the answer is simple: a negative regulator is used to create a negative voltage from an even more negative input source.

So, in my limited neural capacity this means the regulator is sinking current instead of sourcing, which is what I was looking for!

And the best news is, even the el cheapo LM317/LM350 is rated as positive/negative regulator, which would mean they can both source and sink current.

Can any of you experts out there tell me if this indead would work, hooking up a negative regulator over each individual cell and bleeding off the current (via Vin of the regulator) to ground (or something lower)?

I am not able to test this at the moment (have to order the regs...)

note that the component in the image attached has more then 3 leads, but this doesn't matter for the application I'm investigating.
Last edited by kodel; Aug 16, 2004 at 03:22 AM. Reason: Edited on 16/08: this idea is a no/go, forget about it and read the next post
Aug 16, 2004, 02:59 AM
Registered User

prototype ready and tested

Well, after a few months of inactivity with regard to this project I finally found an evening to finish it.
I deviated a little from the original plans but with satisfactory results

- I inspired myself (not to say copied) on Suzanne's design of the Lipo balancers. I did use some cheaper components however since the MAX is quite expensive. (see schematic)

- I ran into a small problem with my design: Since the comparator I'm using has an open collector output, I have to feed the RC filter through a pull up resistor. This causes the "on" switching time for the FET's to be very long, so they operate in the forward active region for some time. Therefore they where getting pretty warm. Since I don't have a temperature probe or scope to measure things, I mounted them on a cooling fin just to be sure. Now they stay cool.
I could have added a buffer to overcome this problem, but the cooling fins where laying around anyway.

- I used a simple variable voltage regulator as a current source. This was my quick way to test the whole of the system setup, but it has the disadvantage of having a high drop out voltage. This means I'll not be able to charge my 3s lipo's from a car battery without voltage booster.
I don't have a need to recharge at the field and our field has a generator anyway's, so to me it's not a problem.
I do have the intention to test my low voltage drop current source design however, but it could take some time (I'm quite busy)

I'll post some pictures / schematics in the following posts
Aug 16, 2004, 03:12 AM
Registered User

finished prototype

Here's a picture of a finished prototype on breadboard.

I didn't do any schematics or breadboard design, I just started soldering since it's so easy. ON the right you see an LM350 with current sensing resistor as a quick current source.

The Multiplex connector is mating to a similar connector on my packs.
I used 4 pins for + and - (2 pins each for lower connector resistance) I used the 2 remaining pins for the cell interconnections.

The back and red plugs go to a laptop power supply (or anything capable of developing 1.5 A @ +/- 16 Volt)

Thread Tools