DIY Li Ion Charger, simple schematic, advice needed

[newInRobotics]

I've got a laptop battery that contained 8 Li Ion cells, each of 2AH capacity. All of them are 3.7V, so they are not ruined. I've been googling and learning a lot about how Li Ion cells operate and how they have to be recharged. Mine ones are SONY FUKUSHIMA  US18650GR G4, min voltage 3V, max voltage 4.2V.

I came up with a very simple design of a battery charger suitable for charging one cell of a type mentioned above. It all works fine in simulation. Can someone (people who are more experienced than myself) have a look at it and tell me if such design would work?



If I adjust R1, R2, and R3 I could use such charger for custom size Li-Ion battery packs up to 12.6V with 19V power supply and 16.8V with 20V(or more) supply.


According to this graph, charging current starts dropping after battery is charged to a bit more than 75%, so LED in my diagram is a good indicator of when charging is going to the end.

In theory, my circuit would not overcharge battery, because regulated voltage is not higher than recommended voltage for one cell. As current is limited by R3, overcurrent problem is solved as well. That's only my opinion. Any thoughts?

[Soeren]

Hi,

I came up with a very simple design of a battery charger suitable for charging one cell of a type mentioned above. It all works fine in simulation. Can someone (people who are more experienced than myself) have a look at it and tell me if such design would work?

Simple is good, but this is a bit too simple I'm afraid.

Don't ever trust a sim unchallenged, simulators are akin to theory, only good as long as all parameters are known and they seldomly are.

In theory, my circuit would not overcharge battery, because regulated voltage is not higher than recommended voltage for one cell. As current is limited by R3, overcurrent problem is solved as well. That's only my opinion. Any thoughts?

No, in theory, your circuit will be able to ruin your (battery) cells and burn down your house. Unfortunately, this is one of the cases where theory and praxis converge.

I'd recommend setting the CV to 4.1V initially, partly because you may not be able to set it spot on, partly because your circuit is temperature dependant and partly because Li-Ion won't tolerate more than 1% over charge (and you won't loose much capacity by doing so).
If all is well, you can always up the voltae a wee bit if the cells can handle it (harder to regulate down on a pile of ashes).

While normal charging is recommended to be 0.7..0.8C, you need to lower the max. current limiting for when the cell(s) are flat (if below 2.9V use 0.1C).

You need to add a circuit for cutting charge completely when charge current falls below 3% (0.03C), as lithium cells are very intolerant towards trickle charging.
If you need stand-by characteristic, re-apply charge only after a week or more.

That was the safety issues and well being of your cells, now onto the component selection...
You won't find a 250 Ohm or a 580 Ohm resistor outside your sim and component tolerances means you need a trimmer.
Your protection diode is supposed to protect what - against wrong polarity?
This will reduce your available voltage by around 0.7V to 1.2V (depending on current through it).
You cannot just use a resistor as a current limiter the way you do, it will F*up your charge voltage as well.

I'm not trying to discourage you in any way, just keep you from ruining your cells and burning down your home

Since you just asked if it would work, I assume you didn't want a solution.

I have a similar 6-cell set of fully good cells from my net book, just didn't care to spend the time it would take me hacking and resetting the electronic "lock" that kept it from working, since I found a 9 cell replacement (i.e. 50% more runtime - close to 12 hours between charges ) at a measly $35,-
For now, I just hand charge the loose cells with a lab supply, but I really have to knit an automatic charger soon... Too many projects ready for soldering up and while they can land a man on the moon and a ROV on Mars, there's still only 24h in a day   

[newInRobotics]

Hi Soeren,

Don't ever trust a sim unchallenged, simulators are akin to theory, only good as long as all parameters are known and they seldomly are.

I don't trust sim, therefore I ask knowledgeable people for advice 

I'd recommend setting the CV to 4.1V initially, partly because you may not be able to set it spot on, partly because your circuit is temperature dependant and partly because Li-Ion won't tolerate more than 1% over charge (and you won't loose much capacity by doing so).

I was thinking myself to set it a bit lower than 4.2V. The drawing in my previous post is just to illustrate initial idea in "perfect" environment.

While normal charging is recommended to be 0.7..0.8C, you need to lower the max. current limiting for when the cell(s) are flat (if below 2.9V use 0.1C).

Initially, I was willing to use 0.4C charging current just to be on the safe side (I might raise it slightly, because charging time on 0.4C would be quite long). I could implement secondary circuit for flat batteries, ore I could use 3.1V cut-off circuit that would go together with battery pack.

You need to add a circuit for cutting charge completely when charge current falls below 3% (0.03C), as lithium cells are very intolerant towards trickle charging.

My design is based on this one. And according to the author:

Note: Some people have commented that this circuit is not "Smart enough" to charge a Li-Ion battery properly. As long as you don't discharge the battery below about 3.0v per cell, this circuit follows the Panasonic recommendations exactly. Below 2.9v/cell, the batteries will need to be trickle charged (0.1 C) until they reach 3.0v/cell. Discharge below 2.3v/Cell will damage the battery. The circuit will not overcharge the battery. When the battery is fully charged, the current drops to zero (actually, the leakage current of the battery.) It will maintain the charged state forever. I have left cells on this charger for months.

I guess I need to prototype my charger and use one test cell to see what happens if I leave it on charge for longer time.

You won't find a 250 Ohm or a 580 Ohm resistor outside your sim and component tolerances means you need a trimmer.

I know  These are values for sim only, in practise I would use resistors in parallel and series to meet required resistance (Ohm) and heat dissipation (W).
What is the trimmer You've mentioned above?

You cannot just use a resistor as a current limiter the way you do, it will F*up your charge voltage as well.

How would it affect voltage?

Since you just asked if it would work, I assume you didn't want a solution.

Solution or at least guidance towards it is always useful. Share Your wisdom please

I'm not trying to discourage you in any way, just keep you from ruining your cells and burning down your home 

Thnaks for Your concern I do not want to burn my home, therefore I always ask first

[want2learn]

What is the trimmer You've mentioned above?

Google variable resistor or potentiometer

You cannot just use a resistor as a current limiter the way you do, it will F*up your charge voltage as well.

How would it affect voltage?

Use ohms law to calculate the voltage drop across the resistor. This'll change as your charging current changes.

Jim

[newInRobotics]

What is the trimmer You've mentioned above?

Google variable resistor or potentiometer

Why not just say so in first place?  Thanks.

Use ohms law to calculate the voltage drop across the resistor. This'll change as your charging current changes.]

I know how to calculate voltage drop across component, however I am not sure if understand what effect it has on charging voltage.
I initially thought that first current goes through battery, as battery charges up voltage drop across it increases because of rising resistance in the battery; input voltage still stays same 4.2V, however voltage that comes from negative pole of the battery drops gradually. As I said before, I do not really understand how resistor (or bunch of them) affects charging voltage?
Explanation (with an example to illustrate) would be appreciated 

[Soeren]

Hi,

Initially, I was willing to use 0.4C charging current just to be on the safe side (I might raise it slightly, because charging time on 0.4C would be quite long). I could implement secondary circuit for flat batteries, ore I could use 3.1V cut-off circuit that would go together with battery pack.

A too low charging current is almost as bad as a too high charging current.

My design is based on this one. And according to the author:

Note: Some people have commented that this circuit is not "Smart enough" to charge a Li-Ion battery properly. As long as you don't discharge the battery below about 3.0v per cell, this circuit follows the Panasonic recommendations exactly. Below 2.9v/cell, the batteries will need to be trickle charged (0.1 C) until they reach 3.0v/cell. Discharge below 2.3v/Cell will damage the battery. The circuit will not overcharge the battery. When the battery is fully charged, the current drops to zero (actually, the leakage current of the battery.) It will maintain the charged state forever. I have left cells on this charger for months.

I'm so glad you call him "author" rather than designer, because he's only got half a clue and that makes him dangerous  

I guess I need to prototype my charger and use one test cell to see what happens if I leave it on charge for longer time.

Don't waste your time until you have a working schematic. You may claim that you based it on the semi-functional one in the link, but you managed to screw it up good, if you pardon my French.

What is the trimmer You've mentioned above?

A trimming resistor, a variable resistor.

How would it affect voltage?

In a bad way!
You asking that question makes me think that you should get lots more experience before toying with potential fire hazards.
All else aside, you shouldn't modify any circuit with a possible outcome as this, if you don't even know basic electronics.

Solution or at least guidance towards it is always useful. Share Your wisdom please

Unfortunately my regular access is down and no stats on when it's fixed, so this is posted on a net book via the cellular network and it'll have to wait 'till my regular is back up, as everything is on the non-microscopic PC  

[Soeren]

What is the trimmer You've mentioned above?

Google variable resistor or potentiometer

Why not just say so in first place?  Thanks.

Because people capable of building a Li-Ion charger calls it a trimmer

Quite frankly, I'd advice against you building your own... Buy one and live to brag about it

[want2learn]

Ok newinrobotics, I know you'll think that soeren is being a little harsh but what he's saying is the truth.

If you make a mistake here were not just talking about a fried circiut. Lithium batteries are brilliant and in most cases so are their respective charging systems.
This is a bit like mains voltage in that it's been around for a while now and made safe, people tend to (or sometimes choose to) forget the dangers involved.

Up until soerens last post I'll be honest I skim read the page and missed this:
"I guess I need to prototype my charger and use one test cell to see what happens if I leave it on charge for longer time."

That worries me a little.
Get a little practice with a few things that don't have the potential to kill people or set fire to things and revisit this then (or even find a better schematic - There's a reason most r/c chargers for lithium batteries cost more!)

Jim

[newInRobotics]

I know that I am playing with fire  I do not blame Soeren for pointing out that I do not have enough knowledge to build what I want to build, I know that myself, otherwise I would not be asking questions about how voltage drop works.

On the other hand, people learn by making mistakes, and even if battery explodes and it's in the safe environment (built to withstand blowing li-ion battery) it is ok, that's the way to learn
In the end, I would know how battery explosion looks like

[waltr]

In the end, I would know how battery explosion looks like Smiley

You may regret that if you see it from too close of a distance.

[newInRobotics]

I am here again with my silly questions  
After some research I've came up with another circuit that in my opinion (of course it might be very much wrong) does not suffer from voltage drop. Here it is:



I wait for comments on this one. Biggest thanks

[Soeren]

Hi,

I am here again with my silly questions  

It's not your questions that's silly!
Your helter-skelter attitude towards the possible dangers of a high temperature lithium fire on the other hand...
Well, lithium is used medical too - too dampen "moon sickness", so perhaps swallow one of your cells     

After some research I've came up with another circuit that in my opinion (of course it might be very much wrong) does not suffer from voltage drop. Here it is:

For starters, you need a trimmer, as the in-range tolerance of the LM317 adjust voltage means that it can be anywhere between 1.20V and 1.30V, which would leave you with anything between 4.0V and 4.3V out.

When the battery is charging at ~1.25A, at an average voltage of ~3.5V, the input to U2 is around 6.5V. This means that U1 has got a voltage drop of ~13.5V @ 1.25V = close to 17W!
A bit much for an LM317, so better lower the input voltage to 10V (saving it from 12.5 of those Watts).

Your circuit could be improved by a BC547 or similar small signal transistor controlling the voltage regulating LM317 (i.e. U1 would not be needed and you could cut the input voltage down to say 7.5V).

You're still missing an end of charge detector and -cut off.

You're still missing a trickle on low cell voltage.

When you have those in place, it should be OK to use on a single cell, but trying to charge a multi-cell (i.e. 2 cells or more) battery without balancing for each cell is a big no-no!


Ready bought Li-Ion chargers have temperature control, max. time control etc. as extra precautions (and they don't add something just for fun).

You are aware of the many dedicated Li-Ion charging chips that implements all the security and balancing for multi-cell batteries, which would get you a better, more secure, charger without you needing to reinvent the wheel, right?

[newInRobotics]

Thanks for Your guidance Soeren.

Your circuit could be improved by a BC547 or similar small signal transistor controlling the voltage regulating LM317 (i.e. U1 would not be needed and you could cut the input voltage down to say 7.5V).

I've made small changes to my circuit (apart from cut off and trickle charge mode) according to Your guide. I am a bit confused with where I should place transistor? Is it in the right place?

[Soeren]

Hi,

I've made small changes to my circuit (apart from cut off and trickle charge mode) according to Your guide. I am a bit confused with where I should place transistor? Is it in the right place?

No, this would mean a voltage drop as well.

Study Q2 of this circuit
The circuit, as it is, won't work for a Lithium cell without changing component values - and adding the mentioned safety circuits, but it shows the current limit by a transistor).

[newInRobotics]

Thanks for guidance again Soeren.

Study Q2 of this circuit this circuit
The circuit, as it is, won't work for a Lithium cell without changing component values - and adding the mentioned safety circuits, but it shows the current limit by a transistor).

I made changes according to the diagram and now have 4.1V constant voltage with 1.3A limited current. As You suggested, input DC voltage is now 7.5V




I was surfing the net and looking for cutoff circuit for charge termination. What I've found was quite useful circuit for undervoltage protection that will be very suitable (with tiny modification) for my project.
I've also found a DIY balancing circuit.
However, it is not that important at this stage, as I was not able to find charge cutoff circuit.

I have a vague idea that I can use PNP transistor, when it's base (at the positive end of the battery) is < 4.1V (voltage created by R5 and R6) then PNP is on also keeping relay close (relay is open by default). When PNP base is > 4.1V then PNP swithes off also opening relay and braking the circuit.
I have a feeling that it is not the way to do it  

[newInRobotics]

Here is a circuit that is a bit more fit for the purpose (at least that is what I think).



R5 and R6 act as voltage divider that create 4.1V. When emitter of Q2 (PNP transistor) becomes higher than 4.1V (higher than base voltage) transistor turns on switching thyristor D1 on as well; current flows through the coil of relay switching it off (opening it), therefore circuit is broken and battery is not charged any more. As D1 gets activated it can be switched off only by interrupting current flow through it (by turning charger off).

Would such cutoff system work?

[newInRobotics]

Ok, I have a sleepless night (just because I think too much if one is better than the other)  

Another design that implement LM358 voltage comparator.



Feedback is realy important to me in this learning stage of mine
Do not be shy and point out what is wrong and why, and how to fix it (if possible).

[newInRobotics]

Ok, I believe this time I've got it right (fingers crossed)  



R7 is a current sensing resistor. According to Lithium Ion Rechargeable Batteries Technical Handbook produced by Sony (maker of batteries that I have) charging must be terminated when charging current drops to "1C/20mA", that means - for 1 cell that is 2000mAh, charge termination has to take place when charging current is 100mA (0.1A).

At 0.1A charging current voltage drop across R7 is 0.1A x 0.05ohm = 0.005V(5mV).
Thanks to Current Sensor tutorial video I've understood how to measure current in a circuit.

Op-Amp (LM358) takes that 5mV and amplifies to 3.5V.

Another Op-Amp (LM358) with a reference voltage of 3.5V (created by R5 and R6) compares it's two inputs and saturates to high side if reference voltage is lower than another input voltage.

Risen voltage activates thyristor (D1), hence letting current go through the coil of relay that in turn breaks the circuit and terminates charging.

Am I correct?

[Soeren]

Hi,

R7 is a current sensing resistor. According to Lithium Ion Rechargeable Batteries Technical Handbook produced by Sony (maker of batteries that I have) charging must be terminated when charging current drops to "1C/20mA", that means - for 1 cell that is 2000mAh, charge termination has to take place when charging current is 100mA (0.1A).

At 0.1A charging current voltage drop across R7 is 0.1A x 0.05ohm = 0.005V(5mV).
Thanks to Current Sensor tutorial video I've understood how to measure current in a circuit.

Op-Amp (LM358) takes that 5mV and amplifies to 3.5V.

Another Op-Amp (LM358) with a reference voltage of 3.5V (created by R5 and R6) compares it's two inputs and saturates to high side if reference voltage is lower than another input voltage.

Risen voltage activates thyristor (D1), hence letting current go through the coil of relay that in turn breaks the circuit and terminates charging.

Am I correct?

Had to redraw the schematic to see what's up and down.

For now, just a few comments...
Why do you use two different sensing resistors which are coupled in series?  (No need for R7 in your circuit, just sense over R1)
Why no gate resistor on the SCR?
With no feedback, your shifting points won't be well defined (temperature will change them a bit as well).
Did you see how I handled end-of-charge termination in the .pdf I referred you to? No reason to use a dual op-amp for that.

It's good to experiment, you'll learn a lot that way, but ultimately, you might wanna use a controller with A/D-C.

[Soeren]

Missed this...

[..] when charging current drops to "1C/20mA", that means - for 1 cell that is 2000mAh, charge termination has to take place when charging current is 100mA (0.1A).

You must have mistyped this, as C is the capacity (Ah or mAh) and  mAh/mA = h.
C/20, which is 5% or in your case and that sound like an OK termination current.
There is no reason for writing 1C as 1C, C and C1 is the same.

Here's some info that may make it easier to read battery and charging texts...
nC means n times C.
Cn means C divided by n.
The termination current can thus be written as C20.

[newInRobotics]

[..] when charging current drops to "1C/20mA", that means - for 1 cell that is 2000mAh, charge termination has to take place when charging current is 100mA (0.1A).

You must have mistyped this, as C is the capacity (Ah or mAh) and  mAh/mA = h.
C/20, which is 5% or in your case and that sound like an OK termination current.
There is no reason for writing 1C as 1C, C and C1 is the same.

"1C/20mA" is an actual quote from the datasheet 

[Soeren]

Hi,

[..] when charging current drops to "1C/20mA", that means - for 1 cell that is 2000mAh, charge termination has to take place when charging current is 100mA (0.1A).

You must have mistyped this, as C is the capacity (Ah or mAh) and  mAh/mA = h.
C/20, which is 5% or in your case and that sound like an OK termination current.
There is no reason for writing 1C as 1C, C and C1 is the same.

"1C/20mA" is an actual quote from the datasheet 

Well, always have an eye out for mistakes, even in data sheets and your interpretation of same.
I made a mistake as well in the previous post, as "[...] or in your case and that [...]" should have read "[...] or in your case 100mA and that [...]".

The real mistake here is a combo. They made a mistake similar to mine, but it was their strange formulation that tricked you...

Standard charge time for charging cells to their rated capacity is 2.5 hours at a
charging voltage of 4.20 V/cell and charging current of 1C. However, when charging
a battery pack the charging time (until the charging current falls to either
1C/20mA) will vary with the internal impedance of the battery pack and contact
resistance of the battery terminals. The differences in charging time for different
internal impedances and charging currents are indicated in Figures 34 and 35
respectively.

The "either" gives it away
They should have (and probably meant to) written "until the charging current falls to either 1C or 20mA

[newInRobotics]

The real mistake here is a combo. They made a mistake similar to mine, but it was their strange formulation that tricked you...

Standard charge time for charging cells to their rated capacity is 2.5 hours at a
charging voltage of 4.20 V/cell and charging current of 1C. However, when charging
a battery pack the charging time (until the charging current falls to either
1C/20mA) will vary with the internal impedance of the battery pack and contact
resistance of the battery terminals. The differences in charging time for different
internal impedances and charging currents are indicated in Figures 34 and 35
respectively.

The "either" gives it away
They should have (and probably meant to) written "until the charging current falls to either 1C or 20mA

To terminate charge at 1C or 20mA is to terminate charge at 100% or 1% (if no parallel cells are used), that does not make any sense.

So, have I understood correctly, charge has to be terminated when current is C/20? Does it mean that if I have battery pack of 4000mAh, charge has to be terminated at 200mA?

And another question, would my design (using 2 op-amps ,thyristor and current sense resistor) work (without taking into consideration that my circuit still does not have circuit to charge overdischarged battery)?