My first post here & it will become clear fairly quickly that I'm no E'ee but I can put together a circuit from a schematic & understand basic electronics.

Now, I want to do a simple LiFePo4 battery charger for the 26650 cells (2300mAh) . I know a bit about these batteries. I know that the ideal charging algorithm is a constant current charge until a set point (3.65V) & then constant voltage (3.65V) until the current reduces to 1/10 of the Battery C (2300mAh). I have looked at a LM317 + TL431 circuit designed for Lithium battery charging & does CC & CV charging. But the LM317 overheats & goes into thermal limitation. I didn't want to have to use heatsinks as there is not much room in the box that this charger is going into & I'm also concerned about the heating effect on other circuits in the box. I

So I don't want to use a switching circuit.

Anyway, I just read that if the charging current is kept to < 0.18C then when the battery reaches 3.6V it will be 99% full. Can anybody confirm this?

 

Here's the CC/CV battery charger circuit but if using a highish current, it overheats the LM317. If using a low current I'm not sure how long it will take to charge the 2300mAh battery as it works by using a reducing current once in constant voltage mode!

So I wanted to use the simpler low current trickle charger with automatic cut-off. I can set this to < 0.18*C which is 0.18*2300 = 414mA current. About 300mA would be fine & would charge a completed depleted battery in 8 hrs or overnight which is my operational target

 

Can someone explain to me what the "minimum load condition" is - as mentioned in the passage below from the LM317 datasheet? I suspect I'm having this problem as I don't get the calculated voltage when I use the LM317 as a simple voltage regulator as per the datasheet. I'm using 2K for the top R (R1) & 4K7 for the bottom R (R2). I calculate 4.19V & read 6.6V with a 2K load on the output.

The LM117/217/317 provides an internal reference voltage of 1.25 V between the output and adjustments terminals. This is used to set a constant current flow across an external resistor divider (see Figure 3), giving an output voltage VO of:
VO = VREF (1 + R2/R1) + IADJ R2
The device was designed to minimize the term IADJ (100 µA max) and to maintain it very constant with line and load changes. Usually, the error term IADJ × R2 can be neglected. To obtain the previous requirement, all the regulator quiescent current is returned to the output terminal, imposing a minimum load current condition. If the load is insufficient, the output voltage will rise. Since the LM117/217317 is a floating regulator and "sees" only the input-to-

 

The "minimum load current" for an LM317 is 10mA. It is provided by the 120 ohm resistor from the output to the ADJ terminal and has 1.25V across it so its current is 10.4mA.
If there is not enough current and the input voltage is high then the output voltage will rise.

 

OK, thanks for that. The 120 ohm you talk about I don't see - is this on the datasheet? But I also see 240 ohm as the top R in the voltage divider - this would give only 5mA - I'm confused again.

Also in the CC/CV circuit the top R is 6K8 but it uses 1K between OUT & ADJ - are these paralleled - still will be below the minimum laod current? Does the TL431 provide the current to the ADJ pin?

 

240 ohms is used for the more expensive LM117 that is shown in almost every schematic in the datasheet.

The circuit you found using a TL431 is wrong. The other circuit is correct and has a 120 ohm resistor from the output of the LM317 to its ADJ terminal.

 

240 ohms is used for the more expensive LM117 that is shown in almost every schematic in the datasheet.

The circuit you found using a TL431 is wrong.

Aha X 2, thanks

The other circuit is correct and has a 120 ohm resistor from the output of the LM317 to its ADJ terminal.

OK, so this explains my issues - I should have checked here first I'll go back to putting together a circuit that I can understand & operates as predicted.

 

Can I ask some more questions if I'm not too much of a pest at this stage?
In the second circuit the charge reduces to a trickle charge when battery is full - is this trickle determined by the 0R56 - Oops I calc this as 2.2A? What am I missing?

Edit: I guess I don't want to use this circuit anyway as it has the same reducing current output & makes it difficult to calculate how long it would take to charge a battery,

 

Here's my final circuit that seems to fulfill all my criteria:
Current can be set to < 0.18C & is constant so it makes charging time easy to calc.
Cut-off voltage is defined by zener.

It looks simple & easy to understand - I hope I'm not missing anything?

Back to my original question - battery charging @ a constant current < 0.18C will result in 99% full battery charge when the voltage reaches 3.65V. Anybody got experience with this?

 

The maximum charge current is set by R1. You can roughly calculate the value for R1 by:
R1 ~= 0.63v/Desired_Current
So, if you wanted 500mA max current, R1 ~= 0.63/0.5A = 1.26 Ohms.

I built something similar to the 2nd schematic to use to recharge a 3.6v NiCD battery at a slow rate. See the attached.

 

Here's similar to my last post, but "tweaked" for roughly 400mA and 3.65v.

I'm not so keen on your schematic with the Zener diode; as Zeners have a tolerance from the manufacturer - and if you're not close to Izt, the voltage may vary a good bit.

 

Thanks Sgtwookie,
Much appreciate the tip - so a zener is not such a good idea. It's this sort of expertise I appreciate here. As I said I'm not an E'ee & have no experience of various parts & their vagaries Pity that circuit is a no-no - it was simple enough that even I could understand it without asking any questions.

OK, so looking at your circuit (& many thanks for tweaking it for me ) - where does the 0.63V come from? - a characteristic of the transistor?

How would I calc the time taken to fully charge a 2300mA battery?

 

Can I ask some more questions if I'm not too much of a pest at this stage?
In the second circuit the charge reduces to a trickle charge when battery is full - is this trickle determined by the 0R56 - Oops I calc this as 2.2A? What am I missing?

Edit: I guess I don't want to use this circuit anyway as it has the same reducing current output & makes it difficult to calculate how long it would take to charge a battery,

A lithium battery charger is supposed to completely turn off when the battery cell is fully charged. The battery cell might catch on fire if it is trickle charged. This old circuit is not good.

The base-emitter voltage of the extremely old BC140 transistor is about 0.7V so when the charging current tries to exceed 0.7V/0.56 ohms= 1.25A then the transistor is turned on and it reduces the voltage and current of the LM317. The current is limited to 1.25A.

 

A lithium battery charger is supposed to completely turn off when the battery cell is fully charged. The battery cell might catch on fire if it is trickle charged. This old circuit is not good.

The Lithium Ferrous Phosphate (LiFePO4) formulation is not anywhere nearly as reactive as Lithium ion batteries & is much safer - it will not catch fire. However, over-voltage will kill it. So it is good to either turn off the charging completely or keep a constant voltage of 3.65V

The base-emitter voltage of the extremely old BC140 transistor is about 0.7V so when the charging current tries to exceed 0.7V/0.56 ohms= 1.25A then the transistor is turned on and it reduces the voltage and current of the LM317. The current is limited to 1.25A.

So this reduces the current to negligible levels, I presume?

 

Thanks Sgtwookie,
Much appreciate the tip - so a zener is not such a good idea. It's this sort of expertise I appreciate here. As I said I'm not an E'ee & have no experience of various parts & their vagaries Pity that circuit is a no-no - it was simple enough that even I could understand it without asking any questions.

Well, Zener diodes have a tolerance from the factory - and there are many different Zeners with different Izt ratings. Better to use a regulator IC that has all of the references internal.

OK, so looking at your circuit (& many thanks for tweaking it for me ) - where does the 0.63V come from? - a characteristic of the transistor?

Yes. The cutoff voltage is generally when Vbe is around 0.5v (the collector current is so small that the transistor is considered to be in cutoff) it starts conducting in the mA range when Vbe (voltage on the base referenced to the emitter) is roughly 0.63v at room temperature.

So, R4 sets the maximum current limit. When the battery voltage comes close to the terminal voltage (3.65v), the regulator decreases the current output by itself. The nominal 10.4mA current keeps flowing through R1/R2/R3/R4 even when the terminal voltage has been reached; this is to prevent the output voltage from rising further, as it provides the minimum load necessary for guaranteed regulation.

Note that the input voltage V1 can be as low as 5.65v. Operating it from a 6v DC supply would be good. More than 6v, and you will start dissipating power in the LM317, which will be inefficient and will require a larger heat sink to get rid of the heat.

How would I calc the time taken to fully charge a 2300mA battery?

It depends how deeply discharged it was, what the charging efficiency is (I don't know; lead-acid batteries are ~70% to 85% efficient), the actual condition of the battery, the battery temperature and possibly other variables.

You should not discharge the battery below 2.8v.
When it reaches 3.65v, it will be charged.

You should set the terminal charge voltage using a small capacitor in parallel with a resistor (say, 1k) instead of an actual battery.

 

So, R4 sets the maximum current limit. When the battery voltage comes close to the terminal voltage (3.65v), the regulator decreases the current output by itself. The nominal 10.4mA current keeps flowing through R1/R2/R3/R4 even when the terminal voltage has been reached; this is to prevent the output voltage from rising further, as it provides the minimum load necessary for guaranteed regulation.

Note that the input voltage V1 can be as low as 5.65v. Operating it from a 6v DC supply would be good. More than 6v, and you will start dissipating power in the LM317, which will be inefficient and will require a larger heat sink to get rid of the heat.

Excellent, I'm looking for lowest thermal footprint - how does this work? I thought 3V was the voltage headroom required for the LM317?

It depends how deeply discharged it was, what the charging efficiency is (I don't know; lead-acid batteries are ~70% to 85% efficient), the actual condition of the battery, the battery temperature and possibly other variables.

Sure but what I'm trying to get at is the worst case scenario i.e I wanted to make sure it could recharge a battery from depletion in about 8 hours. I guess I'm asking how long it takes to put out 2300mA? Does PSpice allow this sort of calculation?

You should not discharge the battery below 2.8v.
When it reaches 3.65v, it will be charged.

You should set the terminal charge voltage using a small capacitor in parallel with a resistor (say, 1k) instead of an actual battery.

Thanks for the tips, much appreciated.

You don't happen to have a simple low voltage sensor circuit? Use a comparator?

 

Excellent, I'm looking for lowest thermal footprint - how does this work? I thought 3V was the voltage headroom required for the LM317?

That's if you're using it specifically as a current regulator. I'm using it as a voltage regulator. In voltage regulation mode, it has a minimum dropout of ~1.7v. I added 0.3v to that due to the current draw.

Sure but what I'm trying to get at is the worst case scenario i.e I wanted to make sure it could recharge a battery from depletion in about 8 hours. I guess I'm asking how long it takes to put out 2300mA? Does PSpice allow this sort of calculation?

Sure, if you program it all in. I don't know all of the variables about that type of battery though. It'll charge at roughly 400mA, decreasing as it approaches 3.65v.

You don't happen to have a simple low voltage sensor circuit? Use a comparator?

Sure, you could do that. I'm all out of time, though.

Perhaps someone else will help you with it. It will increase your footprint.

 

Thank you very much for putting in all this time & the help you've given Sgt. I will build this circuit during the week & report back here!

 

I tried this circuit & it works but I have a thermal issue with it. At any reasonable charge current 0.3 or 0.4A the LM317 overheats. It's probably because I'm running this from a 12V external DC supply - I imagine the heat is the dissipation needed to step down to 3.5V. I can't have this heat in the small box that holds the battery, charger & circuit board that is run from the battery.

So I've decided to go another route: I've decide to use a buck converter based board on the MCP2307 - these can be picked up cheaply & with a bit of modification can be adjusted from an output voltage of 3.3V to 3.6V (the max voltage for LiFePO4 charging). These boards are not reducing current however so I need to implement a high voltage sensor cut-off. The MCP2307 has an enable pin that when high turns on the regulator.

So I wanted to implement a simple comparator based voltage sensor that would turn off the MCP2307 when the battery reached 3.6V. I also wanted one that signaled if the battery fell below 3.0V. I have LM393 comparators in stock

Here's what I came up with:
I'll need two comparators (one LM393) & TL431:
- I'll use the 2.495V reference voltage of the TL431 as the ref input to the comparator.
- I'll use a voltage divider to bring the battery overcharge 3.6V threshold down to 2.5V - this will be fed in on the - of the comparator, the 2.5V ref goes into the + pin of one comparator. So when the battery voltage reaches 3.6V the comparator output goes low - this is connected to the enable pin of the MCP2307 which will turn it off.
- I'll use another voltage divider to bring the battery low voltage threshold 3.0V down to 2.5V & feed this in on the + of the comp & 2.5V ref on the - pin of the orther comarator. This will turn on the enable pin.

Does this sound correct?

I'm not sure how to connect up the TL431 just as a reference voltage?

 

Hi, i am new to this forum.
I have no access to the first picture in post #2. While i can see the second one.
I wonder why.