Simple DIY Li-Ion Rechargeable Charger Part-2

djpark

Enlightened
Joined
Nov 5, 2003
Messages
445
Location
SJ, Malaysia
This is a part-2 of the Super Simple DIY Li-Ion Rechargeable Charger. I have been busy with POP driver and decided to take a short break to cool down the head by making a few chargers.

The basic design is still based on the same Linear LTC4054 single cell li-ion charger controller, but now it has more indicator and I thought I would make a new post. The power is taken from a PC USB port. Please refer to the previous post for the detail.

Whatever was discussed in the previous thread is valid and please refer to it for reverse cell protection by Doug_S and MrAl as well as other charge logging chart and other information.

I will not go through the detail document as the first one, but this needs to be repeated.

I do not guarantee the unit I produce will meet your requirement. I disavow any potential liability for the contents of this document. No responsibility is accepted by me for any loss or damage caused in any way to any person or equipment, as a direct or indirect consequence of following these examples. Use of the concepts, examples, and/or other content of this document is entirely at your own risk.

ltc4054-regular-1-circuit.gif


This new design includes PNP transistor circuit to turn on the charge status LED instead of using the LTC4054 status pin with open drain output. When the charge is complete, the original status pin still let a few uA current to LED and it can be seen as ON when a bright led is used. PNP transistor based inverter removes this and now the charge indicator is ON only when really charging. No mistake..

There is a jumper which select the max constant charging current. With the jumper open, the charge current is 300mA, when closed it is 750mA. 750mA is about 0.5C for Pila and it can charge Pila within 3 hours.

D4 is not used, but it is a good idea to use a schottky diode to drop the input voltage by 0.5V so that the heat generated by LTC4054 chip is less. This reduces the effect of thermal protection by the chip reducing the charging current.

I used a perforated board with 18x5 holes size and arranged the parts as below. Sorry no CAD drawing as I have yet to learn one to do. But this hand drawing should give a good idea. The blue highlighted line is the thread and the red x is to cut.

ltc4054-regular-1-layout.jpg


When the cutting is done, the bottom should look like this (upside down).

ltc4054-regular-1-pcb.jpg


Then arrange the parts and solder. Now I solder the chip directly on the board instead of making an adaptor. This should be much easier for anyone.

ltc4054-regular-1-parts.jpg


Then slot into one side of the 2 AA battery holder and connect the wires. The whole board fits in very nicely. The hole for USB cable is tight, so it doesn't move or come out.

Since I ran out of the blue led, I used Nichia white led (the same one used in Arc-AAA) as charging indicator, it is bright. My fellow CPFer Stanley said this charger is the BRIGHTEST charger of all. How true!

ltc4054-regular-1-done.jpg


Though the holder is meant for AA battery, it can also hold R123. A spacer can be used to charge R123 in the same holder.

ltc4054-regular-1-battery.jpg


I didn't have space to put reverse cell protection circuit using N-channel FET as suggested by Doug_S. So I included just the reverse cell indicator. When the cell is inserted reverse, it turns on the red led.

Be very careful. Don't ever connect the battery reverse when there is power supplied to the charger. It can run as much as 1A (or more) to the charger chip and it may kill the chip, or even affect the PC operation. You have been warned.

The charger has output short circuit protection and it allows only 10% of the max charge current on the short circuit condition.

Any comments are appreciated.

-- dj
 
Just wanna add that the USB port can only supply 500ma max. /ubbthreads/images/graemlins/icon6.gif

Nice post djpark! Me hope I get to build one of these sometime...
 
Thats true legtu, but PC USB are usually configured to accept 2A divided by 4 serial USBs, so if the others devices are close to 500mA you will have troubles, but this is not very common.

nice DJ !!

Pablo
 
from experience.. no 'may'.. it WILL kill the chip.. been there done that.. be careful.

I was going to make a comment about the apparent reverse indicator.. but i re-read what you said.. and it means.. you can plug in the battery when it's not connected to usb.. and you'll get a red LED but won't blow the chip.. that's very smart and simple.

you should get some of the prototyping board for putting SOT-23 devices in SIP .. they would go right on a board like this.. although.. that little spider looking thing is absolutely adorable.

-awr
 
Hello again,

Nice circuit dj, and nice pics too.

Interesting timing, because i was about to embark on a new
Li-Charger circuit too. Features will be: common parts,
total reverse connection protection, ac-dc input, no
heat up even for 1 amp applications, input voltage range
6 to 24v ac/dc (with tiny heat sink), 'fairly' low cost.

I have something else very pressing just ahead of it, but
after that im going to work on it. I hope it works out
as planned.

Sound good?

Right now im using my LM317 based charger to charge my
AA Li-ion cell, because my bigger one is for the larger
cells which take 1 amp charge.

I hope to have pics too when i get my scanner back up and
running.

Maybe i'll build boards up and sell 'em for almost cost.

Take care,
Al
 
It is nice to hear from you MrAl.

The N-FET protection which you and Doug_S brought out really works. When the cell is reversed, it shorts the output of the charger chip and the controller reduces the current output to 10% of the preset max current.

I have an idea of monitoring the charge status with a PIC by monitoring various voltage of the pins and use tri-color led to display with different pattern. This will eliminate all the led related parts in this circuit.

I thought of using PIC to determine the reverse cell or short circuit status to enable or disable P-FET as protection, but decided that having a full time N-FET protection and PIC monitoring is much easier and better option.

This LTC4054 works fine, but it has max 800mA constant charging current limitation which is about 0.5C for 17670. So I consider making another charger for 2A charge capacity with PIC monitoring and controlling. No commercial value I guess, but it is a good pet project.

If I am not mistaken, your LM317 based circuit is originally meant for SLA charge, but the charging characteristic is very similar to li-ion charge except

1) no trickle charge when the cell voltage is below 2.5V,
2) no cut off charging when the charging current is reduced to 10% of max current.

To confirm to the rule of using common parts, you may want to use a voltage comparator to measure the series resistor current to control a P-FET to cut off when charging is done as well as reverse cell protection. (2 comparators?)

I find that the heat generated at regulator chip to drop 10V to 5V is manageable with a small heatsink. But you may want to consider the buck converter.

I found PT6302 from TI with these features. It has many usage which modders will drip sliva.

- 9V-30V input range
- adjustable voltage
- 3A output
- 90% efficiency
- integrated short circuit protection (N-FET reverse cell protection should work)
- on/off control
- 12 pin SIP format

Set it to produce max 4.2V output, then use MAX471 to measure the current and feedback to the voltage adj to control max charging current. Does this sound workable?

-- dj
 
Hi again dj,

Well thanks, and thanks for posting your ideas and stuff. I like
hearing about the PIC based stuff as sometime in the future i plan
on trying one of those out too.
A PIC to monitor charging sounds really decent and would of course
make a really great charger.
Making a charger that goes up to 2A would cover all the batteries
i have and then some /ubbthreads/images/graemlins/smile.gif My biggest battery takes 1A cc charge.
The LM317 based might have been originally for SLA, but it's not
limited to that application. I've used it successfully several
times to charge Li-ion's and it works great. Yes, you cant connect
a dead cell to it (under 2.5v) because there's no trickle charge mode,
but then i havent needed that in all the years i've been charging Li-ion
(about 3 or 4 years now). If you're going to do a PIC based charger
than of course im sure you'll include that function too though.
As far as the cut off current goes, when it gets below 10% it's down to
charging so slow it doesnt really matter if it stays on charge another
2 or more hours so i didnt take the time to design in any cut off.
I think that spec is just to allow users to take their cell off at
a reasonable time (rather than wait indefinitely) so they can use the
cell after two or three hours for example. If there was no spec
how would they know when to take off the cell /ubbthreads/images/graemlins/smile.gif I dont believe
it's for any sort of protection. If i was doing a PIC based charger
i'd still build this in however, just to let the user know it's
ready for use (led or something).

As far as 'common' parts, i see you're talking REALLY generic, and
although i wasnt talking quite THAT generic i think that's a great
idea too! Yeah a few common LM339's to control everything. This way
anybody could build one up because there'd be no problem finding parts.

QUOTE
I find that the heat generated at regulator chip to drop 10V to 5V is
manageable with a small heatsink. But you may want to consider the
buck converter.
END QUOTE

Exactly ! I was hoping to get around the heat generation problems
associated with my linear 1A charger, which connected to +9v gives
off a lot of heat when charging so it's a bit harder to package.
It doesnt get 'super' hot, but still a little too warm for me.
I'm aiming at something really cool running, or just slightly
warm at worst.

PT6302...
Oh ok, i'll have to take a look at that chip. Sounds very good.
So far it sounds like it could work, but i'll check out the chip
a little better and get back with another post. If the chip
isnt too hard to get or too expensive i could try it i guess.

BTW, i meant to ask...
Using the PIC, have you ever been able to monitor voltages
with the analog input that have a small range...say from
3.5 to 4.5v (one volt range) with increased resolution?
Or even say 4.0 to 4.3v (0.3 volt range)?

[ADDED LATER]
Ok, checked out the TI part, and it's not a chip it's
a whole board and it's 18 dollars on Digikey.
I guess it could work, but before i used one i think
i'd check the current backflow from the battery to the
board if the power became disconnected. It would have
to tolerate 4.2v on the output with no power input.
Other than that, the price is a little high...what
do you think?


Take care,
Al
 
[ QUOTE ]

BTW, i meant to ask...
Using the PIC, have you ever been able to monitor voltages
with the analog input that have a small range...say from
3.5 to 4.5v (one volt range) with increased resolution?
Or even say 4.0 to 4.3v (0.3 volt range)?

..

Other than that, the price is a little high...what
do you think?


[/ QUOTE ]

MrAl,

Most PIC ADC are 10 bit resolution and it would measure about 5mV resoultion using 5V Vdd as voltage reference. So you'd get roughly 60 levels out of 0.3V range. There are a few 12 bit ADC PIC as well.

But they all have max 10k input impeadance and it may not be suitable to measure the low current application.

It is possible to use an external ADC with higher resolution which is polled by PIC and also you can add an opamp to increase the input inpeadance.

It will all depend on where it is going to be used.

As for the TI part, they are very generous and you can alwayas ask for a sample chip since you'd want to use it for yourself.

-- dj
 
Hi dj,

10k is probably good enough for my purposes.

Ok, lets say i want 1000 steps (really 1024 but let's make
this more simple) between 4.0 and 5.0 volts. Can i set
the lower voltage range to 4.0 (and set
the upper 5.0 limit) so i get readings like this:

actual/PIC
0.000/4.000
1.000/4.000
2.000/4.000
3.000/4.000
4.000/4.000
4.001/4.001
4.002/4.002
4.003/4.003
.
.
.
4.999/4.999
5.000/5.000
5.001/5.000
5.002/5.000
.
.
6.000/5.000

In other words between voltages of 4 an 5 volts i get
resolution of 1mv (approx) but it cant read voltages
outside this range.
I guess an op amp would do this, but im wondering if the
pic can do it without an op amp.

Any ideas?

Take care,
Al
 
MrAl,

Different has different ADC features. Some offer builtin absolute voltage reference while most relies on Vdd provided. Some offer a single Vref to be used instead of Vdd as reference while many offer 2 Vref which you can set as min and max measurement.

With the last option, you can set Vref-=4.0V and Vref+=5.0V using external voltage reference and you can measure the voltage within the range with the given resolution.

So what you get will be like this. Let's mix 1000 and 1024 here for simplicity. (Really it should have been 1024 instead of 1000.)

actual/PIC
0.000/0
1.000/0
2.000/0
3.000/0
4.000/0
4.001/1
4.002/2
4.003/3
.
.
.
4.999/999
5.000/1000 (actually 1023)
5.001/1000
5.002/1000
.

PIC datasheet reminds that the Max Vain=Vref+0.3V. So you are not supposed to apply higher than 5.3V to the input.

Another thing to note is that the max AC source impeandance is 10K and recommended to be 2.5K which is rather low for sensitive data measurement. It can take as much as 500uA during the conversion. So if accuracy is one concern, you shall consider using an instrument opamp.

-- dj
 
Hi there dj,

Well that sounds good. I just wanted to be able to
measure voltages within a certain range to get more
accuracy for some things without resorting to an external
ad converter.

Take care,
Al
 

Latest posts

Top