« An Engineer in Wonderland - Measuring stuff in the steam age | Main | An Engineer in Wonderland - turning off an LM334-based led driver »

An Engineer in Wonderland - Li-ion charging with an LM317

LiIonCharger250p.jpgSee all 'Engineer In Wonderland' posts

There are an awful lot of Li-ion charging chips out there, and you would have thought there was one for every eventuality.

But it turns out there isn't.

For example, there is a sort of no-designer's-land between 6.5V where linear Li-ion chargers top out - although Linear's LTC1734 works up to 8V - and around 12V where switching chargers start. 

And even without an input voltage restriction, what if you don't want to charge the thing at less than 4.2V?

Why charge at less than 4.2V? - Scroll to the bottom

Or what if you don't want to be tied to one chip maker?

I want to charge a Li-ion cell to 4.0V from somewhere around 7.5ishV. 

So I had a good dig around and found a 1976 Nat Semi app note that uses the good old LM317 variable output voltage regulator to charge a 12V battery.

At this point, you have to promise to read and understand the safety note at the very bottom of this entry.

Thank you.

Now, it turns out that providing you don't short the input or stick a big capacitor there, this well-behaved chip will happily sit with a voltage on its output without blowing up or otherwise going mad.

So if it is fed from through a diode to isolate the input, it is a good constant voltage battery charger - particularly as the accuracy is easily good enough.

Different versions are available with different current limits, so it is a constant current charger as well - Li-ion cells need a CC/CV charging regime.

Admittedly, the LM317 can hardly be described as low drop-out - which is just under 2V at 100mA for the 'L' variant.
But I have 3.5V to play with - room for a couple of volts and the diode.

Part of the genius of this chip is that almost all of its 5mA operating current flows out of its output.
Which does mean the voltage-setting potentiometer needs to swallow all of this if the chip is not to over-volt the battery when charging current drops below this, which it surely will with a Li-ion cell.

LiIonCharger250p.jpgTo prevent the potentiometer draining the battery when no input power is available, the app note recommends an NPN transistor - fed via a separate voltage to prevent back-flow from the 317 keeping the transistor on.

317 and resistor tolerances can add up and may mean the 560 Ohm resistor needs changing to set the output voltage correctly.

The NPN does add a small error voltage to the output, but its saturation voltage at 5mA is going to be 10 or 20mV max.

The NPN base resistors have to be chosen to ensure the transistor is well saturated before the input voltage is high enough for the output to hit 4.0V - or the output could be too high.

Now all I have to do is find time to try it.

THE BOTTOM

The cycle life of a lithium ion cell is strongly dependent on charge voltage - as well as temperature and several other things.
Graphite anode cells are usually charged at 4.2V, giving them a life of around 600 cycles.

According to the 2003 paper 'Current, voltage and temperature govern Li-Ion battery charging' written by National Semiconductor applications engineer Khosrow Khy Vijeh, life goes up to something like 2,500 cycles by stopping at 4.1V, and well over 5,000 cycles at 4.0V.

At the same time, compared to a 4.2V cell, storable energy per litre drops by 15% for 4.1V and 35% at 4.0V.

You don't want to go the other way as charging at 4.3V gains 15% capacity while dropping life to 200 cycles, and will probably trip internal protection circuits.

By the way, charging to 4.1V is an option on a few chips as it is the standard charge voltage for coke anode Li-ion cells, and I can't find the reference but I remember a chip with a lower output voltage specifically for long-life applications.


THE VERY BOTTOM - A SAFETY NOTE
Please don't go making Li-ion chargers if you don't know what you are doing.
The correct maximum current and maximum voltage limits are essential for safe Li-ion cell charging.
The cells are full of:
a, flammable liquid
b, energy
And 'b' is inclined to spontaneously ignite 'a' if you over-charge them beyond a limit that it remarkably close to the normal charging level.
Take a look at www.youtube.com/watch?v=WeWq6rWzChw
They also spontaneously ignite if they get too hot.
The US Consumer Product Safety Commission has put together a particularly useful public advice sheet on safe Li-ion use
www.cpsc.gov/CPSCPUB/PREREL/prhtml06/06271.html

'Alice'

Respond below or to alice@electronicsweekly.com

By the way, EW does not have a mechanism to spam you* if you replay to An Engineer in Wonderland - and that is how EW likes it.

* yet, and hopefully never.

TrackBack

TrackBack URL for this entry:
http://www.electronicsweekly.com/cgi-bin/mt/mt-tb.cgi/34471

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

About

This page contains a single entry from the blog posted on September 4, 2008 2:05 PM.

The previous post in this blog was An Engineer in Wonderland - Measuring stuff in the steam age.

The next post in this blog is An Engineer in Wonderland - turning off an LM334-based led driver.

More posts can be found on the main index page or by looking through the archives.

Don't miss any important developments. Sign up for the weekly editorial eNewsletter. Get highlights of news and blogs straight to your email inbox, no fuss. Tick the option for Weekly News.

RSS Subscribe to this blog's RSS feed
[What is this?]

Archives