polymer li-ion battery charging and charge status estimation

[zilog]

I have an application where I need to detect whether or not I still am in the lower 5% or so of total charge status for a polymer li-ion battery. I know that simple voltage measurement is not that good, but that's all I have to use (the PCB is already in production).

I know that voltage measurement will be good enough to know this part of the charge state, but what happens if I want to perform this measurement while I am charging the battery? (I have electronics that need to be certain that a minimum amount of energy is available from the battery should a cut-out of the charger occur)

If I have the battery sitting without a charger, I feel that I can trust the measured voltage, but what if I have a 0.5C charging current attached? Will the voltage float just like when charging/having charged a lead acid-battery? If this is the case, how long after disconnecting the charger does the condition remain until I can get a good measurement?

[jean-paul]

First let me state that the answers to the questions won't make a difference as you say the PCB is already being manufactured ! Not a safe decision I think. First test the prototype before manufacturing. When the product is tested and when it functions good and safe it can be produced. However you are not alone in this. It seems in fashion to do so and let the customer do the testing. When it does not work as sold the modifications can be done and again the customer has to pay as he/she is the one that operated the device wrongly of course.

Secondly Li-Ion/Li-ploy cells have to be charged and measured carefully (within 1% of the advised curve) otherwise the lithium will become a solid metal again which is a short for the charging circuit. So Boom ! And this is not a joke, see Youtube for the fantastic effects that can have. I suppose you know this but repeating it won't hurt.

Please check the various manufacturers of special charging ICs like Linear Technology. Those are the safest solutions and a few offer the functionality you are looking for.

[zilog]

Originally Posted by jean-paul

First let me state that the answers to the questions won't make a difference as you say the PCB is already being manufactured ! Not a safe decision I think. First test the prototype before manufacturing. When the product is tested and when it functions good and safe it can be produced. However you are not alone in this. It seems in fashion to do so and let the customer do the testing. When it does not work as sold the modifications can be done and again the customer has to pay as he/she is the one that operated the device wrongly of course.

Secondly Li-Ion/Li-ploy cells have to be charged and measured carefully (within 1% of the advised curve) otherwise the lithium will become a solid metal again which is a short for the charging circuit. So Boom ! And this is not a joke, see Youtube for the fantastic effects that can have. I suppose you know this but repeating it won't hurt.

Please check the various manufacturers of special charging ICs like Linear Technology. Those are the safest solutions and a few offer the functionality you are looking for.

You are reading too much into this.

I am hired as the SW guy that puts life into a pile of already manufactured HW. There is a BQ24025 USB charging circuit on the PCB which I can control on/off, and an AD converter that measures across the battery terminals. There is thus no risk of overcharging, as the charging circuit performs constant current/voltage charging according to standard.

What I need to know is how to measure when the battery has enough charge status to be able to fully boot up, and boot down the electronics that it powers. Until this charge status has been reached, I can not switch on the power to the rest of the system, as things will break if I dont have enough coloumbs in the battery to safely shut down once I detect undervoltage.

The time from attachment of the charger, to the time when the system comes alive needs to be as small as possible, thus I can't wait until the charging circuit indicated "charging done" after 3h or so.. and I only need a charge state that corresponds to minutes of charging time. The problem is that if I charge a single cell battery that has 2.6V voltage, it reaches 3.3V within seconds - 3.3V during normal use means that I have lots of power left, but in when I come from a discharged state, it means that I have almost zero power left. This is my problem.

[jean-paul]

OK, I am glad things are safe. I was reading too much in this because of the given information. The important bit of your problem is a standard problem with Lipo cells. Please see the curve here:

Lithium Polymer Cells

Also from the web as I can't put in better in english:

Li-po batteries don't charge/disscharge in a linear rate. NiCD and NiMH have a more or less linear charge/diss slope so that at 50% charge the voltage is ~50% of usable voltage. Li based batteries have a shallow slope until nearly disscharged then there is a sharp drop. Charging is just the opposite. So when charging is at 75% the voltage level is very close to 100%. 80%, 90%, 99% are all nearly the same. The charge monitor IC can monitor these minor differences and at the same time monitor battery temp.


The currently used charging IC is not capable of external monitoring. Seems like you will have design a new Li Ion/Lipo monitoring circuit. Or you could implement the exact curve of the used cells in software combined with the A/D circuit. So you probably will need the specs of the used cells. I think you will want to measure when the cells are not being charged as the voltage differences are already very subtle. Measuring while charging will not give correct information AFAIK. So a temporary disconnect will be necessary.

[zilog]

Ok, looks like I will need to use a timer of some kind for safety then..

[jean-paul]

This chip has some of the features but it also leaves the possibiity of running the device without battery. I do not know if you want that:

http://datasheets.maxim-ic.com/en/ds/MAX8671X.pdf