Battery Charger Design

NO unless you comply EXACTLY with the conditions of that number in the spec. Not likely.

To design a battery charger, you have to disclose a LOT about how the battery is used. Simplest thing is to charge it at 4.2ma. Spec at the link says it will last 6-years being abused like that. But that may not meet your cycle requirements, that you haven't stated.

There is a lot more to charging batteries than that !

70mA is the fast charge max current and you must not charge continuously at that level. Varta recomend voltage controlled charge. Look at and find the LTC4060 data sheet - this may not be suitable for your application but it will give you some ideas. The Varta data on the Farnell web site is very sparse - look at some other battery data and apps notes to find out more.

Michael Kellett

Hi Mike, thanks for the reply.

The battery will be discharged at around 50mA over 3 hours, then re- charged as fast as possible. Its for a portable device running a load of LEDs (like a bike light)

A long life needs to be maintained because the battery will be permanently attached to the pcb.

Does that inspire any more help? Cheers :)

Often, the most difficult part of a project is writing the spec. Words like "as fast as possible" will get you loads of input that may be VERY bad tradeoffs.

Still insufficient information, but... Unless you have some SERIOUS size/weight constraints, your battery is probably too small.

What's your definition of "discharged"? If you mean voltage cutoff at a level that guarantees none of the 5 internal cells gets reversed, you have a chance of success. If you mean run it 'till the light goes dim, your battery will NOT have long life.

Charging is problematic. Depending on the NUMBER associated with "as fast as possible", you may have no option but to use a smart battery charger chip.

I'm assuming this is a production device that requires agency approvals and you care about customer retention. If it's an advertising giveaway device, put lithium coin cells in it and ship it.

Replaceable AAA-cells that the customer has to recharge externally will save you a LOT of grief. Customers cuss the internal battery when the light goes dim at midnight when they're half-way home.

..

. y

The battery is for a research project that involves small plastic glowing cubes, so the size/weight constraint is beyond AAA batteries.

By "fast", I'm only referring to a sensible charge time that would be convenient. e.g. 8 hours is acceptable.

The cubes only need to operate for around 3 hours between charges, and this is an AVR uC application so the system will switch off when the voltage drops below 4.5v. Therefore the battery would not be fully discharged (I assume).

The charger chip approach sounds the most plausible, what would you recommend?

Thanks

--
That site's don for maintenance.

What's the Varta part number?

JF

The datasheet is here:

cheers

Manufacturer Part No: 55615605940

"thorin92"

The battery is for a research project that involves small plastic glowing cubes, so the size/weight constraint is beyond AAA batteries.

** You should consider another option - Farnell have 2/3AAA size NiMH cells (solder tagged too) that are far more rugged than your button cells and have *double* the mAH capacity ( 300mAh against 150mAh ) and four times the max discharge rate.

Size and weight are only slightly more at 45gms against 32gms.

Cost is about the same for either option.

The cubes only need to operate for around 3 hours between charges, and this is an AVR uC application so the system will switch off when the voltage drops below 4.5v. Therefore the battery would not be fully discharged (I assume).

** At 0.9 volts per cell

- they certainly would be FULLY discharged.

.... Phil

One thing to ask is, "how long does the research project last?" Doesn't make sense to go to a lot of trouble if it only has to last for a few months and a small number of recharges. Just charge it at C/10 and use a timer so it doesn't cook too badly. You can use the AVR to implement the timer internally.

There are some considerations you may not have included in your analysis. I'm a PIC guy, so I can't speak for AVR, but assume similar issues apply.

Have you considered that the maximum voltage on the battery during charge may exceed the supply voltage your AVR can tolerate? Don't forget to include the increased internal resistance of an aging cell.

4.5V is 0.9V/cell. That's below what's normally done, but not extreme. BUT, for such tiny cells at high discharge rates, it doesn't take a lot of imbalance to create problems. 4 cells at 1.2V can reverse the fifth one at 4.5V. It's not a matter if IF the cells become imbalanced, just WHEN.

Depends on your actual discharge profiles, but for a timed charge, you might set a flag if the 4.5V shutdown was triggered and apply a longer equalizing charge for the next charge cycle.

If you open up one of the NiMH 9V batteries, you find this type of cell construction inside. I've never had a good experience with rechargeable 9V batteries. They always leak or short.

I don't have any experience fast charging such tiny cells, but you might look into 0deltaV charging. You have a processor, use that to monitor the charge. There are A/D converter resolution issues. You're probably gonna have to turn off the charge current to measure the volts.

/...

am

he

ed.

e

ry

s

away

nced,

Hi Mike. 'Thanks for the detailed reply.

I'm not really looking for a sophisticated method of charging, just a basic Voltage Regulator or Charger IC. Essentially the charger is external and will simply connect to a 9V power supply (wall-wart).

I'm assuming that to achieve a constant current, the charge voltage is variable. I've been looking at:

To implement the constant current source. However, I'm concerned with any problems this method may involve. e.g.

-will the battery accept a voltage higher than 6v across the terminals?

-will charging stop automatically?

My apologies but this area is new to me. Thanks

"thorin92"

I'm not really looking for a sophisticated method of charging, just a basic Voltage Regulator or Charger IC.

** The problem is that " basic" chargers tend to ruin cells and batteries very quickly - by overcharging them and permanently altering the internal chemistry.

I'm assuming that to achieve a constant current, the charge voltage is variable. I've been looking at:

To implement the constant current source. However, I'm concerned with any problems this method may involve. e.g.

-will the battery accept a voltage higher than 6v across the terminals?

** It really has NO choice in the matter .....

-will charging stop automatically?

** No way will that happen, unless a very clever circuit makes it happen.

My apologies but this area is new to me.

** Not too hard to tell that one ..... which makes it near impossible to educate you quickly enough.

So your best bet is NOT to try to be too clever or "high tech" in your choices - rather go conservative, use slow ( ie 10 hour) charging and cells sized to have more capacity that you actually need.

A safety margin can be great asset against the unforseen ...

.... Phil

You didn't answer my questions, so any info is a shot in the dark.

What's your definition of wall-wart? If it's a regulated wall-wart, you can count on 9V if you comply with minimum load specs if any. If you mean unregulated like you get from radio shack, the output voltage can be VERY much different from 9V depending on the load.

I've been looking at:

Putting a constant voltage across a NiMH battery is a good way to make smoke. Stick in current and let the voltage go where it wants. The problem is when to STOP charging. You need to provide some method to stop the process when the cell is charged.

There's a quick and very dirty way to limit the current. Use an incandescent light bulb. For a relatively constant line voltage and a particular wall-wart design, a light bulb with enough volts across it to make it glow red will give you constant-enough current over the range of interest. Takes some experimentation. You can't just look up specs and decide what to do. The parameters you want aren't specified. I use #382 light bulbs just because I have a box of them. This would be considered by most to be a BAD design, but it can work in well constrained cases.

But you still want to STOP charging somehow.

You didn't respond to my caution about overvoltage. You didn't say which AVR processor you're using, but all the ones I looked at are specified for a maximum of 5.5V. Your battery will be over 5.5V while charging...and during the initial portion of the discharge. It's not unusual for a used cell to go to 1.6V during charge. X5 that's 8V. Even at 1.4V/cell, you're still WAY over 5.5V. Unless you're doing something to regulate it, you risk blowing the thing up. If you ARE regulating it, there are issues in that department.

Electronic design is a science. Product design is an ART. There are many, many issues that are obvious to an engineer skilled in the art that will bite you in the A$$.

The battery is one of the biggest problems in a battery-operated design. Sounds like you need somebody local to help you design this.

In the application, the output of the battery is regulated by an LDO Voltage Regulator before powering the AVR, so the AVR supply is of no concern.

The DC supply for the charger will be heavily filtered and regulated, powered from a mains transformer.

I do not wish to use an AVR to manage charging, although if this is a necessity for implementation of a timer then I guess its compulsory.

In the battery specs it states fast charge 70mA for 3 hours. Thus if I build the LM317 regulator with current control to 70mA this should suffice, but how would I know when the battery is charged? Is there some indication from the terminal voltage with this type of battery?

Alternatively, a trickle charge of 4.3mA implemented using the LM317?

Best

Yes it does. If you start with a completely discharged battery, you can charge it at 70ma for three hours...not four or five or six hours...THREE. You tell by measuring the time. If you overshoot by too much, the cell will vent and you'll get poor reliability. If it's not fully discharged, you can't charge it for three hours. If you know the discharge current, you could time it and put back what you took out times some efficiency ratio. Could all be done with internal timers

Charge termination is one of the more difficult parts of using rechargeable batteries. There is no free lunch. If you fast charge it, you shorten the life. If you overcharge it, you shorten its life. I already suggested you look up "zero delta V" charge termination strategies.

You can get clever, but not unless you (we) know all the details.

Thus if I

You're all over the map with your requirements. 4.3ma won't get you

8-hours charge time.

I'm trying to help you, but you're not answering some critical questions. How long does this thing have to work? You say it's a research project. If it needs to function for six months, you have more options than if it is expected to run reliably for years. Expected life under various overcharge conditions are right there in the spec.

Your choice of battery is causing you more grief than if you could tolerate a bigger one.

You're trying to make this much simpler than it is. Wishing won't make it so.

Stuff should be as simple as possible, but no simpler...

Excuse me, but I feel a speech coming on...I can't stop it...here it comes...

System design is an interactive process. You've asked for help on a specific part. That's ok, that's what you're getting.

It is very common for an engineer to design a system that he thinks should work...but creates some difficult problems in one area or another. Often, eliminating the tunnel vision and re-examining the overall solution results in much easier overall implementation.

But not knowing what you're doing prevents us from suggesting things like using three cells large enough to give you the run time you need AND accept overcharge that eliminates most of your recharge issues, eliminating the ldo and using one port on the AVR to implement a boost converter to drive the leds. But there may be other issues that prevent that.

Don't mind me...I'm easily frustrated...

n

le

en

ould

n

Hi, no problem and thanks for the help.

Unfortunately, the Battery has been decided and it cannot be changed. I have tested the discharge rate/capacity for the spec and all is fine. The device will operate for 4-5 hours until the AVR brown-out kicks in and shuts down the system at 4.5v. As I stated before, I assume the battery is not fully dis-charged at this point.

I would like the Battery to last as long as possible, and a trickle charge at 4.3mA is acceptable.

Despite the years of experience I am obviously missing, I do want a simple system, regardless. Since this is for a research project, the charger design is a secondary priority, thus I'm not interested in its sophistication provided that it works, safely, reliably, and charges the battery.

My basic question at this point therefore, is; will the 317 reg current limited to 4.3mA work? even if it takes 24 hours? Or is there something I should be aware of using this method? That is the only question I am asking.

Best

Find some actual manufacturer's recommendations with algorithm for these NiMh cells.

dV/dt charge termination does not work if the current is too low (typically trickle charging is not recommended)

dT/dt charge termination doesn't work if the current is too low (typically trickle charging is not recommended)

You really need to look at cell voltage *and* temperature and control current at least to one of several levels (pre-charge, fast charge, and top-up) to reliably handle these things. If the cell voltage is too high or too low, or the temperature is too high or too low, you do different things, and if the rate-of-rise of cell temperature is high enough you can terminate fast charge.

Perhaps not smoke, but a nice spray of caustic electrolyte. Past a certain point there is an internal "meltdown" and the reaction continues without externally applied current, at least with some cells.

"thorin92"

Unfortunately, the Battery has been decided and it cannot be changed.

** A choice made in ignorance of the consequences is NOT engineering.

You have painted yourself into a very small corner.

I have tested the discharge rate/capacity for the spec and all is fine. The device will operate for 4-5 hours until the AVR brown-out kicks in and shuts down the system at 4.5v. As I stated before, I assume the battery is not fully dis-charged at this point.

** And you have already been told that is a WRONG assumption.

I would like the Battery to last as long as possible, and a trickle charge at 4.3mA is acceptable.

** You have no idea what you are crapping on about.

Is a 60 to 80 hour charge time really OK ????

Despite the years of experience I am obviously missing, I do want a simple system, regardless.

** You are starting to sound like an arrogant fool stamping its foot.

You must in software design - right ?

Since this is for a research project, the charger design is a secondary priority, thus I'm not interested in its sophistication provided that it works, safely, reliably, and charges the battery.

** But have not the tiniest clue what the issues are.

My basic question at this point therefore, is; will the 317 reg current limited to 4.3mA work? even if it takes 24 hours?

** It could take up to 60 hours at that rate and be impossible tell what state of charge the battery is in.

If the pack has dropped to 4.5 volts under load - then it IS fully discharged.

So you only need to TIME the recharge.

Need help to design a 4 hour timer ??

.... Phil