SparkFun Forums

[press9761]

I have a problem I hadn't predicted. Maybe some one has some advice.

I put together a small data logger system. It is powered by a Li-On battery and a solar panel.

The hole thing works well off the battery and off the solar panel. However, when I try and charge the battery (using a Li-On battery charger) some of my IC's brake.

It works like this:

The power from the solar panel (24V) goes to a 7805.

7805 to Vin on Li-on charger.

(Li-on charger pin out = Vin,GND, Vbat)

Battery is connected to ground and Vbat.

The data logger system is powered from the battery,...which means my power line to the data logger is also connected to Vbat.

It seems like the battery charger is sending voltage spikes through the hole data logger as its trying to charge the battery.
So,......what to do...how would you charge a battery and run you IC's off the same battery while its charging. Obviously its not imposable, this is done all the time.

I was thinking of useing an R/C network to smoth the line however, I'm afraide it will damage the battery. The Li-On battery charger is sending out voltage spikes for a reason. If I smooth the line (even after the battery). Won't I be undermining the actions of the charger?

any advice????

Thanks!

JP

[bigglez]

It works like this:

Can you post a schematic?

(Li-on charger pin out = Vin,GND, Vbat)

Is this your own design? An IC? A module?

how would you charge a battery and run you IC's of the same battery while its charging. Obviously its not imposable, this is done all the time.

A Li-Ion is tricky. Better not to charge it
while another load is connected. The other load
could be powered from the same power going to
the Li-Ion charger (with a separate regulator and
load switches as needed).

[teekay_tk]

from what you have described there shouldnt be any spikes.

are you using all sparkfun things (especiallu the MAX1555 charger IC)

can you confirm the output of the 7805 being 5v using a meter

if so, you should see 4.2V out of the charger IC on VBAT. if you suspect spikes then the best thing would be to see them using an oscilloscope.

best thing will be to see if all devices on your end are adequately decoupled.

if you still cant solve the prob then you can post the schematic and probably a picture of the spikes on the scope

[SebaS]

The other load
could be powered from the same power going to
the Li-Ion charger (with a separate regulator and
load switches as needed)

I would do as bigglez said, while charging, power the uC with the external source...

I think you can purchase some plugs wich comes with an internal switch so, when you plug the external source you switch between that external source and the cell to power the uC

Another question... how do you do the charging? wich IC do you use? do you have some sort of battery management? I'm starting a project wich I plan to power with a Li-on battery pack

[emile_b]

Nothing to do with the question (sorry!) but you may want to consider using a DC-DC converter after the solar panel so you dont waste 80% of its precious power!

[metaforest]

I have a problem I hadn't predicted. Maybe some one has some advice.

I put together a small data logger system. It is powered by a Li-On battery and a solar panel.

Take a close look at a Microchip MCP73871

Notice how it internally switches the power source around when an external source is supplied to it's Vin.

when Vin is higher than Vbat, it switches Vin to Vout disconnecting the Vbat from your load.

The key reason for switching the load off the battery when charging is that the charging circuit is trying to pay close attention to the batteries current draw. The charger cannot do that properly if your load is drawing current off the battery during the charge cycle.

The constant current portion of the cycle can supply up to 4.6V to Vbat during charge! if you have a 3.3V or less system with no regulation (hey why bother regulating when nominal Vbat is 2.7V to 3.4V, riiiight ) something is gonna go poof when the charger kicks in!

To protect your load you will need a small buck-boost flying(switched) capacitor regulator(such as a MCP1252) between the Vout on the MCP73871 and your load to deal with the 5V - 6V external source when it is applied to Vout. If you have noise sensitive analog sensors in your project regulate to 3.4V at the and use a low current LDO regulator to supply your analog power to mute the switched capacitor noise coming from the primary regulator(s).

I would seriously reconsider using a 24V solar array to provide external power. Try getting your hands on a smaller one. If you dont have a choice on that seriously consider a DC-DC converter. 7805s dont really do well with 24V on their input they tend to sink that difference into huge amounts of heat, thus wasting the vast majority of energy that a fairly simple Buck converted would convert into useful power.

[metaforest]

I think you can purchase some plugs wich comes with an internal switch so, when you plug the external source you switch between that external source and the cell to power the uC

Another question... how do you do the charging? wich IC do you use? do you have some sort of battery management? I'm starting a project wich I plan to power with a Li-on battery pack

Be careful relying on a mechanical power switch to switch between Vbat and Vext!! As the connector wears it's possible to have Vext appear across the battery before the source actually switches AND there can be a discontinuity where the load sees neither Vbat or Vext!!!

In general that switch is only useful to tell you that an external source plug has been inserted. In fact it's a bad idea to make any switching choice based on the mechanical switch, other than maybe to have your system check for valid Vext....

Those switches were really only useful on analog systems that ran off a non-rechargeable battery source or a external DC source, and had no components that would mind much if there was a short loss of power.... like old tape players or "Walk-Man" type devices. Additionally an alkaline battery isn't going to mind much if it sees a short DC spike on it's positive terminal once in a blue moon. Rechargeable batteries don't like that at all, it seriously shortens their useful life.

[bigglez]

Rechargeable batteries don't like that at all, it seriously shortens their useful life.

Can you cite this claim, or is it your opinion?

[metaforest]

Rechargeable batteries don't like that at all, it seriously shortens their useful life.

Can you cite this claim, or is it your opinion?

an example: https://www.fmadirect.com/support_docs/item_1133.pdf

Lithium Polymer (LiPo) cells and packs can be destroyed during charging by applying a voltage
that’s too high. Safety Guard is a voltage-limiting device that prevents overvoltage during LiPo
pack charging.
What happens when LiPo charge voltage is too high? With as little as 10 minutes of overvoltage
(6V per cell), the cell envelope begins to expand, which severely damages the cell. Continued
overvoltage causes the envelope to swell from increasing internal gas pressure. Under extreme
conditions, the envelope may vent (open) in less than 20 minutes. This destroys the cell. If sup-
ply current is high, venting may be accompanied by flames.

If overvoltage for 10 minutes causes catastophic failure I am willing to bet that even a few ms of it is gonna cause damage.

Li-ion/Li-Po and NMHi are very intolerant of poorly regulated charging, and thats not an opinion its an easily verifiable fact.

Do a google search for "Li-Po and overvoltage" and I think you'll get the idea

Also note that every Li-xx (and quite a few NMHi) charging circuit I have ever seen a schematic for has a thermal safety feed back to stop the charge cycle if the temp in the cell rises too quickly. It's critical for safety reasons to make very sure that a charging system for Li based batteries prevent any condition that might cause damage to the battery.

With older NiCad charging all you riske is killing the cell or causing electrolyte leakage, they aren't inherently a fire hazard the way Li-xx cells are.

[bigglez]

If overvoltage for 10 minutes causes catastophic failure I am willing to bet that even a few ms of it is gonna cause damage.

I'll be happy to take that action. PM with your wager.

[olivier_p]

I too use a solar panel + li-ion + mcu set up. Here is how it goes:

12V solar panel
2 x 3.7V li-ion cells in series
8.4V li-ion charger
5V dc-dc step-down converter

So, Solar panel connects to the charger, charger to the li-ion cells, li-ion cells to the 5V converter.

In my case, I needed 2 cells in series because they are also powering higher voltage devices triggered by the 3.3V mcu. The setup works really well. I also have a 220uF electrolytic cap rigth before the 5V converter.

[2226cc]

Would you be willing to share that circuit layout?

I am super new to this and I'm looking at a setup also running off a pair of 3.7V-4.2V li-po/li-ion cells running in series. This then feeds to a MAX733 to boost voltage to 15V for one part of a circuit. Another part will mimick a USB output (5V/100mA) for some odd reason. I'll have to step it down from 7.4/8.4 in that case.

I am just too much of a newb at this to try and figure out a 8.4V charger circuit which will charge the 2 in-series batteries and provide power to the device at the same time.

Yes, I know I being waaaay too forward in just asking for that circuit layout.

[metaforest]

Take a look at Microchip's battery charging circuits. They support 2 cell charging, and it's fairly painless. They also have some good AppNotes that explain how to get your head around li-po configs.

is there a particular reason you need dual cells? In most apps it's really not required, and it can complicate charging, and many low-loss regulators don't deal with voltages above 5.5V

[2226cc]

I want to feed a step-up regulator to attain 15V and at the same time pack it into a compact package with extended runtime. Portability and extended runtime is what the major factors are ... too often have I seen those terms as oxymorons (do you even get complete "phrase based" oxymorons? )

To feed my regulator with a minimum voltage I'd need 7-8 AA cells which, by weight, would easily outweigh a pair of flat-pack li-ion/lipo cells. They might provide a massive runtime, but it's enormous enough to be ridiculous when compared to 2 cells which meet my voltage requirements and enough runtime to be happy with.

Oh, throw safe charging in as well, so balancing/equalizing is part of the deal too.

Yes, I like to complicate my life even in things I am sub-amateur to.

[metaforest]

there are plenty of ways to get 2.7V-5.5V to 15V DC-DC conversion from a Li-po and get " long runtime" using li-po. the real question with li-po is: "what size battery is your application going need? " And the question has two flavors: volumetric size, and charge capacity. These days a single cell li-po with package dimensions of 55mmx75mmx8mm can easily store 5500mAh.

So how much current do you need at 15V? If it's less than ~200mA there are plenty of simple boost mode converters that can get you there... and they are relatively simple to use.

The complexity is relative: a good dual cell charger is a fairly complex engineering challenge. DC-DC regulator/converters by comparison, for low currents are almost plug and play.

I suggest looking around to see what size Li-Po cell will fit your form factor and what charge storage ratings fall out of that....

Also looking at what technology you'll be building with TH, SMD, or mixed, and see what various devices fit. There are a huge number of solutions...


One look at an old iRiver iPF-390T demonstrates how low you can go: powered from a 1.5V AA it will run for more than 15 hours in record mode. The internal logic is 3.3V with LDC bias and LED back light bias both of which would be at least 7V if not more like 15 - 18V for the LCD bias....

YMMV