external li-ion batt pack for notebook

[tesla2008]

i would like to build external li-ion battery pack to power on the laptop via laptop's power plug.

so, I would like to use the following parts:
-- 5A variable voltage regulator (1 to 37 volts)
-- li-ion batteries (from extra laptop battery packs I have)
-- thermal fuse for protection of battery pack


the idea is to have 4 li-ion batteries in parallel x 6 batteries in series
that would create my battery pack like in this diagram:

+=====- +=====- +=====- +=====- +=====- +=====-
+=====- +=====- +=====- +=====- +=====- +=====-
+=====- +=====- +=====- +=====- +=====- +=====-
+=====- +=====- +=====- +=====- +=====- +=====-

^ ^
positive terminals joined together negative terminals joined together
6 batteries in series would produce 25.2 volts. Voltage regulator would be set to 19V, so battery pack would stop delivering the power to laptop when
batteries are drained past 19v (or so).
Ideally, the circuits should shut down at certain voltage (say 18.5 V)


any input on this idea?


TIA

 

[MOSFET KILLER]

A simple linear regulator would waste lots of your battery power as heat, this means your battery could die quicker and the battery will continue to discharge when it is not plugged in to the laptop. I reccomend a SMPS design using PWM, this will keep the voltage constant but waste less power and therefore be more efficent.

 

[audioguru]

Before you cause your Lithium battery to catch on fire, you should learn about how to charge each cell properly.

 

[tesla2008]

Before you cause your Lithium battery to catch on fire, you should learn about how to charge each cell properly.

can you give me some pointers on how is this done please?

I was wondering if my charge design would be safe as I have analyzed one of those external power packs. I noticed that each cell bank has a lead going to (I presume) charger. I presume that some form of intelligent charging is used, where each cell (or parallel cell bank) is sent specific amount of voltage.

A simple linear regulator would waste lots of your battery power as heat, this means your battery could die quicker and the battery will continue to discharge when it is not plugged in to the laptop. I reccomend a SMPS design using PWM, this will keep the voltage constant but waste less power and therefore be more efficent.

sorry for my ignorance, I know what PWM is, but what is SPMS design?
where can i find a sample of such design?

TIA

 

[audioguru]

Each cell of a series string of Lithium battery cells must have its charging voltage monitored and a clamp must be made when it reaches 4.2V. It is called "charge balancing".

The voltage of each cell must be monitored to reduce the charging current if a cell is less than 3V.

The charging must be turned off when the charging current drops to a few percent of the capacity of the battery, and/or a timer.

 

[MOSFET KILLER]

PWM stands for Pulse Width Modulation, PWM can be used to drop voltage efficently because it varies the duty cycle of a square wave to reduce or increase power to a load. When you have a PWM square wave at 10% duty cycle it means 10 percent less power is being used. Just put filter capacitors to drop the peaks of the square wave to the RMS voltage, the output will be a lower voltage but less power will be wasted. This is an example of a Switch Mode Power Supply, or SMPS. The PWM signal can be generated with an LM555, and it can drive the base of a large transistor.

 

[bountyhunter]

any input on this idea?


TIA

Paralleling Li batteries is dangerous unless you have smart circuitry monitoring each cell that can disconnect them from the string. They also need "balancing" circuits for charging properly.

 

[bountyhunter]

Each cell of a series string of Lithium battery cells must have its charging voltage monitored and a clamp must be made when it reaches 4.2V. It is called "charge balancing".

The voltage of each cell must be monitored to reduce the charging current if a cell is less than 3V.

The charging must be turned off when the charging current drops to a few percent of the capacity of the battery, and/or a timer.

All true, and the cell's temperatures should be monitored.

In Sony packs, there is also series FETs that can open the battery connects in case of overcurrent or a cell that is too low. Sony put a lot of money into the monitor/protection hardware in each Li pack to prevemnt fires and cell damage. Li batteries are not user friendly like the Ni cells which are pretty hardy and forgiving.

 

[tesla2008]

Each cell of a series string of Lithium battery cells must have its charging voltage monitored and a clamp must be made when it reaches 4.2V. It is called "charge balancing".

The voltage of each cell must be monitored to reduce the charging current if a cell is less than 3V.

The charging must be turned off when the charging current drops to a few percent of the capacity of the battery, and/or a timer.

audioguru,
thanks for the tips
that makes sense

PWM stands for Pulse Width Modulation, PWM can be used to drop voltage efficently because it varies the duty cycle of a square wave to reduce or increase power to a load. When you have a PWM square wave at 10% duty cycle it means 10 percent less power is being used. Just put filter capacitors to drop the peaks of the square wave to the RMS voltage, the output will be a lower voltage but less power will be wasted. This is an example of a Switch Mode Power Supply, or SMPS. The PWM signal can be generated with an LM555, and it can drive the base of a large transistor.

MOSFET KILLER,

thanks for explanation

so, if i understand this correctly; if the load on circuit increases LM555 would change the width of the pulse that in turn would provide more amps to the circuit.
is this how it works?

you said "the output will be a lower voltage but less power will be wasted"

so, let's suppose i have a typical power supply from a PC.
if i was to modify the pulse of PWM to go to maximum duty cycle then technically I should be able to get 18V or so on yellow wire that is normally set to 12V.
Capacitors can be used to smooth out the ripples.
I am just trying to understand this correctly.

What is the typical power loss difference between PWM and a simple linear regulator?

I am guessing that in my external laptop battery application it would be substantial difference.

I have seen that IBM ThinkPad uses 2 batteries in parallel x 3, and iRecharge external battery pack uses the same setup (2 cells in parallel x 3) and then they use some kind of voltage upconverter to raise the voltage to 16,19, 20, or 21 V (depending on desired output).

That design seem to be very bad as the external battery pack draws lot more current (I measured at least 1.9A being pulled out of battery pack) and battery pack gets really hot.
I think this is due to the fact that battery pack has to upconvert the voltage with only 3 cells in series and 2 in parallel (10.8V) to double that.
I think that if design had more cells it would be more efficient, not to mention the battery pack would last longer.

So, that is why I was thinking of creating the battery pack that would have enough voltage and amperage and never get more than lukewarm.
I was thinking of using a linear regulator, but I can now understand the problems after all of the input that I got from people here.

your input is much appreciated.

All true, and the cell's temperatures should be monitored.

In Sony packs, there is also series FETs that can open the battery connects in case of overcurrent or a cell that is too low. Sony put a lot of money into the monitor/protection hardware in each Li pack to prevemnt fires and cell damage. Li batteries are not user friendly like the Ni cells which are pretty hardy and forgiving.

Yes, IBM thinkpad has the same kind of setup – there are some very sophisticated electronics inside the battery, there is a main thermal fuse, overload fuse – all as suggested by Panasonic li-ion battery implementation sheet.
Panasonic batteries have internal over current protection as well.

thanks everyone for lessons!

 

[Oznog]

1. The pack must shutdown when any ONE of its cells falls below the cutoff voltage. Otherwise, that cell will be damaged quickly, even one cycle of this is too much. Reading overall pack voltage is inadequate.
2. This won't be all that efficient because laptops utilize power-saving methods only when running off battery. The input DC/DC converter is not required to be super-efficient and will do undesirable things like charge the internal li-ion battery off your external li-ion batt because it thinks it's unlimited wall power. This alone could add a couple of amps to your draw if the internal battery is low.
3. You may want to run a buck/boost DC/DC converter to deliver the power at an optimal voltage. Li-ion battery capacity is expensive and burning 33% of the power (25.2v to 19v) as heat is a huge loss. And, the heat may be huge. Laptops will draw several amps- my supply says 19v @ 4.75A. That could mean 20W of heat loss in a linear reg, which will need a big heatsink and vents to radiate the heat away.

Be aware that li-ion is sensitive to high current draws. They can permanently degrade the capacity. And yes failure to follow the li-ion charging algorithm (Google it) WILL cause a very hot fire. Almost guaranteed, actually.

I'm afraid MOSFET KILLER is misleading you here. PWM is NOT an effective form of DC/DC here, and capacitors will never function as filters as described. The only way to convert power is pulsing (PWM) through an inductance. And a fixed 555 cannot do this either, controlling the duty to achieve a desired output voltage in the face of a varying load is more complicated than this. There are converter chips available from Mouser, Digikey, etc which can do the job- but you'll need to understand the design issues surrounding transistor, inductor, and capacitor selection. It is not really simple unfortunately.

Loss in a linear reg is (Vin-Vout)*I. Loss in a buck converter can be down to 2%-3% or so easily. A linear reg can never raise the output voltage. A buck-boost, however, can, so if the batt voltage has dropped below 19v but all the cells are still above the cutoff voltage it can raise the output voltage to meet the laptop's requirements.