# Battery voltage step up (Booster converter) help

#### Teamawe

Joined Feb 21, 2012
10
Currently my group is working on how to use a 3.7v Lithium-ion polymer battery to power many led's simultaneously.

We use arduino with pwm outputs to control which circuit to activate.
(Total 5 circuits, one active at a time)
Circuit 1. 40 red leds ( 10 in series and 4 parallel)
Circuit 2. 20 yellow leds (10 in series and 2 parallel)
Circuit 3. 30 orange leds (10 in series and 3 parallel)
Circuit 4. 20 red leds (10 in series and 2 parallel)
Circuit 5. 20 green leds (10 in series and 2 parallel)

See attachment for boost converter design.

The question is how to make the Vout range from 18.3volt to 21.3 volt, and Iout range from 40mA to 80mA?

Since connection 40 red reds( 10 in series and 4 parallel) requires 18.3 volt and 80mA.
Connection 20 green leds(10 in series and 2 parallel) requires 21.3volt and 40mA.

The arduino is pwm outputting at a freqency of 100kHz, to the transistor base.

What should the value of the inductor(L1), Capactior(C1) and R2 be?

Is the 3.7v Lithium-ion polymer voltage stable?
Should I consider other battery types?

#### Attachments

• 210.9 KB Views: 180

#### crutschow

Joined Mar 14, 2008
24,742
What is the battery current rating? For 120mA @ 20V, the battery current will be well over a half amp if you include the inefficiency of the conversion circuit.

How would you regulate the voltage? You need to control the duty-cycle of the signal from the arduino for the proper PWM signal to give the desired voltage/current. This typically involves some sort of feed back from the output to the arduino.

If you use resistors to regulate the LED current, then your output voltage must be higher than the 18.3V and 21.3V you need for the LEDs.

#### Teamawe

Joined Feb 21, 2012
10
What is the battery current rating? For 120mA @ 20V, the battery current will be well over a half amp if you include the inefficiency of the conversion circuit.

How would you regulate the voltage? You need to control the duty-cycle of the signal from the arduino for the proper PWM signal to give the desired voltage/current. This typically involves some sort of feed back from the output to the arduino.

If you use resistors to regulate the LED current, then your output voltage must be higher than the 18.3V and 21.3V you need for the LEDs.

Hey most probably the battery will be around 3.7V at 1000mAh or higher amp.
I think of connecting an external TLC 5940 PWM driver IC to the arduino to get high and stable enough frequency, I thought it was possible to adjust the duty-cycle on the Arduino or the TLC 5940, I'm I wrong?

About the output voltage, what do you suggest?
Is 24volt sufficient?
Overtired at the moment, will clarify the rest tomorrow.

#### crutschow

Joined Mar 14, 2008
24,742
Hey most probably the battery will be around 3.7V at 1000mAh or higher amp.
I think of connecting an external TLC 5940 PWM driver IC to the arduino to get high and stable enough frequency, I thought it was possible to adjust the duty-cycle on the Arduino or the TLC 5940, I'm I wrong?

About the output voltage, what do you suggest?
Is 24volt sufficient?
1000mAh is the storage capacity of the battery, not how much current it can readily supply.

You can adjust the duty-cycle but how will you determine what to adjust it to? The required duty-cycle is affected by the battery voltage and the load.

You will need some extra voltage so you can add resistors in series with the LEDs to limit the current.

#### Audioguru

Joined Dec 20, 2007
11,249
An ordinary Li-Po cell is 4.2V when fully charged and its voltage slowly drops to 3.2V as it discharges when its load should be disconnected. Therefore its AVERAGE voltage during a discharge is 3.7V.

Therefore you must use a constant current source to feed the LEDs so that their brightness does not change as the battery voltage changes.

#### bountyhunter

Joined Sep 7, 2009
2,512
Therefore you must use a constant current source to feed the LEDs so that their brightness does not change as the battery voltage changes.
+1, which is why I recommended the LM2731 which is made specifically to do this easily and runs on voltages as low as 2.7V. Make life simple, do it the easy way.

#### Attachments

• 33.8 KB Views: 63

#### Teamawe

Joined Feb 21, 2012
10
+1, which is why I recommended the LM2731 which is made specifically to do this easily and runs on voltages as low as 2.7V. Make life simple, do it the easy way.

I will try the LM2731, expect to hear from me later!

#### John P

Joined Oct 14, 2008
1,808
I think you should replace the NPN transistor with a MOSFET, and eliminate D1 and C1 from the circuit completely. For each LED string, run the current either through the MOSFET or the LEDs according to whether the FET is on or not. Where the current returns to ground, insert a small resistor (1 ohm?) and measure the voltage across it , and based on that, adjust the PWM duty cycle to set the current.

You select the value of L1 by first deciding how much variation in current you can tolerate, and using the typical battery voltage, work out the basic inductor equation, V = Ldi/dt using the time interval of a single PWM cycle.

#### Teamawe

Joined Feb 21, 2012
10
I will try out different methods to see which one is most stable.
Which one should I try? LM2731 or LT1618?

#### bountyhunter

Joined Sep 7, 2009
2,512
Looks like either would work.

#### Teamawe

Joined Feb 21, 2012
10
Looks like either would work.

I ordered the LM2733, probably arriving tomorrow.

If I set the output to 24V.
And the LED's require 18.3V(10 leds red in series).
24V-18.3V= 5.7V

Can I do it the standard way of putting a resistor in series with the leds, (Iout=5.7V/0.02A=285Ohm)?

#### Audioguru

Joined Dec 20, 2007
11,249
Nobody makes a 1.83V red LED.
LEDs have a range of forward voltage. Some might be 1.7V and others might be 2.1V.
Then with a 24V supply and a 285 ohm resistor they might be bright with a current of 24.6mA or be dim with a current of 10.5mA.

Chances are they will average 1.83V each.

#### #12

Joined Nov 30, 2010
18,210
Ummm...yeah. Start with the calculated value then measure and see if everything averaged out to 1.83 volts each and adjust as necessary.

#### bountyhunter

Joined Sep 7, 2009
2,512
I ordered the LM2733, probably arriving tomorrow.

If I set the output to 24V.
And the LED's require 18.3V(10 leds red in series).
24V-18.3V= 5.7V

Can I do it the standard way of putting a resistor in series with the leds, (Iout=5.7V/0.02A=285Ohm)?
Yes, you can boost to a fixed voltage and use a resistor but that's wasteful. There is a way to set up the boost as a constant current boost which will force a constant current into the LED string. Check the LM2733 data sheet and there will be a circuit for that (see below).

#### Attachments

• 35.8 KB Views: 53

#### bountyhunter

Joined Sep 7, 2009
2,512
Nobody makes a 1.83V red LED.
LEDs have a range of forward voltage. Some might be 1.7V and others might be 2.1V.
Then with a 24V supply and a 285 ohm resistor they might be bright with a current of 24.6mA or be dim with a current of 10.5mA.

Chances are they will average 1.83V each.
Also the voltage changes with temperature and increases with more current. That's why it's a good idea to use a constant current desing since that forces the same current all the time.

#### Teamawe

Joined Feb 21, 2012
10
Yes, you can boost to a fixed voltage and use a resistor but that's wasteful. There is a way to set up the boost as a constant current boost which will force a constant current into the LED string. Check the LM2733 data sheet and there will be a circuit for that (see below).

I soldered the circuit.
Connected to a battery, I get the same output voltage as input voltage.
What can the problem be?

#### Teamawe

Joined Feb 21, 2012
10
I soldered the circuit.
Connected to a battery, I get the same output voltage as input voltage.
What can the problem be?

Here is the circuit.
Everything is SMD ceramic components,
except for the Inductor and Schottky Diode.

Someone have any clue what the problem can be?
When I connect 3.7v battery to the input, I measure 3.7V at the output.
This project deadline is friday, I really appriciate help!

#### Attachments

• 65.7 KB Views: 53

#### bountyhunter

Joined Sep 7, 2009
2,512
Check layout and diode polarity.

Is this a PC board layout?

#### Teamawe

Joined Feb 21, 2012
10
Check layout and diode polarity.

Is this a PC board layout?

Is a PCB board air layer(105um).

Suddenly it worked, for a while at least.
I connected a 6.5V battery and got nearly 37 V peak at the output, when measured with an oscilloscope.
Connected 5 leds with 1 resistor in series, worked well.
I really don't understand, shouldn't the output be around 24V?

I disconnected the battery and took a break, now it doesn't work any longer
When measured with an oscilloscope, the output voltage is almost the same as the input with a litte bit of oscillating.

Look at the PCB layout,what can the problem be? interference?
I'm gonna design a new layout with shorter tracks and narrow spaces, to see if it helps.

#### Attachments

• 152.3 KB Views: 38