Small lipo voltage regulator

Thread Starter

Razor Concepts

Joined Oct 7, 2008
214
Hi, I am looking for a voltage regulator that can regulate the 4.2v to 3.7v of a LiPo battery to 3.3v, max current draw of maybe around 100mAh. Smallest form factor is preferred (including any passives like resistors). Any suggestions? Thanks!
 

SgtWookie

Joined Jul 17, 2007
22,230
Do you care about battery life?

Linear regulators are cheap, but waste energy as heat. You could use a LDO linear reg, two resistors, and an input and an output cap, but you'll throw away a good bit of power.

Why don't you tell us more about your project? Are you working on a model boat or airplane?
 

Thread Starter

Razor Concepts

Joined Oct 7, 2008
214
This is for a small data logger project, about one inch square. The microsd card could take up to 100ma, everything else is minimal. After thinking about it, i think i should have enough space for a proper step-down regulator, and i would prefer their higher efficiency. Thanks!
 

SgtWookie

Joined Jul 17, 2007
22,230
See the attached.
The switching regulator is capable of 200mA output at up to 90% efficiency. It is a buck-boost regulator. [eta] A further simulation showed about 88.6% efficiency for the attached circuit. The inductor and cap selections are critical for efficiency.

The comparator has an internal 1.182v reference. The values I've chosen for R1/R2 cause the comparator to turn the regulator off when the battery gets down to around 3.2v. Decreasing R1 to 4.7 meg will drop that a bit lower.

See the datasheet for the LTC3531 for recommended inductors.
http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1003,C1042,C1116,P15981

Datasheet for the LTC1540 is available on this page: http://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1154,C1004,C1139,P1593,D1777
 

Attachments

Last edited:

SgtWookie

Joined Jul 17, 2007
22,230
Don't forget about heat dissipation. Even though the regulator is pretty efficient, you need to provide some heat sinking to keep the regulator within the temp limits. Most of the power dissipation occurs in the regulator itself. Using 2Oz copper will help to conduct the heat away more quickly. If you want the footprint smaller, consider soldering on some small heat sinks to increase the radiating surface area.

There is a suggested layout in the datasheet for the regulator. Your board might not have to be quite that large, as your load is half of the regulator's rating - but I don't know what your environment is.

Choosing an inductor with a low resistance and high peak current (around 2A) will help the efficiency stay high.
L1: Coiltronics CTX10-1 (peak current 1.9a, resistance 44mOhm)
Use low ESR caps.
C1: TDK C2012X5RIA225K (X5R, 10v, 2mOhm)
C2: TDK C3216X5ROJ106M (X5R, 6.3V, 1mOhm)
 

Thread Starter

Razor Concepts

Joined Oct 7, 2008
214
Thanks for the suggestions. I think i can also stick in a bunch of vias for more surface area.

My device is microcontroller based, so can i skip the comparator and use the microcontroller to monitor the voltage? I could add a pullup resistor on the enable pin, and when the microcontroller wants to shut everything down, it pulls it low and everything stops.
 

SgtWookie

Joined Jul 17, 2007
22,230
I suggest just using the uPower comparator how I have it set up.

You won't be getting voltage to the PIC when it's first turned on, so you won't have a way to raise SHDN\ until the PIC gets power. Using a pull-up resistor on SHDN\ will hurt the efficiency quite a bit.

I suppose you could add a bootstrap cap from the regulator's Vin to Vout to give the output an "instant on" feature. Multiply the total capacitance used for bypass caps by 3.3, and that's the size of the bootstrap cap you'll need. This is based on your battery output being 4.1v to 4.3v. You will need to use a Schottky diode across the bootstrap cap (cathode towards Vin) to discharge the bootstrap cap when the regulator is turned off; otherwise the 3.3v rail could drop below 0v which is undesirable.

You will need to keep your sampling frequency of the battery voltage quite low. You should use rather large resistors for the battery voltage divider, as it will be a continued drain even after the regulator is shut down.

You will also need to use dual caps to bypass the voltage divider input to your ADC. Otherwise, it will take too long for the ADC input to settle; you will need to use high-value resistors for the voltage divider to keep the drain on the battery low.

[eta]
During your uC housekeeping, you'll need to set the I/O pin connected to SHDN\ high right away.
 
Last edited:

SgtWookie

Joined Jul 17, 2007
22,230
Use the parallel configuration.

If you're using Cadsoft's Eagle, I have a library with the suggested layout in it.

[eta]
The attached file is inductors.zip; it contains inductors.lbr
You may already have an inductors.lbr.

If you do, then rename the library file in the attachment to inductors-1.lbr or whatever you'd like that is unique from other library names; then copy it into your Eagle\lbr directory.

Start Eagle, and then enter "USE [libraryname.lbr]"
without the quotes and brackets in the command line area.
You should then find CTX01* in the library.

I just happen to have a few CTX01's that are 15uH, which is why I made the library model.

Note that if you wish to make a 1-sided board with SMT components on the bottom in Eagle, you can just use the "Mirror" tool on it. I don't want to say how long it took me to figure that out.
 

Attachments

Last edited:

Thread Starter

Razor Concepts

Joined Oct 7, 2008
214
Thanks, attached is the board layout. How would I add vias to the polygons without eagle trying to go around them? I guess I could go the long way around and add a lot of header pins to the schematic, and attach those to the proper trace.
 

Attachments

SgtWookie

Joined Jul 17, 2007
22,230
You've used the NAME tool on the polygon to make it the same as power or ground, right? To give you a ground plane?

Do the same with a via; name it the same as the polygon. Then use the duplicate tool on it.

You don't have any copper pour area for the regulator. That's going to cause problems.
 

Thread Starter

Razor Concepts

Joined Oct 7, 2008
214
Ah thanks, I was using the command "via gnd" but that doesn't work, instead making the via and using the name command on it does.

I thought I do have the copper pour area around the regulator? The regulator is the one with the white box around it and the polygons are all around there?
 

SgtWookie

Joined Jul 17, 2007
22,230
I thought I do have the copper pour area around the regulator? The regulator is the one with the white box around it and the polygons are all around there?
dOH! I thought the regulator was the item below the inductor. :rolleyes:

Are you going to have forced air cooling (fan) or rely on convection and radiation?

If convection/radiation, you will get much better heat dissipation if you mount the board vertically.
 

Thread Starter

Razor Concepts

Joined Oct 7, 2008
214
Convection/radiation

But will the regulator create lots of heat with, say, a 10mA current draw? The 100ma figure I mentioned before only occurs during short bursts when writing to the SD card.
 

SgtWookie

Joined Jul 17, 2007
22,230
In the simulation, the circuit is about 88.5% efficient. With a 100mA constant current load, the regulator will dissipate about 41mW, which isn't much - you'll probably be fine with the surface area you have.

The inductor is only dissipating 2mW because I suggested that you use one that is rated for nearly 10x the peak current. If you used one that was "just adequate", you would lose quite a bit in efficiency, and generate more heat.

Note that with the 10uF cap on the output, it takes about 260uS for the regulator output to reach 3.3v with a charged battery.

I'm attaching the simulation. If you don't have LTSpice, I suggest you download it from Linear Technology's website (Google: LTSpice download) and experiment with it.

You can add resistors/inductors/capacitors to represent the parasitics of the traces on your board.
 

Attachments

I hunted around for a 3.3V regulator circuit for lipo's for some time on the Internet. I didn't find anything in production but maybe I didn't look in the right places. At any rate this circuit looks wonderful. I'm going to try it as depicted in the image above. I'm assuming the soldering won't be too difficult, and the parts won't be too hard to find. I'll let you all know how I do.
 
Top