The Source of all Good Bits

Using your smartphone as a GPS tracker when biking tends to drain the battery very fast. For long tours this might actually be too fast. So I decided to build a mobile cell phone charger that runs on two standard AA batteries and extends the run-time of my smartphone. (To be honest before deciding to build my own I bought a commercially available one which unfortunately went up in smoke. Fortunately Amazon has a nice return policy and after once more killing the same model again I got my money back).

I’ve previously experimented with the LT1073 converter which is nice because it can provide 5V out of a single cell. However, this time a bit more power was needed and so I got hold of a LT1302 which is able to produce 5V at 600mA from a 2 cell supply.

Since no DIP version of the LT1302 was available I first soldered silver wire to the tiny SO-8 lead tips. After a successful run on a breadboard I soldered together a rather ugly but working package which is actually a bit smaller than a single AA battery.

A first simple test run showed that two 2Ah cells could add approximately a 20% charge to my HTC Desire. That was a bit disappointing but enough for a start. Measuring input voltage and current and comparing it to the output voltage and current I calculated an efficiency of only 58% with 2 cells as a supply source which is far below the 80%+ it should actually deliver. Using a 3-cell setup the efficiency rose to 68% which was still not what I was hoping for.

Since the only suitable casing I had lying around was a 15mm square aluminium tube I had to exchange the inductor with a smaller one (not featured on the pictures). On my first field test of the device I realized that it kept a better charge than calculated and measured above with the old coil. A new measurement with the completed device and the new inductor showed an efficiency of 81% (2.18W input and 1.76W output) which is a very nice improvement indeed. The old coil probably had a much higher DC resistance and a core not suited for power conversion applications.
The casing might receive some brushing to get rid of excess glue and some scratches.

Update (2011/07/31):
Feeding the device with three cells reduces efficiency to about 75% (which is rather odd because the datasheet shows a theoretical increase in efficiency with higher supply voltage), so adding a third cell will increase the total available energy by less than 40% (instead of 50%). Anyway, using three instead of two cells has another disadvantage with the LT1302: If one of the cells is weaker than the other two it might run into a polarity reversal which cannot happen with only two cells because the LT1302 will not work under a voltage of 2V.

Conclusion: if even more runtime is needed it’s likely the best solution to only use two cells at a time but to keep two more reserve cells at hand.