Cheap ebay boost module comparison and performances – 5V from one or two batteries

You ever wanted to supply the 5V voltage to your arduino / electronics project, but didn’t have enough space for 4 or even 6 batteries that are needed to get the correct input voltage? With those cheap (USD 1-2) ebay boost modules you can get the needed 5V from only two alkaline AAA batteries 🙂

(Although the modules are advertised to work from 0.9V input voltage they actually don’t – at least the ones I tested didn’t. You will need at least two normal alkaline batteries for them to work properly.)

The modules convert the lower input voltage into stabilised 5V output voltage. You should however note that the input power is roughly the same as output power, depending on the effeciency of the converter.

Voutput * Ioutput = Vinput * Iinput * effeciency

I have tested two of the boost / step up modules available:

There is not much info available on the modules. Both seem to be bulit for the purpose of supplying a 5V usb device with power from lower voltage source (probably from lithium type battery).

The blue step up module (it also comes in green color) has a mounted usb connector, so it may be a bit bigger than needed if you actually don’t intend to power a USB device with it. It features a big 100uF output capacitor and a 47uH inductor, which probably means that the switcher frequency is low. It has a SS14 high current schottky diode and a unknown shitcher chip (marked with E50D / E5 0D / E5oD / E5oD). From some schematic I found for this module it may be a ce8301 chip.

The red step up module has a bit more compact format and comes without the USB connector. All the capacitors are ceramic, which indicates that the switching speed is probably quite fast, also the inductor is 4.7uH. The number of resistors indicates that the output voltage is most probably programmable (may be set) by external resistors. The switcher chip on my boost module came with AL394 marking. The actual chip used is most probably a SD6271.

For test the ouput of the module was connected to a resistor array, which should consume a given amount of current at the nominal 5V output voltage. For measurements the resistor arrays were exchanged to attain a certain current. Current and voltage was measured at the output of the module.

Input of the boost module was connected to a laboratory power supply. Input voltage was programmed / varied by hand and input current was measured at the power supply.

Additionally the ouput voltage and different points in circuit were monitored by oscilloscope to determine ripple voltage and switching frequency of the converter. From the measurements the minimum input voltage can be determined and also efficencies of the converter were calculated. You can find the whole data here.

The following graphs represent the ouput vs. input voltage of the modules. The red module has much better voltage stabilisation, but has undervoltage protection, which will prevent the boost module from working at a certain input voltage. The blue module doesn’t have the best voltage stabilisation as the output voltage is dependent both on the input voltage and output current:

And the following graphs represent the calculated effeciencies of both modules. Some error due to low precision of measurements can be observed:

At high currents / low value load resistors the additional resistance form wiring can be observed, which is actually quite high – around 7 ohm. There is also some precision error due to low-cost multimeters used.

As expected the effeciency of both boost converters is better when the input voltage is higher. The red module seems to have higher overall conversion effeciencies, but will shut down at a higher input voltage (it can work from ca. 2,5V-3V input voltage, depending on the load). The blue converter starts working at much lower voltage (1-2V) but seems to have a high ripple voltage, which may be problematic in some cases. Also the output voltage seems to be dependent on the input voltage at higher loads – this may be a faulty module in my test or maybe a pcb design error as the chip seems to be switching at frequencies higher than rated. Although the blue module may as well be used in some applications, I really like the red module more 🙂

Future ideas: It didn’t take me much time to connect / make the test resistor and to connect / disconnect additional resistors for varying the current, however it would be much easier to have a “programmable resistor”. One option might be a rheostat (high power potentiometer) or even better a resistor decade. However at these power levels quite a beefy one would be needed – cheap smd decades will start smoking at a few 100mW, while 200mA @ 5V is a whole 1 watt of power, which made even the resistors I tested with quite warm.

Even better would be to bulid a programmable load resistor which could be programmatically changed and which would already contain the voltage and current meter, so external meters would not be needed. This could easily be done with a power MOSFET transistor and arduino, so this may be one future project of mine 🙂

In addition po programmatically controlling the resistor I could also use the computer – control functionality of my power supply (Velleman PS3005D) to automatically vary the power supply parameters to fully automate the whole measurement setup…