maybe alex use this method as stated here
I am going to have to save the circuitry for another post but suffice it to say here that the easiest implementation from the hardware stand point requires us to drive the winding either HIGH or LOW. That requires only three digital I/O pins and a very common H-bridge IC such as SN754410. There’s another possibility, too: we drive only two windings at a time and leave the third one open so no current is flowing through it and no magnetic force is generated, but that requires using 6 I/O pins and 6 discrete transistors. The benefit of the 6-pin scheme is better efficiency and torque due to lack of secondary, smaller magnetic forces pulling in the wrong direction.
http://elabz.com/brushless-dc-motor-with-arduino/
Hexa let me elaborate more: the IC that Alex uses has 2 full bridges - these are called H bridges due to the fact that the letter H is shaped as the electrical configuration of the circuit with the 2 outputs in the middle (out 1A and 2A in the pdf). This enables the 2 outputs to be driven high to low and low to high. This usually happens at the same time (usually they have one input), but the 6205 driver can pilot them seperately so the engineer/developer can drive the outputs as he wishes effectively using them as 2 seperate bridges. The driver used in the open source project is not a full bridge config, it is 3 half bridges, so the single output is driven by 2 FETs not 4 meaning that the output is driven either high or low respect to gnd. This is fine but it is a different setup (Half vs full bridge). A full bridge is more efficient as it totally reverses the potential and current flow while the half bridge only does it respect to ground.
The logical output of the circuits though are principally the same: there are 3 output signals (pins) that are driven in sequence by the MCU to control the brushless motor(s) to keep stability based on the gyro feedback. These output signals go to the power driver ICs that have the necessary circuitry to directly drive the 6 windings (3 phases in star config with center unconnected so 2 couples of windings each). So until this point the logical circuit (Alex and open source) are the same.
The above post refers to 2 different configurations: a discrete driver which is made out of 6 transistors or FETs, 3 couples (3 half bridges), just like the L6234 has except that they are packaged in this IC for integration, and a 2 x half bridge driver (SN754410) which is typically used for stepper motors or 2 phase motors, like Alex style.
I was just wondering how Alex used the L6205 outputs and why he chose this driver instead of a classical 3 phase driver like the L6234. I believe he uses 3 half bridge outputs to drive the 3 phases, 2 at a time effectively using 6 of the 8 half bridges in the L6205, but there are other possibilities. The only way to know is to analyse how he drives the inputs.
ST L6234 datasheet (3 Half bridges):
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00000046.pdf
ST L6205 datasheet (2 full bridges):
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00002345.pdf
L6205 App note:
http://www.st.com/internet/com/TECH...AL_LITERATURE/APPLICATION_NOTE/CD00004482.pdf
TI SN754410:
http://www.ti.com/lit/ds/symlink/sn754410.pdf