Motor Control and Current Measurements

Well, personally I think you should use all three sensors for your implementation if that is not cost prohibitive. You can get away with two sensors as well with few trade-offs. But, if you really have to save every penny at the cost of more development time then you can look at single shunt implementation. There are few publications on how to contruct three phase currrent information from a single bus current measurement but I haven't see anything on converting three phase measurements to a bus current . I guess you can implement the reverse of the first implementation and get the DC bus current. I know TI did some work on this few years back but unfortunately was not able locate the TI reference.

It is quite possible that I will be able to use all three phases.  This has not been decided yet.  However, I am looking to determine over current conditions.  My understanding is that bus current would be better for this check then doing checks individually on the 3 phase currents.  I will continue to research and see if this is correct.

I don't think you should do this with a single set of code, but rather split it into two. You would then need to have the user select (or use some other method like a switch or GPIO) to tell you which board you are using.

For the high voltage kit we are doing this.  We have two of the three legs, plus the single combined shunt.

For PMSM we use the two legs for closed loop and the shunt for over current.

For BLDC we only use the shunt.

We have these as separate SW loads, but we actually have all three (including ACI) burned into the Flash of a single Piccolo controlCARD.  With our GUI you simply select which motor you are hooking up, and it runs the right program.

There will be a controlSUITE update shortly that adds the BLDC project and the GUI...and the Kits are actually going to start shipping next week!

But Arefeen is correct in that it's always better to have as much feedback as possible.  And while a single shunt MAY be slightly cheaper, you also need to use more expensive and faster signal conditioning. There are always trade-offs.

I am currently looking at how a previous product did this.  All three phases were read and the current from one phase was used for over current measurements based on the state of the three HALL switches.  I may be able to take this logic out and reuse.

The number of required current sensors will depend on your system and the level of protect your application demands. There are many cases where two AC current sensors will not meet the reliability, protection and control requirements. You will need to study your customer requirement and standards to determine the appropriate number of sensors.

Your bus current can be calculated indirectly by using power equations. Estimate the output power of the inverter. Then estimate the switching and conduction losses of the inverter by approximations from the IGBT's datasheet. If you are using output filters, estimate losses in the filter (which depend on current and frequency). The sum of the output power and all the losses ends up being the power that the bus delivers to the inverter stage. Divide this power by the bus voltage to give you bus current.