Hi, I have a small capacitor start and run motor that I would like to run with variable speed. I am therefore trying to find an authoritative paper on programming an ATXMEGA chip to give me the required output to drive a commercial 3 phase bridge power chip. There is plently of explanation on SVPWM to generate 3 phase, but I cannot find a paper that gives the vector maths and switching sequence for 2 phase. All the papers I have read go quiet right at the crunch point. The example papers from Microchip are useful but do not give sufficient detail for my limited knowledge.

## generating 2 phase output to control motor 3 leg bridge

It helps to have some fundamental knowlege of AC theory when doing this kind of stuff. Case in point - the capacitor is used to give a 90degree phase shift. Thus the two phases are 90 degrees apart. If one is leading or lagging determines the direction of the motor. You have the choice of generating one phase and using the capacitor or generating two phases.

Since single phase motors are generally small, the capacitor loss can be tolerated so a single phase inverter is adequate. So your generated phases are 180 degrees apart forming a H bridge configuration to give you your active and neutral.

Thanks for that. I have a basic understanding and can understand the d-q transformation and logic behind driving a 3 phase motor but now want to drive a 2 phase motor using the correct generated 90 deg out of phase outputs with different voltages for 'start' and 'run' windings. This I understand is possible using a 3 legged bridge?

I think you might be over complicating things a bit. Either stick with single phase or three phase. Doing two phase means the common driver sees unbalanced currents. Get out your graph paper, colored pencils and your grade school maths book. Do the trig and see what i mean.

Thanks for pointing that out. Am aware of the imbalance but the driver chip I have is more than capable of dealing with the current imbalance. I'm after controlling a small fridge motor (minus capacitor to give me a 3:1 speed control range so do need to produce a decent V/F 2 phase supply.

I have done exactly what the OP proposes

There are a few papers dealing with space vector modulation and the twin phase motor, its easier in some ways but more complex in others

The two windings are 90 degrees apart so the dq transform isn't necessary however the phases are usually unbalanced

The main difference between this and the three phase case in the positioning of the active vectors, they are skewed and two are longer (by a factor of sqrt(200

So the angles arent a constant 60 degrees which is a complication over the three phase case

like I said I have done exactly this, look for papers by Lipo and theres one in particular by Lipo and Eric Benedict

I never did it on an AVR its the realms of ARM chips, I used an STM32F4, having to disconnect the capacitor and taking care to identify the windings makes this not an off the shelf solution

The maximum output voltage you can achieve is (1/sqrt(2))*VRMS

Thats peak!, the max RMS output is only 50% of the AC input Do the math

A H bridge can output what is input, its cheaper to build, theres no disconnecting windings and all that

But the H bridge doesn't cancel harmonics as well as a three phase inverter

Making hardware to switch mains voltages is not a trivial exercise, most people struggle at that point

Regards

I'm after controlling a small fridge motor (minus capacitor to give me a 3:1 speed control range so do need to produce a decent V/F 2 phase supply.

If you are wanting to speed control a hermetic compressor, be advised there *might* be bearing lubrication issues at reduced rpm. I can't say for sure about small refrigerator compressors, but it is a real issue with larger HVAC compressors.

If one really wanted to do this, seems like one could use a voltage doubler at the front end, connect motor 'common' to the mid voltage point of the two filter caps and drive 'start' and 'run' from an h-bridge. You would then be able to drive the motor at full voltage with a lower parts count.

Thanks for you advice Bignoob and Tom. As the motor is running in the sealed gas zone with its compressor piston, I'll take the risk with the lubrication issue. The advice re the reduced voltage and the option of going to a split supply is attractive, but does mean I need 4 bridges which means I'll have to move away from the standard 3 bridge hybrid chip I have driving a 3 phase motor. I'll have to ponder this one - again many thanks to you both.

What Tom suggested means you only use 2 legs of the bridge. Tapping the middle of the capacitors gives you a virtual zero volt point. This also means you'll need to choose your capacitors carefully as the AC current will be significant. Other things to consider is the winding insulation and construction of the motor. Inverters create high harmonic currents that can stress the insulation of the motor windings as well as creating currents that run through your bearings.

If one really wanted to do this, seems like one could use a voltage doubler at the front end, connect motor 'common' to the mid voltage point of the two filter caps and drive 'start' and 'run' from an h-bridge. You would then be able to drive the motor at full voltage with a lower parts count.

A voltage doubler is a commonly used technique from back in the day!, the boost converter is much more attractive these days though

What you propose is perfectly valid and actually a reliable application a split DC bus and half a H bridge per phase with the capacitor disconnected and your good to go, the fixed double DC voltage bus has quite a few drawbacks, the dv/dt stresses in a power converter are bad enough doubling the voltage and running at low speeds requires even better designed hardware to be reliable

does mean I need 4 bridges which means I'll have to move away from the standard 3 bridge hybrid chip

A further complication that arises is the sheer vast amount of different motor constructions!, there are a lot

If you have two separate windings then you could use two separate H bridges per phase so 8 power devices, you could connect the two commons together and supply from the mid point of the DC bus this would use 4 power devices andgive you half voltage out which the voltage doubler could overcome

You can get the midpoint DC voltage with a three phase inverter without using capacitors to split the DC bus. If you use the extra leg by modulating the leg at 50% duty cycle you have effectively spilt the DC bus as proposed and you could use a voltage doubler to overcome the 50% RMS output/input

Most of the two phase motors I have come across don't have separate windings, its usually IME (my experience with these is mostly ventilation motors) one winding with the midpoint tapped at the required point so you don't have the option of separate H bridges and the three phase inverter can be used

Look for that paper I mentioned, it links to many other with all the topologies

As this is a pump application I would forget about all this winding connections and all the different options with voltage doubling etc the best solution here (IMO)is to drop the variable frequency part out of the variable frequency drive!

All the options discussed required removing the capacitor, now I don't know what your machine analysis skills are like?, the capacitor is there to give an (approximate) 90 degree shift between each winding current, this creates a torque from a pulsating magnetic field (three phase has a rotating magnetic field)

The capacitor is sized at mains frequency and if you alter the frequency then the reactance of the capacitor changes in such a way that it spoils the party! if you think of it in the most basic of ways

1/(2*pi*f*C) increases as the frequency decreases and the current will have anything but the 90 degree shift

Staying with this concept as an aside and thinking the other way for inductance 2*pi*f*L decreases with frequency, a motor is highly inductive so as you decreases frequency impedance decreases, decreases the impedance and current increases!

This goes a way to explain why we do V/Hz, less volts for the same current!

Any way if you use a standard H bridge thats four power devices (2/3 of your existing hardware!)

You can simply do sin reference modulation (each leg 180 degree shifted)

Its arguably the easiest strategy to implement, you get full AC output when you need it

No voltage doubling

Less hardware

But you cant alter the frequency (you can to an extent but unless your Einstein_2 you need to experiment to find the SOA)

So fix the frequency and just alter the magnitude of the output, its an electronic transformer and your doing glorified stator voltage control.

You will be able to do all the things you need, for a pump application this is perfectly adequate

I could type all day about this stuff but I got to go now

Regards