DC link voltage for 3-phase inverter

[robiwan]

Hi, I aim to build a 3-phase inverter (PWM) with output 3x230VAC (Y, 3x400VAC Delta), but I'm confused as to how calculate which DC link voltage is needed for the required outputs ?

TIA
/R

 

[FvM]

Usual VFD inverters run from 560V, in other words they are powered by a three phase bridge rectifier. You need a "balanced" sine modulation to get along with 560V, otherwise bus voltage must be 650V.

 

[robiwan]

You need a "balanced" sine modulation to get along with 560V, otherwise bus voltage must be 650V.

Not sure what you mean by "balanced" but I'm planning to create the three phases with 3 half-bridges, I suppose this would be considered "non-balanced" ?

 

[FvM]

An example showing unbalanced sine(-.-.-), balanced sine (blue) and superimposed neutral point voltage (magenta).

 

[iop95]

You need 3 legs to create 3 phase not 3 bridge. DC link must be minim equal with peak of AC voltage, so for 230V ac you need DC link of 325Vdc. And if you want o create 3-phase balanced system, need also to keep 120degree phase shift. Some inverters use a chopper in DC link or a controlled rectifier to control output voltage at various load (current) to overcome voltage loses.
In motor 3-phase inverter this voltage control is mandatory, to keep voltage/frequency aprox. constant on entire speed range.

 

[robiwan]

You need 3 legs to create 3 phase not 3 bridge.

Which I'd assume is the same as 3 half-bridges ? (i.e. a half bridge consist of 2 semiconductors)

DC link must be minim equal with peak of AC voltage, so for 230V ac you need DC link of 325Vdc. And if you want o create 3-phase balanced system, need also to keep 120degree phase shift. Some inverters use a chopper in DC link or a controlled rectifier to control output voltage at various load (current) to overcome voltage loses. In motor 3-phase inverter this voltage control is mandatory, to keep voltage/frequency aprox. constant on entire speed range.

This is for a UPS style system, so frequency is fixed at 50/60Hz. I also see that with the 3 half bridge system I'd need, just as FvM pointed out, 2 * 230 * sqrt(2) ~= 650 VDC in the DC bus. The virtual neutral point would then be at ~= 325 VDC giving a p-p voltage of 650 V on each phase ( = 230 VAC rms )

 

[iop95]

Yes, sure 3 half bridge; sorry...
Regarding voltage: if phase voltage = 230V, result line voltage = 400V; maxim inverter line voltage = DC link voltage, so you need that peak line voltage that is 400x1.41 = 564V = DC link voltage.

 

[robiwan]

Yes, sure 3 half bridge; sorry...
Regarding voltage: if phase voltage = 230V, result line voltage = 400V; maxim inverter line voltage = DC link voltage, so you need that peak line voltage that is 400x1.41 = 564V = DC link voltage.

Ah.. but therein lies my confusion, you've captured in nicely. The 564VDC would be correct if the line voltage was driven by two phases 180 degree apart, in essence like having a full bridge driven by 564VDC resulting in peak to peak voltage of 2 x 564 V = 400 VAC rms. But the phases are 120 degree apart so there will never be peak to peak voltage of 2 x 564 V. So I think the DC link voltage needs to be 2 x 230 * 1.41 = 650 VDC as suggested by FvM... :???:

 

[FvM]

So I think the DC link voltage needs to be 2 x 230 * 1.41 = 650 VDC as suggested by FvM...

Now you arrived at the point I addressed in my posts.

I didn't suggest to use 650 V. I suggested to use balanced modulation to get along with 560 (or more exactly 565) V.

 

[robiwan]

Now you arrived at the point I addressed in my posts.

I didn't suggest to use 650 V. I suggested to use balanced modulation to get along with 560 (or more exactly 565) V.

Ok, sorry for putting words in your mouth ;-)

However, I'm still confused. The balanced sine you showed show only one phase, with a neutral changing in voltage. With a symmetric 3-phase system, the (virtual) neutral doesn't change in voltage (with a Y-coupled load)