Modified sinusoidal PWM

[sam781]

What is the relation between DC source voltage and sine wave AC RMS voltage of a Modified sinusoidal PWM inverter?

 

[goldsmith]

Dear Sam
Hi
What do you mean by AC RMS voltage , here ? is your mean after power stage ? can you tell me a bit more explanations about your exact problem ?
Best Wishes
Goldsmith

 

[sam781]

@Goldsmith: For example, I have 200V dc source. If I apply Modified SPWM technique what will be the output AC voltage?

 

[Tahmid]

What exactly do you mean by "Modified SPWM"?

In normal cases, when you apply SPWM, the peak voltage = √2 * RMS. Since your peak voltage will be 200V, your RMS will be (1/√2)*200 = 141V approximately. Then you have to factor in the losses you get. This is assuming maximum duty cycle.

If you're talking about quasi sine: If you have 80% duty cycle for modified sine wave (quasi sine), you get approximately 160V. If you have 90% duty cycle, you get approximately 180V. And so on.

Hope this helps.
Tahmid.

 

[goldsmith]

Hi sam
The maximum out put wave of each half cycle will given by voh=1/T*integral over Vs dt from zero up to ton . it means the out put amplitude depends on your modulation depth .

What exactly do you mean by "Modified SPWM"?

Hi Tahmid
Did you see hise attached picture ? it showed a modified SPWM .



Best Wishes
Goldsmith

 

[sam781]

Hi sam
The maximum out put wave of each half cycle will given by voh=1/T*integral over Vs dt from zero up to ton . it means the out put amplitude depends on your modulation depth

But in this Modified SPWM case, there are so many Ton and Toff. Are you talking about the middle (large Ton, 60 to 120 deg time period) Ton period?

Inverter output will be 50Hz. Hence, time period 20ms, half time period 10ms. So, middle Ton will be 3.333ms. Am I right? What should be the peak output voltage?

 

[goldsmith]

Hi sam
Don't confuse yourself . i told that you can take average from that wave . but of course it is not reasonable to take it for each cycle . it is enough that you know M=Em/EC EC is amplitude of your triangular wave and Em is amplitude of your message signal . if you analyze it's spectra you will understand what i'm trying to say . by the way , do you know how is the spectra of an SPWM signal ? how about three level SPWM ? if you know there won't be any problem . but if you don't know it , it will be a bit hard to talk about averaged out put .
Anywhere . with a filter you can justify the out put signal with two method :
1- it's action as an LPF to remove other side bands ( this approach needs knowing about it's spectrum )
2- it's action as an average taker circuit . ( it will be simpler for you ) . so you can analyze your signal instead of each cycle . ( D.C of each cycle*VS ) . i agree that it is a bit time consuming but it is the simplest way for you .
By the way , did you think about your power stage for inverters such as this ? i have some doubts that you know about it's power stage . isn't it ? perhaps i'm wrong and you know what kind of power stage here needed .
By the way , may i ask you that what happened that you have decide to use a modified SPWM ? there are many other ( and simpler ) topologies to go through . do you need pretty high precision in your aim ?
Best Wishes
Goldsmith

 

[FvM]

Apparently your "modified" SPWM scheme is assuming specific parameters although they aren't explicitely mentioned in your post. Generally the fundamental and harmonic levels can be exactly calculated for a given set of parameters, using well-known fourier series expressions or integral calculus.

 

[sam781]

@goldsmith: My intention is to make a model inverter smile having the following qualities -
1. Efficient
2. Light weight
3. Pure sine wave
4. Power range 400-500w

I know, there must have a tradeoff among the above features to make a practical inverter. But my aim is to make it as closer as possible practically. It is not mandatory to make a Modified sinusoidal PWM.

I've started this discussion considering only the inverter part. I wanted to know the DC voltage required in inverter input to get 230V AC. I'm planing to use bridge converter to make high voltage DC. I've chosen Modified SPWM to get a more accurate sine wave than conventional SPWM. I'll make switching signal using MCU. It should not be any major issue.

As told earlier, it is not must to use Modified sine PWM technique, could you please help me choosing a topology to make my desired model inverter? :smile:

 

[goldsmith]

Hi Sam
You can use this way : at first design an H bridge to achieve 230*sqrt(2) ( an H bridge converter which converts you low voltage to a 325v DC . ( this section does not need feedback ! the feedback loop will be closed from main out put to this section)) ( this section will have a simple HF transformer ) . then you will need another H bridge . but without any transformer ( with just a simple LPF ) ( butterworth family )
this section will need SPWM as driving signal ( i.e the first stage drove with PWM ) . the carrier frequency can be something around 20 KHZ or higher or a bit less . the message signal will be a pure sine wave (50 HZ ) . ( in my country frequency of power line is 50HZ ) . then you will need a DC feedback from this section to the first stage .
By the way , the out put filter can be selected on 5KHZ frequency ! but it will give you 50HZ because the lowest harmonic is 50HZ !
Best Wishes and Good luck
Goldsmith

 

[FvM]

I've chosen Modified SPWM to get a more accurate sine wave than conventional SPWM.

The waveform in post #1 is just the opposite: Less accurate sine or in other words higher harmonic content. An advantage of the low switching frequency (fpwm = 15*fsine) is reduction of switching losses, it would be primarly meaningful for high power (>100 kW) inverters.

 

[sam781]

@FvM:
I found the attached image from a Power Electronics book (H. Rashid).

 

[goldsmith]

Hi Sam
That out put called three level SPWM . one of the ways to achieve three level SPWM is comparison between a triangular wave and two sine wave . first one in phase and the other one out of phase .
And then after your power stage the out put will be that waveform . see below , please :
https://www.edaboard.com/blog/1523/
Best Luck
Goldsmith

 

[sam781]

Hi goldsmith,
Sinusoidal PWM signal can be generated easily using a microcontroller. Hence, we can avoid comparison between triangular wave and sine wave.

 

[goldsmith]

Sinusoidal PWM signal can be generated easily using a microcontroller. Hence, we can avoid comparison between triangular wave and sine wave.

Hi again
Yes i'm agree it can be generated with a micro controller . but i told the basics . of course there are many ways to achieve it !

 

[sam781]

Okay, I understood.

In this case, if frequency of triangular wave is increased notch will be increased. Probably this three level SPWM is standard one. Is there any other way to improve the waveform more?
If three level SPWM is better other level, I need to know the timing of each edge (rising & falling). How can I get these value to be used in microcontroller program?

 

[FvM]

A standard sine PWM with suffcient high PWM frequency will achieve better sine quality with less hardware effort. I didn't yet hear a plausible motivation why you should use a "modified" SPWM or multilevel design for a 500 W inverter.

 

[goldsmith]

Hi sam
Are you looking for other ways ? sure multilevel inverters such as 24 level !!! but i don't think in your applications there would be needed to use those ways . a simple two level will be enough for you ( HF DC Link inverter)
Good luck
Goldsmith

 

[sam781]

Hi goldsmith,
No, I don't want to make multi level inverter which require large number of switching device.

 

[goldsmith]

Hi Sam
Well , now what is your problem ? is your problem solved , or still you have problem ?