Your image link did not work.
Please attach the image to your post using the "Go Advanced" and "Manage Attachments" buttons below the reply box. .PNG format images are preferred.

You want a low-pass filter with a sharp cutoff below 60Hz.
If you want it to be usable on 50Hz systems, then you need an even lower cutoff frequency.
A Chebyshev filter will provide a brick-wall-like response, if you allow for significant ripple in the passband; perhaps 2dB-3dB.

Here is a 5-pole Chebyshev-type filter whos' values were rounded to the nearest 5% tolerance:

A plot of transmission is attached.

 

It takes all day for ImageShack to wake up.
I wish people would post their schematics here instead of over there.

 

Hi there,
I'm sorry for the broken link.I've attached again the circuit here. As you can see, the inverter is connected to the point of common coupling through a low pass filter between a source and a load. It actually functions as a STATCOM supplying reactive power to the power system by injecting reactive current. As you can also see that the inductor of the low pass filter functions as a coupling inductor between the inverter and PCC. What i want to do here is to make Vinverter to be sinosoidal at the side of PCC by the use of low pass filter. All fundamental frequencies are at 60Hz so i'm passing this 60Hz. I need to ask what cutoff frequency should i put if the first harmonics would occur at 180Hz? and how do i know my input and output impedence for this case for network matching? And again do i have to look into the S-plane to design my filter? If yes,why? How do i go about designing this filter?Please advice.Thanks.

 

Sorry,
i forgot to mention that i'm designing a butterworth filter.Thanks

 

What is the output impedance of the inverter?

[eta]
If you insist on a 2nd order Butterworth filter, it won't work very well.

A 2nd order inductor input 150Hz lowpass Butterworth with a 1 Ohm input impedance will need a 1.5mH inductor, 1500uF cap on the output, will have about -17dB return loss @ 60Hz, and only be about 4.6dB down at 180Hz.

A 2nd order inductor input 85Hz lowpass Chebyshev with a 1 Ohm input impedance will need a 4.6mH inductor, 1124uF cap on the output, will have under -30dB return loss @60Hz, and be about 16dB down at 180Hz.

Far more effective at passing the fundamental frequency, and far more effective at blocking harmonics.

[eta]
Just to show you what I'm talking about, here's the Butterworth design:

The plot of transmission (S-12) and return loss (S-11) is attached. Note the return loss at 60Hz, and the transmission at 180Hz.

 

Assuming i'm using a 2-level 3-leg inverter supplied from a capacitor having a on resistance of 1e-4 ,is that the output impedence? Sorry,i'm not sure of how to calculate that.

 

What is the total impedance of the output?

The capacitor, the switches you're using, the resistance of the wires, etc?

 

OK, here's the Chebyshev 2nd order filter. Note that the schematic is very similar except for the values, but there is a world of difference in the plots.

Schematic:

Plot is attached. Compare return loss @ 60Hz to the Butterworth design.

Compare transmission @ 60Hz and 180Hz to the Butterworth design.

For the simplicity of a 2nd order filter, there is simply no comparison.

I could've easily increased the passband ripple of the Chebyshev and made the rejection at 180Hz even greater.

 

Here's an 85Hz Chebyshev with more bandpass ripple:

Compare it to the last one. Return loss hasn't changed, but look at transmission at 180Hz vs 60Hz. It's not a brick wall, but it might as well be.

 

Hi SgtWookie,
Thanks lots for your detailed explanation. It's definitely eye opening
Unfortunately,i have still 4 issues to deal with.
1).Assuming my output impedence of the inverter R1 is 1 ohm,looking at my attached thumbnail, should the output impedence be 1 as well?
2).Should the output impedence R2 be connected in this manner?(how do you calculate the output impedence on your previous examples?
3).My STATCOM is suppose to supply reactive power to Vpcc based on the Vrms magnitude of the Vinverter, but with the presence of the capacitor and the output impedence resistor, there will be additional current supplied by this 2 branches since the Vpcc is in parallel with them which is undesirable since i only want the current supplied to the converter be controlled by the inverter voltage.Any ideas?
4).I know that LC circuits will introduce a phase shift between input and output voltage. In my case,is it possible to make Vinverter to be in phase with Vpcc?


Regards,
alvin

 

Hi SgtWookie,
Thanks lots for your detailed explanation. It's definitely eye opening

I figured those plots would get your attention.

Unfortunately,i have still 4 issues to deal with.
1).Assuming my output impedence of the inverter R1 is 1 ohm,looking at my attached thumbnail, should the output impedence be 1 as well?

It's ideal if input and output impedances are matched. If there is a mismatch, you wind up with reflected energy. That makes the return loss go up.

2).Should the output impedence R2 be connected in this manner?

Your output R2 should be in series with the load, not in parallel with it.

(how do you calculate the output impedence on your previous examples?

The input impedance is supplied to Elsie, the program that performs the calculations for the values of L and C in the filter. The output impedance is calculated by Elsie.

3).My STATCOM is suppose to supply reactive power to Vpcc based on the Vrms magnitude of the Vinverter, but with the presence of the capacitor and the output impedance resistor, there will be additional current supplied by this 2 branches since the Vpcc is in parallel with them which is undesirable since i only want the current supplied to the converter be controlled by the inverter voltage.Any ideas?

Try connecting R2 in series with the load.

4).I know that LC circuits will introduce a phase shift between input and output voltage. In my case,is it possible to make Vinverter to be in phase with Vpcc?

I didn't bother to plot phase shift, but that's a definite concern. If you're out of phase even a small fraction of a degree, you will see substantial power loss.

You should download Elsie and experiment with it. Try reproducing the results I've shown for starters. I've already posted many of the input parameters.

One thing I can tell you with absolute certainty is that designing and building truly well-performing LC filters is an iterative process; you will almost never get it right the first time, even using sophisticated software tools. There have been many huge volumes written on the subject, and I've barely scratched the surface.

One thing working much in your favor is that you are working with very low frequencies; parasitics won't much come into play. However, you will get "bitten" by tolerances. Even a small change in L or C can thrown everything in the toilet.

When you see Elsie giving values to 5 decimal places, there's a reason. Look at the plots, and then try rounding the values to the nearest 5%, and look again. 5% often makes the plots look like spaghetti.

It seems like it should be simple. It isn't.

 

Hi,
there's still something i would like to clarify.Assuming a DC-AC inverter supplying a three phase load,assuming if the AC voltage output supplied has very low THD values due to the elimination of lower order harmonics , would conecting a inductor between the inverter and the load be sufficient to act as a filter since i'm expecting it to filter off high order harmonics? And does the input and output impedance matter for this case? If yes,how do i calculate it?
Thanks..