How a 98%eff Inverter is built

mike95490

An interesting article in Electronic Design, spills the secrets about building an inverter.

This is NOT a DIY or kit.

http://electronicdesign.com/Articles/ArticleID/21575/21575.html

excerpt:
Build An Efficient 500-W Solar-Power Inverter Using IGBTs

Combining optimized low- and high-side high-voltage IGBTs in a full-bridge topology, a dc-ac inverter uses a solar module to efficiently generate 500 W at 120-V/60-Hz single-phase ac output from 200-V dc input generated by a photovoltaic panel.
//
8 pages, and they do it without transformers, smoke or mirrors, just 2 smallish inductors
per fig #5, it's a sine output .

GreenerPower

Re: How a 98%eff Inverter is built

This is MSW design, you can easily get high efficiency with this. For true sine, it is much tougher.
GP

mike95490

Re: How a 98%eff Inverter is built

per fig #5, it's a sine output.

They modulate a 20KHz signal to create the sine. It may mess your AM band up.

GreenerPower

Re: How a 98%eff Inverter is built

You're right. I missed the part that they use PIC controller to do PWM for the sine output. Wonder why the 600W true sine is still much more expensive than the MSW counterpart.
GP

Solar Guppy

Re: How a 98%eff Inverter is built

Most GT inverters are exactly this for the AC stage ... nothing new here

What this is missing is galvanic isolation and variable DC input. For example, the works ok for 120vac but you need 400 vdc for 240vac ... and no provisions for a variable input ( like 200-600vdc ).

The Xantrex GT for example is a buck converter to make the 600-190vdc a constant 190vdc, which then goes thru a 1:2 stepup ( h-bridge ) which takes care of the galvanic isolation and creation of the HV needed ( 400Vdc ) , the similar h-bridge for the DC to AC step

And it does all of this ( 3 conversion processes ) with about 4% total losses

Also Mosfets are substantially better than IGBT's even in the voltage class. I'd say the Xantrex DC-AC section is about 99% peak efficiency.

Solar Guppy

Re: How a 98%eff Inverter is built

GreenerPower wrote: »

Wonder why the 600W true sine is still much more expensive than the MSW counterpart.
GP

Sine is two steps, the up converter ( 12 to 200 vdc ) and then DC to AC conversion

MSW is just a direct step up ( H-bridge )

GreenerPower

Re: How a 98%eff Inverter is built

Solar Guppy wrote: »

Sine is two steps, the up converter ( 12 to 200 vdc ) and then DC to AC conversion

MSW is just a direct step up ( H-bridge )

Well SG, what I see is a 12V-200V DC-DC on the true sine vs. 12V-150V DC-DC on the MSW (not much difference). Then the MSW H-bridge is switched at 60 Hz whereas the true sine is switched with PWM. The PWM switching is probably around a few hundred KHz. So, the difference is the adder for PWM controller (PIC, for example) and the faster MOSFETs or IBGTs. I don't imagine the cost adder is that much. Am I missing something ?
GP

Solar Guppy

Re: How a 98%eff Inverter is built

Yes, your missing the LC filter on the sine, in both cases the transformer is the voltage conversion , the sine then needs the smoothing transistion to make the 20-40khz a smooth voltage

GreenerPower

Re: How a 98%eff Inverter is built

Solar Guppy wrote: »

Yes, your missing the LC filter on the sine, in both cases the transformer is the voltage conversion , the sine then needs the smoothing transistion to make the 20-40khz a smooth voltage

I know there is an LC low-pass filter to remove the PWM 20-40KHz or so carrier, wouldn't imagine this filter costs that much at this high frequency. The OB and XW inverters are heavy with a transformer inside but I think this is for the DC-DC step up which should be the same for true-sine or MSW.
GP

RCinFLA

Re: How a 98%eff Inverter is built

I am curious on the Outback (and apparently the XW series) if the low frequency silicon-iron transformer is fed directly with PWM waveform.

I saw pictures of the inside of the Outback FX inverter and I did not see any large PWM filter chokes. Can the low freq transformer somehow be used to remove the high frequency PWM component?

mike95490

Re: How a 98%eff Inverter is built

RCinFLA wrote: »

Can the low freq transformer somehow be used to remove the high frequency PWM component?

Thats what we do at work, it's called a choke. You plan the core material so it stops
responding at a desired frequency, and then the coil inductance attenuates the high
frequencies. The selected core couples the 60hz to the other winding.