Transformer-less Inverter Design Discussion

[thannara123]

Pleas discuss A up to 500 watt transformer-less Inverter's design structure , advantage ,disadvantage etc ..
I am planning to make it ?

 

[Warpspeed]

This has been done before, its not a new idea.

One way to do this is to directly switch a high voltage bank of batteries to produce a stepped sine wave output voltage.
It works, and its very efficient, but also very dangerous.

Maintenance of the batteries and battery wiring could present an interesting hazard, as some cells might be up at 350 volts dc or more.
Not for me, the whole idea terrifies me.

 

[chuckey]

What are your input and output voltages? You could do it by charging capacitors in parallel and discharging them in series. But as the voltage output goes up the ripple through the capacitors increase so the ESR of the capacitors become of great concern.
Frank

 

[thannara123]

Input Voltage is 12volt and output voltage is 230 volt

 

[Orson Cart]

So for transformer-less you still need some type of boost ckt to get your 12v up to 350VDC, usually this would be using a push pull ckt - which has a kind of transformer (tapped push pull inductor) to do this.

The draw back of transformer-less is the output is connected to the input, so for a big bank of solar panels on a roof the DC wiring (up to 500V sometimes) has to be rated for 230Vac mains as the panels are connected to the inverter o/p so the insulation system from the silicon in the panels to the frame etc has to be good and rated for 1000V generally.

 

[BradtheRad]

This idea is what you'd call half-baked, but it looks as though it might have potential.

An LC arrangement has the effect of turning switched DC into a sinewave, going back and forth through it at resonance.

This simulation converts 12 VDC into 100 VAC sine.



The op amp switches an H-bridge in everyday fashion. A real circuit will need a more sophisticated control method. (The simulated switches are a 'cheat' component available in the simulator.)

The op amp detects zero crossings on a small value resistor. This automatically occurs at the resonant frequency of the LC combination. The alternating action may stall if it is interrupted.

This method must be operated with a load. If the load is light, or absent, then massive current builds.

The L & C values need to be selected for best efficiency with a given load. The values can be selected for a target 50 Hz resonant frequency, or 60 Hz, or any desired frequency.

Two loads can be driven. They will divide the power output between them, however.

My simulation shows 83 W peak output. Efficiency is debatable. Parasitic resistance takes a bite out of performance. If you can minimize these resistances, the reward will be greater efficiency as well as greater amplitude at the load.

 

[Warpspeed]

Very clever Brad !!

Something like that should work much better off a higher dc input voltage.

So how about first stepping up the incoming +12v with a boost converter up to a respectably higher dc voltage.

If the resonating inductor was designed to work just below magnetic saturation, that might reduce the no load over voltage problem.
Ferroresonance of sorts....

 

[FvM]

Using resonant converters can be advantageous, particularly to deal with the unavoidable leakage inductance of a high frequency transformer.

I doubt however that using a single inductor boost topology (either conventional or resonant) is a good idea. The basic problem is that the 12:230 or 12:300 boost ratio also multiplies the switched power (Imax*Vmax) by a large factor. Means that the 50 A switches which must be rated for e.g. 40 V in case of a transformer converter must now handle 400 or 600 V. Switching losses are increased respectively.

That's why the classical topology 12:300 V DC/DC converter with high frequency transformer and high voltage output bridge can be hardly beaten in terms of efficiency, size and component costs.

The difference between a transformer with isolated windings and a "tapped" inductor (autotransformer) isn't that important for 12:300V ratio.

 

[Warpspeed]

I doubt however that using a single inductor boost topology (either conventional or resonant) is a good idea. The basic problem is that the 12:230 or 12:300 boost ratio also multiplies the switched power (Imax*Vmax) by a large factor.

Very true.
but how about 12v x 5.3 to about 64 volts dc.

then 64 volts dc x 5.3 to 340 volts peak ?

An in/out ratio of about five is not too harsh for boost topolgy.
And it would have to be a lot more efficient than trying to do it all in one go.

 

[Orson Cart]

Hello Brad, your technique is used a lot in resonant power converters too, 200 - 450kHz, in your ckt you can regulate the o/p to 230Vac by varying the applied freq (up for less power, down towards res for more) but of course you can't stay at 50Hz with this method. A very famous paper was posted once by HP (~1982 I think) showing the same thing but they designed a variable inductor (varied with applied DC) to give a variable resonance freq in the power ckt with constant applied freq - this worked very well and would be great for your ckt...!
p.s. take the o/p across the C...!

 

[BradtheRad]

Thanks for comments. I was happy to think maybe I found an inexpensive, simple way to do the job. However I realize it's not suitable to put on the consumer market.

The resonance principle can also boost sagging AC levels, in areas where the power company is overburdened (for instance, India). An LC filter can easily increase it 100 percent.

The fact that the inductor value can be smaller with greater current draw, gives the concept even more appeal.

However I guess it is impractical at this point. It's crucial to adjust the L & C values for the load and frequency. And if a load is absent, current levels build until they can become destructive.

A very famous paper was posted once by HP (~1982 I think) showing the same thing but they designed a variable inductor (varied with applied DC) to give a variable resonance freq in the power ckt with constant applied freq - this worked very well and would be great for your ckt...!

A variable inductor... Such a thing would certainly make it easier to create a usable AC voltage booster.

 

[Warpspeed]

Brad,
It could still have application if each resonant inverter only had to drive one very specific load.

I built something like this many years ago that ran at 20 Khz for an induction heating furnace project.
I used voltage feedback from the tank to phase lock a VCO that in turn drove the full bridge tank switching circuit.

The advantage was that the switching to the tank was always at the exact tank resonance peak, even when reactive loads tended to pull at the tank resonance frequency around.

The tank was current fed (via a 40Khz phase locked buck regulator) and that had feedback from the tank voltage amplitude.
You could set the desired tank voltage with a potentiometer, and it stayed pretty constant, regardless of load.

Current feed also made it far less prone to self destructing if the VCO momentarily lost lock, or during start up and shutdown.
The tank drive current was commutated during voltage zero crossings.

It would certainly be a lot more challenging to do at only 12v.