Which topology is good for 12V to 300V?

[arvinfx]

I am building an inverter from 12V to 310V 1kW. and I am really confused about topology .
I thing push-pull is good for low voltage inputs, but some people told me Half-bridge is better. But in half-bridge input voltage is divided by 2! and it will double input amps!

What should I do?

 

[tcmtech]

Standard push pull.

No need to over complicate a simple power handling circuit.

 

[ronsimpson]

You could use a boost or transformer boost topology.
You will be pulling 100A from the 12V.

 

[()blivion]

The word "Inverter" heavily implies AC, not rectified DC. Knowing which you are doing would be helpful. (300v @ 1kW sounds like tube land to me.)

 

[arvinfx]

You could use a boost or transformer boost topology.
You will be pulling 100A from the 12V.

How are you ron?
Have you had experience on boost at 1K and low voltages? I heard that core will be huge and Imax too

 

[MrAl]

Hi,

You havent yet answered the required questions. Is this a DC to DC converter or are you going DC to AC, and if AC, what frequency?

A bridge gives higher output with no additional work and it works with non center tapped transformers.

You can also use a boost converter to get the required higher DC voltage, whether or not it is AC output.

 

[arvinfx]

Hi,

You havent yet answered the required questions. Is this a DC to DC converter or are you going DC to AC, and if AC, what frequency?

A bridge gives higher output with no additional work and it works with non center tapped transformers.

You can also use a boost converter to get the required higher DC voltage, whether or not it is AC output.

Yes, this is a DC to Dc converter.

You say this style is better than simple push-pull?

 

[tcmtech]

The word "Inverter" heavily implies AC, not rectified DC. Knowing which you are doing would be helpful. (300v @ 1kW sounds like tube land to me.)

Sounds like European standards domestic 220 VAC 50 Hz peak voltage to me.

As far as switching circuits go almost all factory made power inverters, along with any other type of power electronics devices like audio amplifiers, that use a DC - DC step up stage to raise the input voltage to something else at a high wattage a center tapped type primary driven by two banks of switching devices is standard design.

Its the simplest cheapest and most effective way to handle high current and power switching at a low DC input voltage.

Why are you bothering to try and build your own inverter any way? A standard 1 KW factory built unit can be had way cheaper and be far more reliable than you will ever design and build one for.

Take your pick, https://www.ebay.com/sch/i.html?_od...0.l1313&_nkw=1000+watt+inverter+220+&_sacat=0 , Buy it now from around $33 and up for a 1000 watt 220 VAC power inverter.

 

[arvinfx]

Tcmtech you right, but I need a pure sine wave and they are too expensive , another reason : this project is a hobby for me

 

[tcmtech]

The easiest and most cost effective solution to that is to still start out with a cheap modified sine wave type power inverter and just change out the output H bridge driver system to a PWM shaped sine wave system.

It still saves you all the headache of having to design and build the step up power stages and related systems and you get a nice neat case to put it in. Plus to be honest you can buy a compete sine wave inverter driver board for around $10 and use that to modify any common modified sine wave inverter to be a pure sine wave inverter!
**broken link removed**

Specs.

5V single power supply.
4 pin set of pure sine wave output frequency: 50Hz pure sine wave of fixed frequency ， 60Hz pure sine wave of fixed frequency， 0-100Hz pure sine wave frequency adjustable， 0-400Hz pure sine wave frequency adjustable.
Unipolar and bipolar modulation,
Comes with deadband control, pin 4 dead time: 300nS dead time 500nS dead time 1.0uS dead time 1.5uS dead time
External 12MHz crystal oscillator.
PWM carrier frequency 23.4KHz.
Voltage, current, real time temperature feedback.
Overvoltage, undervoltage, overcurrent and overheating protection.
Soft-start mode pin setting the response time of 1S.
Serial communication to set the output voltage, frequency and other parameters.
External Serial LCD Module 1602 displays inverter voltage, frequency, temperature and current information .

But beyond that personally I feel that if you don't know enough about low voltage high current switching systems to be at the point of having to ask what type of layout you need I have strong suspicions that a full hand crafted sine wave inverter in the 1000 watt class is way beyond your present skill levels.

So there you go a semi DIY 1000+ watt pure sine wave inverter system (With a LCD display and full over current and voltage protection built in) that can be built for under $100 without the problems and hassles of designing the complex switching systems and step up stages.

BTW I just ordered four of them!

 

[MrAl]

Hi again,


Since you want to build this yourself i'll still post some circuit topologies so you can think about all this for now.

Since you now have also revealed (since last time i read this thread) that your final project will have an AC output, i think i should add that in some of these topologies you can go right from DC to synthesized AC, where the synthesized AC is a much cleaner AC pulsed wave that is not as simple as the kind you buy on the web. That means the output AC filters to near perfect without too much trouble. The kind you buy on the web usually put out a single pulse with output amplitude equal to the average AC output, but using multiple pulse techniques allows going from DC to AC (with a transformer and full wave bridge as in E in the diagram but with no rectifiers) and the AC output is quite clean. The only thing that changes is the switching pattern of the four transistors.

Note that there are some other little things that come up like switching spikes across the transistors which requires snubbing networks, one for each transistor. You also have to keep an eye on the input capacitors ripple current ratings and things like that.

In the drawings in the attachment you'll see that some topologies require a more complex transformer and some less diodes. The one in A is not a good idea because of the high input current and the loss of 1/2 of the input voltage due to the capacitors. E is well suited for DC to synthesized AC (without the diodes of course).

 

[arvinfx]

Hi again,


Since you want to build this yourself i'll still post some circuit topologies so you can think about all this for now.

Since you now have also revealed (since last time i read this thread) that your final project will have an AC output, i think i should add that in some of these topologies you can go right from DC to synthesized AC, where the synthesized AC is a much cleaner AC pulsed wave that is not as simple as the kind you buy on the web. That means the output AC filters to near perfect without too much trouble. The kind you buy on the web usually put out a single pulse with output amplitude equal to the average AC output, but using multiple pulse techniques allows going from DC to AC (with a transformer and full wave bridge as in E in the diagram but with no rectifiers) and the AC output is quite clean. The only thing that changes is the switching pattern of the four transistors.

Note that there are some other little things that come up like switching spikes across the transistors which requires snubbing networks, one for each transistor. You also have to keep an eye on the input capacitors ripple current ratings and things like that.

In the drawings in the attachment you'll see that some topologies require a more complex transformer and some less diodes. The one in A is not a good idea because of the high input current and the loss of 1/2 of the input voltage due to the capacitors. E is well suited for DC to synthesized AC (without the diodes of course).

What is the bold point on E? why not C?

 

[MrAl]

What is the bold point on E? why not C?

Hello,

C is fine for DC to DC, i meant that E was better for going from DC directly to AC without having to convert to DC first.

You can go from DC to DC first, then later to AC, but this requires handling the power twice. It is done that way in some commercial products, but going from DC to synth AC means only having to handle the power once.

 

[arvinfx]

Excuse me! How DC to AC works? its really new for me!

 

[()blivion]

Pulse-width modulation <--- DC to AC

 

[tcmtech]

Excuse me! How DC to AC works? its really new for me!

Well if you have to state and ask that then building a 1000 wat inverter with a DC - DC -AC or any other similar conversion process is way over your skill level.

What exactly do you need a 1000 watt sine wave inverter for anyway?

 

[arvinfx]

No no, mral and me were talking about Full-Bridge and he/she said to me:

i meant that E was better for going from DC directly to AC without having to convert to DC first.

I know what the SPWM is, I have already simulated this in spice.

Dear tcmtech , I live in Iran ( unfortunately it wasn't my choice) and there we haven't good power grid , it goes off every 2 week for 1-3 hours. This is the reason that I decided to make an inverter.

And about experience I have made a boost 150W 12-24V . a push-pull 350W 12 to +-35V and a flyback 220v to 50V 2A with a 12V 0.5A.

 

[MrAl]

Hi,

First off, very sorry to hear about your power line problems going off quite a bit, i know how troubling that can be because we had a bad storm here recently and we were without power for 4 or 5 days and nights. So i understand where you are coming from.

There are some problems though, and that is that when you use a 1000 watt inverter for an hour that means you need a battery set that can put out 1000 watts for an hour. You probably realize this. And that means you could be dealing with 100 AHr battery sets. It's possible but it wont be super cheap.

You live in Iran, not by choice, well i live in US also not by choice. It's funny that if we subtract all the governments we all becomes friends

Also to note that building one of these units may take some time for testing and making sure certain things work properly. The last thing you want to do is find out that it fails when you need it most. But there are important safety issues too that go beyond the usual home electronics project. One of these is that when dealing with the testing of these units parts can draw huge current levels and literally blow up like "ash cans" or very big fire crackers. And when they blow they shoot out hot molten metal in every direction as well is sharp bits of plastic or ceramic. And the electrolytic caps also can blow up for various reasons, sending chemicals spewing in all directions. It's not pretty at all and is very dangerous to humans standing nearby. Safety goggles and even a body shield is always necessary and very intense care when making measurements because the unit could be working fine one minute and then a second many of the parts blow up. So you really do have to be very very careful. You have to be aware every second during testing that the high current parts could explode any second. I've seen it happen wayyy too many times in manufacturing plants.

Direct DC to synthesized AC is almost the same as DC to DC when using a transformer, but the difference is that the pattern of transistor switching is changed to reflect a changing output DC level that is really the AC now. So with a DC to DC you have a constant (say) 100v output, but with a DC to AC you have DC output that is first 0v, then maybe 10v, then maybe 20v, then maybe 40v, then maybe 70v, then 100v, then back to 70, then back to 40, then 20, then zero, then -20, then -40, then -70, then -100, then -70, then -40, then -20, then back to zero and that's one complete AC cycle. So you are really making the "DC" change levels all the time. In fact, doing it in steps like that is one way to do it. But with PWM a simple strategy is to make the average PWM signal equal to the average AC at that point in time over the entire cycle. Because we use a circuit like "E" in the drawing (which is just called a full bridge) we can actually make the output go negative as well as positive. Of course this means we get an AC output and of course we dont use rectifier diodes because we dont want DC we want AC output directly.

So for the DC output we use a constant switching pattern, but for the AC output we use a changing pulse width that varies like a sine wave (the widths that is).
There is quite a bit of theory on this but the simplest explanation is that the average of the PWM matches the average of the sine wave at certain points in time. We choose some number of points like 10 per half cycle, then try to keep the sine average correct at those 10 points in time for both halves of the cycle. The result is a sine wave that is fairly clean and the more points we use the cleaner it gets. There's a limit, but that's another story.

A real simple example with just five pulses per half cycle, at 50Hz:
Each half cycle is 10ms long, and we have:
The sine of 30 degrees is 0.50,
the sine of 60 degrees is 0.87,
the sine of 90 degrees is 1.00,
so if we make the pulse at 90 degrees equal to 2ms then we make the other pulses:
00 degrees: 0.00*2ms=0ms
30 degrees: 0.50*2ms=1ms
60 degrees: 0.87*2ms=1.74ms
90 degrees: 2ms (1ms either side of 90 degrees)
and then repeat that for the other half of the half cycle at 120, 150, 180 degrees.
So you see from this rough example that the pulse widths change over the time period of one cycle.

The output filter can then be a small inductor and capacitor set.

You may not need that many switch pulses either, depending on what you intend to drive with this inverter.

 

[arvinfx]

Dear Mral thanks for your sympathy.yap I saw the Sandy on the TV and net . I was awful. I hope everything goes good. I got the point and I have an idea :

Can I use a special feedback for IC driver ? For example we can change the FB of TL494 and add a sine wave to it with Vpp=2.5v and 50Hz. so we wil provide pure sine wave. ( I am not thinking about negative stage ...)

 

[Timescope]

Here is a complete inverter design DG 5 SGA inverter design

Timescope