[MOVED] Puresine wave inverter

I plan to build this inverter but I am seeing some comments about it having a high standby current draw like 6 to 8 amps, just by looking at the circuit do you think this is true and if so how could that be fixed.

The current will be high when the output is shorted or heavily loaded. Without a load then the current is low.
You have an LM7812 being fed with 12V so it will not regulate. Its minimum input is 14.5V.

I consider it rather generous to describe it as "pure sine wave". The output will be AC but the waveform certainly won't be a pure sine and it will change according to the load current.

I also have serious reservations about the drive available to the output transistors. There are four PN junctions (two in the TIP122) to overcome before the PIC output voltage can possibly saturate the output transistors. From the data sheets, it needs something around 4.7V drive and I suspect under load the PIC can't produce that reliably. It would make far more sense to use Q1 and Q2 as common emitter inverter stages so their collectors can make use of the 12V supply.

Brian.

I was thinking that the Pic is generating a stepped sinewave and the transistors are a crude linear amplifier. Then the transistors will heat as much as the load.

Six amps at 12v is only 72 watts, which is quite low for a 5Kw inverter with a low frequency output transformer.

To do better, you will need a much more efficient output transformer that has a much lower no load magnetising current.

Try connecting the secondary of the transformer you plan to use direct to the 220v mains supply and measure the current it draws under no load.

Usually no load power is not an important design consideration for most transformers.
The designer tries to design it to produce the rated output power using as little iron and copper as he can get away with. He is not designing this for fun, but to make something he can sell for a profit. To stay in business he needs to make it as cheaply as possible.

So you get something that always runs warm, even with no load, and always sucks a fair bit of power, even with the output disconnected.

Only way to do any better is to design and wind your own scratch built transformer, where cost and size is of no concern. It will be better, but probably still way short of what you would really like to have.

Warpspeed said:

Six amps at 12v is only 72 watts, which is quite low for a 5Kw inverter with a low frequency output transformer.

To do better, you will need a much more efficient output transformer that has a much lower no load magnetising current.

Try connecting the secondary of the transformer you plan to use direct to the 220v mains supply and measure the current it draws under no load.

Usually no load power is not an important design consideration for most transformers.
The designer tries to design it to produce the rated output power using as little iron and copper as he can get away with. He is not designing this for fun, but to make something he can sell for a profit. To stay in business he needs to make it as cheaply as possible.

So you get something that always runs warm, even with no load, and always sucks a fair bit of power, even with the output disconnected.

Only way to do any better is to design and wind your own scratch built transformer, where cost and size is of no concern. It will be better, but probably still way short of what you would really like to have.

Click to expand...

SO your telling me if I have an inverter powered by 12v and has no load running that 6 amp standby current is nothing how can that be efficient, I thought that would be considered a lot of wasted energy that should drain batteries fast. Shouldn't it be pulling milliamps when its not under load?

- - - Updated - - -

I wouldnt say 5kva either even though the person who built it didnt specify how much output transistors they used and what size transformer, but for them to be complaining about power draw it could be that they didnt build it all the way up to 5kw what if that current draw is from them building it up to 1000w, but ill build it and test it out.

- - - Updated - - -

betwixt said:

I consider it rather generous to describe it as "pure sine wave". The output will be AC but the waveform certainly won't be a pure sine and it will change according to the load current.

I also have serious reservations about the drive available to the output transistors. There are four PN junctions (two in the TIP122) to overcome before the PIC output voltage can possibly saturate the output transistors. From the data sheets, it needs something around 4.7V drive and I suspect under load the PIC can't produce that reliably. It would make far more sense to use Q1 and Q2 as common emitter inverter stages so their collectors can make use of the 12V supply.

Brian.

Click to expand...

So in other words do you mean you think it would be better to remove the tip 122 transistors from the circuit and use the bc547 to feed the 2n3773 transistors? By the way Is there anything in the T0-3 package that I could use instead of the 2n3773 transistors even if its a lower power rating.

- - - Updated - - -

Audioguru said:

The current will be high when the output is shorted or heavily loaded. Without a load then the current is low.
You have an LM7812 being fed with 12V so it will not regulate. Its minimum input is 14.5V.

Click to expand...

It says 12v input to the regulator but it wont be getting 12v remember a fully charged battery should be giving atleast 13.4v or a deep cycle will be giving 14v fully charged or somewhere there but it would be more than 12v.

SO your telling me if I have an inverter powered by 12v and has no load running that 6 amp standby current is nothing how can that be efficient,

Click to expand...

Never said it was efficient, only that is pretty normal for this type of inverter.

I thought that would be considered a lot of wasted energy that should drain batteries fast.

Click to expand...

Yes it is a lot of wasted energy, and it will drain batteries fast.
But its all relative to the maximum power output.

Shouldn't it be pulling milliamps when its not under load?

Click to expand...

Haha, you will never see that from a low frequency transformer type of inverter.

Just get a large transformer, maybe of a few hundred watts, and plug it into the mains without any load connected on the output.
Measure the mains supply current going into it with a multimeter, and work out what that is in watts (VA's actually).
Its going to be maybe tens of watts, not the few milliwatts you say you expect.

I was thinking that the Pic is generating a stepped sinewave and the transistors are a crude linear amplifier. Then the transistors will heat as much as the load.

Click to expand...

I don't think so, the only way that kind of PIC can do that is to generate the sine voltage using PWM and even then doing it in software as that type doesn't have PWM hardware. Even if it was using PWM, the circuit would be incredibly inefficient to use an iron core transformer to filter HF PWM.

Considering the transistors have no bias and therefore can not operate in class A region and that the output of the PIC can only reach ~0.5V more than needed to make the transistors conduct anyway, it has to be digital anti-phase signals leaving the PIC. In other words it is being used to produce two square waves at a frequency derived from the crystal. It might have dead band timing included which might help a little but it seems a waste when a much smaller (8-pin) and cheaper PIC would do the job just as well.

So in other words do you mean you think it would be better to remove the tip 122 transistors from the circuit and use the bc547 to feed the 2n3773 transistors? By the way Is there anything in the T0-3 package that I could use instead of the 2n3773 transistors even if its a lower power rating.

Click to expand...

No, I'm saying to use the BC547 transistors as inverters. ALL the voltage reaching the output transistors comes from the PIC at the moment but it loses ~1V as it passes through each transistor, that doesn't leave enough to ensure the output stage is full switched. In fact you could remove the 12V regulator completely and connect the BC547 collectors to 5V and it would still work exactly the same. If you make a minor change it is all fixed - just connect the emitters of the BC547s to ground, put the 1K resistors and 100 Ohm resistors to the collector and feed them from 12V, also add 470 Ohm resistors between the BC547 bases and the PIC. If you do that, only two extra resistors are needed and it ensures the output stage can be hard driven.

There may be some advantage to connecting the TIP122 collector to +12V instead the transformer but if you do that you will need to add further current limiting resistors.

The 2N3773 is OK for small loads but it's saturation voltage is as high as 4V (33% of supply!) by the time it reaches 16A collector current so it doesn't make a good switching device. A MOSFET is a far better choice in that kind of circuit.

I think the whole design is "concept" rather than practical, clearly it won't work properly even at low output power at the moment. You should also consider that at the full suggested load, it will draw about 800 Amps from the 12V supply!

wow 800 amps, ok so do you think I should work on modifying this design to something thats good enough or should I use a different circuit all together

First consider that if it was 100% efficient, the input power would equal the output power. The power is calculated as the voltage (12V) multiplied by the current so to get 5000W out you would put 5000W in and the current would be 5000/12 = 417 Amps. In practice, that kind of inverter might at best be 50% efficient so the current would be about twice that.

The only way to efficiently convert 12V DC to high power 220V AC sine wave is to use a digitally generated PWM signal and low Rds power MOSFETs to do the switching. The circuits are quite complicated though and I wouldn't advise you to try building one unless you are experienced in power switching circuits. A circuit using that technique might achieve 95% efficiency.

There is a middle method that produces a "modified sine wave" which is nowhere near as good as pure sine but is much easier to generate. Basically, you use an efficient high frequency boost converter to lift the 12V up to about 190V DC and feed it to a H-bridge circuit. You then switch the bridge in one direction, then switch it off, then in the other direction, switch it off again then repeat the cycle. The output AC is like a square wave but there is a gap at each rising and falling edge where the voltage sits half way. It is nearer to a sine wave than a square wave but can still be troublesome if you drive reactive loads (motors etc). The circuit is efficient at producing power but the poor wave shape may make the load draw more current so the overal efficiency may be low.

Brian.

I did alot of readings and also built several pwm and modified sine wave inverter circuits none of which where to my satisfactions interms of driving a load but I saw that inverter circuit I posted up in instructables and the guy who designed it tested it and it works as he said and with his scope the wave looked pure but I know that if the scope has a high enough resolution it would be able to zoom in on the wave to show that its actually pwm closely together but I was ok with that Idea but showing you guys the circuit now I am wondering if its good enough to even build, even if the efficiency isnt that high I would still build it also I wouldnt build it all the way up to 5000watts at first I would try a smaller version to see how it performs.


I see people on youtube building puresine wave inverters with arduino uno but I can never get the code or schematic for those specific versions the only ones I see codes and circuit for are the pwm inverters do you guys happen to have any codes for projects like that?

First thing you need to do is think very seriously about what you want to use this inverter for.
If you really do need a very high peak power output, the current required at only 12v is just too high to be practical, no matter how many or how large your batteries are.

If you are stuck with 12v, maybe in a vehicle, its much better to run individual small inverters to run each individual load. Its extremely inefficient to continuously run a 2Kw inverter just to run a single small 50 watt load, or maybe no load at all.

If you are going solar, start off at least at 24v or even 48v. Its much easier to build an efficient inverter at the higher voltages because the current required is far lower than at only 12v.

I would want to use 24v to run my inverter I already have a 24v transformer for the project I was thinking of using a 24v regulator in place of the 12v regulator and I would build about 2 inverter 1 with about 1500 watts and 1 with about 2000 watts

Enzy said:

I would want to use 24v to run my inverter I already have a 24v transformer for the project I was thinking of using a 24v regulator in place of the 12v regulator

Click to expand...

The circuit does not need a 24V regulator and it will not work anyway since its minimum recommended input is 27V.
Your 24V transformer voltage is too high since the transistors have some voltage loss. You might need a 20V transformer.

how did the minimum reach 27v and the persons testing this circuit only use a regular battery to test it which would be 13v or 14v, remember its a 12v inverter.

- - - Updated - - -

Audioguru said:

The circuit does not need a 24V regulator and it will not work anyway since its minimum recommended input is 27V.
Your 24V transformer voltage is too high since the transistors have some voltage loss. You might need a 20V transformer.

Click to expand...

how did the minimum reach 27v and the persons testing this circuit only use a regular battery to test it which would be 13v or 14v, remember its a 12v inverter.

- - - Updated - - -

in other words if I was to use a 24v transformer I would be using a 24v supply also but in reality you know 2 fully charged 12v batteries in series will give more than 24v which would be roughly 27v as you said but why wouldnt it need a 24v regulator. Is it that it only needs enough voltage from the bc547 to turn on the tip122 so that the 24v from the center of the transformer can pass through it to ground and back again?

Audioguru is quite right.
You are going to need a 20v to 220v transformer, otherwise you are not going to reach 220v at full load under normal full battery operating conditions.

You cannot just stick any transformer into this and expect it to work well. the transformer design is absolutely critical for success.

The other thing is that a 24v nominal input inverter may possibly see over 30v if the battery is being charged while the inverter is still connected.
Most common voltage regulators are only rated for 30v maximum input and could possibly fail causing huge damage to your inverter.
It would be better to use regulators with a much higher rated input voltage such as an LM317HV set to give the required +5v and +12v dc outputs.

The circuit you posted has 5 voltage losses. If you feed a 20V sinewave to a 24V to 230V transformer then its output will be (20/24) x 230= 192V.
But the 24V to 230V transformer is actually 230V to 25V then its output drops at full load. Then the 20V sinewave is actually feeding a 25V transformer and the 230V output will be (20/25) X 230= 184V.

I understand that, In my case its 110v that we use in my country but since there does seem to be much issues then differently from what your saying ill just wait to get the crystal then build it even if the voltage is low thats ok I just need it to work to get transformers I have to order them online so once i know its working with my 24v then if its low then ill get another

do you think I should work on modifying this design to something thats good enough or should I use a different circuit all together

Click to expand...

I hear most replies suggesting to use a different circuit.

Apart from this clear statements, modifying the present design to something better has prerequisites. You must understand the circuit completely, identify it's issues, estimate voltage and current waveforms, calculate component values, determine if all components are operated within their maximum ratings.

Well I wanted to make changes based on what I was told by persons with the understanding more than I do, I got aid by one person so far but since I will be building a pcb for it I think I will just build it as is and see how it works I wnt be able to test all aspects of it like how you guys would but what ever is to be tested you could direct me and I would after I gain all the results then parts can be changed.


Personally I would preffer to be using Fets only.