Hi,
I have been working on a project that generates dual output voltage (+15/-15) from a 12v battery. The circuit I have made consists of a mosfet H-bridge across a transformer's primary coil. The transformer is wound with 2 secondary coils. The purpose of the H-bridge is to produce a square wave( Vpp = 24v ) from a 12Vdc battery. The turns ratio is selected to produce a square wave ( Vpp - 30v ) across the secondary coil. I have selected the turns ratio such that the transformer's core doesn't go into saturation.

Iam using IRF540 mosfets in the H-bridge. The mosfets are driven at 1Khz switching frequency from high and low side drivers as shown in the schematics.

The problem is the voltage across the tranformer's coils doesn't maintain a square wave shape. It is like a ramp. Voltage starts at the maximum and keeps falling during the switching period. Can someone explain the reason behind this.

Below Iam attaching the schematics and the voltage waveform picture.

Thank you.

Coil voltage waveform

Mosfet H-bridge

High-side driver

Low-side driver

 

Is that the input or the output of the coil?
It would appear that the transformer is saturating, causing the droop.
When you calculated the saturation point, did you include the flat portion of the square-wave, which has a low frequency component?

Edit: The magnetizing current is roughly 1/2 the square-wave period times the supply voltage divided by the inductance.
This must be below the transformer saturation current.

If you calculate the saturation current from the square-wave RMS value using a sine-wave equation, then you need to use an RMS value about 15% above that calculated for a square-wave.

 

The waveform I have posted is the output secondary coil. Same is the case with the primary coil. The voltage waveform across the primary is exactly the same.

Does it still appear that the transformer is saturating? Or is there some problem with the switching circuit ?

Thanks.

 

Hi,

The upper portion of the wave is curved slightly indicating that the current increases with time. That would be a sign of saturation, or more exactly probably just going into saturation.

The way to know for sure is to look at the current not the voltage, or the current at the same time as you are looking at the voltage. If the current spikes up, that means the transformer can not handle the volt seconds for some reason. The reason could be that the transformer does not have enough turns, or there is a DC offset current in the primary. The effect of a DC current is the core starts to enter saturation even with a voltage that would normally work just fine with the transformer. You do need to look at the current in the primary to see this.
DC current in the primary normally comes from a voltage offset in the bridge circuit. This can come from one or more transistors being different than the others, where when one turns on it has a higher or lower turn on characteristic. This may be very hard to spot because it could be a very small difference, so looking at the current in the primary should tell more of the story. Usually some form of dynamic compensation is employed if it really is a flux imbalance, and i am not sure what you are willing to do here if that's the case.
Also be aware that if you use a small resistor to sense current in the primary it must be very small as the effect of added resistance reduces the DC current effect. Without expensive equipment though sometimes it's the only way to see the current.
You can see how the effect changes by keeping an eye on the voltage too.
Sometimes you can see the effect "not be there" when you first start the converter, then after a few seconds of run time the effect starts to show up. That is a sign of flux ratcheting.

 

The pink waveform is transformer primary voltage and the green waveform is the current in the primary coil. Since there are no peaks in the current wave can we safely assume that the transformer isn't running into saturation.

Thanks.

 

What is the peak current value and how does that compare with the calculated saturation value?

If that's below the saturation point then there would seem to be something wrong with the driver circuit.
Look at the voltage at all the bridge MOSFET gates.

 

Hi,
since I don't have a LCR meter I didn't calculate the primary inductance earlier and decided the turns solely based on magnetic field calculations instead of magnetizing current. I assumed that the core saturates at 1.2 tesla and wound the primary such that the maximum field would be 0.6T.
I guess the proceedure was wrong. I wound the transformer primary again with four times the turns of what used to be earlier and it solved the case. This is the final result


Pink waveform is the input voltage and green is the transformer primary voltage.

Although there are some ripples, I hope they will be gone if I increase the input capacitance. But the ramps in transformer voltage are gone. Saturation is the culprit here.

Thank you all again for you time.

 

ink waveform is the input voltage and green is the transformer primary voltage.

Aren't those the same?
Do you mean secondary voltage?

 

Green waveform is a square wave and pink is the input dc voltage. They are alternatively overlapped at the top. If you look closely you can see that pink waveform is 12VDC with some ripples similar to the transformer voltage.

Does increasing the input capacitance deal with the ripples?

 

The small dips in the waveform are due to snubber capacitance, which is more than required. Once it is decreased to its appropriate value everything is set. It is a complete square wave now.

Thanks.

 

Hi,

Sounds good

So the main problem seemed to be just not enough turns on the primary.

As the transistors heat up the flux balance could change a little so that's worth checking too.

 

Hi
I think you can greatly improve fronts if you put the grid resistance to much smaller MOSFETs. The manufacturer recommends for low switching times Rg = 9.1 ohms

 

Hi
I think you can greatly improve fronts if you put the grid resistance to much smaller MOSFETs. The manufacturer recommends for low switching times Rg = 9.1 ohms

Sorry but you English is a little mangled.
What's a "fronts"?
Only vacuum tubes (valves) have a grid.
MOSFETs have a gate.