500VA sine wave inverter , problem in HBridge and filter designing

FvM said:

The error in simulation setup: It apparently uses MOSFETs without backward (substrate) diodes. They don't exist as technical devices.

In a real inverter, the diodes clamp the inductor flyback voltage to the power supply.

Click to expand...

Yes. I neglected to install these.

After revising my simulation, the spikes are now reduced (as you state).



At switch-Off, the coil kicks current back up through the power supply. Hence the power supply needs to contain a component in its output stage, which can offer low impedance to this reverse current.

At switch-Off, the coil kicks current back up through the power supply. Hence the power supply needs to contain a component in its output stage, which can offer low impedance to this reverse current.

Click to expand...

Good point. In an inverter, the regular bus capacitor won't have problems to absorb the energy. But at least part of the capcitance must be placed near the output stage, otherwise current transients could cause voltage spikes at the bus.

1) if i want to use external schottky diode in my mosfet irf840,what should be the specification of that diode?
2) what is the role of body diode in mosfet? i mean how it protects the mosfet from any transient in h bridge?
3) bradtherad , which software are u using for the simulations?

--BawA-- said:

3) bradtherad , which software are u using for the simulations?

Click to expand...

This is Falstad's interactive animated simulator. Free to download and use, at falstad.com/circuit.

Or click the link below. It will open Falstad's website, load my schematic (post #41) into his program, and run it on your computer.

https://tinyurl.com/lgu4sh4

You can alter values by right-clicking on a component, and selecting Edit.

To save the schematic to disk: Click File, select Export, copy the window contents, paste it into a word processor program, and save from that program.

Never heard of it before.i'll try it soon

1) if i want to use external schottky diode in my mosfet irf840,what should be the specification of that diode?
2) what is the role of body diode in mosfet? i mean how it protects the mosfet from any transient in h bridge?

Click to expand...

External diodes in MOSFET H-bridges are mostly effectless and thus unusual. In an ideal circuit, they aren't required because the body diodes serves the same pupose. With real MOSFETs, you want to have diodes with better specification than the internal body diode, but external diodes take only part of the current, so their effect would be rather limited. For bus voltages above 100 V, no diodes with sufficient low forward voltage are available.

An inductive load current is "commutated" to the body diode of high-side transistor when the low-side transistor turns off. When the low-side transistor turns on again, a slow body diode doesn't immediately stop conducting, a reverse recovery diode current is build up. After reverse recovery time, the diode current "snaps off" quite fastly, injecting current transients into the DC bus. If a critical dV/dt rate (e.g. 2V/ns) is exceeded during reverse recovery, the highside transistor might trigger a parasitic internal PNP transistor into latchup, ususally causing catastrophic failure of both output MOSFETs.

Possible solutions to prevent reverse recovery transients upto transistor failure are:
- use MOSFET with fast body diode, unfortunately not presently available above Vds of 500 or 600V
- avoid body diode conduction by synchronous control schemes
- slow down transistor turn on to limit reverse recovery dV/dt below critical rate, at least avoiding transistor failure. In a bad circuit layout, current transients might still disturb control signals, e.g. cause false switching of gate drivers.

FvM said:

avoid body diode conduction by synchronous control schemes

Click to expand...

what are the synchronous control schemes for driving H bridge? how to generate that synchronous SPWM?

FvM said:

After reverse recovery time, the diode current "snaps off" quite fastly, injecting current transients into the DC bus. If a critical dV/dt rate (e.g. 2V/ns) is exceeded during reverse recovery, the highside transistor might trigger a parasitic internal PNP transistor into latchup, ususally causing catastrophic failure of both output MOSFETs.

Click to expand...

i didn't get this, can u explain it in detail?

A synchronous control scheme switches high-side and low-side transistors alternatingly, instead of only switching the low-side transistor as in your design. With proper adjusted timing, almost no current flows into the substrate diodes.

Googgle "Peak Diode Recovery dV/dt"

is it possible to generate UNIPOLAR SPWM from single microcontroller? if yes , then how?

is it possible to generate UNIPOLAR SPWM from single microcontroller? if yes , then how?

Click to expand...

Some microcontrollers have pwm units that can control a H-bridge or three-phase bridge, e.g. STM32. They are supporting center aligned PWM and deadtime control for complementary outputs, so it's quite easy to setup unipolar modulation (3-level PWM).

i have to connect the LC filter at the output , i am attaching two images of filter,


1) which type of orientation would be better to use?
2) I have calculated the value of damper resistance,and it comes out to be 75ohms, but how to decide the wattage of that resistor?

The symmetrical filter is better in terms of supressing emitted EMI and reducing the capacitive load of the output bridge, the difference matters i the load is connected by longer cables.

Where do you want to place a damper resistor?

I want to place 75 ohm resistor in series with capacitor.

FvM,
i have done an arrangement in which the internal body diode of mosfet will be disabled .i have connected a Schottky diode in series with the Drain Cathode toward the Drain and an antiparallel high voltage high speed connected between the Anode of the Schokky and the Source of the FET -- this is done on each FET.
will it protect my circuit , ?

is it good to use 75ohm resistor in series with filter capacitor?

I expect that the diode circuit helps in thoose cases where reverse recovery of slow substrate diodes causes the problems. I also thought about this topology but it seemed akward to me compared to other solutions.

A 75 ohm resistor partly cancels the filter effect. If damping is actually required, I would refer to a C parallel RC combination in place of the capacitor.

FvM, i want to try that diode on each mosfet, if it will not work then i 'll change my scheme to synchronized SPWM scheme.
but i am confused in selection of diode, .. according to you which diode is suitable for this application? can u tell me any diode model no. ?

is it possible to generate UNIPOLAR SPWM from PIC SERIES microcontroller?
as i dont have idea of 32bit microcontrollers, so i wanna use PIC series microcontollers to generate the UNIPOLAR SPWM,
if even two PIC mcu are reqired , i'll be ok with them . but i don't know whether it can generate UNIPOLAR SPWM or not?

I presume it's possible with external circuits for dead time generation.

Is it not possible to generate dead time in software?

Yes, but the PIC processors don't have enough independant PWM outputs,