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

[--BawA--]

hello ,,
I am working on microcontroller based sine wave inverter,
till now i have successfully designed my dc dc converter stage , i have also designed my h bridge circuit which is driven by microcontroller.
now the problem is as i connect my LC filter to the output of hbridge ,my hbridge mosfets as well as the mosfets that are used in DC DC converter stage gets damage.
if i don't attach the LC filter ,,and simply put a resistive load (100w bulb), bulb glows ,and circuit works normally.

if any harmonics which has frequency equal to the resonant frequency, then XL =XC ,,so total reacance will be equal to 0, the only resistance that will left is the resistance of wire used in designing inductor , so it will short my output,,so to avoid this condition, I have put a 100W bulb in series with capacitor ,so that it can act as a damper.
since on attaching the filter ,my DC DC stage 's mosfets also blown it clearly means that there was a short circuit in my hbridge output stage, beacause i have checked my DC DC converter stage separetly with a load of 200w bulb and the circuit worked normal,,i have also checked my hbridge (without attaching filter) , it also worked normally,
one more thing if i apply a low voltage dc to input of hbridge + LC filter , i get a nice sine wave,,
so i think there is a problem in my filter design,
i have used 5MH INDUCTOR ,wound on ETD 39 (1mm gapped) ferrite core,
carier frequency is 16khz, cutoff is 5khz


and 0.22uf CAPACITOR (non polarized box type)
i am attaching some images of my inverter
this is my DC DC stage


this is my Hbridge


below is my LC filter




please anyone help me,, i think there is a problem in my filter design ,,either my core selection is not good or some other problem,,please guide me,,i am working on it since a large time,, and i have made my circuit 31 times again n again ,,but i got only failure,,,, i have checked my spwm waveform after Hbridge on DSO and it was as expected ,,so where is the problem..

 

[pradeep.g.belchada]

which type of mosfet you used mention both in dc to dc & ac side

 

[--BawA--]

irfz44 in DC DC stage
and IRFBE30 in Hbridge,
as i already mentioned that if i dont attach LC filter to my hbridge, my circuit works fine on resistive load,,so i think there is no problem in my DC DC stage as well as HBRIDGE,,,
perhaps there is a problem in filter ,

 

[pradeep.g.belchada]

i think their is a problem in currnt limiting circuit in inverter. 0.22 should be 275vac x2 cap

 

[--BawA--]

i have used 0.22uf capcitor (box type non polarized),,is there any problem in my filter design? if yes , plz tell me how to proceed?

 

[pradeep.g.belchada]

can you upload your schematic

 

[BradtheRad]

If you are turning square waves into sine waves, using LC in series with the load, or as a second order low-pass filter...

The cap and coil values need to be matched to the load, or else you are liable to get intense spikes, or excessive voltage swings, etc.

 

[--BawA--]

Hi breadrath
What should i do to match the cap n coil values to load??

 

[BradtheRad]

I could be wrong, because I have assumed your filter is carrying full power at 50 Hz (as the resonant frequency).

In case you are using a series LC filter...

These simulations show the effects of various coil and capacitor values.

Notice that the correct values (left) result in a near sine wave going at the load.

Furthermore current in the loop is at zero when the power supply voltage changes state. This minimizes the stress on your switching components.



The simulation below shows two LCR loops. The one on the right is powered by sine waves. It provides a pattern for finding the correct coil and capacitor values. These values are then used in the left hand loop which is powered by square waves.

Also shown are the formulae for calculating values.



---------------------------------------------------------

On the other hand, if your filter is a low-pass butterworth...



It is not so easy to calculate values for the coil and cap here. I think I remember using a calculator for speaker crossovers.

Notice that current is non-zero when the power supply changes state. This is liable to stress components.

 

[FvM]

There has been a detailed discussion about possible reasons of failure with inductive (LC) filter load of your inverter. https://www.edaboard.com/threads/290299/

 

[--BawA--]

thanx bradtherad
i got the idea, your simulations were really helpful.
actually i am using SPWM wave to drive my hbridge, can you guide me complete step to design the filter for SPWM,,?
carier frequency is 16khz
resonant frequency is 5khz
fundamental frequency is 50hz
inverter rating is 220v 400watts approx

 

[--BawA--]

Is back emf from the filter is responsible for damaging of mosfets?
Should i connect external schotkky diode to every mosfet of h bridge for protection from back emf??

 

[FvM]

Is back emf from the filter is responsible for damaging of mosfets?
Should i connect external schotkky diode to every mosfet of h bridge for protection from back emf?

Not particularly back emf. MOSFET backward diodes are already cutting it in a H-bridge. Slow switching of the backward diodes may be part of the problem, but it can't be easily improved by parallel connected external diodes.

 

[--BawA--]

Thanks fvm
But what should i do now..
I have followed everythig that u had said in my other thread. I changed my gate resistor to 100ohms
Bootstrap capacitor to 47uf
Gate source resistor is 4.7k
Also i fabricated my pcb..
I attached a 100 watt bulb as a load to HBRIDGE(filter not connected). And bulb glows without damaging the mosfets. But as i apply a LCR filter , my mosfets blown out also mosfets of dc dc stage blown out..i used a 100watt bulb (R) in series with capacitor so that it can act as damper..


Also if i test my whole setup at 12vdc input with LC filter ,i get a clean sine wave...
Now tell me what should i do ..? Where is fault in my design ...?

 

[FvM]

I summarized my thoughts about the problem in this previous post. https://www.edaboard.com/threads/290299/#post1249345
Unfortunately I'm unable to get further insights from a distance.

 

[--BawA--]

Plz tell me if anybody have idea about where the fault exist?

 

[pradeep.g.belchada]

why cant you upload the schematic other members will also get benefit from it

 

[FvM]

You find a schematic in the previous thread linked above. The OP can tell if it's still valid. According to recent posts, bootstrap capacitors and gate-source resistors have been adjusted to avoid triggering of the undervoltage lockout which could happen with the original 22 µF/1k values.

 

[BradtheRad]

I think your filter is conducting excessive current to ground. This is possible if (a) the series coil is too low a Henry value, and (b) the capacitor value is too large.

This can be cured if you raise the coil value, and reduce the cap value. It is possible to keep the same rolloff frequency, if the changes are done in equal proportion.

 

[--BawA--]

i am using 0.22uf capacitor and 5mh inductor,
i have used ETD39 core (1mm gapped) for making inductor,
inverter rating is 220v 400VA approx. 50hz,,, carrier frequency is 16khz ,
is this not a correct value?