1.5kw DC to DC Boost Converter Circuit

[Easy peasy]

Are you using a DC blocking cap and the same transformer? if no cap and same transformer then the magnetising current will peak up for 24V and kill your fets - what proof have you that you are applying too much voltage to the gates of your fets?

 

[Powerglide]

Are you using a DC blocking cap and the same transformer? if no cap and same transformer then the magnetising current will peak up for 24V and kill your fets - what proof have you that you are applying too much voltage to the gates of your fets?

What do you mean by same the same transformer?
My scope trace shows that the 50% duty cycle increases on both high side mosfets and if I measure the voltage on d1 at VB you can see it increase from 15 volts to over 25 volts or until mosfet fails. They fail immediately when I connect full power.

 

[Easy peasy]

Yes - but you are measuring the high side fet gate with respect to gnd - so you see Vrail ( 24V ) + V boobstrap ( 15V ) = 39V

to see the actual gate drive volts on the high side you need an isolated probe - if you answer my questions above you will see that you are putting too many volt seconds on the main transformer - which then saturates causing high current and dead fets - also check your bootstrap diodes aren't dead

If you have 15V only going to the boot-strap supply then you can't over-voltage the gates on the high side -

- - - Updated - - -

the same transformer as for 12V operation ... ?

 

[Powerglide]

It's the same core e55 but with different windings for 24 volt and I'm pulling a 100 watt light bulb which takes about 10 amps when I pull it from 24 volts. So the transformer is different.

 

[Easy peasy]

have you got twice the turns on the pri for 24V?, also an 100W light buld is a bad idea unless you have a slow soft start - or a soft 24V supply - as the cold resistance is only a few ohms and the current to heat it up is very large - hence the 24V side current will be >> 100A ...

- - - Updated - - -

Do you have a DC blocking cap - I assume it is full bridge per Thamid's sketches ?

 

[Powerglide]

Yes, on the 12 volt I had 1 turn and on 24volt I have two turns with 0.5mm copper foil covering complete cross section of core. The secondary turns are 39 with 2mm litz wire to give 400 volt. I also have a small soft start on the sg3525.
I have irfp 7430 on the 24 volt side which can handle 100 amps .

I don't have a dc blocking cap as I didn't needed one on the 12 volt and yes it's a full bridge with feedback.
If I'm right the dc block cap is on the one side of the primary blocking voltages from the reverse diodes of the mosfets?

 

[Easy peasy]

What is the core size of the Tx? - the blocking cap ( which would be quite large in this case ) stops DC current saturating the Tx - if you don't have a blocking cap then you need to run current mode control or have at least 5% dead time to allow core reset.

 

[Powerglide]

The Ae of the core is 354 and I have adjusted the dead time also with no success. On the 12 volt I had a 4 ohm resistor on the dead time of the sg3525 .
I think maybe I should add more turns on the primary?
Also why I'm thinking my gate voltage is high , is because my low side mosfet is charging the bootstrap cap so high to VB and that voltage is fed to the high side gate which I think is to much for the gate.

 

[Easy peasy]

OK, so, 50kHz = 170mT pk, 20kHz = 425mT - so hopefully your freq is 50kHz, and the dead time is 1uS at least ...?

 

[Powerglide]

Yes I'm running at 52khz but I'm not sure about dead time, I used a 4 ohm resistor on the 12 volts but having 10 ohms at the moment. Gate resistors are 5 ohms.
When I add resistance to the power load so that full power is not engaged, than everything runs normal and mosfets are not even heating up, it's just when I connect it straight to full load than it fails the fets

 

[Easy peasy]

dead time or discharge resistor should be about 5% of the timing resistor, try 47 ohm, 100 ohm - and a large soft start cap ... and a 40W light bulb or better still 2 x 40W in series for a load ...

- - - Updated - - -

it does sound like the core is saturating as there is no flux control preventing flux walking / staircasing ...

 

[Powerglide]

The voltage increase from 12 to 24 does not do anything to the dead time right?
Because the mosfets heated up with 47 ohm discharge resistor on the 12 volt.

 

[Powerglide]

Thanks easy easy, I only had to change the soft start cap, from 1 uf to 22uf, I left the discharge resistor on 10ohm becuase the fets still get hot on 47 ohm.
All that still bothers me a little is that q1 and q2 gets little hotter than q3 and 4 .

 

[Easy peasy]

This is the trouble with not posting a full & accurate schematic - because with full information the fixes can be posted all at the same time - i.e. results a lot quicker.

A good idea to have a turn off R, say 2.2 ohm in series with a back facing schottky ( 40V 1A ) to give a fast turn off, and then a 47 ohm turn on resistor in parallel with the series combination - this gives 47 ohm turn on which makes less RFI and is easier on the other fet and the o/p diodes.

Yon need to measure the dead time between the turn off of one fet in the totem pole and the turn on of the other - this is tricky without a true isolated input scope.

Please post some D-S wave-forms from the two sides of the H bridge as differing power levels ... kind regards,

 

[Powerglide]

A good idea to have a turn off R, say 2.2 ohm in series with a back facing schottky ( 40V 1A ) to give a fast turn off, and then a 47 ohm turn on resistor in parallel with the series combination - this gives 47 ohm turn on which makes less RFI and is easier on the other fet and the o/p diodes.


I will definitely give this a try and will post some wa forms over the weekend.
Kind regard

 

[Powerglide]

This is both HO signals from the full bridge pulling 600watt and on about 25% duty cycle.
Is it normal for mosfets to heat up when I pull small loads and it's running on 15 or 10% duty cycle?
Efficiency is good through all duty cycles.

 

[Easy peasy]

Is it normal for mosfets to heat up when I pull small loads and it's running on 15 or 10% duty cycle?

It depends on the wiring inductance ( layout ) and de-coupling on the power stage ... and the leakage in the Tx, and the o/p choke, and the diodes, and any snubbers used, and how good the heat-sinking is ...

 

[Powerglide]

If I wanted to add snubbers to my circuit, would it be a good idea to have them from the source of my mosfet feeding back to the positive supply rail?
After inspecting closely I saw that there was ringing, a specially on low duty cycles and increasing gate resistors only made mosfets run hotter.

 

[Easy peasy]

Start with 470pF 1kV cap and 47 ohm 3 watt R across each fet, D-S

- - - Updated - - -

and, 330pF, 1kV cap and 680 ohm 3W across output diodes ...

 

[Powerglide]

Start with 470pF 1kV cap and 47 ohm 3 watt R across each fet, D-S

- - - Updated - - -

and, 330pF, 1kV cap and 680 ohm 3W across output diodes ...

Resistor is in series with cap?
What do you mean by output diodes, on my secondary side of transformer of the bridge rectifier?