Device power rating for push pull

[krish2487]

Heres a stupid noob question.

I am presently trying to design a 36 v to 330 v push pull converter using 3525

Heres where i got stumped

If i want to deliver 100w from this converter then what should be the switching devices power ratings?

Obviously they should be low voltage high current types
I was looking at a 60 v device
But the current rating got stumped
I am not trying to operate in a linear fashion that if i try to deliver 100 w of power then my devices should also be rated at 100w power dissipation.

or is it??

how should i arrive at the switching devices power ratings?
The voltage is dependant on the supply voltage but how about the current rating??

and while we are on the topic
heres another one
the 3525 can be used to drive power mosfets
but i do not exactly understand the relation between the gate drive capability of the 3525 and the MOSFET gate capacitance
does anybody have a white paper or a application note explaining the gate capacitance to the gate drive.

[sawreyrw]

krish,

You need to do some homework on push pull SMPS, but here are some hints. The voltage swing on the collectors (or drains) of the primary switches is 2 times the supply voltage PLUS any L*di/dt voltage due to leakage inductance of the transformer. Snubbers will help limit the voltage spikes, but you need to use at least 100 volt transistors; 150 volts would be better.

The current rating is not exactly what you are looking for. Instead, you need to be concerned about power loss in the transistors. This is determined by Rdson and the switching losses. Again, do some homework on this. Same comment for driving the gate of the MOSFET. Many vendors have good information on these subjects, so that would be a good starting point for you in your search for more information.

Rick

[Enzo]

What Rick just said.


I have to react to your statement that if you need 100w then you use 100w parts. NO, not at all. WOuld you drive your car around town revved up to the red line all the time? I don;t want to come very near to my red line, I'll say.

If for whatever reason you need to dissipate 100 watts - and refer again to what rick said about what the parts have to do rather than the power that results - your parts ought to be good for at least twice that. Another analogy, if you weigh 200 pounds and need to climb on your roof, would you buy a 200 pound rated ladder? Maybe a 250 pound ladder is a better idea?

[krish2487]

as i said i am only a beginner trying to learn these things

as per the suggestions
i did my homework.
and this is what i found out
a good paper which i hope will be help to beginners like me

[cbdb]

Heres another way to look at it: If your supply puts out 100 watts and is 50% efficient the supply (mostly the mosfet) will burn 100 watts. If its 90% efficient the supply (mostly the mosfet) will burn about 12 watts.

[TechGuy]

A better option would to use a half-bridge topology for this PS, especially since your a beginner. With a half bridge your less likely to run into core staturation problems and you can limit the current applied into the transformer using smaller switching capacitor between your switching transistors and the transformer. For a Fixed switching frequency PS, the smaller the switching cap the lower the energy input per switching cycle. This also avoid blowing transistors when you accidently saturated the core. The only issue is that you have to avoid shoot-through current if you accidently switch both transistors on at the same time. Although there are MOSFET half bridge gate drivers that have built-in shoot through protection.

It would be a good idea to find articles or App notes for the SG3525 on Push-Pull or which ever topology you decide. I suspect that you may find more app-notes and articles with the Half-Bridge Topology using the SG3525 then other topologies.

FWIW: I would recommend using MOSFET Gate Drivers instead of driving them directly using the SG3525. Using a external gate driver will make it easier to construct your PS.

The issue is the amount of current that the SG3525 can deliver to drive MOSFETs. MOSFETS with low Rds-ON(sat) typically have high gate charge. If your output gate current is low it will take a considerablly more time to bring up the gate to saturation. During the period when you apply gate current to the mosfet and before it reaches saturation, Rds will be very high, causing excessive switching losses which will dissipate a lot of heat into the switch transistors. Its important to keep the switching time from off to saturation low to avoid this issue. its also possible that the built in BJT transistors can be blown if the output current exceeds there maximum current limit. One of the several reasons why gate resistors are required is to prevent exccessive gate current from blowing the drivers internal transistors.

BTW: You will need to use a MOSFET Half-bridge Driver for the Half-bridge or a gate transformers since you can't drive the High-side MOSFET directly using the SG3525.

There are also numerious books about designing SMPS. Many of which, are available on Amazon or other online book stores. It would also be a good idea to design a PS with a much lower output voltage for your first design 330VDC is lethal. It make more sense to experiment with a low voltage system first.

[krish2487]

thank you all

especially techguy

that really clears up my doubt

i m presently experimenting with low voltage versions
open and closed loop and its compensation.
it ll be a week or two before i can do some high voltage/power versions.


presently for testing i am not using dedicated drivers but i plan to use IR2113 as the power requirement goes up.

[bountyh]

Originally Posted by krish2487

Heres a stupid noob question.

I am presently trying to design a 36 v to 330 v push pull converter using 3525

I was looking at a 60 v device

The switch transistors must have a voltage rating greater than TWICE the maximum supply voltage PLUS the voltage spike from the turn off switch. I recommend at least 100V for your app.