schootky diode vs ultrafast one

Status
Not open for further replies.
S

Snub

Newbie level 5
Joined
Feb 24, 2015
Messages
10
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Visit site
Activity points
84
I want to reduce the heat generated in a circuit where the IRF3205 mosfet drives a flyback transformer with a pwm signal (fmax=50KHz) provided by the TL494 chip ,so optimal performances, Must I use a schootky diode (SBL3040PT VF=0.55V ) or a ultrafast one ( MUR820 VFmax= 0.97V trr=35nS) ?

p.s the circuits works perfectly,but I'd like to improve the efficiency..the irf3205 mosfet has inside a wheeling diode with a typical trr= 69nS and trr max=104nS VF=1.3V)

if I used the MUR820 diode,the internal diode will be bypassed,so less heat generated by the mosfet?

very thanks
 

I don't understand....please be more accurate...thanks
 

Transistor current in flyback is unipolar sawtooth waveform. Where do you see current through body diode?
 

I'm talking about a pwm signal ( a square wave signal) not sawtooth one and in the moment that mosfet turns off,the body diode allows the current to pass in an alternative path without any damage due to voltage transients..
 

You have been talking about flyback topology. Please clarify the circuit. Is it actually push-pull?
 

Snub said:
I'm talking about a pwm signal ( a square wave signal) not sawtooth one and in the moment that mosfet turns off,the body diode allows the current to pass in an alternative path without any damage due to voltage transients..
Click to expand...
Unless you're talking about a FET for synchronous rectification, this isn't true.
 

sorry I had given poor pieces of information: I'm driving a line output transformer , also called CRT, used in old TVs, with a single mosfet without any push pull circuit...I have no problem with the gate driver,because the switching frequency is low and the mosfet ( irf3205) has a modest/low gate charge..I used a simple SBL3040PT schottky diode as snubber, and i think about the power dissipation...is it clearer now ? thanks
 

With "snubber" you mean a diode parallel to MOSFET drain-source? It doesn't serve a purpose in a flyback circuit ("single MOSFET driving a transformer").
 

Snub, The gate drive has a big impact on temperature. The TL494 has only 200ma drive, that may not be enough to drive the gate capacitance of your FET. If the FET does not switch fast enough you operate longer in the linear region.

Take a good look at your gate drive waveforms and check the rise and fall times.
 

Some old CRT TV sets had what was called a "damper diode" across the CE terminals of the power transistor which drove the flyback transformer.
Click to expand...

The first point is if the flyback output actually fully swings back to forward bias the diode. If it does, the oscillation energy is pretty low and won't cause excessive recovery losses. A diode might be necessary for BJT protection, but not for a MOSFET.

For illustration you can refer to recent high voltage MOSFET technology. It's designed for the PFC boost and flyback converter demand, in both topologies there's no need for fast body diode. Unfortunately designers who want to make high voltage MOSFET bridge inverters are still waiting for MOSFETs with fast body diode.
 

but a high voltage mosfet ( High VDSS ) implies a high rdson,so the power dissipation is high...right?

the frequency is about 10KHz and maximum one is 50 KHz , I estimate a maximum gate current of about 30mA,thanks to the low Total gate charge and low switching frequency ... I repeat it , the problem is not the driver..
 

but a high voltage mosfet ( High VDSS ) implies a high rdson,so the power dissipation is high...right?
Click to expand...
I didn't suggest a high voltage MOSFET for you, just made a side remark about slow body diodes that aren't a problem in flyback converters. Sorry for confusing you.

Back to your design. There's no reason to add a fast diode parallel to the MOSFET.
 

The body internal diode is ridicolous because it has , in the case of irf3205, a voltage forward of 1.3V, this means more heat dissipated vs an external diode with a Vf lower....
is there a way to bypass the internal diode in favor of the external diode ?

I've done some tests today with 24V , by a couple of 12V sealed acid battery (12V 7Ah), and with a 8V gate voltage,the consumption of the circuit was about 15A and then irf3205 finished to live for some casual reasons ( medical bulletin : drain shorted with source)

the behaviour of my circuit is strange: it seems that works better in linear mode that switching mode ! I'm joking, it's actually impossible this situation
The optimal performance is at these points:

12V power supply

Vgs= 4.2V consumption= 1.9A the mosfet and snubber diode dissipate low heat

24V power supply

Vgs= 3.2 V consumption= 2.4A

the mosfet and snubber diode are warm...

this happens because I've wounded 8 turns on the primary side of the flyback transformer, more turns means more inductance so higher voltage spikes while less turns means less inductance so more current absorbed....

what a crazy world
 

You don't get it. In a properly working flyback, the body diode does not conduct any current (except in the case FvM mentioned, and even then it won't cause significant dissipation). It doesn't matter what it's forward voltage or recovery time is. It doesn't affect anything. Your problem is somewhere else.
 

It's strange what you say because otherwise there is no need to design a diode with lower Vf ( or simply schottky diodes are useless)....in my project,I undestood that recovery time is pratically useless because the switching frequency is low.

I can't still understand: the body diode conducts in the moment of the mosfet is switched off..I will post my circuit later..
 

Snub said:
It's strange what you say because otherwise there is no need to design a diode with lower Vf ( or simply schottky diodes are useless)
Click to expand...
In many converters, like the simple flyback, this is true with regards to the FET's body diode. Of course the output freewheeling diode, on the other hand, definitely should be fast and have a low Vf.
 

You are kidding me ?
So I was right
how can I post my scheme ? is it necessary an image or a fidocad scheme ? thanks
 

I believe not so many people are using Fidocad, but it can export pdf, that would be a preferred format for the attachment. Or export PNG and add as image to your post.
 

Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…