Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Using ir2153 during smps test FET and IC are burned out after 30 seconds after

Status
Not open for further replies.

eonscom

Newbie
Joined
Oct 22, 2017
Messages
2
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
29
Good morning
30 seconds after power is applied during SMPS test
FET and IC are burned out
I do not think there's anything wrong with the circuit, so let me know which part is wrong. 2017-10-22 03.19.49.jpg
 

Attachments

  • 2017-10-22 03.14.04.jpg
    2017-10-22 03.14.04.jpg
    550.7 KB · Views: 800
Last edited:

Hi,

I assume your schematic is correct.

But this is a high frequency, high current switching circuit.
It needs a carefully designed PCB layout. Extremely short traces, fast decoupling capacitors at any DC node..and they all need to be connected to a solid GND plane. Single GND traces have too high stray inductance. This causes ringing and high voltage overshot during switching.

These overshots are short in time, maybe only a couple of nanoseconds, but they hurt the semiconductor internal isolation barrier. Not immediately. Such circuits may run a couple of seconds ...or up to several month until the isolation barrier totalky brakes down.

The board you are using is not useful for such circuits. At least a two layer PCB is needed.
An experienced user may be able to modify your circuit...a capacitor here, a diode there, a resistor to reduce rise and fall times...and it may work reliably.

I recommend to read some design notes, application notes, datasheets of MOSFETs and their gate drivers. Usually they explain how to design an appropriate layout.

With a fast scope you may see the overshots.

Klaus
 
also... you might consider using a faster diode for the boost rail (pin 8), the 1N4007 isn't much use above a few KHz.

Brian.
 

I wonder if the switch-off delay caused by the large gate resistor might completely eat-up the inbuilt IR2153 dead time, ending up in shoot-through?

Using slow 1N4007 for bootstrap diode is inappropriate.

What's the actual load impedance?
 

Hi Eonscom,
Since your mosfets are getting burned, then it is very clear that your mosfet is doing something more which can't be tolerated by it.
Obviously it is draining more current
Think for reason which creates this scenario

Since your IC has a built-in high side driver, you are using two N-channel mosfet instead of a P-channel mosfet to drive high side.

Imagine when both of your mosfet are turned on simultaneously.
Obviously then it is short circuit to ground right?

Since you are operating at high frequency it takes some time to burn fully, else it would happen within a couple of seconds.

So now the debugging.

Slightly reduce the operating frequency

Check for both mosfet gate drive overlap

A sufficient gap between turn on signals will do good and check for some dead time control in the IC datasheet, since it is not a pwm controller I guess, it will not be facilitated with that, but have a check.

Slightly improve the value of the gate resistor

Use a flyback diode with a small resistance in series across the mosfet gate resistor


This should do the job!!!!!

Congrats!!!!

I could understand how it pricks when a circuit promoted to be simple in online but not working for us.

Many people promote circuits without testing, and if the circuit is working for someone it is not mandatory for everyone who test the circuit will work in the same way
 
A good heat sink is also needed.
It's difficult to read the MOSFET numbers from the picture but it looks like 2SK2950. If it is, that device is completely unsuitable for this application, it has an RDSon of 23 Ohms!

Brian.
 

I read 2SK2850, which has considerably lower Rdson. But we don't have a load specification, I'd say any NMOS with sufficient Vds rating can work in this circuit, with appropriate load. Similarly I can't determine from the given information if a heat sink is needed.
 

thank you for answering

I pointed out things I could not think of
I will try the test again.
 

It appears you are feeding AC to the circuit without rectification?

What is the inductance of the 50 turn coil?

What is the selected switching freq?

(your basic idea is not wrong but there may be small mistakes).

Reduce the gate resistor (5-10 times!)

I pointed out things I could not think of I will try the test again.

I cannot imagine how much money I have burnt this way (and thank everybody for their understanding...)- so take heart!
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top