IR2110 pin 1 and 3 short in high-side

Thread Starter

Lordcx

Joined Nov 28, 2018
8
Hi
I want to build a half bridge inverter using Arduino Mega 2560, IR2110 and 6N60 MOSFETs. So I generated a sine wave in Arduino and its invert as attached photo where the blue signal is original and grey one is its invert . The schematics of the circuit is also attached. The problem is when I apply the gating signals (a 1k resistor is connected at the out put as load), the mosfets drain about 2A. After some review of the circuit and its component connections, I found that IR2110 pin 1 and 3 getting short at the first few seconds. I also tried to create dead time by subtracting few values (2) from inverted signal means when the original signal is at 175, the inverted one is at 255-(175-2)=82. The frequency of the gating signal is about 1hz (can be changed from 0 to 50hz using 10k pot in Arduino programming ), and all GNDs are connected to each other. How to solve the problem? Should I replace the sine gating signal with square one or other type signals?
Thank you in advance.

Capture.JPG IMAG0273.jpg
 

Attachments

Alec_t

Joined Sep 17, 2013
14,280
Should I replace the sine gating signal with square one or other type signals?
Definitely. If you are driving the gates with a 1Hz sinewave the FETs will spend a lot of their time only partly turned on and will generate a LOT of heat.
For generating, say, a 50Hz sinewave at the inverter output, the usual method would be to use PWM, with a high pulse frequency (many kiloHertz ).
 

ebp

Joined Feb 8, 2018
2,332
Why are you using optocouplers?

Your "sine" waves are severely distorted and asymmetric. Something is wrong in the algorithm to produce them.

It is not at all clear to me what you are attempting to do. Typically in something referred to as an "inverter" a sine is synthesized by using PWM to produce a half cycle of one polarity then a half cycle of the opposite polarity. With a half bridge, this is done by using a "split" power supply - one side of the load can be switched to a voltage either positive or negative with respect to the other side of the load. Your load isn't shown. 2 amperes may be reasonable or completely outrageous, depending on what the load is.

I have doubts about your "dead time" method, but again because I don't really understand what you are attempting. Dead time must be produced by assuring BOTH switches are turned off. Optocouplers are very much the enemy of precision timing, but ... Once more, I don't understand if by 1kHz you mean the PWM counter is running with a full cycle period of 1 millisecond or if the period of your sine wave is to be 1 millisecond, so I don't know how much the delays, and more importantly, the uncertainty and asymmetry of rise and fall times, through the couplers will affect things. If the PWM period is 1 ms and dead time is properly managed, the optocouplers most probably are not detrimental.

I'd get rid of the diodes across the gate resistors. They probably aren't necessary and make a very high current path if a FET fails in a way that allows a DC path from the gate to other terminals.
 
Top