Maker Pro
Maker Pro

another bad circuit

pgib8

Jul 26, 2015
107
Joined
Jul 26, 2015
Messages
107
Hi. I came across another bad circuit at my work. It's producing a ridiculous amount of noise and actually failed conducted and/or radiated emissions during UL certification.
This project is not under my responsibility and I'm sharing it because I would like to hear your opinion, to provide you with a little challenge, and so that someone may also learn from it.

bad2.png
At the power rails from the left there is 330VDC. The output towards the right will be 3.5V. But that's not important.
The area highlighted in yellow apparently cuts power off during a surge test to protect the circuit from over-voltage.
Below that is the switcher IC. It has another mosfet inside which is mostly off but pulses on every now and then.

The reason I think this design is bad because when the switcher is off, the Source of Q1 will rise, effectively turning Q1 off. When the switcher turns on again, one way or another Q1 will also turn on but within this transition I believe there is significant oscillation. I suspect the switcher doesn't turn on instantly and that the Source of Q1 compared to the Gate will oscillate until the switcher is turned on fully.
I think the lesson for this designer should have been that it's not wise to use a N-Channel MOSFET as a high-side switch.
In this case Q1 should have been below the switcher and always kept on, or alternatively Q1 would have had to be a P-Channel MOSFET and wired so it would always stay on.

Next are some scope shots of the awful noise being generated here. I was using a battery-powered scope, completely isolated and instead of the ground lead I used the little spring tip, to increase measurement accuracy.

bad1.jpg

Feel free to share your opinions :)
 
Top