One of the most common causes of damage to our driver is ringing on the output causing the output pins to be driven above VDD or below GND. Is it possible to get scope shots of the output signal from our driver with respect to our GND pin?
Also, I only see 0.1uf capacitors being used for decoupling our drivers. Our recommended decoupling is 0.1uf caps placed as close as possible to the VDD/GND pins and a 10uf capacitor that can be a little farther away. We have seen this practice remedie a number os these types of issues.
I hope this is helpful,
Silicon Labs Technical Support
Attaching photos of waveforms taken directly at pin 6,7 of IC 8262 w.r.t. GND. NOTE- Ceramic disc capacitors (blue capacitor) c14 and c23 of value 0.1uf / 500 volts connected very near to the ic supply pins. Film foil capacitors (yellow) c4 and c35 value 0.47uF / 250 volts also connected to the ic supply pins of the respective driver ics. (c4, c35, are marked as 0.1uf in the circuit diagram- actual value is 0.47uF) Regulator ics u9, u11, u2 and u3 have 100uF/25V electrolytic capacitors (c43, c32, c11 and c1 resp) at their output +15V and -7V for each ic We have been using driver ic TLP 25O Iwith the attached circuit since year 2011. As an upgrade we switched over to Si8262 in May 2016. 3 Nos ics have failed on site within 2 months of supply of our equipment. Hence we have immediately reverted to using TLP 250 as we have no failures of this ic. Since I plan to use Si8262 in high speed resonant inverters above 100 khz. please give us a solution to the problem. Please note- I have about 100 nos of Si8262 in my stock.
You mention that you are using 500V 0.1uf ceramic disk capacitors for decoupling. These types of capacitors have a significant amount of lead inductance that could be rendering them less effective. Do you really need to have 500V capacitors here? Our supply specification only goes to 30V. By using a capacitor with a maximum voltage of 50V, you should be able to select a capacitor that is physically much smaller and thus has much lower lead inductance. It would also allow you to place it much closer to the supply pins to make it more effective.
Can you confirm that the scope you are using is set to 10v/div? If so, then it looks like the signal in the blue trace has excursions below ground up to -4V. This drastically exceeds our absolute maximum range of -0.5V to 36V. If this is real, it could be the cause of your damage. You might want to consider adding a fast diode between the output and the ground pin to clamp this reverse voltage. You could also do this with a diode between the output and the supply pin. If this fixes the issue, then you could decide how best to implement the fix. If it doesn’t fix the issue, we can keep looking.
I hope this is helpful,
The layout and the choice of components is hilarious in your application. Kindly redo the layout and switch to SMD MLCC X7R capacitors.
The layout and the choice of components is hilarious in your application
anyone can design a switcher., very few can do the right layout