Hello Ellen,

I will try to answer to your specific questions. It is hard if not impossible to substitute the design and laboratory testing time with general answers from a remote place from PI experts. I can identify obvious mistakes but I can not give you assurance your product will be perfect. We take the same path with our products and we go trough all the design and testing steps like you do. Nothing can replace extensive bench tests with the real product. I have to assume you already optimized and tested your design for efficiency, EMI, stability, regulation, etc.

For the first question, the trigger frequency is three times higher than the real switching frequency because of shape of the waveform used for trigger. If you move the trigger up or down, change the bandwidth of the trigger condition, etc. you will get a match between the switching waveform and trigger display.

I am not sure when you get a burst mode problem. Usually, burst mode or pulse mode are just two different types of the switching pulse grouping, both of them normal. In some special conditions, because of very weird burst grouping is maybe possible to get some audible noise issues, but not necessarily design or functional problems.

I do not see any dangerous thermal stress in your table. I know from experience you can get a lot of trouble using thermocouples. Sometimes I have to use real SMD thermistors, like this one from DigiKey: PTS120601B1K00P100 (or the smaller 0603 version) in order to avoid noise interference. With a thermocouple you can get false readings and errors up to 50% function of the noise levels. With a thermistor you will know for sure you can count on the specified precision, with unmatched stability. You have to use thermally conductive sticky tape (like Bond-Ply 100 from Bergquist) around your thermistor to isolate the body from the electrically exposed surfaces. The sensor should be soldered in a minimal volume with thin triple isolated wires and some extra thermal grease. You must provide very close and tight thermal contact.

There are some other tests you can run. If your product will be handled by the consumer, maybe you need an ESD test, and maybe a vibration test. You should check with your regulatory engineer to see if a corrosion test (or other environment test) is required.

Very careful engineers are not testing the product to specifications only. You can try destructive tests, like HALT tests or parts of this specification. This is a proved method to identify many hidden weaknesses before production launch. As an example, let’s say your product must withstand 55C ambient temperature. You can increase (or decrease) the ambient temperature (remove any plastic parts) up to the destruction point, anywhere between 55C and 200C (or -90C). Your product must be capable to shut-down in a non destructive way and restart when you return to nominal conditions. If you get a failure at 60C for example, you can identify what component failed, why, and maybe fix a hidden problem.

There is no easy way to improve the Schematic and/or the PCB. Everything is interdependent and only trough many tests you can get a better understanding of how to optimize your design. After one quick look, I do not see really bad problems in your design.

Cheers,
PI_Crusher