Solution Finder Get Tech Support

TinySwitch with a Bug?

Posted by: luchao on 07/05/2022

Hello, sir.

In the application document, the chip current limiting point will change with the load change.

In the document, it is introduced that the chip will have three overcurrent points as the load increases, such as 30%, 50%, 100%; As the load is reduced, there are also three overcurrent points, such as 100%, 70% and 30%, but in the actual test, I slowly increase the load, and I can see four different current limiting points, namely 30%, 50%, 70% and 100%. What is the reason for this? Is it an abnormal phenomenon?


Submitted by PI-Enderman on 07/05/2022

Greetings luchao,

Thank you for using our product.

The device is operating as intended and what you observed is not a bug. We have implemented a Current Limit State Machine functionality which you can read more about in the TinySwitch-4 datasheet, on page 4, under the paragraph entitled ON/OFF Operation with Current Limit State Machine.

We hope this resolves your inquiry.

Best Regards,


Submitted by luchao on 07/13/2022

Thank you for your reply.

I have read this document. And this is described in detail on page 26 of the application note "an-90_tinyswitch-iii_family_design_guide".The so-called light load, medium load and heavy load are vague concepts, and we need to refine them. In the last discussion, I said that when the load is increased very, very slowly, it will lead to some problems. As shown in the picture, under a constant load, the current limiting point jumps from "limit 2 (ILIMIT=50%)" to "limit 4 (ILIMIT=100%)", and then back to "limit 3 (ILIMIT=70%)". There seems to be no problem with this, but according to my debugging situation, there may be the problem I mentioned above: The Current Limit State Machine of the chip will produce different current limits according to the increase or decrease of the load.

One case (assuming A) is,"ILIMIT=40%" -> " ILIMIT=50%" -> " ILIMIT=100%" or " ILIMIT=100%" -> " ILIMIT=70%" -> " ILIMIT=40%"; One case (assuming B) is, "ILIMIT=40%" -> " ILIMIT=50%" -> " ILIMIT=70%" -> " ILIMIT=100%" or "ILIMIT=100%" -> " ILIMIT=70%" -> " ILIMIT=50%" -> " ILIMIT=40%". This will result in different output ripple and efficiency. The above are the problems I found in the test. My purpose is to ensure that my power supply system can be stable under any conditions during mass production, as in case (A) above, rather than as in case (B) or as in the figure. I found that the transformer and some peripheral parameters will affect the Current Limit State Machine. Please give me some suggestions on how to ensure that the Current Limit State Machine only works in a certain situation.

I'm looking forward to communicating with you on this issue.

Attachment Size
1_6.png 59.1 KB
Submitted by PI-Enderman on 07/19/2022

Hi luchao,

Thanks for the clarification.

The cases that your circuit experienced is normal, speaking from experience with our ICs. Rest assured that the feature will not cause any instability as it has been extensively evaluated by our engineers to function correctly as long as you followed the Application Notes and Datasheets as well as the PI-Expert/XLs in creating your design.

Best Regards,