SHUTDOWN_REQ pulled low before reaching Thermal Throttling

It’s still happening without R329.

Got it, thanks. We are checking this internally.

Hi, can you please probe the VDD_5V and VDD_IN voltage level when issue happen (removing R377 and R329) to check if they are in correct status?

Also, how many modules have you tested? Do you have another one to validate in your test setup?

Both voltages look good to me, here is a snapshot:

I can see the same behavior with other modules, they are failing at similar temperatures. But most tests were done with one module only.

I still think that the problem is caused by the temperature in the upper left corner of the module, specifically Q38 or Q39.

Are you sure you probe correct point of shutdown_req? The pull-up of shutdown_req is VDD_IN on module and carrier which is 5V here. Why is it only <3V in your waveform?

You are right, the SHUTDOWN_REQ signal was actually getting pulled low to about 2.7V by our mainboard circuitry. After adding a series resistor to the signal, I didn’t see this problem anymore. The module now reaches thermal throttling without shutting down. I will do some further tests, but I think the problem is solved.

Thanks a lot for your help and your patience!
Ralf

Great to hear that. But I am a little curious how your 2.7V shutdown_req is pulled to low when temperature is high while won’t be low if it is 5V?

No, it’s always pulled low to that voltage. Maybe there is something on the module with temperature dependency. For example the latch connected to SHUTDOWN_REQ that is mentioned in the Design Guide.

Sorry I don’t get what you mean. You mean your original design will cause SHUTDWOWN_REQ to be 2.7V always, and will be pulled to 0 when temperature reach some point, right? And “After adding a series resistor to the signal, I didn’t see this problem anymore”, which means the voltage level of SHUTDOWN_REQ is 5V then? And won’t be pulled to 0 when with same temperature point?

No, we don’t pull the signal actively to GND. We have connected the signal to the base of a NPN transistor through a 2.2k series resistor. The 5k pull-up on the module results in 2.7V which I guess is an undefinded level for the latch.
I have now simply increased the series resistance to solve the problem.

OK, can you please share this part schematic design?

Screenshot 2023-07-10 103548

The base resistor is 2.2k and the base-emitter resistor is 47k.

Got it. So after increasing the series resistor, the voltage level is 5V and no drop seen, right?

The voltage is now at about 4.5V, no drop as far as I can tell.

請讓我用中文補充一下
電晶體特性很容易受溫度升高影響而飄移,在環境溫度升高時,我之前遇過相似的狀況
請多測試不一樣的環境溫度驗證,提供參考,謝謝

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.