STUSB4500 damaged during reconnection

[LFjær.1]

Hi,

We have a custom implementation based around STM32H7 and STM32F1.
The application is powered internally by a rechargable 4s li-ion battery.

We use USB-PD to charge the device, and USB 2.0 to communicate with the STM32H7 through an external ULPI-PHY. STUSB4500 handles the USB-PD negotiations and logic. VBUS is then passed through a power-MOSFET (STS6P3LLH6) to a battery management IC (BQ24179).

We use the following power profiles for PD:
PDO1: 5V, 1.5A
PDO2: 9V, 1.5A
PDO3: 20V, 1.5A

We use a suitable off-the-shelf USB-PD charger (OKDO FJ-GN236CAN) which supports all the mentioned profiles.
We use a custom magnetic contact for the USB (VBUS, GND, CC, D+, D1).

Generally it works well, and we are able to charge at PDO3.

However, we have discovered that the STUSB4500 is damaged beyond repair if the magnetic connector is connected then reconnected rapidly (typ. less than 500ms).

In some of the cases we've had, STUSB4500's VDD pin is burnt right off and its pad on the PDB is also burnt.


Do you have any idea what might be causing this?

Best,
Lars

[NBALL]

Dear Lars,

I'm suspecting some ringing on VBUS which generate overshoot above our AMR (Absolute Maximum Rating)
Is there any possibility to capture some waveforms (VBUS/CC)  ?
Source (charger) should have discharge VBUS during this 500ms of unplug.

USBPD standard requires source to reach vsafe5V (i.e 5V) within 275ms. So if it is properly discharge re-attach should not provide voltage overshoot due to cable impedance.

Best regards

Nathalie

[LFjær.1]

---

Hi, thanks for the reply.

The only waveform I am in possession of at the moment is a unplug-plug event where the charger was broken (not the STUSB4500).
Unsure if this can provide some information.
Will try to recreate STUSB4500 failing and capture waveform of VBUS/CC.

From our (numerous) experiments with unplug-plug, it does not seem to cause spikes on VBUS going above AMR (28V for VDD).

Due to the nature of our application (magnetic connector), we might have unplug-plug events in a shorter time-span than Vsafe5V (275ms). How would the STUSB4500 react to this?

A different question that came up:
When connected to computer (USB-C, 5.0V VBUS no PD) we read back PORT_STATUS_1 -> ATTACHED_DEVICE and it sometimes report 0x02 (reserved). Do you have any explanation for this?

Thanks,
 Lars

 

[NBALL]

Hello 
Sorry for late feedback.
Did you succeed to probe VDD, VBUS, CC during the issue ?

Having @x0E=0w02 is symptomatic of unlike reset. May I suggest to use 1k or less (instead of 10k) for Reset pin ?
I do not see neither 3V3 generation for Vsys. Can I suggest to add 1uF cap ?
Best regards

Nathalie