Discharge circuit - Unwanted activation of VIN_PWR_BAD



We have designed a custom carrier board for Jetson TX2 module.
It has a lot of ports (USB,CSI Camera, GPIO, …) and functions and everything is OK on the board.
We followed the OED DG as far as possible. Especially, the power input circuits, power up sequence, and discharge circuit are almost a copy of circuits on the DevKit.

Recently, we have performed an EMC (Electromagnetic Compatibility) test.
Fortunately, the combination of TX2 and our custom carrier board passed most of the standard tests.

The circuit fails (suddenly shutdown) in a test in which a ripple with an amplitude of 0.7Vpp is applied on the nominal 12V input supply. The frequency of the ripple is swept from 30Hz to 100KHz. The circuit fails (suddenly shutdown) only in a low frequency range: 30Hz to about 300Hz.
We investigated the circuit and realized by simulation that the Discharge circuit activates (pulls down) the VIN_PWR_BAD_L signal when 0.7Vpp ripple with 30Hz-300Hz frequency is present on the 12V input power rail. According to OEM DG, activation of the VIN_PWR_BAD_L when Jetson is in ON state, triggers a controlled shut down.

Here is my simulated schematic in LTSPICE:

And the waveform of the input positive supply, and the gate of NMOS transistor (Q13 in P2597_C02 schematic):

(100Hz ripple)

(200Hz ripple)

(320Hz ripple)

As is clear from the waveform, the gate voltage goes high for 100Hz and 200Hz frequency, while remains low for 320HZ and above.

According to OEM DG, the discharge circuit respond to a 20% instantaneous reduction of the input power rail. The 0.7Vpp ripple is very lower than that limit.


1- Is this a normal behavior? Is the discharge circuit designed to respond to such a ripple on the supply line? or this is an accidental response to a specific frequency range?

2- What can I do to solve my problem, while sticking as much as possible to the recommended circuits in DevKit schematic, and not deteriorating the power down sequence? I realized that manipulating R and C values can solve the problem…

Thanks in advance for your help.

Hi, please refer to the chapter 5.5.1 DV/Dt Circuit Considerations in Xavier DG (same to TX2 on this).

The portion of the circuit that asserts VDDIN_PWR_BAD_N when power is removed is designed to start the shutdown as soon as a voltage drop of ~0.5V is detected…

1 Like

A very to the point answer.

We have no “reserve” footprint in our pcb to add that diodes.
We replaced the Bipolar transistor MMBT4403 with a P-channel mosfet having a threshold voltage of about 2V. This increased the activation level of the discharge circuit from 0.5V (bjt) to about 2V (FET).

Thanks for your help Trumany.

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