No voltage on 5V power rail

I’m currently trying to activate/deactivate a relay from the Xavier NX’s GPIO. I use pin 29 which is GPIO01, the signal go through an inverting buffer CD74HC04 that command a BS170 N-channel MOSFET, letting the current from the 5V rail through the relay coil. However I measure no voltage on pin 2 and 4 which are the one I use to power the coil.

Is there a way to enable/disable the 5V power rail ? I’ve tried with to Xavier NX, one with Armbian 20.04, one with Ubuntu 18.04. Both show no voltage on 5 V rail

hello maxime.bay,

could you please refer to NVIDIA Jetson Xavier NX GPIO Header Pinout - JetsonHacks for Xavier NX’s J12 pinout, it’s pin-2, pin-4 to supply 5V by default.

Thanks a lot, I indeed found 5V measuring this morning, but I verified multiple time yesterday and found nothing. Is there some kind of resetable fuse protecting the 5V voltage regulator thay I may have triggered while measuring ?

hello maxime.bay,

I don’t think so, please try to reproduce the issue again and share the steps for reference.

So I mounted my circuit on a breadbord like that:

When GPIO01 is LOW, the MOSFET let current through and the 5V power Rail drop to 0V. When GPIO01 is HIGH, MOSFET block current and I can observe the 5V on the power rail with my multimeter.

My little circuit perform well when used with a lab power supply and show a consumption of 40 mA for the coil which is well in the limit of the Xavier 5V supply.

Even when my 5V power rail drop to 0, I can use normally my other GPIO peripheral: an I2C IMU, and a Emergency Stop Button using the same inverter buffer array. However they are powered by the 3.3V power rail, which is separated from the 5V, I guess

Please refer the 40-pin header docs:

https://developer.nvidia.com/jetson-nano-developer-kit-40-pin-expansion-header-gpio-usage-considerations-applications-note

Based on the provided documentation, I reproduce Figure 6’s LED example. Voltage from 5V power rail drop to 2.1V (measured between GND and 5V) when current pass through the transistor. When I power the relay, it drops to 0.3V so the coil it’s no energized

I assumed it may be due to not enough current available however carrier board specification say that the 5V regulator can provide up to 1A. Carrier board specifications

Is there a way to maintain the 5 V and avoid the drop ?

Why did the 5V drop to 2.1V? It is impossible because 5V supply of 40-pin header can be up to 8A output…there should be some issue in your design.

The 5V drop when I reproduce figure 6’s example of Jetson Nano Developer Kit 40-Pin Expansion Header GPIO Usage Considerations or when I use the previously shown design on the schematic with the relay.
I used a green LED, and a BS170 MOSFET, instead of the 2N3904 shown on the figure 6’s example. The rest is unchanged. While the LED is lighting up when I apply a tension on my MOSFET gate, the voltage is indeed dropping to 2.1V
When I use an external regulator (R-785.0-0.5) to get the 5V from the battery of the full system instead of the Jetson supplies, or use a lab supply, it works just fine, I have no drop at all

I don’t think the 5V will drop if you follow figure 6 correctly. It has series resistor which will keep the current not exceed the limit and so the 5V will not drop. But in your design, there is no series resistor for relay.

The relay could be modeled as inductance and a resistor. I did not measure the relay coil inductance, however its resistance is 114Ω, which mean with 5V, it limits the current to 0.043A in steady state.
If I take my schematic as a whole, with the serie LED and its resistor, and the RDS(on) of the BS170, assuming the LED is an ideal diode and D3 allow no current to flow, there is a resistance of 89Ω between the 5V and GND. Which mean it should draw 0.055A, which is limited enough I guess ?