GPIO unstable Issues

From: Jeff Chinni <jeff@sigma-usa.com>
Sent: Thursday, May 21, 2020 11:02 AM
To: ‘Rick Curcio’ <Rick@sigma-usa.com>
Subject: Edit and send back

Hello

We are using the Jetson Nano and we are having some very unstable issues

What we are using the board for is to tie in 4 photo eyes.

We have designed a board and we are pretty confident in our board.

We have a 12 volt power supply on the Sigma Board that is powering the photo eyes, this is an isolated power supply that is only used for the photo eyes.

Our board connects to the Nano as follows, and please note we have tried various pins on the Nano and have had similar unstable results.

Nano Sigma

Pin 1 3.3 pin 1

Pin 39 GND Pin 39

Pin 29 GPIO A Pin 12

Pin 7 GPIO B Pin 38

Pin 33 GPIO C Pin 40

Pin 32 GPIO D Pin 35

On the Sigma board, we are doing the following for each port

We have a 1K Pullup resistor

We have a relay Omron G3VM-31DR https://assets.alliedelec.com/v1580814023/Datasheets/6d625c28730683987544fd81220783f5.pdf

We have a 1K current limiting resistor that is used so we don’t burn up the diode.

We have written our program that simply reflects when the pin changes to High (we call it Rising) and when it goes Low and we call it Falling. This code is written in C# using .NET Core 3.1 and the Microsoft System.Device.Gpio 1.0.0 library It’s published for ubuntu-arm64.

What we are seeing is the following

GPIO A seems to always work

GPIO D seems to always work

GPIO C seems to partially work however a lot of times it will just Fall or it will Fall and show the other pins reacting as well.

GPIO B Seems to never work. When we introduce B we see results for ACD along with B. It makes the system very unstable. If I have A and D working and I introduce B (by plugging the photo-eye in or even the wire connecting 7 to 12

It is as if the pins will latch and not release. If you give them a minute they seem to release.

If I just use A and D it will work great

With C or B in the mix it will eventually reboot the Nano.

We have changed B from pin 31 to pin 7 and it acts the same Instable way

I would also like to add that we started out wanting to use pins: 12 , 35 , 38 , 40

When we use these pins , after just a few events the Jetson Nano hard crashes and resets.

Also when I monitor these IO pins with an oscilloscope the outputs switch only from approximately 1.5 volts to ground each time I trigger the banner sensor and occasionally switch from 3.3v to ground very cleanly the way I would expect.

hello jeff1,

could you please provide more information about “photo eyes”?

may I have confirmation that pin-7 and pin-33 you’re talking about were from Nano J41 Header?

BTW,
could you please setup an uart console to gather kernel messages.
we should also check what’s the failure to cause Nano hard crashes and resets.
thanks

Jerry we have shelved the testing with thew photo eyes but here is a link to what we are using.

https://www.sick.com/us/en/photoelectric-sensors/photoelectric-sensors/gr18-inox/grtb18s-p2412v/p/p521574

I have put together a PLC that has 4 solid state relays on it and the PLC goes 4 seconds then the relay makes for .5 seconds.

Yes pin 7 and 33 are on the 40 pin header on the Nano

I am not the developer but Doug can chime in on this.

“could you please setup an uart console to gather kernel messages”

Jeff

We were asked if we could bread board something that utilize only pull up resistors ( 1.5K ) to eliminate our custom board. So we removed our board and the photo eyes ( Sick ) and built a bread board driven by a PLC with photo relay to pull the pull up resistors to ground. The results are very much the same. With only one channel connected we seem to get very stable results with more than one channel connected it is not to long before two of the channels will start to run away. You can see this happen on an oscilloscope , the voltage will brake out into an oscillation. In our case with this test it will do that for the duration of the PLC idle time. As soon as the PLC pulls the pull up resistor to ground everything is normal again for a duration of time. Unfortunately it will happen again. If we were using photo eyes that runaway would continue until the next time the photo eye pulled it low.

I have included a small video. The voice description starts about midway ( 30 second video ).

xxhttps://www.adrive.com/public/96fnMg/IMG_0100.mov

Take out the XX in the front and paste into a browser to see the video.

hello jeff1,

may I know what’s the power supply of Nano, you may refer to below for recommendations,
thanks

Jerry we are using a Mean well IRM-30-5ST
https://www.arrow.com/en/products/irm-30-5st/mean-well-enterprises

hello jeff1,

may I know had you configure pinmux spreadsheet to set pin-7 and pin-33 as GPIO types.
there’s Jetson Nano Developer Kit 40-Pin Expansion Header Configuration for your reference.
thanks