There are seven relays on the board and a ULN2003 relay driver to control these relays. The BCM pins to the relay driver are marked in yellow below. BCM pins are IO pins from Jetson Nano Development Kit.

enter image description here

When two relays are ON, a different relay that should be OFF automatically turns ON. It does this sometimes, not always.

I've posted this issue here before.

As seen below, I added a resistor in parallel to the inputs of the relay driver. This solution fixed my problem, but it came back. Can anyone help with this?

enter image description here

enter image description here

  • 1
    \$\begingroup\$ Why have you got diode D1 in the circuit? \$\endgroup\$
    – HandyHowie
    Dec 14, 2021 at 15:47
  • \$\begingroup\$ There is nothing in the circuit drawn that would explain the problem. Maybe bad connection, incorrect wiring, or software issue. What is controlling the ULN2003, how it is connected and what software it runs? Have you measured voltages at ULN2003 input for problems? \$\endgroup\$
    – Justme
    Dec 14, 2021 at 15:51
  • \$\begingroup\$ No decoupling capacitor ? \$\endgroup\$
    – Antonio51
    Dec 14, 2021 at 15:52
  • 1
    \$\begingroup\$ What are "BCM pins"? \$\endgroup\$ Dec 14, 2021 at 16:31
  • 1
    \$\begingroup\$ Where is your chip bypassing? With D1 blocking, there is little to no bypassing for that chip. \$\endgroup\$
    – Aaron
    Dec 14, 2021 at 18:33

1 Answer 1


Without a complete view/picture of the setup it's difficult to know what's wrong.
The diode D1 should not be here. This is not the reason of the issue. All the other diodes are not necessary, the ULN2003 has internal free-wheeling diodes (connect the COM pin to +5V).
The pull-down resistors on the inputs are not necessary, the ULN2003 has internal pull-down resistors (and NPN transistors don't need one anyway).

I suspect the source of your issues might be the grounding.
You should have a solid power ground between your +5V supply, the relays and the GND pin of the ULN2003.
The signals coming from the BCM (Raspberry Pi?) should have another ground. Both grounds should be tied at the GND pin of the ULN2003.

You're probably connecting 5V and ground from the Pi to the extension board.
Do you know how much each relay coil draws?
Are you sure this is within the Pi's limits?

Increase/add more ground wires, or use an input stage before the ULN2003 to clean up the signal, like a Schmitt-trigger before feeding it to the ULN2003.

  • \$\begingroup\$ There is only one Ground on the system, and everything that is Ground is connected to each other polygonally. The +5V supplied to the relays comes from the 5V-5A regulator output. What do you mean by using the entry november into ULN2003? It seems unlikely that I will use Schmitt-trigger for each input signal. \$\endgroup\$
    – Enes Orhan
    Dec 15, 2021 at 9:34
  • \$\begingroup\$ 74HC7541D or equivalent single IC with 8 schmitt-triggers. \$\endgroup\$
    – romain145
    Dec 15, 2021 at 13:38
  • \$\begingroup\$ I can add this to my design but are you sure you won't have the same problem with schmitt-trigger's inputs? What if one input is triggered and another is triggered? \$\endgroup\$
    – Enes Orhan
    Dec 15, 2021 at 14:34
  • \$\begingroup\$ What are the IO capabilities of the Jetson Nano Developer Kit? \$\endgroup\$
    – romain145
    Dec 15, 2021 at 18:24
  • \$\begingroup\$ IO is actually GPIO (digital inputs and outputs). 0-3.3V logic output or input. \$\endgroup\$
    – Enes Orhan
    Dec 16, 2021 at 14:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.