Arduino Uno, ULN2803, Relay, and a 24VDC Pump

[KeepItSimpleStupid]

This is the first time I have come across the issue. A professor in school had a story of a computer that was affected by the contactor and fan motor of the A/C duct in the room. One day while pondering the issue they heard the contactor close and noticed the computer crash. They moved the computer 6 inches and the problem went away.

Been there, done that. In the days of an 8" floppy drive and a 40 kV lamp ignition pulse, the drive activity light used to light when the arc lamp was started. What do you expect from a poorly shielded computer in those days and 12" of essentially antenna? The manufacturer eventually came up with a retrofit to move the lamp housing.
Fixing, really wasn't an option for something that occurred 1x and at your discretion about 1x per day. Removing the floopy prior to ignition worked.

 

[be80be]

I install door locks that have to open the door. The maker of the latch recommends to use 1 pair unshielded wire and a diode at the latch works great. Cat5 and shielded alarm wire
Turned it into a antenna I guess the coil tuned it

 

[KeepItSimpleStupid]

My first comment is moving the diode across the coil as close as possible to the relay coil. What diode did you actually use? I would use one with a PRV of at least 50 V.

Check cable routing too. Long parallel lengths of say the 24 V pump line running in parallel with some thing that matters. One system I set up, I didn;t wait for a problem to happen, but used a shielded power cord for the IR heating elements that were in a 13.56 MHz induced plasma.

The power supply to the CPU may also dip. Can you temporarily power the pump from another supply?

 

[ADWSystems]

My first comment is moving the diode across the coil as close as possible to the relay coil. What diode did you actually use? I would use one with a PRV of at least 50 V.

1N4001

Check cable routing too. Long parallel lengths of say the 24 V pump line running in parallel with some thing that matters. One system I set up, I didn;t wait for a problem to happen, but used a shielded power cord for the IR heating elements that were in a 13.56 MHz induced plasma.

The 24V lines inside the panel are not in parallel with anything (or at least not bundled with anything) and outside the panel, there is nothing to be in parallel with except for mother earth.

The power supply to the CPU may also dip. Can you temporarily power the pump from another supply?

The CPU power supply and the pump power supply are separate supplies already.

Does the moment of freeze coincide with when the the relay is energized or de-energized?

Must freeze when the pump turns on, because the pump is always running when the arduino is locked up. Being a relay means the contacts bounce, so starting the pump actually means off-on-off-on-off-on (standard contact bounce, hence the reason for debouncing switches used as inputs). I'm going to try a snubber and see if I can get a hold of a DC-DC SSR.

 

[KeepItSimpleStupid]

Must freeze when the pump turns on, because the pump is always running when the arduino is locked up. Being a relay means the contacts bounce, so starting the pump actually means off-on-off-on-off-on (standard contact bounce, hence the reason for debouncing switches used as inputs). I'm going to try a snubber and see if I can get a hold of a DC-DC SSR.

There's a thought: Change to a contactor. Are the relay contact ratings good for DC?

An old one-stop-shop: https://www.omega.com/pptst/SSRDC100V.html

 

[ADWSystems]

The relay contacts are good for 8A at 30VDC or 250VAC. I have a contactor, but (a) I think it is too tall to fit in the enclosure and (b) the coil is 24VDC.