3
\$\begingroup\$

I'm designing a pcb relay circuit to switch on/off power to an inductive load on a 12V line. I'm using an SPST-NO relay (omron relay) to switch the circuit open and closed, and I'm having trouble tracking down a high-frequency (~10MHz) +5V voltage spike seen when the switch closes. Any advice on how to clean up the initial power on signal? Oscilloscope attachedenter image description here, note the voltage divisions are 5V and the time divisions are at 500ns.

I've tried connecting a TVS diode in parallel with the inductive load with no success. Happy to include a schematic if it's useful. Any advice would be much appreciated.

Thanks

\$\endgroup\$
  • \$\begingroup\$ It could be the contacts bouncing. Is it a single spike or ringing around 10MHz? If you tried to include a picture of your scope trace, it didn't work. \$\endgroup\$ – evildemonic Oct 18 '17 at 15:13
  • \$\begingroup\$ Whoops, edited to add the image of the scope trace. It's a single spike followed by a ring down which stabilizes after a few hundred ns. I hadn't considered by the back emf of a contact bounce. I do have a flyback diode to protect the coil side of the circuit, but it's not clear to me what is needed to protect the contact side. \$\endgroup\$ – pcb_noob Oct 18 '17 at 16:08
  • 1
    \$\begingroup\$ Do you have a flyback diode on the load? \$\endgroup\$ – Jeroen3 Oct 25 '17 at 11:22
  • 2
    \$\begingroup\$ Possible duplicate of Relay voltage spikes \$\endgroup\$ – winny Oct 29 '17 at 16:19
  • \$\begingroup\$ @winny The problem with this duplicate is that it applies to an AC load, and the main advices given is to switch during zero-crossing, which is not applicable to OP's case. The only applicable advice is, I quote: "... After that it is inductors, resistors and capacitors to construct filters to damp any spikes left". Which certainly is a bit laconic. I'm therefore leaving this question open. \$\endgroup\$ – dim Nov 9 '17 at 14:55
0
\$\begingroup\$

I would try to add a MOV(metal oxide varystor) parralell to load if MOV response is fast enough, It might suppress the over rising

Note : this is only an idea , not tested yet but i think it worth to be tried

\$\endgroup\$
0
\$\begingroup\$

Try this previous answer.

https://electronics.stackexchange.com/a/81438/166508

Note: If it works it would be polite to upvote his answer not mine.

\$\endgroup\$
  • \$\begingroup\$ This is a link-only answer. Please add at least summarized content to your answer in case the link content changes. \$\endgroup\$ – Bort Sep 6 '18 at 17:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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