1
\$\begingroup\$

I am looking for a few techniques to reduce contact chatter in a relay controlled by a 120 VAC voltage signal. The circuit uses a bimetallic thermal switch to turn a resistive heating element on and off. During transitions, the switch chatters and passes along a rather dirty control signal to the relay coil. Based on my testing, this appears to be less about setpoint hysteresis and more about the design the switch, as it doesn't have clean make/break performance. All components are rated for AC voltage at the specified levels. The relay is an Omron LY1F-AC100/110 and the bimetallic switch is one of those old-school mechanical rotary types.

In the waveforms below, you can see the fluctuating input to the relay coil, which causes the contacts to rapidly chatter. The heating element is rated for 750 W, so the contacts produce visible arcing and I am concerned about premature relay failure (as well as the rather unpleasant machine-gun-like sound).

In a low-voltage DC control circuit, there are a number of methods I could use to filter the relay coil signal to produce a clean switching action, but I am less sure about a mains-level AC application. What are some ways that I could solve this problem without overly complicating the circuit?

EDIT: Based on the comments and answers below, I thought I would share the reason for not using the bimetallic switch to directly operate the heating element. The real-world circuit has a mode selector switch between the voltage source and SW1. This switch enabled other banks of heaters and carried the entire current load. Over time, the contacts became worn down due to arcing. Using a relay reduces the current through this mode selector switch to preserve the life of the contacts.

schematic

waveforms

components

\$\endgroup\$
9
  • 1
    \$\begingroup\$ How about a snubber? \$\endgroup\$
    – Justme
    Nov 6, 2021 at 13:23
  • \$\begingroup\$ You may want better THERMOSTAT, with "hysteresis". Use Snoober for "debouncing". \$\endgroup\$
    – jay
    Nov 6, 2021 at 13:26
  • \$\begingroup\$ Aliexpress still sells some hermetical contacts with mercury tip. It was used in old thermostats. But orientating of contact important, gravity used. Not sure for high voltage. \$\endgroup\$
    – user263983
    Nov 6, 2021 at 13:43
  • 2
    \$\begingroup\$ @user263983 please no mercury. \$\endgroup\$
    – jay
    Nov 6, 2021 at 14:19
  • \$\begingroup\$ @jay So only electronics thermostat can be chosen. All mechanical contacts shatter. \$\endgroup\$
    – user263983
    Nov 6, 2021 at 19:01

2 Answers 2

1
\$\begingroup\$

Rectification with RC filtering should help. With this configuration it needs about 3-4 of full periods to coil reaction (about 6 transient/incomplete periods). The C will hold the coil during transient.

The R1 will slightly decrease the coil voltage so choose the R1,C1 according the relay coil operating voltage range/resistance.

It is also possible to use relay with 24V coil if the voltage decrease is a problem.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
4
  • \$\begingroup\$ Could you share how you determined the amount of periods required for the coil to react with this rectifier? I can't test your solution at the moment, as I don't have a ~100 uF cap rated for mains voltage. I can upvote and add comments when I have a chance to test it. \$\endgroup\$
    – higrafey
    Nov 9, 2021 at 0:50
  • \$\begingroup\$ Follow up: I simulated your circuit and understand what's going on. Before I mark this as a solution, are there any issues operating the AC relay coil with DC? \$\endgroup\$
    – higrafey
    Nov 9, 2021 at 4:12
  • \$\begingroup\$ AC coil will work with DC but the voltage on coil must be reduced. It is an advantage in your situation because you are reducing the voltage with R1 already. I think the R1 must be even higher than 220. Look for exact calculation. \$\endgroup\$
    – user208862
    Nov 9, 2021 at 4:46
  • \$\begingroup\$ That makes sense! I can adjust the value of R1 in my LTspice simulation to tweak the voltage as you described. Adjusting C1 changes the number of cycles ("ramp" and timing I guess) before the relay coil sees its activation voltage. I found this appnote explaining how to drive AC relays with DC and vice-versa: uk.rs-online.com/euro/img/task-requests/135515/guide_2.pdf \$\endgroup\$
    – higrafey
    Nov 9, 2021 at 14:50
1
\$\begingroup\$

A relay with a higher coil current should solve the problem as it would not respond to the thermostat chatter.

Bi-metal thermostats, normally intended for direct switching of heaters, would be capable of carrying higher coil currents.

\$\endgroup\$
3
  • \$\begingroup\$ Could you clarify what you mean by a higher coil current? Usually coil current goes down as voltage rating goes up, so I think you're suggesting using a relay with a lower coil operating voltage? I would think that'd burn up pretty quickly. \$\endgroup\$
    – higrafey
    Nov 9, 2021 at 0:55
  • \$\begingroup\$ Hi higrafey, A larger relay, with a larger armature travel would require a higher coil VA for its operation. The relay you have chosen has a 1.1 VA coil. A contactor, for example, would have a 10 VA coil. A contactor would be a better choice for your application. \$\endgroup\$
    – vu2nan
    Nov 9, 2021 at 5:17
  • \$\begingroup\$ Gotcha! Putting that in terms of VA makes sense now. :-) \$\endgroup\$
    – higrafey
    Nov 9, 2021 at 14:51

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.