In the context of building PID controllers for very large heater systems (ovens, kilns, etc.) nearly all the designs I encounter use an SSR to switch the load, a choice I don't understand.

The power requirements can top 50A at 250V, where even a reasonable-quality SSR would be expected to dissipate something like 40-70W. By comparison a mechanical relay for the same capacity and quality is dramatically more efficient, doesn't come with inconvenient cooling requirements, and is 5-10x cheaper even on a bad day. Plus, EMRs generally fail open, which is a plus for heaters.

You could argue that SSRs have the upper hand on longevity and moving parts, but certainly not to a degree that justifies the price differential and special cooling requirements.

So... what's the benefit for SSRs here that I'm missing? Why are they preferred in so many designs?

  • \$\begingroup\$ Most heater designs have temperature controls which implies at a minimum power cycling to control the temperature. For a relay this means arcing....which means wear and eventual failure. With an SSR you avoid arcing, and can control the cycling down to less than a cycle depending on your design …..so the benefits of SSR control far outway the power dissipated. \$\endgroup\$ – Jack Creasey May 6 '19 at 23:27
  • \$\begingroup\$ A MOSFET-based SSR can dissipate much less than a thyristor based one. especially if the switching rate is slow. \$\endgroup\$ – Jasen May 7 '19 at 3:44
  • 2
    \$\begingroup\$ "Plus, EMRs generally fail open" - Do they though? Contact welding is a real risk, especially at higher power. I would say failing closed is a real concern with mechanical relays in heaters. \$\endgroup\$ – marcelm May 7 '19 at 9:35

It's not very different from why you always have solid state switches and never electromechanical relays for variable speed motor drives.

A electromechanical relay can only ever be 100% on and 100% off with nothing in between, and it can't do anything about it because it can't switch fast enough and it can't switch very many times before it completely wears out.

As a result, a mechanical relay can only turn on when the element is too cold and turn off when the element is too hot. (Think your house thermostat or your toaster oven where it turns on for some period of time, then turn off another period of time with the temperature fluctuating wildly). All you can do is hysteresis control. You can't do PID because all you have are off and on.

Technically an SSR can also only ever be 100% on or 100% off too, but SSRs do not wear every time you switch them and they can switch fast. Therefore, you can switch it fast enough and often enough that you can simulate something like running the heating element continuously at 50% power and keep the temperature range of the heating element nice and tight.

An SSR can switch dozens or hundreds of times per second without wearing out so it can keep the heating element in a much tighter temperature range essentially making it a variable temperature element.

|improve this answer|||||
  • \$\begingroup\$ I honestly never would have expected to see someone compare heater control with VFDs, but I suppose now I have! \$\endgroup\$ – Hearth May 6 '19 at 23:41
  • \$\begingroup\$ @Hearth It doesn't even need to be a VFD (in the industrial sense of the word). Just a single transistor or a single relay driving a brushed motor unidirectionally works too. Or an LED. Or an H-bridge for a brushed motor that uses relays vs one that uses transistors. \$\endgroup\$ – DKNguyen May 6 '19 at 23:45
  • \$\begingroup\$ Consider also the repeated clunking of a large number of contactors, or the buzzing that tends to happen with contactors left energised long-term. Not good if your control cabinet is near anyone. \$\endgroup\$ – SomeoneSomewhereSupportsMonica May 6 '19 at 23:54
  • 1
    \$\begingroup\$ @Toor Interesting thought; kilns have so much thermal mass that sub-second cycling seemed a but unnecessary -- I'm somewhat adverse to switching at x% of each cycle because you get that dimmer circuit whine and corresponding vibration in your otherwise delicate heating elements. But I could see the value in doing a zero-crossing detection on the way in, and then power N of M cycles from zero-to-zero. Is that a thing? Or am I making up something new here? It'd cause awful flicker in a lightbulb, but with a kiln.... \$\endgroup\$ – tylerl May 7 '19 at 1:24
  • \$\begingroup\$ @tylerl With something like a kiln, the speed might not be the limiting factor but the number of switching cycles might be. 100,000 cycles switching once a minute is only 69 days of continuous operation. \$\endgroup\$ – DKNguyen May 7 '19 at 5:13

enter image description here

Figure 1. SSRs allow rapid switching while allowing varying duty cycle for heating loads. Source: Opto-triacs, solid-state relays (SSR), zero-cross and how they work.

Relays can be used for heater switching but when the switching period gets down to about a minute or 10 s or so mechanical wear becomes a problem. The duty cycle is determined by the the thermal response of the system so when this is in the order of seconds then a higher frequency switching is required. SSRs are capable of this as they don't wear. Zero-cross types eliminate both audible and electro-mechanical noise from the switching.

Further reading:

|improve this answer|||||

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.