1
\$\begingroup\$

Below is the functional schematic that I am planning to use to control an AC load (air conditioner in this example with max current draw of 32A). Since I am still relatively new to mains power and relays, would appreciate if you guys can comment on the design and point out if I am missing something and whether this design is safe or not.

enter image description here

  1. I plan to use a DPST Latching relay with coil voltage of 5V and max load current of 40A. The relay would operate on both the Live and the Neutral line for added safety
  2. There would also be a 40A fuse on the Live line to protect against over current
  3. A 12V Schottkey to protect against back EMF
  4. A Hall sensor on the Live line to give feedback to the microcontroller on weather the current is going to the load or not.
\$\endgroup\$
  • \$\begingroup\$ Which kind of latching relay are you using? Dual or single coil? \$\endgroup\$ – Majenko Oct 29 '14 at 16:09
  • \$\begingroup\$ @Majenko-notGoogle I plan to use dual coil to keep it simple (using 2 micro-controller pins. \$\endgroup\$ – Ankit Oct 29 '14 at 16:11
  • \$\begingroup\$ Ok, that's fine then. \$\endgroup\$ – Majenko Oct 29 '14 at 16:14
  • \$\begingroup\$ I also assume you're not directly driving the relay's coils with the MCU's IO pin, but through a transistor? \$\endgroup\$ – Majenko Oct 29 '14 at 16:16
  • \$\begingroup\$ If your micro can drive enough current to make the relay work, then that part is okay. Most can't, so they need an amplifier of some kind in between, usually a single transistor per coil for what you're doing. Also, the current sensor needs to be isolated from the mains and not become a transformer either. Other than that, I think you're good. \$\endgroup\$ – AaronD Oct 29 '14 at 16:18
2
\$\begingroup\$

Well a couple of things. First, if you've ever looked at the control box in a whole house AC, you'd see that the relay is a pretty hefty open frame unit, driven by a solenoid, with contacts that open quite wide ( sometimes as much as 1/2 inch). That's because the inductive quality of the load is sufficient to produce significant arcing at the contacts when the relay opens. If you're relay is not rated for such inductive "kick-back" you'll want to add a capacitor between L and N, so that this arcing is suppressed. Among other advantages, your contacts will last a lot longer. The exact value of such a capacitor is hard to calculate, but a 'start" capacitor (similar to what might be used in the compressor motor circuit) might be a good starting point. 6uF at 400VAC should be a good ballpark estimate, and observing your contacts as they open with and without the cap will be a good confirmation. The difference can be staggering! You might also consider some capacitance across the relay coil too, as an added protection for the coil driver electronics against the back EMF you mentioned.

Second, I would strongly advise reconsidering the use of a latching relay to control such a load. From a safety standpoint you are inviting too many fail-mode situations where the load could be left energized, when you really want it shut down. If the controller malfunctions or even just loses power, wouldn't you want the load power shut down? If you have an MCU involved, at least a non latching relay could be operated by a fail safe circuit, where the absence of an occasional pulse would indicate the MCU is not running, and trigger a shutdown. Any and all additional safety sensors or features you add to your current or future plans will be easily bypassed if the controller fails, if the relay is a latching type. And BTW, such a mechanically latching power relay will likely be expensive. If you really do want a latching type, it might be both better and cheaper to find a relay that has an extra "holding" contact. These allow circuit arrangements where a pulse closes the relay, and the coil then remains energized through the holding contact.

\$\endgroup\$
  • \$\begingroup\$ Depending on the application, it might be okay to leave it on. If it's designed to be plugged in all the time and he just wants to unplug it on occasion without actually unplugging it... \$\endgroup\$ – AaronD Oct 29 '14 at 17:05
  • \$\begingroup\$ Also, for the holding contact, you'll probably have to make that using a 3-pole relay. They do exist, most likely as motor starters/contactors (starters have circuit protection built-in, contactors don't) with auxiliary contacts that are designed to tell an industrial control system what state it's actually in. The main contacts of these are designed for motors with no additional arc suppression. The trouble there is that the coils are typically designed for either line voltage or 24V DC, but that's okay if you have a suitable supply, separate from the MCU, and a transistor. \$\endgroup\$ – AaronD Oct 29 '14 at 17:12
  • 1
    \$\begingroup\$ I was actually planning to use a magnetic latching relay with dual coils.The idea behind using a latching relay instead of non latching one was to save on the constant current required to hold a non latching relay compared to a momentary pulse in the latching relay case. I would be using the control unit to operate my air conditioner through commands over bluetooth from my phone. So even if the control unit fails (in the closed position) I would realize this and simply turn it off. \$\endgroup\$ – Ankit Oct 29 '14 at 17:51
  • \$\begingroup\$ As long as its intended for human supervised operation, that's fine then. If you've already picked out this magnetic latching relay, I wouldn't mind seeing a link to the product. I'd still bet its pricy compared to a relay with a holding contact or other arrangement that does require the constant current, but I still wouldn't mind seeing what you find. \$\endgroup\$ – Randy Oct 29 '14 at 18:32

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.