5
\$\begingroup\$

I have looked at several questions on EE.SE as well as searching the internet and nothing I came across has answered my question in a way that makes me feel safe dealing with mains.

Having read many questions and answers on EE.SE I know that when dealing with mains you want to apply galvanic isolation to your circuits to minimize the possibility of someone getting electrocuted/injured. Below I have a circuit that is designed to indicate the phase rotation for a 3 Phase 120V 400Hz power line. It was designed and simulated via https://www.falstad.com. While the design is not the focus of this question any design input is highly appreciated since the simulation (probably) works as an ideal environment.

So on to my question...how would you connect the optocouplers to the three AC lines without damaging them or injuring anyone who uses the PCB? The entire circuit would be housed in a plastic case.Falstad Circuit Simulation

\$\endgroup\$
6
  • 2
    \$\begingroup\$ Is there a reason you can't use transformers? \$\endgroup\$
    – Theodore
    Apr 26, 2021 at 21:35
  • 1
    \$\begingroup\$ On a related note you might be interested in this simple phase-sequence tester: electronics.stackexchange.com/questions/551340/…. \$\endgroup\$
    – Transistor
    Apr 26, 2021 at 21:41
  • \$\begingroup\$ @Theodore I should have mentioned in the main body of the question...a goal is to keep this piece [as small as possible] about 1"x1.5" board space. Three transformers would take up substantial area. \$\endgroup\$ Apr 26, 2021 at 22:57
  • \$\begingroup\$ @Transistor I am unsure why that page never came up in my searches but I am thankful that you brought it to my attention so thank you for that! I saw versions of this but was unable to simulate them to verify if they would do what I was looking for. If there is a way that I can use something other than lamps and indicate simply which leg is higher voltage then that may work. \$\endgroup\$ Apr 26, 2021 at 22:59
  • 1
    \$\begingroup\$ 1" x 1.5" seems very small for a device with 3-phase mains connection. I'd be concerned about how you will achieve creepage distances and adequate isolation between mains and LV side. \$\endgroup\$
    – Transistor
    Apr 27, 2021 at 6:53

2 Answers 2

5
\$\begingroup\$

Though not ideal, each 100 kΩ resistor, by itself, would limit fault current to ~1.2 mA, generally considered safe on skin contact. Of course, a resistor failure could make the the circuit "live"... but if the optocouplers are in an effective enclosure, and the only wires or traces leaving enclosure are from the output side, the 5,000 V isolation specification should make this safe.

That said, typical on current for the LED is 10 mA to turn the output completely on. At 120 VAC, two 100 kΩ resistors would limit current to ~600 µA, which might not provide sufficient drive.

\$\endgroup\$
1
\$\begingroup\$

This claims an isolation of 2.5kV RMS. Maybe that's fine, but it's difficult to say unless you're working against a specific requirement.

Adding some kind of surge protection is not a bad idea. Consider an R1 + C1 series, MOV in parallel, R2 + C2 + input diode in series. Caps are nice because they reduce the amount of thermal dissipation. Analyse the impedance at 400Hz. Use self-healing, high-voltage metallized polyfilm caps. Set the resistors so that if the mains are connected at a peak, the instantaneous surge current through the caps will not exceed the diode's rated 80mA. 1.3k * 2 would have a max current of 65mA peak.

Let's assume that your desired current is 10mA RMS. The following assumes 10mA RMS sine even though the reality is half-rectified, but anyway:

|120 / (2*1300 + 2/2pi/j/400/C)| = 0.01, so C ~ 68 nF which are common and cheap.

With left and right legs equal impedance, choose a MOV rated for the midpoint of your voltage.

schematic

simulate this circuit – Schematic created using CircuitLab

simulation

If there is an over-voltage event or R1 or C1 fail short, the MOV will try to protect you. If C2 fails short, the diode will not burn - it'll just be too bright.

One thing limiting the practicality of this circuit for your application is that it will impose a phase shift, which you might not want to have to adjust for.

\$\endgroup\$
1
  • 2
    \$\begingroup\$ Phase shift may not be an issue; if all three phases shift equally then it shouldn't have any impact on determining the rotation. \$\endgroup\$ Apr 27, 2021 at 12:59

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.