Hi,
I've started a little project to detect an RCD trip time. To actually introduce the RCD fault, I'll use a commercial tester. My intention for this topic is to talk about detecting AC line phase from a transformer secondary winding.
I'm thinking a zero crossing detector might work, only if it was detecting from peak to peak, from a neutral wire. Otherwise if it detected every half cycle, many zero crossing detectors don't seem to change their output based on phase or polarity. They just give a signal when zero is crossed.
I've attached a picture to help understand the circuit I have at the moment. It's made from a wall adapter, with a slight circuit modification to output AC instead of the original DC. The transformer is double insulated. Only the 3 wires to the Arduino leave the wall adapter.
In order to detect the AC phase, I'll probably need to reference the mains neutral or earth. But i'm thinking it's also possible to measure with a multi meter to find which of the 2 outputs covers 0° - 180°, and which covers 180° - 360° from the 10v output. Once this is known, there might not be a reason to involve mains neutral or earth after the transformer secondary, simply to find the phase.
The output waveform is odd, possibly because of the small capacitors in parallel with the diodes. The original circuit had these, so I've left them on the internal wall adapter PCB
I'm hoping someone might know how to find the phase with a multimeter or scope, or if necessary by referencing neutral or earth after the circuit secondary winding
I must be dense or something, but I cannot understand what you think you are doing. I see you have a full wave rectifier attached to the secondary of a transformer. But what is the purpose of all this?
Paul
The theory is that one output rises when the mains AC is at 0°, and the other rises when it's at 180°. I wanted the 2 diodes to force the center tap to be negative, to use it as ground. The capacitors probably shouldn't be there though
The intention is to have two outputs above ground potential and phase separated, as one possible way to measure mains AC phase
You’re tilting at windmills. Sorry if you’re unfamiliar with the expression.
That circuit can be an Arduino destroyer. The only thing that prevents the atMega input ESD diodes from instantly blowing are the resistors. The peak voltage seen by the input pins could be 10 / 2 * 1.414 = ~7 volts. Fortunately, the resistors are limiting current to ~ 1 mA so that keeps the smoke inside the processor but that is right at the limit so you’re taking a chance there. The Xc of the caps matters too which could be limiting the current to a safe level, not sure since you did not provide values.
The diodes are backwards, you’re putting a rectified negative voltage on the analog inputs and you’re just measuring the current through the capacitors, not the diodes. Remove the capacitors and reverse the diodes and the signals may make more sense.
I fail to understand why phase matters and why you need to detect it. You do need zero crossing if you’re measuring operation time. But tripping the device requires unbalanced current bet
The idea for testing both phases is common in RCD testers, since trip time can slightly differ. I've been trying to save a dollar, but now starting to think it might be better to spend $100 on a tester that already does this. I tend to start ambitious projects and never finish them
I'll open the wall adapter again to correct the secondary PCB with the diode and capacitor setup, then will be able to get a good waveform.
Thankyou all very much for helping and guiding me in the right direction.
van0014:
Hi,
I've started a little project to detect an RCD trip time. To actually introduce the RCD fault, I'll use a commercial tester. My intention for this topic is to talk about detecting AC line phase from a transformer secondary winding.
I'm thinking a zero crossing detector might work, only if it was detecting from peak to peak, from a neutral wire. Otherwise if it detected every half cycle, many zero crossing detectors don't seem to change their output based on phase or polarity. They just give a signal when zero is crossed.
I've attached a picture to help understand the circuit I have at the moment. It's made from a wall adapter, with a slight circuit modification to output AC instead of the original DC. The transformer is double insulated. Only the 3 wires to the Arduino leave the wall adapter.
In order to detect the AC phase, I'll probably need to reference the mains neutral or earth. But i'm thinking it's also possible to measure with a multi meter to find which of the 2 outputs covers 0° - 180°, and which covers 180° - 360° from the 10v output. Once this is known, there might not be a reason to involve mains neutral or earth after the transformer secondary, simply to find the phase.
The output waveform is odd, possibly because of the small capacitors in parallel with the diodes. The original circuit had these, so I've left them on the internal wall adapter PCB
I'm hoping someone might know how to find the phase with a multimeter or scope, or if necessary by referencing neutral or earth after the circuit secondary winding
I doubt one would need to examine any of the waveform if all you want is time.
Simply measure the time from when a button is pressed until you loose transformer ac secondary(for safety).
Other than that I think you will find any approved rcd tester will have a price well above $100.00.
I'm not an electronics expert, but this may give an indication of when the phase changes. The "bridge" is just a couple of diodes to produce just enough volt drop to light the respective opto-coupler diodes. the circuit software didn't have an opto-coupler symbol.
More to the point, although it should be possible to home bake a tester, it can be hazardous to apply the test voltage by introducing a line to earth fault. It must be assumed that the RCD is not going to trip, and that the whole ground /earth system in the property will become live for the duration of the test. Then there is the issue of setting the the trip current which (in the UK AFAIK) should be at 0.5x, 1x, and 5x the rated trip current for both half phases. When diagnosing RCD issues it is also very useful to be able to ramp the test current, There's a reason why the testers are fairly expensive. 
Steve
van0014:
The idea for testing both phases is common in RCD testers, since trip time can slightly differ. I've been trying to save a dollar, but now starting to think it might be better to spend $100 on a tester that already does this. I tend to start ambitious projects and never finish them
I'll open the wall adapter again to correct the secondary PCB with the diode and capacitor setup, then will be able to get a good waveform.
Thankyou all very much for helping and guiding me in the right direction.
I had to Google what the device is you are trying to mimic. re you aware the phase information you are looking at on the transformer secondary is NOT the voltage phase of your mains. Since a transformer operates on current, not voltage, the voltage output of your transformer is the Current phase of the mains, not the voltage. There may be a difference.
If you are looking to use a transformer to do what you are trying, you need to find a transformer with phased windings. The data sheet or the actual transformer housing will have a dot, ".", on the beginning of the primary winding and the beginning of the secondary.
Paul
@blewtobits Thankyou for this circuit. I believe that will work, and like how it doesn't use LDRs. I heard somewhere that LDRs don't have a fast response. Also, your reply is helping me realize it's best to simply buy an existing tester
@Paul_KD7HB My research suggests only the current will have a different phase, since voltage should lead by 90° in an inductor. There could be a 180° difference depending on coil winding. The current will indeed be phase shifted, but my opinion at the moment is that there's a negligible voltage phase difference caused by the transformer
@bluejets
bluejets:
I doubt one would need to examine any of the waveform if all you want is time.
Simply measure the time from when a button is pressed until you loose transformer ac secondary(for safety).
Other than that I think you will find any approved rcd tester will have a price well above $100.00.
I think your right about the price. Though I found the CEM DT-904 does this, I can only find clones of it. And this clone might not be calibrated.
I think the phase matters, since the AS/NZS 3760:2010 standard says "Some RCDs may have a different result (approximately 10 ms) depending on the point on wave of the test current. In case of doubt, the operating time at both 0° and 180° should be tested"
The standard specifies tolerance for both trip current and time. The CEM DT-904 product appears to meet the criteria, but the cheaper clone device probably isn't calibrated
I still hope to buy this clone DT-904 device, unless it's likely to be too far out of spec. The next cheapest was $170 for a UNI-T device, which is hopefully up to scratch