AC Zero crossing detect, H11AA1 troubleshooting

[aeberbach]

I'm building an Arduino controlled water heater. A DS18B20 1-wire sensor is giving a nice reading for current temp. I will write some PID code to control the heater element of a 400W Birko jug via a MOC3052 opto and Q6015L5 triac. This will hopefully cook eggs, sous vide style and maintain a nice even temperature for a hide glue pot in the shed when it isn't cooking eggs.



The resistors are IRC GS-3 3W 1% 47K. Both measure spot on. On the Arduino side it doesn't seem to matter if pins 4 and 5 are connected as shown or left hanging.

When I apply power the house RCD trips, so it's a journey to the fusebox to reset. The circuit is definitely as shown here - have rewired 3 times, tripped power 3 times. Resistance measured across the mains terminals is around 16.3M. I have two H11AA1 chips here and both do the same. Any suggestions as to what is going on? I should mention that the RCD seems sensitive, sometimes it will trip when the AC or motor in the central gas heater switches on also but this shouldn't happen.

[mij59]

If the rcd trips there's probably a earthing fault in the electric installation or in one or more appliances.

What is the nominal trip current of the rcd ?
The 240 Vac side of the circuit you use can not cause the rcd to trip, what is connected to arduino side ?

[wraper]

No way this circuit could trip RCD unless you counting optocoupler pin numbers incorrectly. Therefore likely there is a leak somewhere else outside this circuit or you do not understand what you are doing. What happens if you disconnect one lead of 47k resistor? Does RCD still trip? Maybe stupid question, but is it really neutral? Don't you connect your circuit between Line and Earth by chance?

[Ice-Tea]

You could have a grounding error if as you have both the net and GND connected to that opto. Wiring fault? Dead opto? Disconnect your setup from the net, measure resistance between pins 1 and 2 and 4.

[jlmoon]

I notice in your schematic you have a 240VAC source.. I hope that isn't a (N) for neutral, but a L2.. right?
In this design what are you using as a return reference?  Ground -- I hope not. 
If you're using 240VAC might I suggest you use a control transformer (step down) to manage those dangerous voltages a little, you only need a few volts to bias those diodes.  Adding a transformer might provide you the isolation to faults that your RCD requires too.

[Kjelt]

+1 recommend small fixed 6VAC orso transformer as isolation.
Then on pin 4 the ground symbol is drawn as electrical safety ground instead of a low voltage return (ground), what is it, I hope the 0V belonging to the +Vcc of the arduino and not the earth ground?

And another thing, a standard water heating element is almost a short when it is cold, it only has real resistance when hot. You might want to start applying power very slowly till it is a bit warm.

[wraper]

I notice in your schematic you have a 240VAC source.. I hope that isn't a (N) for neutral, but a L2.. right?
In this design what are you using as a return reference?  Ground -- I hope not. 
If you're using 240VAC might I suggest you use a control transformer (step down) to manage those dangerous voltages a little, you only need a few volts to bias those diodes.  Adding a transformer might provide you the isolation to faults that your RCD requires too.

Likely he is not in US, for example, Australia have 240V single phase.

[aeberbach]

"or you do not understand what you are doing"

Fair comment, I'm a software guy.  Thanks all for the comments. I am in Australia.

I do have brown/blue connected, not green on the AC side. The Vcc/GND on the right of the opto is Arduino, no AC connection other than shown.

My only explanation is that I have two faulty H11AA1 - perhaps I did have the connection wrong at one time and destroyed both. I will start again on a new board and if it trips again with only the AC connected I guess I'll be ordering more. It looks OK because there is around 15M between pins 1 & 2...

WRT transformer - this looks suitable? http://au.element14.com/block/vb0-35-2-6/transformer-0-35va-2-x-6v/dp/1131608?MER=MER-BPR-4PPM-PD-BLK-TF&

If I am going to be driving a triac with the arduino via MOC3052 will adding a transformer on the zero crossing opto add much?

(I will be getting an electrician in to look at the RCD, it's nice that it trips when there is a fault but is a real pain most of the time - either it is a unit designed for less power than we use or there is some real fault in house wiring.)

[wraper]

My only explanation is that I have two faulty H11AA1 - perhaps I did have the connection wrong at one time and destroyed both.

Very unlikely to be so. There is very strong insulation between emitter and coupler with Isolation test voltage: 5300 VRMS. You can destroy them, so they no longer work but not damage the insulation.

[wraper]

I do have brown/blue connected, not green on the AC side.

The question is, are those brown/blue wires connected to the proper pins in the mains plug? If you disconnect Arduino and there is only optocoupler with resistors left, does RCD still trip?

[Kappes Buur]

In any case, it would be advisable to use a stepdown transformer.
Such as 240 VAC to 9 VAC at 500mA. They are not too expensive.

[Psi]

Most likely someone will spot the problem in under 5 minutes if you post a photo of your actual circuit with everything visible.

There really has to be a wiring error for that circuit to be triggering the RCD, that or the H11AA1 is dead.
You sure you didn't mistakenly pick 47ohm resistors

[aeberbach]

I pulled everything apart. Checked the cable (it's a molded computer cable I chopped for the project, so not likely to be bad). Rewired everything carefully. And here's my zero crossing pulse with no trips to the fusebox:



For this project I had planned to use an AC-DC converter, mains input, 3.3V @300mA out. Instead I will use a Block 6V PCB mount transformer and detect the zero crossing on the 6V output. Then use a diode bridge and LM317 with 6V AC input to get 3.3V for the arduino.

For 6V, aiming at 10mA across the diodes in the H11AA1 I would use 600R resistors, 680R would be close enough - sound right?

[wraper]

If using transformer you do not need optocoupler at all. It can be made with one diode and resistor. No other parts needed.

[Kjelt]

And then ofcourse in the software you have to compensate the delay between detecting the pulse and switching the optotriac else you are still to late with the switch on.

[jlmoon]

Just another thought... someone correct me if I am wrong.. A properly working RCD can detect microamps of differential leakage current in reference to ground and phase differential between neutral (if used) and and line side.  Is it possible there is enough leakage current in the material of the HC11 causing a fault current created by a loop from that arduino supply and your mains?  Just another path to ground .. per sè

[mij59]

Just another thought... someone correct me if I am wrong.. A properly working RCD can detect microamps of differential leakage current in reference to ground and phase differential between neutral (if used) and and line side.  Is it possible there is enough leakage current in the material of the HC11 causing a fault current created by a loop from that arduino supply and your mains?  Just another path to ground .. per sè

A microamp rcd is not practical, you would not be able to switch it on.
In a domestic environment 30 mA is considered to safe.

[eneuro]

I have two H11AA1 chips here and both do the same.

What about low reverse voltage of its input diodes? 
I used even LEDs as 230VAC indicators, but they were protected by 1000V 1N4007 in series  with current limiting resistors
However in spot welder project I used SS495A hall sensor inside cut toroid core with a few turns to monitor current under resistive load and detect zero crossing in ATTiny85 MPU by moveing averages processing based on sensed analog response from Hall sensor...so noise can be easy filtered digital, etc...

[wraper]

I have two H11AA1 chips here and both do the same.

What about low reverse voltage of its input diodes? 
I used even LEDs as 230VAC indicators, but they were protected by 1000V 1N4007 in series  with current limiting resistors
However in spot welder project I used SS495A hall sensor inside cut toroid core with a few turns to monitor current under resistive load and detect zero crossing in ATTiny85 MPU by moveing averages processing based on sensed analog response from Hall sensor...so noise can be easy filtered digital, etc...

There are 2 LEDs in reverse-parallel arrangement in H11AA1 optoucoupler. It is specially designed for AC/zero crossing.

[eneuro]

It is specially designed for AC/zero crossing.

Checked its datasheet and no reverse voltage specs or I missed something, so maybe it is capable to work even at 230VAC, but it looked very strange 

[wraper]

It is specially designed for AC/zero crossing.

Checked its datasheet and no reverse voltage specs or I missed something, so maybe it is capable to work even at 230VAC, but it looked very strange 

There is NO reverse voltage. This optocoupler does not care about voltage polarity. Don't you see how LEDs are drawn in the datasheet?

[Richard Crowley]

So follow the usual troubleshooting (AmEnglish) or fault-finding (BrEnglish) procedures.
Does it trip with the brown and blue wire hanging in the air?  With the two resistors in series across the line?  With an ordinary incandescent lamp?

Although, it is not clear why you need ZC detection at all? Its not as if you are dimming lamps here.
You are regulating the temperature of something with a substantial amount of thermal mass.
Typically, power to heating elements is controlled in integer cycles, phase-angle is of no consequence.
Use a ZC opto isolator (or SSR) and you don't even need to know or care about ZC.

[eneuro]

Don't you see how LEDs are drawn in the datasheet?

Everybody see this, but this is what is seen on its input-each diode inside must have some reverse voltage and probably is low.

Never tried something like this with crappy common LEDs-lets say this bule: 3Vf   5Vr only .
http://www.tme.eu/en/Document/06c97902c4f80b821a05430ac96f9a0d/OSG3DA5121A-TU.pdf
Really only with current limiting resistors such two diodes in this configuration can subsist 230VAC 50Hz without any additional elements? 
Lets find it out tomorow: 2 x 50k resistors  2 x such diodes @ 230VAC -> around 2mA with 50% duty on each diode and 0.26W loses.

[eneuro]

For 6V, aiming at 10mA across the diodes in the H11AA1 I would use 600R resistors, 680R would be close enough - sound right?

In setup with 240VAC and 2x47k resistors (input Vf in H11AA1 is typical 1.2V ) you had 3.6mAmax only 

[wraper]

Don't you see how LEDs are drawn in the datasheet?

Everybody see this, but this is what is seen on its input-each diode inside must have some reverse voltage and probably is low.

Don't you get that reverse voltage is limited by another led to approximately 1.5 V?