How to shield a live wire at 240V ?

[_q12x_]

I have a electronic board with leds (made by me) that is driving a relay. When that relay is closing it’s contacts, is actually switching the 240V for the light bulb switch in my room. This board is quite long, like 50cm (half a meter). The relay is in the left corner, and the live wires from it goes in behind the board to the hole in the wall for the mains switch. I also have a mild steel sheet behind my board that is grounded. It is shielding the interference of the 50Hz from the live wires from the wall to my sensitive circuit. But even If I have this grounded metal shield behind my board, the live wires from the relay are still affecting my entire circuit board, keeping it ON all the time. If I am disconecting the live wires from the relay, the board is functioning very well. Another IF, is if I disconnect the ground from the metal shield, the circuit board goes nuts. So the shield is doing it's job fine, but only for the live wires inside the wall !!! But not for the wires from the relay to the live switch.
- I want a way to shield these wires !
Thank you !

 

[_q12x_]

Good morning. Some very helpful references - thank you. I was gathering together the components yesterday, many of which were well hidden in various drawers. Today is the day to start constructing my SCM.

You are gaining a good understanding of how the mains and parts of your own circuitry are interfering and disrupting the operation of your proximity switch.

The little video was useful. It does seem to work well but we both can see this circuit, when demonstrated, is not near any unearthed mains equipment or mains wiring which would interfere or disrupt its operation, which your is near.

Good morning minster @marconi
Indeed, mine is a special case, a challenge and a pain in my but.
What I've tried today.
My other friend asked my why I dont keep the pot in the SCM, as the original movie shows. And I give him a vague response at that time, "because it works without it". And today I tested both with a 10k and 1k pot, and both just shut off completely the entire circuit. Like a shunt. Nothing lit. Just all dead. When I disconected the pot, everything got to the normal. And Im talking experimenting on the wall directly. What a pain in my back from my awkward position on it to solder those fine wires, and not to hang some other wires.
Another good idea was to work directly on the wall and I did it for a couple of hours of switching back and forth components and wires. In the end, the conclusion, all the filtering we are adding, is kind of helping but not really. I removed all the caps and resistors from the signal wire in the end.
Next, I changed the signal wire with a coaxial wire, very small diameter one. And it worked normal as before with the copper wire, but when I put the ground to the mesh of the coaxial, everything got lit to maximum. And also nothing responded. Aaah what a disappointment. I noticed that signal wire is very sensitive, pretty much like my antena, not quite, but too sensitive anyway. So I wanted to shield that sensitivity. But it didnt work.
I noticed something else very important. The white led from the SCM is ON all the time. It is off while everything is off. When I activated to OFF mode, and also the light in my room is off as well. BUT, when I switch to ON mode, that white led should get On when my hand is near it, which it does, but immediately as I retract my hand from it, it should go off. Instead remains ON all the time. I didn't pay attention to this detail before. Im sure it was the same in my previews videos and attempts. SO in a word, no matter how many "smart" filtering we will add to the signal wire or to the rails of SCM, he is functioning normal !!! but the source of interference is outside of it and is keeping it open. That's the reality. We must find that source and neutralize it.
Small modifications : I eliminated the 1000uF because they were delaying too much the normal operation of the entire board circuit. I also replace one to 100uF but then I realized its not really necessary. Sorry to be against the "normality" but hear my reasoning : If on the table is working flawlessly, then is in very good condition and very well made so far. The only problem is the wall itself. Well, the 50hz that I suspect from inside him. I tested all the caps so far and no real visible improvement from them. So I took them off. They were test subjects and nothing more.
What else? Im so tired of it not working. Its the truth.

 

[_q12x_]

You may have two problems. One is that your relay is located to close to the electronics board. You can put this relay a couple feet away in its on grounded metal box and stop a lot of the interference. The control wiring for that relay may have to have a filter of some type on it to prevent the interference from traveling back on those control wires.

Since you stated it is a 5vdc suppling you electronics device then I will ask where is the power for that power supply coming from. If it is the same 220 circuit then you need a better power supply or to filter that dc5volt more.
The interference can be coming into your electronics at more then one place.

I understand your point. Its a very good point. But is not applicable in my case. Here is my home, and not some industrial place. It is the normal light switch. I have no knowledge of the wiring but I imagine it is wired to the rest of 8 floors as common or something. Im not sure at all, just imagining. And also I believe is kept away from normal 220 socket that as you say, other motors in the building can interfere.
It is my believe that is simply as I said , only the 50hz interference, because this is a very sensitive circuit/transistor, and very close to that power line, it is just too loud to him not to hear it.
You are right, I should make a faraday cage to that relay. I will put your idea in queue.
As for the power supply 5V for the board, it is from a socket and not from the light 220 in the hole in the wall that is exposed right now. It is also very far away, on top of the door. And I drag 2 very long wires for 5V from it to power my board.
Thank you for your intervention ! Interesting points nonetheless.

 

[_q12x_]

The time has expired to edit my previous post
I made a movie in this time and I upload my picture right now.
Notice the wire going to the grounded metal shield from the coaxial mesh.
Everything should stay OFF at this point, but it is ON. In short, is not working.
I will upload a short video very soon how is behaving on my working table, when I connect a 220 light bulb to it.

 

[marconi]

After pressing 'post reply' there is only a ten minute window' to edit.

I cannot comment right now and indeed owe it to you to start constructing my SCM. I shall turn the heater on in my shed and then set to work in half an hour while watching the snow fall outside.

Happy days.

 

[_q12x_]

A new test on the workbench. Very short video of 3min.

 

[marconi]

Here is how I made a small Faraday shielded box to put my SCM inside. The foil is backed with grease proof paper and then glued to the box of matches.

 

[_q12x_]

Wonderful ! I didnt manage to make one yet, I was thinking too complicated, to make it from my iron sheet. But yours is way more cheap, that's why I like it. Bravo.
Now test your SCM near the wall. See if you can get the same strange result as mine. I will very soon make a Faraday cage for only the SCM and experiment with it on my wall. I am very curious if you have the same strength as mine here (for the same sensing circuit I mean). Instead of the match stick in the video for the 1G ohm, I used an inverse bias diode.
Now that I remember... I already did tests for faraday cage for my SCM. And I think they blocked my field as well.... I dont remember very well since it passed some months from then, but I remember I chosed not to use it on the SCM itself and I build the metal shield in the back instead. I made houndreds of modifications so far, I start to forget them.

 

[marconi]

The story so far. Not finished of course. I made a safe box with 230 V ac inside with which to test the final circuit next to it.

 

[_q12x_]

Watch my video at #140 and see my experiment that I made on my workbench. Everything worked perfectly there. I am sure it will work fine on yours as well. Best to try it exactly how I did, next to the light switch.
Or do what you have to do, but get close as much as you can from my weird situation, if possible.
I actually had another idea to try on my workbench, to disconnect the entire iron table underneath from earth and attach the earth to my shield. I think the 220 interference from my lamp was also diminished by my table. It might.

 

[Dreaded]

I understand your point. Its a very good point. But is not applicable in my case. Here is my home, and not some industrial place. It is the normal light switch. I have no knowledge of the wiring but I imagine it is wired to the rest of 8 floors as common or something. Im not sure at all, just imagining. And also I believe is kept away from normal 220 socket that as you say, other motors in the building can interfere.
It is my believe that is simply as I said , only the 50hz interference, because this is a very sensitive circuit/transistor, and very close to that power line, it is just too loud to him not to hear it.
You are right, I should make a faraday cage to that relay. I will put your idea in queue.
As for the power supply 5V for the board, it is from a socket and not from the light 220 in the hole in the wall that is exposed right now. It is also very far away, on top of the door. And I drag 2 very long wires for 5V from it to power my board.
Thank you for your intervention ! Interesting points nonetheless.

Place does not matter friend interference happens every where. Yes it is all common wiring as you think. The interference can be coming from any where in the house or even from outside you home from the power company.
There is usually some interference on all electrical power. This usually does not make a difference in most electronic devices because of shielding and filtering of electronic devices produced by manufactures. If you device is electronic without the input powering the device being sufficiently filtered then you will have this problem.

This is what I would do. First move the relay and place it in a metal grounded box away from the electronic board. If this did not correct it then I would get a put a ferrite bead around the AC wires powering the DC power supply. If that did not fix problem I would look closer at filtering the wires going to the relay.

 

[_q12x_]

Shielding Problem Resolved !

 

[marconi]

Good morning. We have snow settled on the ground - not thick but enough to form a blanket.

How have you succeeded? Really looking forward to finding out.

I could not replicate your problem with my copy of your SCM in my shed - I even tried using the reverse based diode. I could use my SCM to detect the presence of my hand when it was within a few centimetres or I touched the antenna but the usable output was in on/off form - no usable variation with which to gradually illuminate the wings which I believe is still a key requirement for the operation of your wings? - so it could not be connected to my ADVM and control the 8 yellow LEDs - my 'wings'.

This could be - I cannot be sure - because my shed is wooden and all the electrical fittings are metallic and earthed apart from the overhead pendant lamps - so there is little ambient mains electric field for my body to pick up Or because I did not have a 5V relay switching 230V ac like you have during my experimentation with the SCM circuit although I did try my safe source of 230Vac inside the white box.

I think that the proximity detector you used and I copied is actually just that - if something is close to it which is energised by the mains and unearthed then the LED will glow. The current that is flowing through the LED (and MOSFET) is actually half-wave rectified alternating current. If our eyes could see it the LED would be flashing on and off at 50Hz with maybe a little persistence produced from the MOSFET's gate-source capacitance being charged and discharged. There is a little bit of LED brightness variation but I think this is due to the uncertain level of quiescent gate bias of the MOSFET. It maybe because I am using 2N7000s but they are from the same family as the 2N70002. What these MOSFETs have in common is a wide variation in gate threshold voltage Vgs(th) - typical low is 0.8V, typical high is 3.0V - quite a variation - see 'ON tests page 2' of:

https://www.onsemi.com/pub/Collateral/2N7000-D.PDF

- this may be a factor too.

I have other 'criticisms' of this circuit for the purpose of your wings project and have been thinking about an alternative circuit design which I hope to have to time to construct today.

 

[_q12x_]

Thanks for the early response and interest. Especially you being 1h behind me, Mine is 6 now but yours is 5 in the morning. Haha.
So, from the super advice of my friend on another forum, I got the idea to ground the 0V rail. That was the entire trick I didnt considered at all.
Here is a 9min video.

But I brake it again.
I made the test on my workbench after that. It worked exactly until I changed the antena wire.
I started to change the antena with a larger wire, and I started to get weird behavior. 9'th led is staying lit, no sensing activity whatsoever. Also SCM led is staying lit as well. I took out the 2n7000 and test it with my DMM on diode, and I was getting readings all over it G-S G-D S-D, everywhere. I change it with a new 2N7000, test this new one as well and got readings only from S-D but not on the rest. I solder it and the same weird behavior as before. Im starting to think maybe is the IC gone? Or parts of it gone? It is just impossible to keep it alive too long.
I also replace the 2n3904 from SCM but no difference at all.
Ive checked every connection in SCM and resoldered eveything, again, no difference. An the SCM led is staying lit as well
Everything on the working bench.

 

[marconi]

Have courage - you know now it can work as you desired....so for that reason I have not changed my 'Winner's star' award to you!

 

[marconi]

Just to check something which may remotely be the problem. With the light switch in place and you wings removed, take a digital voltmeter on the ac Volts range and then measure the potential difference to earth of the wall in the region (shown in yellow in my diagram) where the wings would fit. There might be a damaged line conductor in the cable run to the switch box. Start on the highest ac Volts scale and then work downwards in scale. Are those two holes above the switch in a region where the cable might be run in the wall?