Unwanted triggering by mains

Where is that switch located? In the box or outside?

--Spehro Pefhany

Terry Pinnell wrote

Your circuit basically is a RF detector Either re-design it so it has a capacitor to ground from pin 12. or just add 1 nF from pin 12 to ground, might just be enough.

Tin is not ferromagnetic and makes a crappy magnetic field shield, unlike steel? Replace with steel box. Your junction impedances at IC inputs are really good receptors of magnetic field induced voltages.

Most tobacco tins are made of coated steel. Tin is an expensive metal and rarely used if there is an alternative.

The input side into pin 12 is asking for trouble! I would swap the GND side and switch side connections around. And raise the cap to 0.1uF.

An alternative is to put 0.1uF and say 100R in series from the switch input to ground.

He should use magnet attraction test.

It may not explain the unwanted triggering but isn't a flyback diode across the relay coil to suppress voltage spikes a generally good idea?

You could get rid of the one-shot, something like this:

There is no need for a switch de-bounce circuit as all that bounce will do is lengthen the on-time by a few tens of milliseconds.

Follow the suggestions of a capacitor from pin 12 to ground as this will get rid of most RF noise. To achieve this: Delete the 0.01uF capacitor from its current location and connect it between pin 12 and ground instead. Reduce the 1Meg resistor to about 100 ohms. This is just to limit the peak current in the switch so that the contacts do not burn. Add a low value capacitor in parallel with the 100uF power supply decoupling to improve decoupling at RF. Maybe add a zener diode or TVS rated at about 24V across the relay coil to reduce the back EMF (connected so it does not normally conduct). Consider a very small capacitor from pins 8&9 to ground if there are still problems. Make sure the case is connected to ground. John

I think I'd just go for a simple make break switch in the power line to the relay coil, or even just to drive the transistor, unless there is a real need for a monostable.

The EMI is conducted voltage spikes, not magnetic fields. Besides, a "tobacco tin" is probably not made of tin; most likely plated steel.

I know! I realized I was being an idiot two seconds after hitting send. I even typed the word monostable in my reply without it sinking in!!

I think that's a momentary, used as a trigger for the NORs which are a mono stable on the order of 3 minutes? The timing resistor is labeled 1M8 for 10 Meg?

That does not help the magnetic pickup around any external component from p in 12- thru pin 12 and internal cmos gate to Vdd || Gnd (parellel) thru Gnd || Vdd back to external component. All the induced voltage is across the h igh impedance internal CMOS gate.

That resistor is to discharge the ac-coupled triggering cap off the switch.

Okay, that's good.

or

=1

e steel? Replace with steel box. Your junction impedances at IC inputs are really good receptors of magnetic field induced voltages.

I disagree. Where are voltage spikes coming from unless he has really small gauge wire connected to his lamps that cause big line drops. Now magnetic field surges due to inrush on turn-on and inductive interruption at turn-of f make much more sense. He possibly could improve situation, fairly immense ly, by converting to LED lamps, but that doesn't directly address the probl em.

Yep- he either omitted it from diagram, or has really lossy relay.

I would say a very good idea. Some relays have them internally, but not something I would take for granted.

ing

t or

aw=1

ike steel? Replace with steel box. Your junction impedances at IC inputs ar e really good receptors of magnetic field induced voltages.

ll gauge wire connected to his lamps that cause big line drops. Now magneti c field surges due to inrush on turn-on and inductive interruption at turn- off make much more sense. He possibly could improve situation, fairly immen sely, by converting to LED lamps, but that doesn't directly address the pro blem.

Fluorescent lamps can make a lot of RF, especially when they are starting. He said the circuit malfunctions when he turns other lighting on or off. I'd certainly suggest a ceramic caps in parallel with the 100 uF power deco upling cap.

Everyone focuses on the pin 12 input to the CMOS gate, but how do the noise spikes get past the switch? If they capacitively couple through the switc h even a small cap from pin 12 to ground would fix that, say 330 pf.

Rick C.

- Get 6 months of free supercharging - Tesla referral code -

On Sunday, January 13, 2019 at 4:08:24 PM UTC-5, snipped-for-privacy@gmail.com wr ote:

e:

ering

p.

it or

?raw=1

nlike steel? Replace with steel box. Your junction impedances at IC inputs are really good receptors of magnetic field induced voltages.

mall gauge wire connected to his lamps that cause big line drops. Now magne tic field surges due to inrush on turn-on and inductive interruption at tur n-off make much more sense. He possibly could improve situation, fairly imm ensely, by converting to LED lamps, but that doesn't directly address the p roblem.

. He said the circuit malfunctions when he turns other lighting on or off. I'd certainly suggest a ceramic caps in parallel with the 100 uF power de coupling cap.

The fluorescents with electronic ballasts switch at something like 36KHz, b ut the current is pretty well smoothed with a series inductor. It does stri ke an HV arc at turn on but not a surge current. Any kind of grounded metal lic box enclosing the circuit shunts those kinds of frequencies due to elec tric field right to ground. You will have a real tough time measuring elect ric field induced currents in his circuits.

se spikes get past the switch? If they capacitively couple through the swi tch even a small cap from pin 12 to ground would fix that, say 330 pf.

Yep, I have one in place but forgot to include it in my schematic, sorry.

Terry

Many thanks for all suggestions, which I'll follow-up tomorrow.

BTW, it's an LED lamp, mounted outside the shed. The lights inside shed are a mix of fluorescent and legacy filament.

The light switch is mounted close to the tobacco tin.

Precise lamp timing is not important; currently about 60 secs to find my way at night after leaving shed.

Terry, East Grinstead, UK