Solid-state relay not fully isolating load

Thread Starter


Joined Jan 18, 2016
Hi all,

Recently I decided to dabble in home automation, so I built a mains switch controllable over WiFi.
It worked great for an electric kettle and a (presumably - it wasn't mine) incandescent lamp.

However, today when I was trying to set up some crude current monitoring, I noticed a weird problem with the LED table lamp I plugged in.
Even with the relay control input completely disconnected, the lamp would briefly and dimly flash, about once per second.

First I suspected a faulty relay, so I pulled out another one and connected the AC terminals in the same way, with nothing on the control input. The same phenomenon occurred.

I connected my scope to the AC output terminal of the relay, and instead of showing zero volts like it's supposed to, I saw a 40V, 50Hz AC signal which over the duration of 1 second increased to about 85V, causing the lamp to flash and dropping back to 40V. The shape was not sinusoidal but more like a trapezium shape, distinctly clipped at the top and bottom.

The relay is a Fotek SSR-25, supplied from 230VAC.

I'm just wondering if anyone has any idea what might be causing this, or if they've encountered the same thing?

Thread Starter


Joined Jan 18, 2016
oops, went over the edit deadline...

EDIT: OK, so my post was a bit premature. The problem only seems to be associated with the LED table lamp. I tested it with a fan and there was no AC appearing on the outputs when the relay was switched off.
Still, I'm curious as to just what might be causing this effect. The LED bulb contains what looks like a switchmode converter, powering 7x 1W LEDs...

EDIT...AGAIN: I connected a USB power socket to the relay output, and into that I plugged a USB mini LED light.
Guess what? The /exact same/ thing happened!
This has me legitimately perplexed, but it definitely has something to do with switchmode converters.


Joined Nov 30, 2010
I see a snubber circuit in the SSR. That will allow a tiny bit of current.
The basic design is an opto-isolator to a zero-crossing triac (with a snubber).
The triac should be immune, but the snubber isn't.

And just for nit-picking, the SSR doesn't have a voltage output. It's a shorting switch. You supply the voltage and the load and the SSR shorts its output when you tell the input to "go".


Joined Aug 1, 2013
I agree with #12.

It looks like the switcher input is rectifying the AC trickle current and integrating it into its filter cap, and when the voltage on the cap gets high enough its control chip comes out of undervoltage lockout and runs until the cap is discharged below the UVLO threshold. Congratulations! You've discovered the world's most complex relaxation oscillator.

Someday down the road someone else is going to have this problem, and I'm going to look like a fricking genius.


Thread Starter


Joined Jan 18, 2016
And just for nit-picking, the SSR doesn't have a voltage output.
Well, there are two AC terminals. One is connected directly to the active line of the mains, meaning that the current flows into the relay from this terminal; i.e. the 'input'. By default this makes the other one the 'output'.

But I get your point, it's the "load side terminal" ;)