Simulating noise in Falstad

[farmhacksca (Kody)]

This is a quick question about simulation and how (or if) there is a good way to simulate noise.
Here is the circuit I am working with and my attempt to inject noise into it:

It can also be found here https://tinyurl.com/284qumuk in the falstad simulator.

For some context this project is something I am working on to implement an alarm system for my current employer. The first design did work (I used one or gate and had each switch connected individually) but it @ChrisGammell brought to my attention that with all the welders and air compressors around significant noise could be banging around.

It turned out there was. I used my analog discovery to view a 1V sine wave and when a welder started welding there was a couple of tenths of noise added. Earlier this week I posted on electronics stack exchange (was going to post here too, but didn’t get to it) and someone responded with this circuit and suggested I research what a fully supervised loop is.

Now I do like this circuit and the idea of a fully supervised loop, but I am not sure if I’ll stay with it since I need to talk to my employer how much more time they want me to work on this.

My main question is more general on what is the most proper way to inject noise into a simulation? It could be for this circuit or others.

[farmhacksca (Kody)]

I hope I don’t take this too off topic by posting about my first breadboard testing of this circuit.

So this is the circuit that I started testing:

Now the two comparators and NAND gate will be after the 1.5k (actual values may change when I pick those ICs). The wire being used (the 2k soldered at the end) is 100ft of a 18 GA twisted pair wire.

I put channel one of my analog discovery 3 (using the x10 probe), and the 12V is being supplied by a power supply.

Now I did take some measurements of when no coupled inductor is there.

This one is when the welder was not working.

And here it is when the welder was welding.

As you can see the welder introduces a couple of hundredths of volts in noise.

Here is when I put two ferrite beads of the same value in-place of the transformer shown. I do not recall whether a welder was welding or not.

Now for some more context here is my AD3 output from awhile back with no power supply attached but not using a twisted pair (using the x1 probe).

So simply using a twisted pair aided in getting a clean signal a lot!

If anyone has more advice with dealing in noisy areas or about my original question of simulating noise I’d be ecstatic to here from you.

[farmhacksca (Kody)]

I’m going to share a couple screen shots of my AD3 once more. This time I used a couple of common mode chokes.

The first up is a 4.7mH one from Bourns (Part # DR331-475BE).

And here is from the logger

Next is a 1mH also from Bourns (Part # DR331-105BE).

And here is from the logger

So I have two questions after doing these tests.

1. It appears to me that the chokes do not reduce the noise very much at all. Are they needed as long as I use a twisted pair?
2. Second question will require some more info. When I start the logger I see a big spike in voltage before falling down (the max value will be shown below, 6V). Is this concerning? What would be causing it?