Mains powerline comms.

[dr pepper]

First off I'm an industrial leccy, so power distribution and safety is something I'm familiar with.

I want to communicate over the mains network (just one phase and within 10m) with either a pic or arduino digispark.
So far I'm thinking generate a hf squarewave maybe 100kc using a software loop or 555, and use a ne567 tone decoder or similar at the other end (simplex comms) to pick up the data again (I could use a choke to seperate the hf from the 50hz) just so I can fire a relay in one of half a dozen remote devices from a central control unit, isolation isnt required all devices can be at mains potential.
Any ideas?

 

[Tony Stewart]

I might have done it that way in the 70's or use the CMOS PLL for FM encoder/decoder with LC filtered current transformers.

The current state of the art, may interest you https://www.maximintegrated.com/en/app-notes/index.mvp/id/5445

It depends what protocol and user interface you want.
ANdroid?, Duo? PC, Rasp. Pi? or roll your own.

 

[AnalogKid]

This can be done off the shelf with X-10 modules if building something from scratch is not part of the project.

ak

 

[KeepItSimpleStupid]

How about powerline Ethernet? To expensive probably?

 

[Nigel Goodwin]

First off I'm an industrial leccy, so power distribution and safety is something I'm familiar with.

I want to communicate over the mains network (just one phase and within 10m) with either a pic or arduino digispark.
So far I'm thinking generate a hf squarewave maybe 100kc using a software loop or 555, and use a ne567 tone decoder or similar at the other end (simplex comms) to pick up the data again (I could use a choke to seperate the hf from the 50hz) just so I can fire a relay in one of half a dozen remote devices from a central control unit, isolation isnt required all devices can be at mains potential.
Any ideas?

I seem to recall there's an application note at MicroChip about this exact thing?

Quick google found these as the first two hits:

https://ww1.microchip.com/downloads/en/AppNotes/00236a.pdf

https://vonkonow.com/microcoder/projects/x10/x10_hw_en.php

 

[Tony Stewart]

If you just want to toggle AC powered units, an offline simple DC supply (20mA) to power an FM DTMF decoder , toggle mode flipflop and Triac driver , then you can choose 12 functions to toggle each or have 10 with master all on and all off.

Sending DTMF would be pretty simple to copy existing designs.

 

[dr pepper]

Thanks Nige.
I think I'll come up with a combination of the x10 tx and part of the x10 receiver with a ne567 pll.

 

[Tony Stewart]

no problem
BTW here is a 2^12 number encoder chip used to remote control >4000 different addressable garage doors by RF or LED .
https://www.engineersgarage.com/sites/default/files/HT12E_0.PDF

 

[dr pepper]

I just looked up dtmf over the mains, found a circuit that uses a standard dtmf encode and decode chips, I was expecting some filtering and coupling to the mains, nope, a 0.1uf cap to neutral, yep thats it straight out the encoder through a 0.1, and into the decoder with a 0.1.
I'll try that, I have some decoders, maybe theres a library for the 'duino to generate tones.
From memory the decoder chips I have the last valid code remains on the chips o/p, so with a little thinking of bit patterns I can control my 4 devices at least at the receiving end with a transformerless power supply, the decoder and a tiac.

 

[AnalogKid]

I just looked up dtmf over the mains, found a circuit that uses a standard dtmf encode and decode chips,

Link?

 

[Tony Stewart]

0.1uF to 230Vac 50Hz mains is around 25k Ohms reactance., so it must be plastic 400V rated

while a 10us lightning impulse is around 25 Ohms reactance @ 50kHz and drops with rising spectrum to 1MHz.

Make sure you have a low ESR TVS installed and a X2 line rated plastic Cap. with a line filter.

 

[Tony Stewart]

Beware of EDN schematics. These are not tested.
The article illustrates a technique; it is not a blueprint for a finished product.

 

[dr pepper]

I have some x7r caps, and a few salvaged transient suppressors, not sure if they are mov's or 'zorbs, will have to dig them out.

One thing that concerned me, the circuit applies dtmf across the earth and neutral, here the earth and neutral are strapped together in the substation, maybe they rely on inductance, I'm 1/4 mile away from the sub so it might work, I'll try it.

The circuit is in an electronics mag from the 80's, the covers gone so not sure which.

 

[Tony Stewart]

I have some x7r caps, and a few salvaged transient suppressors, not sure if they are mov's or 'zorbs, will have to dig them out.

One thing that concerned me, the circuit applies dtmf across the earth and neutral, here the earth and neutral are strapped together in the substation, maybe they rely on inductance, I'm 1/4 mile away from the sub so it might work, I'll try it.

The circuit is in an electronics mag from the 80's, the covers gone so not sure which.

X7R will fail from switched CM transients on Neutral. Must use Plastic (PU,PE type) pref, X1 or Y2, unless you plan on trying to transmit via Ground current.

 

[Nigel Goodwin]

One thing that concerned me, the circuit applies dtmf across the earth and neutral, here the earth and neutral are strapped together in the substation, maybe they rely on inductance

It's a common technique, bear in mind you're using an RF frequency and it doesn't really travel much out of the house (due to the said impedance of the wiring).

 

[Tony Stewart]

Actually I think it is universal that all DT's are neutral ground (at every distribution transformer.)

 

[dr pepper]

Yes thats what I thought, however the article I'm talking about applies dtmf tones to the mains wiring, and as you know dtmf is audio.
Some ht9020's turned up yesterday so I'm going to have a play with them, I allready have some decoders.

 

[Tony Stewart]

I noticed everyone uses the hot LINE to interface the 125kHz carrier and Microchip specify X2 rated caps.

I think your ceramic coupling choice will work just as well between Neutral and Ground vs Line and ground without the 240Vac stress on the cap. It wont trigger ground fault current interrupters either.

Although the impedance at 50Hz from Neutral to Earth is lower than Line to Earth , but at 125kHz the impedance of Line<>Gnd vs Neutral<>Gnd due to dielectric impedance (C) of oil in DT coupling to earthed tank. So coupling signal between Neutral and Ground ought to work same as Hot to ground. I guess it depends on Neutral inductance to earth and distance to that connection.

 

[Nigel Goodwin]

I noticed everyone uses the hot LINE to interface the 125kHz carrier.

Certainly not 'everyone', however if that's what the X10 standard specifies, then you obviously follow the specification in order to be X10 compliant.

Plenty of other gear has used neutral and earth, or even live and neutral, it really makes little difference as long as you stick to the same standard. But even if you use neutral and earth, it's still essential to use an X2 rated cap.

 

[dr pepper]

Must rely on the inductance of the lines back to the sub then for neutral & earth.
At least with live & neutral you have the inductance of the power trans in the sub, from phase to star point, which for a 500kw or so tranny found in a suburb street will be fairly high, the thing then would be the impedance of any equipment running off the mains, telly's & stuff probably will be fine, washing machines and low impedance stuff might be more of an issue.
I'm going to have a play as soon as I've made my new led panels work over my workbench.
I have some x's, x2's and y's, the latter for the earth side.