Crude Powerline FSK comms literally shorts the power bus at a certain frequency?

[cupoftea]

Hi,
Is there such thing as a powerline FSK comms system where multiple points along a long cable (bus) are
supplied power by the cable, but also the cable is literally shorted/opened at a certain frequency so as to give an
FSK powerline comms system? All the load circuitry doesn't get affected by the bus getting shorted because the loads
are supplied by capacitor banks which are downstream of high resistances connected to the bus so that the capacitor bank is
effectively "isolated" from the bus.
Obviously there just needs to be some initiation time so that the capacitors can get charged up at power up time.

 

[dick_freebird]

Powerline comms are short haul from what I've seen.
You throw C-blocked HF carrier onto the (effectively)
long-wire-L-blocked 50/60Hz line and you pick it
off C-blocked, elsewhere (before it's attenuated too
badly by L / distance).

So "AC short" (but at an AC frequency well
out-of-band). Since a useful modern data rate
is > MBPS carrier will be way up there, C will be
small, "short" will escape notice (just more
small signal noise, picocoulomb blips)

 

[cupoftea]

Thanks, i see what you mean , you are thinking of coupling the signal to the powerline via a blocking capacitor. I have seen that before too. Also, AYK, you get transformer coupling. However, the schem i have in fornt of me has none of that...it just has a BCX56 SOT89 NPN shorting the rail.....(24v rail)...well, its not shorting as its got a 75R in series with it.
But i guess due to the powerline L, that will give a bit of a dip in the rail which will be detected by the detector?
Have you seen that type of FSK system before?

I cannot see any blocking cap or signal transformer here.
The schem i have is on multiple pages with cross references going off all over the place but i deffo dont see any C or Transformer.

 

[dick_freebird]

So this is DC power line I guess. Same basic idea only now you can have as big a DC choke as you please (properly situated outside TX / RX link).

 

[cupoftea]

Thanks, It seems odd to do it our way with FSK?....i would have thought just ON/OFF keying would be the one for the way we are doing it?
What frequency would you say (highest) is possible with our way?

 

[dick_freebird]

Well a frequency detector would be one way
to trade bandwidth for spurious rejection (a
single pulse might be indistinguishable from
other line-crud)?

 

[FvM]

However, the schem i have in fornt of me has none of that...it just has a BCX56 SOT89 NPN shorting the rail.....(24v rail)...well, its not shorting as its got a 75R in series with it.

Sounds like you are not talking about a powerline modem but a special bus powered communication interface, most likely base band rather than modulated, e.g. MBUS or similar. Bus power is current limited, and pulling down the voltage by a certain delta or even short it temporarily is intended operation.

 

[cupoftea]

Thanks , whereas with DALI comms you have a current clamped comms bus, and you literally pull it to ground with a transistor to do the signalling....i think the one we want to use just kind of puts some small dV ripple on the bus, which can be read somehow. The ripple is the bits so to speak. And since its only small ripple its done by pulsing at a certain frequency say 10 times for a ONE, and a different frequency ten times for a ZERO. Is there such a comms system as this?
Its good because it doesnt require the full current limiting, just a bit of current limiting.......and probably a decent size inductor just downstream of the PSU that supplies the bus.

 

[dick_freebird]

Your "tone detector" ICs would "snap to" in about
10 cycles.

National used to have many "special function"
ICs, like DTMF encode / decode. Put the tones
on some HF carrier of convenience / legality,
put it to the wire, decode and detect elsewhere.

With enough selectivity in a "channelizer" you
could have FDM room for many "channels" on
one wire. Cable company manages it. They sure
are not using baseband shunt switches to do it.

How about everybody gets a "factory-set"
bandpass filter and matching oscillator and
PA w/ blocking/matching network? With a
home office control-head that has N of same,
in card-rack?

 

[cupoftea]

Thanks, as is well known, there are two ways to do Powerline comms for low voltage DC power cable...
1...one is where you have lots of capacitance along the power bus and so you have to transformer couple to it.
2....The other is where you have no capacitance on the bus, because the caps are "behind" high impedances through which they slowly charge up.

We did method (2) about 10 years ago and it is very problematic with noise when cable is long, due to no decoupling capacitance along the actual cable...however, (2) is the easiest to do powerline comms with as you can just literally just pull some current through a resistor in series with the start of the bus in order to get a voltage drop on the cable...and this can be detected as a 1 or a 0.....obviously not pulling as much current through the resistor gives the other logic state. Or, as we did 10 years ago by copying the common method at the time...literally pull current through a resistor with pulses at one frequency for a "1"...and at a different frequency for a "0".
This method is the well known hobby signalling method, on the web all over the place as we found 10 years ago.......but it is rubbish in terms of noise and EMC performance, and Radiated Immunity is non existent.
So why does anyone do (2)?

Why else is (1) so much better than (2)?

 

[FvM]

1. There are more methods of "powerline communication" than two. I guess there are more commercially established methods than you ever heard of.
2. Everything starts with a specification. Without it, you can't decide about suitable method for your communication problem.
3. Parameters of interest are e.g. powerline Un, In, AC impedance, noise spectrum, required data rate, acceptable modulation effort.

 

[cupoftea]

Thanks, the attached is the way we did it when we were kids (PNG schem and LTspice sim as attached) ....communicating from house to house to house with a single power/signalling cable.
The question is, why doesnt eveyone do it like this?
Its so simple.
Obviously you cant do this when you have loads of power being drawn from the cable, but if its just a lot of light loads at each "Point" (as shown here)....then why isnt everyone doing it like this?...because you can find every powerline comms in the book on the web...but not this beautiful noddy_land simple one. Why not?
Whats wrong with it?

--- Updated Mar 22, 2025 ---

______
Also, one of the attached one's uses a Loop Power supply regulator with a VERY slow feedback loop crossover frequency.....As such, even the pulsation of fairly light loads on it brings the power rail down significantly......as such, powerline comms is very easily achieved as shown. ...absolutly no isolation comms transformers needed......all thats needed is a simple FET with a pulsing drive...and the load it switches on to the bus is very light.....This is noddy level electronics in the extreme. Nobody could deny this. Why is no-one doing it like this?
Surely this is madness?
The Website has amalgamated this post with my last post so if wish to see the sim or schem then will need to read the title of it.

 

[cupoftea]

Woops, my apologies, i messed up the attachment showing the noddy level powerline comms.
So its attached here in LTspice and PNG.
Why does no-one do this?
We did it like this when we were 10 year old kids.

All you do is follow a PSU with a low dropout , extremely slow feedback loop linear regulator. As per countless websites which have depicted it over the decades. Then you can get a pulse pattern on the powerline by simply switching in/out a relatively light load. (as shown).
Why is nobody doing this any more?

 

[dick_freebird]

There has been much advancement since the
telegraph.

Not that advanced is necessarily "better".

But it might well be.

Might spend time in FMEA and decide whether
a long wire where every blip is a bit, is robust
enough comms for the service (and those who
will punish excessive falsing - over here it's a
money maker for the Sheriff / local PD).

 

[daiki96]

Yeah, that's basically the brute-force method behind older FSK powerline comms—intentionally creating a brief impedance change or “short” at a specific frequency to transmit data. It’s crude but effective in low-bandwidth setups. Definitely not ideal for modern interference-sensitive environments, but fascinating how simple physics can still be used for data transmission. Old-school tech doing its thing! youtube vanced