Power-line communication and noise filtering

[Francenthusiast]

Hi everyone !

I have a small problem with my garage bench: I used to repair a lot of stuff there and I always had a bad internet connectivity with my power-line communication adapter. I just realized that two things are messing with the PLC bandwidth: my soldering iron and an overload protection plug (just 3 MOVs between each wire of the mains and a neon lamp with a resistor) . If either of the two is plugged, my bandwidth drop about 60% (when I ping google.com, I get 53% loss).

Is it possible to modify the overload protection (i.e. with a cap) or add some filtering to the soldering iron to prevent those bandwidth drop ?

F.

[Marco]

In the case of the protection plug it's clearly not interference, I think the capacitance of the MOVs is eating the signal. Bourns suggests placing GDTs in series with the MOVs.

Not sure what the soldering iron is doing ... putting it on a long extension cord might be enough to minimize problems.

[madires]

A GDT in series with a MOV? Wouldn't it be more sensible to use just the GDT? Another method for hiding the capacitance of a protection component is to use a diode bridge.

Has the soldering station a X filter cap (between L and N)?

[Francenthusiast]

Hi Marco and Madires !

Marco: You were right about the cord extension, the iron is no longer affecting the PLC bandwidth. One problem solved.

Madires: Sadly, this is just a cheap Conrad soldering iron, not a fancy soldering station, so I don't think there is any discrete component inside (maybe just a heating resistor)

[TheMG]

Sounds like the signal strength may be marginal to begin with, and the characteristics of the surge protector and soldering iron are such that it further attenuates the signal lower than the device is capable of reliably communicating with.

Is this an attached garage or detached? If there are other circuits available you may want to try relocating your powerline device to a different circuit and see if it helps. Also, the less wiring between the two devices the better. Avoid plugging it into a power strip, extension cord, etc. Choose an outlet on the circuit that is as close as possible to the breaker panel if possible.

You may also want to consider other options such as WiFi with directional antennas to form a point-to-point link.

[T3sl4co1l]

A GDT in series with a MOV? Wouldn't it be more sensible to use just the GDT? Another method for hiding the capacitance of a protection component is to use a diode bridge.

No, because the GDT will stay on after the pulse, until a fuse blows (or it absorbs too much energy and explodes..).

The bridge is a neat idea.  Beware that it may not withstand MOV sized transients; or if it does, it'll be pretty big anyway, i.e., not a bad capacitor itself, at least near the AC line peaks.  It could also make it worse, because the variable capacitance acts to parametrically modulate the communication waveform at 100/120Hz.

The MOVs may not be necessary at all.  Depends on your electrical environment, what equipment is plugged in nearby, and what kind of survival, over a period of years, you expect from it.

Tim

[Kleinstein]

One could replace the over-voltage protection with a version that also includes filtering by inductors, not just an X cap.
There may be other parts (e.g dimmer, fluorescent lights, fans, ...) in the house that do to much filtering and thus have a negative influence on PLC. Another problem could be if the signal has to go from one GFI circuit to an other. Also a neighbor using PLC as well might interfere.

LAN over powerline is broken by design anyway: to much other units with simple filtering can stop it from working. Also emissions are sometimes to high (above legal limits)  - so in principle you might be liable for interference with other radio communications.

[Francenthusiast]

TheMG: that's an attached garage, plugs are 15m cables away from the breaker panel and the other PLC is 5m in the opposite direction. I just changed the configuration to connect the PLC on the plug tha tis the closest from the breaker panel and used a 5m RJ45 cable to the PC - we'll see if it works that way because I can't change the computer place nor the bench's, it's bolted on the wall. (I don't know if I'm clear enough, have to enhance my english skills)

T3sl4co1l:

No, because the GDT will stay on after the pulse, until a fuse blows (or it absorbs too much energy and explodes..).

From what I learned on wikipedia :

a GDT once triggered will continue to conduct at a voltage less than the high voltage that initially ionized the gas; this behavior is called negative resistance. Additional auxiliary circuitry may be needed in DC (and some AC) applications to suppress follow-on current, to prevent it from destroying the GDT after the initiating spike has dissipated.

So for small voltage circuits, GDTs have to be used with other circuitry to reset it after it triggered ? One can't use it alone like a MOV ?

The bridge is a neat idea.  Beware that it may not withstand MOV sized transients; or if it does, it'll be pretty big anyway, i.e., not a bad capacitor itself, at least near the AC line peaks.  It could also make it worse, because the variable capacitance acts to parametrically modulate the communication waveform at 100/120Hz.

I didn't got it... What kind of bridge ?

The MOVs may not be necessary at all.  Depends on your electrical environment, what equipment is plugged in nearby, and what kind of survival, over a period of years, you expect from it.

In the garage, not much: water tank, a radio, a long power cord for the vacuum cleaner. That's it. The thing is I live in Provence (south of France) when from March to June, when there are lightnings, there are huge loads of lightnings !  (I have 1 or 2 electrical equipments that blows their power board or MOVs or fuses each year !) 

Kleinstein:

One could replace the over-voltage protection with a version that also includes filtering by inductors, not just an X cap.
There may be other parts (e.g dimmer, fluorescent lights, fans, ...) in the house that do to much filtering and thus have a negative influence on PLC. Another problem could be if the signal has to go from one GFI circuit to an other. Also a neighbor using PLC as well might interfere.

I forgot that: every ligths in the house are LEDs (IKEA, Philips, not chinese one with crappy PSUs inside ! For those ones, I don't turn them on, I take them apart !)

LAN over powerline is broken by design anyway: to much other units with simple filtering can stop it from working. Also emissions are sometimes to high (above legal limits)  - so in principle you might be liable for interference with other radio communications.

True. But I don't want to flood the house with 10W wifi repeaters nor 40m long RJ45 cables running on every walls (already done that at work, done with it !) 

[madires]

If I got that right, a GDT with a MOV in series to deal with the follow-on current of the GDT. And since the GDT is slow a parallel diode bridge with a TVS?

[T3sl4co1l]

So for small voltage circuits, GDTs have to be used with other circuitry to reset it after it triggered ? One can't use it alone like a MOV ?

Correct.  You need:
1. A limited-energy circuit (e.g., ELV AC/DC supply, telecom, or other TNV type circuit), or
2. A series element to increase the "on" voltage, allowing it to self-extinguish (that's what the MOV does), or
3. Some sort of protection or reset circuit (at least a fuse?).

I think GDTs are used most often on telecom lines, protection for test equipment (e.g., oscilloscope input), stuff like that.

IEC 60384-1 talks about situations for protection elements, namely what's required to achieve a certain insulation class (for example, reinforced can't have ground leakage, so you must use an MOV + GDT to absorb line-to-GND transients).

Tim