Small SSTCs, RFI range & safety considerations / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Small SSTCs, RFI range & safety considerations

previous 1 2 3 next

Move Thread

LAN_403

flannelhead

Sat Jul 11 2015, 02:54PM

Registered Member #952 Joined: Mon Aug 13 2007, 11:07AM
Location: Finland
Posts: 388

Sigurthr wrote ...

You could always build a faraday cage and run the coil inside it. Size your mesh right and you'll have nothing to worry about.

That's definitely worth considering, thanks for the reminder!

Sigurthr wrote ...

2) mains ground may be high impedance to RF, and thus poorly conduct to earth. Or it could be high impedance to even DC, and be a really poor earth anyway.

So, as I understand the matter, the high impedance to RF will cause some potential differences in the mains ground conductor, right? If I add an EMI filter on the side of the coil like in this schematic (so there's a common mode choke in addition to the class Y/X capacitors which are from an ATX PSU) and connect the secondary base to the mains ground, the RF current from the base should get equalized on all the three conductors. As a consequence, nearby appliances don't "see" the current as all the three potentials are shifted equally by the RF current.

Now, let's say this filtering happens to be enough so that I don't notice any issues with my appliances (remember, I'm not able to install class Y capacitors anywhere in the house wiring). As the RF current travels along the three conductors, could the possible high RF impedance and differences between the conductors cause some phase shifts / potential differences between the conductors so that somewhere else along the line (like another apartment) there could suddenly be much more interference due to the current than in my apartment? In other words, am I always the first to notice if there's an RFI problem due to the grounding or could it happen somewhere else (another apartment) without me noticing anything? My intuition says that at least the interference wouldn't get worse further away from the coil. I've yet to properly check that the mains ground actually connects to mother Earth, but I'm fairly confident it does. At the very least my DMM shows quite a small DC resistance (< 1 ohm) between PE and the cold water tubes. I'll do the check with a coathanger and a wire anyway! The distance to earth is quite long, though â€“ I'm living on the 7th floor.

Sorry for all the (probably simple) questions â€“ but I'm really trying to understand what I am about to do before actually doing anything (possibly stupid). While I'm quite familiar with general electronics and safety (enough to safely build and run something like the mini SSTC at relatively low powers), the RF and EMI stuff are something I've really not that much experience with. I'm studying physics at the university for the third year so I've got a grasp about electricity, but the radio frequencies seem more like a special niche so there hasn't been much of that on the courses.

Sigurthr

Sun Jul 12 2015, 02:49AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Just a note: RFI filters at the coil are useless; the capacitive coupling of the EM field to the environment causes orders of magnitude more EMI than what travels through the power cord to the wall outlet (unless you're operating the coil in the center of an airport hangar, lol!). That's why you have to make equalization nodes at the wall outlets. I just made little RF equalizer boxes that plug into the outlets and have the caps and wiring inside them, then placed them around the house.

I'm not sure if the impedance mismatches will cause maxima nodes at different points down the lines relative to each conductor, but it seems reasonable to assume so if the wiring's conductors are of uneven inductances with respect to each other. Maxima nodes will form on the mains wiring even with equalization throughout, but unless two points connect to a singular device, no device will see the potential difference.

Radiated EMI should be low, since most TC wavelengths are still much longer than wiring length. A HF TC could approach the 1/4 wave wire length in a distant circuit though, but it's unlikely.

flannelhead

Sun Jul 12 2015, 02:46PM

Registered Member #952 Joined: Mon Aug 13 2007, 11:07AM
Location: Finland
Posts: 388

Thanks Sigurth. I really appreciate the time you've taken to answer my questions. Thanks to you, now I think I understand the EMI/RFI problem at least to some degree.

It seems my best bet really is to build a Faraday cage to reduce the interference, as I don't even accidentaly want to cause harm to my appliances, let alone my neighbours'. You said the mesh size should be chosen right. As I understand it, the diameter of the apertures should be smaller than the smallest wavelength I want to block (up to some harmonic, probably? Which one?). Are there any other things I'd have to take in account? Also, the cage and the secondary should be both connected together AND to the mains ground to prevent them from climbing up to some high potential, right?

I couldn't find specific guidelines for designing a Faraday cage for a TC (despite googling), but I'm thinking of something like shown in this thread. I assume leaving the top open for convenience doesn't pose too much of a problem as it seemed to work very well for WaveRider's coil.

Dr. Slack

Sun Jul 12 2015, 04:55PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

Building a full Farady cage is rather a big undertaking. You may not need to. If you do build one, and get other things wrong, you can still have a problem.

Step by step ...

You must connect all the mains grounds together, and to the secondary ground. It doesn't matter whether this is connected to soil at ground level, this is irrelveant, even though it will be, it matters for lightning and for system-wide fault currents, it doesn't matter for TCs. The point is that all metal buried in the walls, floors, ceilings should be connected to coil ground. This reduces EMI carried on one bit of grounded cable with reference to another bit.

You must have a RFI filter at your coil supply. It is a very good idea to 'class Y' all of your mains sockets. As a tenant you can't do this from behind the sockets, but you can put caps in plug tops, and plug them in. This reduces EMI carried between the wires and ground.

It's an improvement to have an extended ground plane under the coil (counterpoise), which must be connected to the secondary ground. This intercepts many of the electric field lines coming from the coil, reducing its far field radiation efficiency (ability to upset the neighbours).

At any point, if this reduces emissions to a satisfactory level, you can stop there. A TC is a very inefficient radiator of RF to start with, if you wanted to design an antenna for these frequencies, it would not be the size and shape of a TC. There's no point in over-engineering a solution, when it's good enough, stop. Are you able to ask your neighbours 'does it interfere with anything *now*?'

If it still does, then you could extend the sides of the counterpoise up the walls with chicken-wire, and continue closing a Faraday cage. This is where the difference between a counterpoise, and a partial Faraday screen begins to become one of degree. As you go up the walls, you intercept more and more of the field, until with a fully closed screen, it is all stopped.

As a rule of thumb, as long as the dimensions are << wavelength/10 (which they are for room dimensions and TC frequencies), a circular hole of diameter d in the screen that's more than 3d from the radiator is going to leak damn-all. That said, the same goes for a slot of length s. So if you have a door of height 2m, width 800mm, and you don't run conductors across to maintain continuity, the critical dimension is the 2m, not the 800mm. There is little point putting mesh on your door, even if it is all connected at the hinge side, unless you can make multiple connections across the opening side when you operate the coil, croc clips would be fine.

flannelhead

Sun Jul 12 2015, 08:28PM

Registered Member #952 Joined: Mon Aug 13 2007, 11:07AM
Location: Finland
Posts: 388

Instead of turning a whole room into a Faraday cage, I was thinking of building a smaller, partial cylindrical Faraday screen of diameter ~40cm with walls and bottom made of mesh. I'd make it just large enough to accomodate the coil (see the link in my previous post). The walls of the cage would extend some tens of centimeters above the topload, with the top left open. This seems less of an effort than making a large enough local ground plane, and would intercept a good amount of the electric field lines as well (to my understanding).

Ash Small

Sun Jul 12 2015, 08:44PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

As a last resort you 'could' hire a small generator and take it somewhere else to run it, if you can't take it somewhere with mains power?

flannelhead

Mon Jul 13 2015, 12:49PM

Registered Member #952 Joined: Mon Aug 13 2007, 11:07AM
Location: Finland
Posts: 388

A generator would be a good last resort, as you said, but I've got no means of transporting it as my only vehicle is a bicycle. And of course I'd prefer a more permanent solution.

I created a schematic illustration of what I'm thinking of doing based on this discussion. Does this make sense in terms of reducing the coupling of RFI via the EM field to the mains wiring? A small partial Faraday cage with the top open would be quite easy to construct and probably would also "complete" much of the secondary LC circuit for RF and thus reduce emission by a good amount. There would still be some interference coupling to the power cord, but that would be filtered by class Y capacitors outside.

I don't see this as over engineering but rather just playing safe (is there a difference, though?) At the same time this would be a nice experiment.

Sigurthr

Mon Jul 13 2015, 10:54PM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Looks fine to me, but have a removable top mesh you can easily place and remove. No reason not to include that, really. Given that the cage is so close to the secondary it will significantly load the coil and reduce visible output, bear that in mind.

flannelhead

Tue Aug 11 2015, 12:06PM

Registered Member #952 Joined: Mon Aug 13 2007, 11:07AM
Location: Finland
Posts: 388

The SSTC is now working

Thanks for helping me understand the RFI issues.

I decided to test the coil without the Faraday cage at first. I have 100nF class X/Y2 capacitors coupling the three mains wires together on the extension cord to which the coil is also connected. The coil works very well and doesn't seem to cause any interference to my electrical appliances. The only sign of interference was when I had an AVR in-circuit programmer connected to my laptop and to the ATtiny based interrupter. When the coil was on, the ATtiny would repeatedly reset itself. I suppose the programmer cable was picking up interference. That was about half a metre from the coil and the interrupter is unshielded.

So radiated interference doesn't seem to be much of a problem. I'm still a bit confused about interference further down the line caused by possible high impedance of the mains wiring. As I happen to live in a 7th floor apartment, the RF current has a long way to go to the ground, and I'm puzzled if it could cause some havoc in the lower floor apartments.

I measured the DC resistance between the mains ground and some different pieces of water pipe going through my apartment. The resistance was near zero (shown by a multimeter). Shouldn't that be a good sign of the mains ground having a good connection to earth?

Sigurthr

Mon Aug 17 2015, 12:26AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

That means there's a good DC connection from mains earth to ground, but not any indication of a low impedance RF connection. Sounds like what you have going should be ok though. If you get any complaints, then you know you've got to change something.

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob