"Invisible fence" for dogs

On 14/03/16 16:54, snipped-for-privacy@yahoo.com wrote: [...]

The field of a single wire drops off as 1/d. The total field is the sum of several of those, so while it doesn't drop off as fast inside as outside the perimeter, it's certainly not constant.

Jeroen Belleman

I don't know? I was hoping for feedback from someone smarter than me! I think there is an arrangement like this that will cancel the signal.

Mikek

Did you ground the cylinders, or just let them float electrically? (with transmitter sending wrt the same "ground".)

George H.

Maybe it's such a bad idea no one can figure out how to tell me it won't work. Anyone have an idea about this? Or a transformer drive cancelling circuit.

Mikek

No, because it's conductive, ferrite won't show the same current, copper will.

Woah! did you just repeal kirchoffs node law?

bullshit.

--
  \_(?)_

Not going to work consider Kirchoffs node law.

--
  \_(?)_

true, you only need to double any perimeter that has an opening in other words if you want an opening in the perimeter you need to double all of it.

You're making shit up.

one option that might work is to split the current many ways until it's too diluted to trigger the collar,

.----------------. | | | | | D | | | `-e-e-e-e-e-e-e | ............. | . . | . . | . . | ............. | e-e-e-e-e-e-e-. | .---------------' | | | | `----------------' D driver e earth peg ... house

seems risky though, if some plumbing picks up most of the current that could cause a peak in the signal.

--
  \_(?)_

(A) [ edit (A),(B),(C) and (D) added ]

where is the safe path from (A) to (B) ? (A) to (C) is possible via the "passage" or via (D) but not (A) to (B)

--
  \_(?)_

closed

That's what happens in coax cable, NOT with a length of iron pipe that hasn 't any current carrying contacts at its endpoints.

pipe

will.

This is about a steel pipe, not ferrite; it does NOT have high frequency ma gnetization response, there will be a small depth (skin depth) on the inside of the pipe that gets magnetized, and no induced current on the outer surface of the pipe. Laminations for 60 Hz are 1 mm, pipe walls are maybe

3mm thick, and the frequency of interest was described as kilohertz-to-mega hertz.

There's no dB/dt near the outer surface of the pipe, so no EMF there.

The pipes were floating. A single-point ground wouldn't make any significant difference, but ground both ends and you'll open up a whole can of worms to keep this group busy for months. ;-)

Jeroen Belleman

Jasen is right, this is exactly what happens: The current on the inner surface of the pipe (almost) cancels the current on the wire, and the circuit is completed by that current flowing back over the outside surface of the tube. Nett result: The B-field outside the tube is the same as if the tube hadn't been there.

My measurement confirms that well enough.

Jeroen Belleman

How about a sleeve made with lossy Ferrite beads (large), toroids or tubes. Mikek

a n

el

I think the double back concept is this...

you make a loop around the perimeter, at the end of the loop, you turn back and make a second concentric loop, so you have two concentric loops spaced apart by say 10 feet. The dog cannot cross the inner loop, the outer loop is far enough away so it does not cancel the inner loop.

Now you should see where this is going. At the location where you want to allow passage, you simply bring the two loops together, so they do cancel.

Two concetric loops, current flowing in opposite directions, located apart where you want the fence to work and located closly where you want an openi ng.

Mark

I believe the MAGNETIC field INSIDE a loop is essentially constant in the plane of the loop.

Yes H drops as 1/r outside the loop or relative to a long straight wire, but INSIDE a loop, as you move away from the perimeter, you are moving closer to all the rest of it.

They use two loops to make a volume that has a constant H field in 3 dimensions.

I think the dog fence cannot be based on MAGNETIC fields alone but rather uses EM fields.

If it was based on MAGNETIC fields alone, the entire area inside the loop would active the collar.

Mark

ng a

ion

ncel

at

ck and make a second concentric loop, so you have two concentric loops spac ed apart by say 10 feet. The dog cannot cross the inner loop, the outer lo op is far enough away so it does not cancel the inner loop.

o allow passage, you simply bring the two loops together, so they do cancel .

t where you want the fence to work and located closly where you want an ope ning.

That sounds potentially wasteful. He can use an exterior SPST switch /withi n/ the fenced area to break the loop open. The lead-in to the switch is a t wisted pair so he can walk the animal up to it on a leash. And you have a 1

00% protected perimeter when the switch is turned on.

You already said that. My measurement contradicts your belief and so does the theory. (Which I also already said...)

Surely you have no trouble accepting that the H field would be much stronger very close to one wire rather than near the centre of the loop? This 'constant field inside a loop' is only an approximation, and a pretty rough one at that, valid only for certain geometries and if you're not too picky.

Jeroen Belleman

Hmm I'd think a single point ground would make a big difference.. maybe I'll have to do my own experiment... I've got this image that once you are thicker than the skin depth the inside and outside of a metal shield are decoupled. With separate current flowing around on each surface. I don't see how grounding on both ends would make much difference. (Assuming the length of the shield is a lot less than the wavelength of the AC signal on the wire.)

George H.

[...]

On thinking this over, I still don't think a single-point ground on the tube would make any difference. The signal voltage on the wire is very small because the impedance of the loop is so low. Nothing changes if there is no current flowing through the ground wire.

On the other hand, I admit that you are likely correct that grounding both ends won't make much of a difference as compared to a single-point ground. The major part of the current flowing over the outside surface of the tube will keep doing so because that's by far the lowest impedance path available.

Now, where did I put that miniature dog fence...

Jeroen Belleman

Woff woff,

You might be right. I've had very little success at trying to shield magnetic fields. Whereas electrostatic fields are easy. I did this hand-wavy skin depth argument, but to be honest I'm not sure it applies when you have a near field situation as is the case here.

George H.