Ghost Hunter

A meter tells you how much. It does not tell you where, and where is the only question anybody standing in a dark room has ever actually wanted answered.

The tracking set answers it with arithmetic that is not clever and does not need to be. Your phone reports a compass heading many times a second. Near anything that disturbs a magnetic field, that heading stops being able to settle, and the scatter in it rises — this is the same measurement the field dispersion meter puts on its needle, and it is real.

What this instrument does with it is bin that scatter by the direction the phone was pointed in when it was measured. Thirty-six sectors, ten degrees each. Sweep the room and the sectors fill in behind you. When enough of the circle is in, the set takes the worst one and calls it a bearing.

Then it sends you there, and this is the part that is worth the page: the reading climbs as you approach. A magnetic dipole falls off roughly as the cube of distance, so the difference between pointing at a loudspeaker from across a room and standing in front of it is enormous. That steep curve is what makes a hunt possible. The set is not guessing that you are getting closer; it is watching the field it is sitting in get stronger.

What is measured and what is presentation

The Survey is fastidious about this line and this instrument has it drawn in three places.

• The bearing is measured. It is the sector with the highest real reading in it, and the set refuses to name one off fewer than five swept sectors or fewer than eight readings in the winner. A direction nobody swept cannot be locked onto, and a sector nobody pointed at is drawn on the rim as an empty tick rather than as a reading of zero — unmeasured and measured-quiet are different facts.
• The range is measured. It is the live reading referenced against what that same bearing read during the sweep. Walk away and it falls. There is no script and no timer anywhere in it.
• The words are presentation. COLD, COOL, WARM, HOT, BURNING are a five-step rendering of one number between nought and one, and the number is printed next to them. So is the hold before a contact is declared, and so is the idea that any of this is a hunt.

There is no random number anywhere in the signal path. A room with nothing in it never gets hot, the set says so, and the session is filed saying so. An instrument that finds something every time has found nothing.

What it does not do

Find ghosts. It does not find ghosts. Nothing on this bench does, nobody has ever demonstrated an instrument that does, and the Survey has been careful about this since 1974 and sees no reason to stop.

It is also not a metal detector, a stud finder or an EMF meter in the sense the trade uses that phrase. It reads one thing — the confusion in a compass — and reports it in one direction at a time.

Walk the room first, with the set off. Note where the wiring runs, where the appliances are, and where anybody has left a phone charging. Every one of those will take a bearing, and an operator who has not looked will spend the evening being walked into a fridge.

Throw the master switch and stand still. Eight seconds. The set is measuring what this room does when nothing is happening, and everything it reports afterwards is a comparison against that figure. It is four seconds shorter than the meter’s window, because a bearing is referenced against the set’s own sweep rather than against the floor — and it is a slightly worse floor for it. There is a recalibrate on the controls.

Then turn on the spot, slowly. Slowly is the operative word. The rose fills a sector when the instrument dwells in it, and a fast spin gives every sector one glance and the set will refuse to name a direction off that. A complete, deliberate turn takes about half a minute and is worth every second of it.

Follow the marker. The display is head-up: straight up on the glass is wherever you are pointed, and the marker slides round the rim as you turn. Bring it to the top, then walk. The range climbs and the click train speeds up as the distance closes.

Then look at what you are standing in front of. This is the job and it is the only part of it that matters. Nine times in ten there is an obvious cause and it is dull — a router on the other side of the wall, a doorbell transformer, a radiator pipe. Tell the register what it was. An operator who logs a contact without walking to it has logged nothing, and an operator who walks to it and does not say what they found has logged slightly less.

The tenth is why the Survey still exists.

If the set keeps revising the bearing

It does that when another sector genuinely overtakes the one it was on, and the commonest reason is that you have walked past the thing. Stop, turn on the spot again, and let it re-take the bearing from where you are now standing. A bearing is a line, not a position; two bearings taken from two places are a position, and that is how the Survey has always fixed a source.

If nothing ever gets hot

Then nothing in that room is disturbing a compass enough to find, which is a result and should be recorded as one. Try it again in a kitchen. If the set will not get above COOL anywhere in the building, recalibrate — a floor established while you were walking about is worth very little, and the set will have taken your own movement as the room’s normal state.

Head-up, not north-up

The bearing rose on the field dispersion meter is drawn north-up, because it is a map: you read it afterwards to work out what the room looks like, and a map that spins while you read it is useless.

This is not a map. It is a thing you hold in front of you and walk behind, and the only question it has to answer is which way to turn. So the ring is drawn head-up: straight up on the glass is wherever the phone is pointed right now, and the bearing slides round the rim as you turn. Centre the marker and walk. Nobody has to think about north — and on a device with no north reference nothing is lost, because a relative bearing is exactly as walkable as an absolute one. The set says which it has.

The ring

Thirty-six sectors of what has actually been measured. A sector that has been swept is drawn as a bar whose height is the highest reading taken in that direction. A sector nobody has pointed at is drawn as an empty tick on the rim, outside the plot. The display will not let you mistake an unsurveyed direction for a quiet one.

The marker and the gutter

The chevron outside the rim is the locked bearing. The lit arc around it is the tolerance — the set counts you as on the bearing anywhere inside it, because a hand is not a theodolite. The chevrons at the left or right edge of the glass say which way to turn when the marker is off screen behind you.

The core

The range, as light. Distance is not drawn here as a bar or a dial; it is how brightly the middle of the instrument is lit, which is the one thing on the face that is legible at arm’s length in a dark room. The number is printed in it for anybody who wants the figure.

The ring that sweeps round the core is the dwell. The set will not declare a contact off one frame at the top of the range — that is a knock against a table — so the range has to be held. The dwell drains rather than resetting, at a third of the rate it fills, so a cough two seconds in costs you a fraction of a second and you can see on the ring that it cost you almost nothing.

The click train

It follows the range rather than the raw reading, which is the one thing this set deliberately does differently from the meter. A tracking set that clicks faster when you point away from the thing you are tracking would be no use to anybody. While it is still sweeping there is no bearing yet, so it falls back to the raw reading and behaves exactly as the meter does. There is a mute on the panel.

Contacts

A contact is the set saying: you are on the bearing and the reading has been at the top of its range for long enough that this is not an accident. It is not a detection of anything. It is a statement that the field here is much stronger than it was where you swept from, which is the definition of having arrived.

Every contact is filed to your service record along with the bearing, the reading, the sectors swept and the room’s own floor, and the bearing rose goes with it — so a session card from this instrument is a genuine picture of the room you just walked around.

The Survey keeps a list, in descending order of how often each item turns out to be the answer, and it is filed with the meter rather than here because it belongs to the measurement rather than to the instrument. It starts with loudspeakers and ends with the magnetic clasp on somebody’s own notebook.

Read it before you write anything up: where excursions actually come from. The Survey’s files are more useful than most because the boring explanations were eliminated first, and an incident record that does not say what was ruled out is not a record of anything.

If you are new to this, the two pieces worth reading before your first evening are how to take a baseline and what a haunted house usually turns out to be. The second one leads on carbon monoxide, which is the honest ordering.

Ghost Detector Lab is an entertainment site operated as the public terminal of the Thinplace Survey. See the disclaimer, the privacy notice, and the method for every instrument on the bench.