Getting started
How to Read a Metal Detector Signal: Tone, Number and Repeatability
Quick answer
Judge a signal on three things: whether it repeats from two directions ninety degrees apart, whether the target ID holds steady or jumps, and whether the tone is sharp and symmetrical. A good non-ferrous target usually holds its number from both directions. Iron usually does not.
A detector gives you three streams of information at once: a tone, a number and a depth estimate. Beginners read the number, intermediate detectorists read the tone, and people who find things read the repeatability. That progression is the whole subject of this page. A single reading on a single sweep is the weakest evidence a machine offers, and treating it as a verdict is why so many holes come up empty.
None of this replaces knowing your own machine in your own soil, which is what a test garden is for. What follows is how to interpret what you hear once you have that baseline, and how to decide whether a signal earns the twenty seconds it takes to cut a proper plug.
Top pick
Gray Ghost Original Platinum Headphones
Aftermarket detector headphones are built around hearing faint deep signals rather than music, which is the distinction that matters when a target is at the edge of range.
Price was accurate when this page was compiled and changes often.
On this page
What are the three things a signal tells you?
Read them in this order. Repeatability first, then tone, then the number. Reversing that order is the most common reason for a disappointing dig rate.
A target that repeats cleanly from two directions with a steady ID and a sharp symmetrical tone is worth digging, and a signal that changes character when you turn ninety degrees is usually iron or two targets.
| What you hear or see | Usually means | What to do |
|---|---|---|
| Repeats identically from two directions | A single round non-ferrous target | Dig it |
| Good one way, iron or silent the other | Iron, or an irregular ferrous object | Usually pass, unless the site is worth digging iron |
| ID steady within a point or two | A clean target within range | Dig it |
| ID swinging widely between sweeps | Iron, corrosion, or two targets close together | Try a smaller coil or approach from another angle |
| Sharp symmetrical tone, short | A small shallow target | Quick recovery, small plug |
| Soft broad tone, low volume, steady ID | A deeper target, often the good one | Worth digging carefully and slowly |
| Loud broad tone with unstable ID | A large shallow iron object | Usually pass on a park, dig on a relic site |
| Double beep on one pass, single on another | Often a coin on edge, or two adjacent targets | Reposition and sweep again from several angles |
| High tone that will not repeat at all | Electrical interference or a hot rock | Re-run noise cancel, or check ground balance |
These are the patterns owners consistently report across machines. Your own machine in your own soil is the authority, which is what a test garden establishes.
Why does sweeping from two directions matter so much?
Because most iron is not round. A nail presents a completely different profile to the coil depending on which way the coil crosses it, so it commonly gives a respectable non-ferrous response along its length and an obviously ferrous one across it. A coin is round and presents the same profile from every direction, so it reads consistently. That asymmetry is the most reliable discriminator available on any machine, and it costs you two seconds.
The practical routine is simple: get a repeatable signal, step ninety degrees around it, and sweep again. If the tone and the number survive the move, cut a plug. If they collapse, you have saved yourself a hole. This test is more reliable than any discrimination setting, and unlike discrimination it does not silently remove targets you wanted.
How much should you trust the number?
As a band rather than an address. The master target ID chart sets out the conductivity ordering that holds on every machine, but the exact figure belongs to one machine, one mode and one patch of ground. Depth pulls IDs downwards and makes them noisier, corrosion shifts them unpredictably, and a target on edge can drop out of its band entirely.
The number is most useful for telling you what else lives in the band you just heard. A signal in the pull tab range might be a tab, a small gold ring or a piece of lead, and no machine separates those reliably, which is the honest trade discussed in the chart. Use the number to understand the odds, then let repeatability and tone decide.
What does tone tell you that the number does not?
Size and depth, roughly. A small shallow object gives a short sharp response as it passes quickly through the coil field. A deeper object gives a softer, broader response because it is in the field for longer and the signal is weaker. Machines with proportional audio, such as the Garrett AT Pro with its published proportional audio and tone roll, vary the volume with signal strength so that you hear depth directly rather than reading it off a meter.
This is why headphones matter more than most beginners expect. A faint deep coin at the edge of detection is a change in a whisper, and a speaker in wind will not deliver it. Wired sets such as the Garrett MS-2 Headphones cannot run out of battery, aftermarket sets such as the Gray Ghost Original Platinum Headphones are built around blocking ambient noise, and low latency wireless sets such as the Minelab ML105 Wireless Headphones avoid the delay that makes ordinary Bluetooth useless for pinpointing.
When is a good signal still not worth digging?
When the ground tells you to be suspicious. Bottle caps produce excellent coin numbers and are the reason experienced hunters check from several angles on picnic ground. Aluminium screw caps sit right on top of the cent and dime range. Large iron at depth can wrap around into a high non-ferrous reading, which is a well known behaviour rather than a fault. And on a site with heavy iron, two nails close together can imitate a single good target convincingly.
The other case is when you cannot dig it properly. A signal under a tree root, in a flower bed you were asked not to disturb, or at the edge of a manicured sports pitch is not worth the damage, whatever the number says. That judgement is part of the code of conduct, and it is the reason detectorists keep access. Recovery technique itself is covered in the hole filling guide.
Sources
- Garrett AT Pro product specifications, proportional audio and tone roll
- Minelab Equinox 700 and 900 instruction manual
- DetectorPro Gray Ghost product specifications
Frequently asked questions
Why does my detector give a good signal and then nothing when I dig?
Usually because the target moved into the plug you lifted, or because the signal was two pieces of iron imitating one good target. Check the plug itself with a pinpointer before widening the hole, then check the spoil. If neither contains anything and the signal is gone, it was almost certainly a false produced by nearby iron or by the ground itself.
Should I trust the depth meter?
Treat it as an estimate calibrated for a coin sized target. It works reasonably on coins and misleads on anything else, reading a large deep object as shallow and a tiny shallow object as deep. Use it to decide roughly how much to cut, not as a measurement. Realistic ranges by target and soil are set out in the depth chart on this site.
What does it mean when the target ID jumps around?
Most often iron, corrosion or two targets sitting close together. A clean non-ferrous target within range holds its number within a point or two across repeated sweeps. Wide swings mean the machine is getting inconsistent information, which on a trashy site usually means masking. A smaller coil frequently resolves it into two separate and identifiable signals.
Are headphones really necessary?
For deep targets, effectively yes. The faintest signals are small changes in volume that a speaker in open air will not deliver, especially in wind or traffic noise. Headphones also keep your audio to yourself in a busy park, which matters for how detectorists are perceived. Wired sets are cheap and cannot go flat, while low latency wireless sets avoid the delay of ordinary Bluetooth.
How slowly should I actually swing?
Slower than feels natural, and low and level throughout the arc. A slow overlapping sweep gives the machine several looks at a deep target while a fast sweep gives it one, and letting the coil rise at the ends of the arc loses depth exactly where you are covering new ground. If you are covering ground faster than you can listen, you are not detecting, you are walking.
Researched, not professional advice. This page is compiled from published specifications, manuals and owner-review consensus, not hands-on testing. Target ID numbers, depth figures and settings vary with soil chemistry, moisture, target size, depth, orientation and firmware version, so treat every figure here as a starting point and verify it against your own machine's manual and your own ground. As an Amazon Associate we earn from qualifying purchases.