Tutorial — guides
How to Test Gold at Home: Five Methods, Ranked by What They Prove
By NorwegianSpark Editorial — written with AI assistance and reviewed by the NorwegianSpark SA editorial team | Last updated: August 2026
Before You Start: What You Are Actually Testing For
There are three different questions people mean when they say they want to test gold, and the right method depends on which one you are asking.
**Is this gold at all, or is it base metal?** The easiest question. Magnet plus a surface acid test will answer it in most cases.
**What purity is it?** Harder. Acid testing gives a rough karat band, and XRF gives an actual composition. A hallmark is evidence but not proof, as our [international hallmarks guide](/journal/international-hallmarks-guide) explains.
**Is this solid gold all the way through, or is it a shell over something else?** The hardest, and the one that matters most for bullion. No surface test answers it. Density gets you close for most fakes and fails specifically against tungsten.
Be clear which question you are asking before you pick a method, because the commonest error is running a test that answers question one and treating the result as an answer to question three.
And the most useful preventive measure is not a test at all: buy from a reputable dealer, in recognised forms, with assay packaging intact. [Silver Gold Bull](/go/silvergoldbull-gold) and comparable dealers stock recognised-refiner product precisely because provenance is cheaper than verification.
Method 1: The Magnet Test
**What you need.** A strong neodymium magnet. A fridge magnet is too weak to be informative.
**Steps.**
1. Place the item on a non-magnetic surface such as a wooden table. 2. Bring the magnet slowly toward it from a few centimetres away. 3. Watch for attraction. Then try to lift the item with the magnet. 4. For a chain or a bracelet, test along its length — clasps and jump rings are sometimes a different metal from the body.
**What a result means.** Strong attraction means significant iron, nickel or steel content, and the item is not solid gold. That is conclusive.
No attraction means only that there is no ferrous metal. Brass, copper, lead and tungsten all pass. So does a gold-plated brass ring.
**One caveat.** Some clasps and spring mechanisms in genuine gold jewellery contain steel springs, so a faint attraction at the clasp of an otherwise genuine piece is not necessarily a problem. Test the body.
**Time: under a minute. Cost: negligible. Proves: only a negative.**
Method 2: The Density Test
This is the most informative test you can do at home without chemicals, and it works because gold is unusually dense.
**What you need.** A scale accurate to 0.01g, a container of water, a stand or thread to suspend the item, and a calculator.
**Steps.**
1. Weigh the item dry. Record this as **A**. 2. Fill a container with water and place it on the scale. Tare the scale to zero. 3. Suspend the item fully submerged in the water on a thin thread, without letting it touch the bottom or sides. Record the reading. This is **B**, the weight of the water displaced, which in grams equals the item's volume in cubic centimetres. 4. Calculate density: **A ÷ B**.
**Interpreting it.** Pure gold is about 19.32 g/cm³. 22k lands near 17.7–17.8, 18k near 15.2–15.9, 14k near 12.9–14.6, and 9k near 11.3. The ranges overlap because alloy composition varies — white gold and rose gold of the same karat have different densities.
Brass and bronze land near 8.5, lead near 11.3, silver near 10.5. Any of those is far enough from gold to be obvious.
**Where it fails.** Tungsten at 19.25 is within measurement noise of gold. Density cannot distinguish them. It also gives misleading results on anything hollow, anything with stones set in it, or anything with a clasp or filling of a different metal.
**Time: ten minutes. Cost: the price of a scale. Proves: a great deal, except against the one counterfeit that matters most for bullion.**
Method 3: The Acid Test
⚠️ **These are corrosive acids. Wear eye protection and nitrile gloves, work on a non-reactive surface in a ventilated space, keep them away from children and pets, and never add water to acid. If any of that is impractical where you are, use a jeweller with XRF instead — the test is free or nearly free in most shops.**
**What you need.** A gold testing kit containing graded acid solutions for different karats, and a black testing stone.
**Steps.**
1. Choose an inconspicuous spot — the inside of a ring shank, the back of a clasp. 2. Rub the item firmly on the stone to leave a visible metal streak a couple of centimetres long. 3. Apply a drop of the acid for the karat you expect — start with the lowest and work up. 4. Watch the streak for about twenty to thirty seconds. 5. Repeat with the next acid up if the streak survives.
**Reading it.** If the streak **dissolves quickly**, the metal is below that karat. If it **fades or partially dissolves**, it is near that karat. If it **stays put**, the metal is at or above that karat, so move to the next acid to bracket it.
A streak that vanishes instantly under even the lowest-karat acid usually indicates base metal or plating.
**Where it fails.** It is a surface test. A gold-plated or gold-filled item with a thick layer will read as solid gold, because the streak you made is all plating. Scratch deeply enough to reach the core if you are testing for a shell — and accept that this leaves a real mark.
**Time: five minutes. Cost: modest. Proves: purity at the surface only.**
Method 4: The Ceramic and Visual Checks
Two quick supplementary checks worth knowing.
**The ceramic scratch.** Drag the item across an unglazed ceramic surface — the unglazed foot ring on the underside of a plate works. Gold leaves a **gold-coloured streak**. Pyrite, brass and most gold-coloured base metals leave a **black or dark grey streak**. It is quick and it will mark the item, so use an inconspicuous edge.
**Visual inspection under magnification.** With a 10x loupe, look at:
- **Wear points** — the underside of a ring shank, the edges of a clasp, the high points of engraving. Plating wears through here first, and a different colour beneath is decisive. - **The hallmark itself.** A struck mark displaces metal and has a slight raised lip. Cast or laser fakes look flat. - **Colour consistency** across the whole piece. Plated items often show slight colour variation where the layer is thinner. - **Seams and joints.** A hollow item has them; a solid one may not.
**The one test to skip: the bite test.** Gold is soft and pure gold will take a tooth mark, but so will lead, and you will damage both the item and possibly a tooth. It proves nothing useful.
**Time: two minutes. Cost: a loupe. Proves: often more than the chemical tests, in experienced hands.**
Method 5: Professional Testing, and When to Stop Guessing
For anything of real value, stop testing at home and pay someone.
**XRF** reads elemental composition without damaging the item, in seconds, and gives you actual percentages rather than a band. Most jewellers, pawnbrokers and bullion dealers have one. Its limitation is that it reads only a shallow surface layer.
**Ultrasonic thickness testing** measures how sound travels through a bar and detects a core of different material. This is the test that finds tungsten, and it is what serious dealers use on large bars.
**Electrical conductivity testers** work on a similar principle and are increasingly common in dealer hands.
**Specific gravity measurement on laboratory equipment** does the density test properly, with precision a kitchen scale cannot reach.
**Assay** — physically sampling and chemically analysing the metal — is definitive and destructive, and is what refiners do when they buy.
**A sensible sequence.** For jewellery: magnet, then loupe, then ask a jeweller for XRF. For bullion: buy from a reputable dealer in assay packaging and do not open it, because an intact sealed card from a recognised refiner is worth more than any test you can run, and breaking the seal costs you resale value. Our [counterfeit bullion guide](/journal/counterfeit-bullion-how-fakes-are-detected) covers the detection methods dealers actually use, and the [purity guide](/journal/understanding-gold-purity) covers what the numbers mean once you have them.
Related Collections
Frequently Asked Questions
Is the magnet test reliable?
It is useful as a first filter and it proves almost nothing on its own. Gold is not magnetic, so a piece that jumps to a magnet contains a significant ferrous component and is not solid gold. But the metals most commonly used to counterfeit gold — tungsten, lead, copper and brass — are not magnetic either, so passing the magnet test is entirely consistent with being a fake. Treat a fail as conclusive and a pass as meaningless.
Why is tungsten the dangerous counterfeit?
Because its density is almost identical to gold. Gold is about 19.32 g/cm³ and tungsten about 19.25, a difference too small for a household density test to resolve reliably. A tungsten core with a genuine gold shell passes magnet, density and surface acid tests. Detecting it requires either ultrasound, electrical conductivity measurement, or drilling — which is why bars from recognised refiners in sealed assay packaging matter so much.
Will acid testing damage my jewellery?
It leaves a small mark. The standard method scratches the item on a testing stone, so the acid touches the scratch rather than the piece, but you have still abraded the surface. Do it on an inconspicuous area such as the inside of a shank or the back of a clasp. For anything valuable, antique or set with stones, use a jeweller with XRF instead — it is non-destructive and takes seconds.
What is XRF and where can I get it done?
X-ray fluorescence measures the elemental composition of a surface without damaging it, and it reads out the actual percentages of gold, silver, copper and other metals. Most established jewellers, pawnbrokers and bullion dealers have a unit, and many will test a piece for free or for a small fee. Its one limitation is that it reads the surface only, so a thick plating or a filled core can defeat it — which is why it is often paired with a density check.