Why your charcoal ashes badly — and why the percentage isn’t the answer

Your supplier quoted an ash figure and met it. Your customers are still complaining that the bakhoor dies halfway through, or that there is grey dust on the table. Both things can be true at once, and this page is about why.

§1 · What you are actually complaining about

Ash percentage is a lab number. Ash behaviour is what your customer sees. They are not the same specification, and the second one is the one bakhoor cares about.

Every coconut charcoal supplier quotes an ash figure, because every laboratory report has a line for it. It is a mass: burn the cube, weigh what is left, express it as a proportion. That number tells you how much ash there will be and nothing whatever about what it does.

Not one supplier specification we have seen describes whether the ash stays compact or lifts, how quickly it builds on the top face of the cube, or whether it sets into a crust. Those are the properties your customer actually experiences, and a cube can hit a good figure on all of the first kind while failing on all of the second. That is not a supplier being dishonest. It is a specification written for a different burner.

§2 · The differential

Find the row that matches what you or your customer described. Only the first row is about quantity at all.

What you see	Probable cause	Buyer test
A lot of ash, but fine and white, and it stays where it forms	High ash content from clean feedstock — most often soil or sand picked up when the shell was collected, rather than anything in the carbonising	Ask for the ash figure and the method behind it. Then weigh three cubes, burn them, and weigh the ash on a kitchen scale: you now have your own figure to hold the quoted one against.
Ash that lifts off the cube and drifts	A low-density ash structure, which is a formulation property rather than a quantity	The still-air test in §6, which measures how far it travels rather than how much of it there is.
Dark grey, brown or black ash rather than white	Incomplete carbonisation, or feedstock that was not all coconut shell	The snap test on the cracking page — a dark core and dark ash are the same fault seen from two ends.
Ash forms a crust that chokes the burn	Ash behaviour, not ash content. A cube can produce very little of it and still seal itself over	Note when the bakhoor stops giving scent and compare that with when the cube actually goes out. A wide gap is the crust.
Gritty residue with visible particles	Sand or mineral contamination carried in with the shell	Rub the cold ash between your fingers. Ash is soft. Grit is not ash, and it did not come from the carbonising.

Rows two and four are the ones that will not show up in any specification you are sent, because neither has a number attached anywhere in the industry. They are also the two most likely to be behind a complaint about the bakhoor rather than about the coal, which is why they reach you as a fragrance problem.

§3 · The mabkhara geometry

This is the whole argument, and it is a matter of which way up the burner is.

In a hookah the coal sits on top, above the bowl, usually under a heat-management device. Ash forms on the upper face and either stays there harmlessly or falls away into a tray. It is moving away from the product the entire time. Ash quantity is therefore a housekeeping question: more ash means more cleaning.

On a mabkhara the arrangement is inverted. The bakhoor sits on the coal, or on a slip of foil directly above it, and the heat travels upward through whatever is in between. Ash forming on the top face of the cube is forming inside the heat path, between the coal and the product. It does not fall away, because there is nowhere for it to fall. It accumulates into the one place you need to keep clear.

Two consequences follow, and they pull in opposite directions. A thin, even ash layer is useful: it is the insulating buffer that moderates the coal's peak down toward the band the resin wants, which is the entire reason for the white-ash wait before you place a chip. A dense crust is the same material doing the opposite job — throttling the heat until the chip cools and stops giving scent while the cube is still visibly alight.

So low ash as a percentage is necessary and nowhere near sufficient. What a bakhoor line actually needs is ash that forms as a coherent thin layer and then either holds there or sheds cleanly — and no figure on any specification sheet distinguishes that from a cube that seals itself over.

§4 · What an ash figure is actually telling you

When a number does arrive, read the basis it was measured on before you compare it with anything. Laboratory figures are usually quoted dry: the sample is dried before it is burned, which is correct practice and makes results comparable between laboratories. The cube you receive has not been dried. It has crossed an ocean and sat in your warehouse, and charcoal takes moisture back out of the air it sits in.

The ash percentage itself does not move much for that — ash is what survives the burn either way. What moves is everything around it: a cube holding moisture ashes differently, builds its layer at a different rate, and behaves unlike the one the laboratory tested. So a figure quoted dry is not wrong, it is simply answering a narrower question than the one you have.

Our own ash figure is 2.3%, published with its basis, its method and its report number on the specification page and restated nowhere. It is also not what this page is about: everything above concerns how the ash behaves once it is there, and no proximate analysis measures that.

§5 · The question to ask your supplier

What is your ash figure, on what basis — dry or as-received — and can you describe how the ash behaves on an open burner rather than in a hookah?

The second half is the discriminating one, and it is worth asking even though you can predict the outcome. A supplier whose product line was built for shisha has no answer, because nobody has ever asked them: in their application the ash falls away and the question does not arise.

How to read what comes back

A figure with a stated basis. Good. Ask for it on the certificate of analysis for your lot rather than as a catalogue claim.
A figure with no basis. Ask which it is. Dry and as-received are not interchangeable and a supplier who does not distinguish them is quoting a number somebody else measured.
Any real answer to the second half. Rare, and it tells you they have burned their own product in the application you are selling into.
Silence on behaviour, or a restated percentage. The most common response. It is not evasion so much as a category error — and it means the cube was specified for a burner that works the other way up.

§6 · The still-air test — our proposed method

This one is ours. It has not been validated against an instrument and it is not an industry procedure. We publish it because ash drift is the one property in the differential with no number anywhere in the trade, and because it costs an evening.

Set a mabkhara on a dark, flat surface — a tray, a sheet of dark card — in a room with the doors shut and no fan or draught.
Burn one cube through a full session, as your customer would, with bakhoor on it.
Do not move the burner, and do not walk past it more than you must.
When the cube is out and cold, photograph the surface from directly above, with the burner still in place for scale.

What you are measuring is the radius of ash fall: how far from the burner the ash has travelled. A cube whose ash stays compact leaves a tight ring or nothing at all. A cube with a low-density ash structure leaves a wide, faint scatter, and that scatter is what lands on your customer's table and their furnishings.

Run it on the cube you buy now and on any candidate, in the same room on the same evening. As with every test on this site, the comparison is the result — the absolute radius means nothing on its own, and it will not transfer between your room and ours.

That drift tracks a low-density ash structure is our reasoning rather than a measured correlation. We have not burned a population of cubes against a laboratory density figure, and §9 says so.

§7 · What cannot be fixed

Grit is in the feedstock. If the residue is sandy, that sand was on the shell when it was collected, it went through the kiln, and it is now part of the cube. No storage regime, no handling change and no way of burning it will take it out. That is a supplier conversation about where and how shells are gathered, and if the answer is unsatisfactory it is a supplier decision rather than a batch claim.

Crusting and drift are formulation properties and equally beyond your reach — you cannot change how a finished cube ashes. What you can change is what you buy next, which is why §5 is the section that actually does something.

§8 · Questions buyers actually ask

My supplier's ash figure is low. Why are customers still complaining?

Because the figure and the complaint are about different properties. See §1 and §3: a low figure produced as a dense crust still sits between the coal and the bakhoor. Run the crust check in row four before you go back to them, so you have something specific to describe.

Is white ash always better than grey?

Colour is a reasonable signal, not a grade. Fine white ash generally indicates clean feedstock taken properly through carbonisation, and dark ash generally points at mixed material or an incomplete burn. But a white ash that drifts across the room is still a problem, which is the whole point of separating behaviour from appearance.

Can I just brush the ash off during the burn?

Your customer can, and many do. Whether they should have to is the question you are selling against — a cube that needs tending mid-session is a worse product than one that does not, and they will notice even if they never mention it.

The ash seems worse in summer. Am I imagining it?

Probably not, though the mechanism is moisture rather than ash. A cube that has taken up water from humid air burns differently and builds its layer at a different rate, as §4 describes. The ash did not change; the cube did.

Reviewed by Mohamed Noor, Director — 23 September 2026

§9 · What supports the differential above

Three of the claims below are ours rather than established findings, and the page's main argument is one of them. Here is the basis for each and the form we state it in.

Claim	What supports it	How we state it
Ash behaviour matters more than ash percentage for bakhoor.	The geometry in §3: on a mabkhara the ash forms inside the heat path instead of falling away. Reasoning about the application, not a measurement.	Our position, with the geometry shown so you can check it rather than take it.
Elevated ash comes from soil or sand in shell collection, from binder or filler, or from incomplete carbonisation.	Trade consensus across several accounts.	Stated as fact.
Fine white ash indicates clean feedstock and complete carbonisation; dark ash indicates mixed material or an incomplete burn.	Trade consensus. A signal rather than a grade, and §8 says so.	Stated as fact, with its limits.
A thin ash layer moderates the coal's peak toward the band the resin wants.	Standard practice plus physical inference, set out in full on the charcoal page.	Stated as fact, with the mechanism shown there.
Ash drift correlates with a low-density ash structure.	Our reasoning. We have not measured drift against any density figure.	Our reasoning, labelled as such in §6.
The still-air test measures drift usefully.	Our own reasoning. Not validated against an instrument, and untested at scale.	Our proposed method, labelled as such wherever it appears.
A dark core and dark ash are the same fault seen from two ends.	The carbonisation mechanism behind the snap test, which is documented trade practice rather than ours.	Stated as fact, and the test is credited to the trade on the cracking page.
Sand contamination cannot be corrected after carbonisation.	It is mineral matter through the body of the cube.	Stated as fact.

What we do not publish here: an ash figure. Ours will carry a test method and an issuing laboratory or it will not appear, and we do not hold it yet. Published sources disagree about what ceiling is even correct, which is the same reason the moisture figure is absent everywhere on this site.

§10 · If you want a cube to compare against

Drift and crusting only mean anything side by side. If the still-air test on your current cube leaves a wide scatter, the useful next step is to run the same evening again with a different cube on the same tray.

Request the sample, and see what arrives with it — then run §6 on ours and yours, same room, same night.
Why your charcoal cracks or sparks, if the ash is dark and you want the snap test that confirms the core.
What changes if you have it made under your own brand, for buyers who have diagnosed this and want the formulation changed rather than the batch replaced.