Almost everything written about cleaning a diesel particulate filter treats soot and ash as the same problem. They are not, and every decision you make about the filter follows from the difference.
Start here: a regeneration is not a clean. It burns the soot and leaves the ash behind, and the ash is what eventually takes the filter out of service.
A DPF fills with two different things.
Soot is carbon — unburnt fuel. It is combustible. That is the entire basis of regeneration: get the substrate to around 600 °C, either passively over a catalyst or actively by injecting fuel late, and the carbon turns to gas and leaves up the pipe.
Ash is what remains once everything combustible has gone. It comes mostly from engine oil — the calcium, zinc, magnesium and phosphorus in the additive package — plus wear metals and whatever the fuel carries. It does not burn. There is no regeneration temperature that removes ash, because anything hot enough to affect it has already destroyed the substrate.
So from the day the filter is fitted, ash accumulates and never leaves on its own.
Ash takes up room permanently, so there is less space left for soot. Less soot capacity means the filter hits its regeneration trigger sooner, so regens come closer together. Each regen burns fuel and washes some unburnt diesel past the rings into the sump.
Eventually the filter is ash-limited: back-pressure is still high immediately after a completed regeneration. At that point the restriction is not soot, and burning it again changes nothing. No amount of forced regen will fix it.
On a fleet this shows as a pattern before it shows as a breakdown — regen intervals shortening across a group of units of similar age and hours, then a derate.
Ash accumulates in every filter. What varies is how fast, and it comes down to two things: whether the exhaust gets hot enough to regenerate, and how much oil is reaching the filter.
Ash is burnt oil additive. So anything that puts more oil past the rings puts more ash in the filter.
Four checks, in the order worth doing them.
The last one catches people out. A filter that has been through an uncontrolled regeneration — or a badly run bake oven — can look fine from the face and be cracked inside.
Filter into a cabinet, air pulsed back through the channels, usually with the flow reversed. Fast, dry, cheap equipment.
It moves loose soot. It does very little to ash, because packed ash is not loose — it binds against the channel walls and towards the closed ends. Air does nothing about oil either.
Held in a controlled oven long enough to oxidise soot and hydrocarbons, then usually blown out afterwards. It handles oil-soaked filters that air alone cannot.
It still leaves every gram of ash, for the reason in section 1. And it carries substrate risk: cordierite has a lower service temperature than silicon carbide, and an unevenly loaded filter can develop a local hot spot during the bake. A cracked core is the usual failure.
Water, usually with a detergent, pumped or pulsed through the channels. Better than air on ash, because at least the ash is wetted and some of it is carried out.
The limit is that flow follows the path of least resistance. Water takes the channels that are already open and leaves the blocked ones blocked — which is the opposite of what you need. Drying matters too: a filter returned damp to a hot exhaust is a thermal shock event.
The filter is fully submerged and cleaned in two stages, because the two contaminants need opposite chemistry. Section 7 walks through it.
Because the filter is under the liquid, every internal surface is in contact with the bath — including channels that are completely blocked. Nothing has to find a path through first.
| Method | Soot | Oil | Ash | Substrate risk | Evidence it gives you |
|---|---|---|---|---|---|
| Compressed air | Loose only | No | No | Low | Visual |
| Thermal bake | Yes | Yes | No | Cracking, hot spots | Visual, sometimes weight |
| Hydro flush | Yes | Partly | Open channels only | Thermal shock if damp | Weight |
| Two-bath ultrasonic | Yes | Yes | Yes, dissolved | Low | Flow number and weight |
Read a dozen DPF cleaning websites and three things are consistent.
Those three gaps are the same gap. If you do not separate soot from ash you cannot pick the right chemistry, and if you do not measure flow you cannot tell whether it worked.
One of those claims is worth pulling out. Non-caustic is not automatically a virtue. Alkalinity is what lifts carbon and oil. Ash is mineral and comes off on the acid side. A process built around a single "safe" solution is doing one of those two jobs, and the filter tells you which one was skipped the next time the back-pressure is read.
This is the sequence, and the reason for each step.
Record the part number, the weight, and the condition of both faces. The weight is half of your evidence later. Anything cracked or melted stops here.
On the bench, at a fixed test point. Write the number down. Without this, nothing you do afterwards can be proven.
Heated, with ultrasonics running. This stage takes the soot, the unburnt hydrocarbons and the oil. Oil has to go first because it binds the ash into a paste, and ash that is glued in place will not respond to the second bath until the binder is gone.
Carry-over between an alkaline bath and an acid bath neutralises both. Rinse properly or you are paying for chemistry that cancels itself out.
This is the stage the other methods do not have, and it is the one that matters. RegeniK is the chemistry for it: it dissolves the calcium ash, the zinc phosphate and the magnesium oxides that the oil additive package leaves behind.
The distinction is worth being precise about. Air, water and heat try to dislodge ash and move it somewhere else. Acid-side chemistry dissolves it, so it leaves as part of the solution rather than as a particle that has to be pushed through a blocked channel. That is why the method works on the channels that are worst, which is exactly where the others fail.
Dry thoroughly. A filter returned damp to a hot exhaust is a thermal shock event and you can crack a core you just spent an hour saving.
Same bench, same test point. Two numbers, before and after, plus the grams of ash that came out. That is the paperwork that goes back with the filter.
A cleaned filter with no number against it is an opinion. Three measurements make it a result.
If a filter will not come back to an acceptable flow, that is also a result. It means the core is finished, and finding that out in the workshop beats finding it out on the road.
Every DPF decision ends up as a comparison between cleaning and replacing. Here is what is publicly quoted, with the source attached, because a lot of the figures circulating in this trade have no source at all.
| What | Quoted range | Where the figure comes from |
|---|---|---|
| DPF replacement, Australian light vehicle | $300 – $1,000+, median about $500 | Australian trade-pricing guide |
| Professional clean and refit, Australian light vehicle | $350 – $500+ | same source |
| Dealer replacement, diesel pickup (Ford 6.7L, Duramax L5P, Ram 6.7L Cummins) | USD $2,400 – $6,000 | US model-by-model breakdown |
| Professional clean, diesel pickup | USD $600 – $1,500 | same source |
| Heavy vocational and mining plant | No credible published figure | — |
That last row is deliberate. Replacement filters for heavy vocational engines and mining plant are not sold at published prices, and the numbers repeated online for them do not trace back to anything. We are not going to make one up.
The number that matters is on your own invoice. What you paid for the last send-out, and what you paid for the last replacement filter. Those two figures, times the filters your fleet goes through in a year, is the entire business case — and they are figures you already have.
Most fleets start by sending filters away. Four lines decide whether to stop, and the price of the machine is not the first of them.
The case is usually made by the third line, not the first. A workshop that can clean a filter overnight does not need the float of spares a two-week turnaround demands, and on a remote site that float is real money sitting on a shelf.
Where the numbers do not support it, they do not support it. A workshop doing a handful of filters a year should keep sending them out, and we will say so.
A DPF is awkward to clean: heavy, awkward to handle, and it has to be fully submerged with the bath free to move through it.
The EvoSonic EVS range is where this work usually lands. The two-bath DPF case study documents a set run end to end, with a flow number against every filter.
Tell us the fleet, the filters a year, and what you currently pay to send them out. Technical sales will tell you whether cleaning them in your own workshop stacks up at that volume, and what it takes to do it properly.