What actually kills a UV printhead

A printhead is the most expensive replaceable part in a UV machine, and almost every way it dies is preventable. Here are the seven, in roughly the order they actually happen — starting with the one nobody in this market talks about.

1. Stray UV light, the failure mode unique to UV machines

A solvent head dries out. A UV head gets cured — and it can be cured by its own lamp, reflected back off the job.

Roland, verbatim: “ink affixed on the surface of the print heads might be cured by the reflection of the UV-LED lamp light and by ink mist… Continued use of the product in this state may cause a failure that requires the replacement of the print heads.”

And the substrate warning that follows from it: materials “that are likely to reflect the UV-LED lamp light, such as mirrors and stainless steel… cannot be used as the object to be printed on.”

Mimaki names the mechanism: “Stray light is a phenomenon in which UV light intended to irradiate the ink on the media is reflected and scattered, causing it to hit unintended areas such as the head surface.” And on letting it run: “If deterioration of the head due to reflected light is neglected for a long period of time, it may cause malfunction to the point where the head may have to be replaced.”

Jig and substrate rules that follow

• Jigs should be black and matte. Never white, silver or bare metal.
• Transparent media needs thought. UV passes straight through clear acrylic or glass, hits whatever is under it, and comes back up at the nozzle plate. A black matte underlay is not decoration.
• Polished metal and mirror are the hard cases. If you print them, you are managing reflection deliberately — see printing on metal and printing on glass.

The loop worth understanding

A warped board arrives. The operator raises the head gap to avoid a strike. A bigger gap means the drops fly further, so more of them turn into airborne mist. That mist lands on the nozzle plate. The lamp reflects off the job and cures it there.

The raised gap, adopted to protect the head, is what kills it. The real fix is flatter stock and a correct gap, not a bigger one.

2. The nozzle plate coating, and why wiping it wrong is fatal

Most modern heads have a non-wetting coating on the nozzle plate. Xaar: “a special coating, called a non-wetting coating (NWC)… Contact with this surface must be avoided.” Fujifilm brands theirs REDIJET. Ricoh markets a head on the strength of an “enhanced non-wet coating” — a manufacturer telling you plainly that the coating is a wear item.

Here is the nuance every maintenance blog gets wrong: not all nozzle plates are non-wetting, and the two types need opposite treatment. Xaar’s XJ128 and 200 Plus use a wetting plate where “a thin film of ink is required on the surface of the nozzle plate at all times” and you must “ensure that the nozzle plate surface is not wiped dry.” Wiping that head the way you would wipe a coated one damages it.

Four rules that hold across the coated heads:

• Never wipe dry. “Normally a purge is performed prior to wiping” — the ink is the lubricant.
• One direction only. “Wiping in both directions would re-deposit this debris.”
• A soft blade, not a hard one.
• Non-shedding material. Which brings us to cotton buds — covered in the nozzle blockage article, because they are a common and entirely avoidable way to destroy a head.

3. Particles, and the only filtration figures anyone publishes

Xaar publishes the numbers: the printhead’s own integral filter is a 15 µm stainless mesh, the primary ink filter “should be a 5 µm replaceable filter”, and — the part that matters — “the printhead filter cannot be replaced or cleaned.”

For scale, pigment particles run around 50–200 nm. A 5 µm filter is twenty-five to a hundred times larger than the thing it is letting through. It is not there to filter pigment; it is there to catch everything that should not be in the ink at all — fibres, dust, cured flakes, seal debris.

We deliberately do not publish a filtration table across five printhead brands. Only Xaar’s figures are documented; the micron numbers circulating online for other heads trace back to ink vendors’ blogs rather than to the head manufacturers.

4. Air and cavitation — and the case against turning the voltage up

Xaar gives one hard mechanical limit: “The maximum acceleration/deceleration the printhead can tolerate before the ink meniscus breaks… is 4 g.” Beyond that, the ink surface at the nozzle tears and air gets in.

Once air is in, a synchrotron X-ray study of drop-on-demand jetting describes what it does: “The entrained bubble compromises the channel’s incompressibility… and can even stop the entire jetting process.” The same paper delivers the finding every workshop should know: “the probability of bubble entrainment increases exponentially with the driving amplitude.”

That is the sourced argument against the standard field fix of raising head voltage to revive a weak nozzle. It may fire today. It raises the odds of drawing air in — exponentially, not proportionally — and air is what takes the channel out altogether.

5. Electrical damage, and the rule technicians break constantly

Xaar: “The printhead must not be connected or disconnected to the electrical supply whilst the power is on. This is known as ‘hot-plugging’ and will cause damage to the printhead.”

Electrostatic protection tops out around 2 kV — which sounds like a lot until you see what a dry workshop generates. A person walking across a vinyl floor at low humidity can carry 12,000 V. That is covered, with the measured table, in setting up the room, and it is the practical reason the humidity floor in the manufacturer’s specification is not optional.

6. Head strikes, and the honest limit of gap detection

Mimaki is unusually candid here, and it is worth quoting exactly because it contradicts what buyers assume they are getting:

“The gap check function is not a function to avoid the head hit completely by detecting obstacles… Even when the GAP CHECK is ‘ON’, be sure to check the work height visually before drawing.”

Automatic height detection reduces head strikes. It does not prevent them. The operator is still the last line, which is why checking the bed for debris is on the daily list.

7. The wrong cleaning fluid

Short version: isopropyl alcohol does not merely attack seals — it crashes UV ink out of solution into a gel, in the worst possible place. The mechanism, the manufacturers’ wording and what to use instead are in nozzle blockage: reading the test print.

What a replacement actually costs you

We publish no printhead price, for the same reason we publish no machine price: it depends on the head, the count and the configuration, and a headline figure would describe somebody else’s machine.

What we can give you is the planning frame. Trade press quoting Nazdar puts consumables — heads included — at 10 to 15 per cent of machine cost per year, with heads lasting one to three years. Treat that as a reserve to fund from the first month, not a surprise in year two. The payback calculator has a field for exactly this.

And note what genuinely moves that number: not brand loyalty, but whether the head recirculates, whether the room is clean and humid enough, whether the jigs are matte, and whether anyone shakes the white. Every one of those is a decision, not luck.

Sources

• Xaar, Guide to Inkjet Printing — nozzle plate coatings, wetting vs non-wetting plates, filtration, 4 g meniscus limit, hot-plugging, ESD
• Roland DG user documentation — reflected UV and head replacement, reflective substrate warning
• Mimaki technical documentation — stray light, gap check limitation
• Ricoh, Fujifilm printhead product pages — nozzle plate coatings
• arXiv:2205.06635 — synchrotron X-ray study of bubble entrainment in drop-on-demand jetting
• Wide-Format Impressions / Nazdar — consumables budget (trade press)