Count what sits between the ink bag and the nozzle plate on a typical wide format machine: filter, sub tank, valve, ink tube, damper, O rings, manifold, cap top, wiper, pump. Ten parts. Eight of them can destroy the most expensive component in the list, and the head is the one that gets replaced. That imbalance is why printhead sales stay healthy while the actual root causes rarely change. The same five failure modes account for almost every dead head that comes out of a shop, and four of them leave evidence you can see before you buy anything.

This is not a list of symptoms. Symptoms are what the customer reports. Failure modes are what the metal and the ink actually did, and knowing which one you are looking at tells you whether a new head will last a month or a year.

Failure Mode 1: Ink Starvation and Dry Firing

A piezo head fires by flexing a ceramic element that pushes ink out of a chamber. The ink is also the coolant and the mechanical load. When the chamber is empty, the element keeps flexing against nothing, heats up, and slowly loses its displacement. The head still fires, just weaker, and the weakness is permanent.

This is the most common way a healthy head is killed by a healthy machine. It happens during a long print when supply cannot keep up with demand, after a cartridge change when air enters the line, or when a sub tank float sticks and the machine keeps printing on a nearly empty chamber.

  1. Print a solid patch at production speed and at half speed on the same media. If the patch is clean at half speed and streaked at full speed, you have a supply limit, not a head limit.
  2. Check the damper ink level immediately after a long run, not after a cleaning. A damper that is low at that moment is being outrun.
  3. Look at the ink filter and the sub tank screen. A restricted filter starves the head at exactly the flow rate where the print is most profitable.
  4. Confirm the ink cartridge is venting. A blocked vent creates a vacuum in the bag and the pump pulls air instead of ink.

What it looks like when it is too late: a nozzle check with a fading band of weak nozzles that never recovers, and a solid patch that is uniformly light in one color. That is element damage, and no flush will restore displacement.

Failure Mode 2: Air and Foam, the Great Impersonator

Air does not kill a head directly. It kills it by making everyone around it do the wrong thing. A head with air in the manifold prints dropouts, the operator runs powerful cleanings, each cleaning pulls more air, and after a day of that the head is genuinely clogged from ink drying in the nozzles between cycles. The original fault was a two dollar O ring.

The signature of air is inconsistency. Dropouts move around the nozzle check instead of staying in one place. The first few centimeters of a print are clean and then the color fades. A print that was fine at nine in the morning is bad at eleven with no settings changed.

  1. Hold a light behind each damper and look for a bubble at the top of the ink column. A bubble that grows between checks is an air leak, full stop.
  2. Check every O ring in the path, especially the ones at the head manifold that get pinched during a head swap.
  3. Look for ink weeping at a fitting. A fitting that weeps out will also breathe in.
  4. Inspect the cap for cracks at the lip. A split cap lets the pump pull air from the room instead of ink from the head.

Stop cleaning once you suspect air. Every cycle you run without fixing the leak pushes the head closer to a real clog and dumps ink into the waste bottle.

Failure Mode 3: Physical Damage at the Nozzle Plate

This is the mode that shows up in photographs. The nozzle plate is a thin, coated surface, and it does not survive contact with media, a swollen cap, or a careless swab. Head strikes happen when a sheet lifts, when a roll edge curls, or when the head height was set for thin film and nobody changed it for canvas.

Look at the plate under a bright light with a loupe. What you are looking for is a scratch that crosses the nozzle rows, a dent with a ring of displaced coating, or a patch where the coating has lifted at the edge. Crust around a row is cleanable. A scratch across the rows is not.

Two behaviors make this worse. First, printing after a strike without checking: a damaged plate can spray mist for weeks, and the mist lands on the encoder strip, the carriage board, and every nearby connector. Second, wiping the plate with a dry swab or a shop rag, which drags dried ink across the coating and leaves micro scratches that collect pigment.

A cheap habit that catches this early: run a fingertip along the edge of the media before every long job. A lifted corner or a wave in the roll is a head strike waiting to happen, and it takes ten seconds to flatten or trim it.

Also check the cap for swelling. A cap that has absorbed solvent can rise slightly and touch the plate when the carriage parks, which marks the nozzles without any media involved at all.

If you find impact damage, photograph it before you do anything else. It is the one failure mode where the evidence is easy to destroy and the warranty conversation depends on it.

Failure Mode 4: Electrical Failure in the Cable, the Board, or the Head Rank

The head is a passive device. It fires what the board tells it to fire, when the board tells it, using an ID that the board must recognize. A break anywhere in that chain produces the same result as a clog: no ink, no error message, and a nozzle check that looks like a hardware failure.

The most common version is the flat cable. It flexes thousands of times a day, it sits in a mist rich environment, and solvent ink migrates along it. A cable with a conductive ink bridge can short adjacent signal lines and take the head with it, which is why a burnt connector is a replace both situation, not a clean and reseat situation.

Do not overlook the simple version either: a connector that is fully seated but not latched looks installed and behaves like a break the moment the carriage moves.

The second most common is a head rank mismatch. When a board is replaced and the rank of the installed head is never written, or when a head is installed with a rank label from a different unit, the machine can print, but it prints with the wrong waveform. The result is weak firing, poor dot placement, and a head that dies early for no visible reason.

  1. Power down and unplug before touching any cable. This is not caution theater. The board holds charge.
  2. Reseat both ends of the FFC and look at the contacts for ink film or discoloration.
  3. Read the head rank in service mode and compare it with the label on the head.
  4. If the machine has two identical heads, swap them between slots. Fault follows the head, the head is guilty. Fault stays in the slot, the cable or board is guilty.

Failure Mode 5: Chemistry, Meaning the Wrong Ink or the Wrong Flush

Ink is a suspension, not a dye solution. Pigment settles, solvent evaporates, and the balance between them is what keeps the head alive. Three chemistry mistakes do most of the damage: mixing ink brands in one line, using a flush that is not matched to the ink chemistry, and leaving a head parked with solvent ink in it for weeks.

The mixed brand problem is subtle. Two inks can look identical on a print and still be incompatible in the line, because the resins and the carrier solvents differ. The result is a gel or a precipitate that forms in the damper screen and the manifold, where you cannot see it, and shows up as a color that slowly starves.

The flush problem is worse because it is invisible until it is expensive. A water based flush pushed through a solvent line leaves water in the head, and water plus solvent ink equals a coagulated mess. Always confirm the flush is specified for the ink family in the machine, and never assume that a general purpose cleaning fluid is safe for every head.

Long shutdown is the third. A head parked dry for a month with no cap seal and no preservation fluid will have ink polymerized in the nozzles. That is not a clog you can clean out. Before any shutdown longer than a couple of weeks, cap the head, fill the line with the correct storage fluid, and seal the cap station.

Every one of these five modes is cheaper to prevent than to repair, and four of them can be identified in under an hour with a light, a loupe, and a nozzle check. The head is almost never the first part to fail. It is just the one that gets the invoice.