Tuesday afternoon, the 3 meter wrap job is twelve meters in and the operator sees it: a pale horizontal band that starts at meter twelve and repeats all the way to the take up. The first eleven meters are clean. The reflex is to hit head cleaning and keep going. Stop instead. Before you touch the maintenance menu, cut the print, let thirty centimeters dry, and measure the distance between the center of one band and the center of the next. That single number tells you whether you are about to spend five minutes or five hundred dollars. Banding is not one fault, it is an inconsistency in one of four places: how the nozzles fire, how far the media advances, how the two carriage directions line up, or how the ink reaches the head. Each of those has a different price tag, and three of the four are free to test.
Read the Banding Interval Before You Touch Anything
Banding has a signature, and the signature is geometric. Measure the interval, look at whether the line is light or dark, and note how many color channels it crosses. Write those three things on the back of the job ticket before you change a single setting. The table below is the fastest triage you can do on a production floor, and it costs nothing but two minutes.
| What the defect looks like | Most likely source | First thing to spend |
|---|---|---|
| Evenly spaced bands down the whole length, all colors affected | Media advance error | Twenty minutes of calibration |
| Bands only in one color, or in one tonal range of one color | Missing or deflecting nozzles | A nozzle check and one normal clean |
| Dark bands, image looks compressed and slightly soft | Passes overlapping, feed too short | Feed calibration, then pass count |
| Pale gaps, thin white lines between passes | Feed too long, or starved nozzles | Feed calibration, then ink supply |
| Soft double edge on vertical detail, ghosting | Bidirectional alignment | Unidirectional test, then alignment |
| Fine regular lines that drift position between jobs | Dirty encoder strip or sensor | Alcohol and a lint free cloth |
The light or dark distinction matters more than most operators think. A light band means the printer moved the media slightly too far, so there is a sliver of unprinted space between passes. A dark band means it moved too little and the passes overlapped, doubling the ink. That tells you which direction to nudge the feed correction instead of guessing. And if the bands are not evenly spaced at all, if they wander or appear in clusters, stop looking at the feed system. Irregular intervals point to nozzles firing intermittently, air in the line, or a failing pump, not to calibration.
The Two Checks That Cost Nothing but Five Minutes
Print a nozzle check pattern before anything else. Not a test image, the actual nozzle check pattern from the maintenance menu, the one that shows every channel as a block of thin lines. Read it channel by channel, at reading distance, with a light behind the sheet. What you are looking for is broken lines, gaps in a block, or lines that lean instead of running straight.
Then run exactly one normal cleaning cycle and print the pattern again. The behavior of the pattern after that single clean is the whole diagnosis. If the gaps moved to different positions or partly filled in, you have wet ink or a light blockage and two more normal cleans will usually finish the job. If the identical nozzles are missing in the identical positions, cleaning is not going to help, and every additional cycle just burns ink, fills the waste pad, and puts thermal and mechanical wear on the head. Three identical patterns in a row is the signal to stop cleaning and go look at ink supply: dampers with air in them, a cap top that is not sealing, a wiper that is dragging debris across the nozzle plate.
The second free check is pass count. Take the file that banded and reprint one meter of it at double the passes, so a 4 pass mode becomes 8 pass or a 6 becomes 12. Do not change anything else. High pass modes spread the ink delivery across more nozzle positions, which averages out individual nozzle variation. If the banding mostly disappears at the higher pass count, you did not fix a broken machine, you found out that the job was being run in a mode that leaves no margin for a slightly tired head. That is a scheduling decision, not a hardware problem, and it is legitimate to run the rest of the week at higher passes while you plan the real repair. If the banding looks exactly the same at double the passes, the cause is mechanical or calibration related and no amount of cleaning will touch it.
Feed Calibration Belongs to the Media, Not the Machine
This is the single most misunderstood point in wide format troubleshooting. A feed calibration is not a property of your printer. It is a property of one media, at one thickness, at one pass count, in one set of temperature and humidity conditions. A machine that is perfectly calibrated on a heavy scrim banner will band on a thin matte film with the same file and the same profile, because the film grips the rollers differently and stretches differently under tension. So when a job that ran clean last month now bands, the first question is not what broke. The first question is what changed on the roll.
The test is straightforward and you can do it without a service tech. Load the actual production media, not a scrap of something similar. Print a one meter long calibration ruler or a file with a known 1000 mm dimension marked on it. Let it settle, then measure with a steel rule. If the printed length is off by more than about a millimeter on a meter, feed compensation is out of range and every long run will show it. Adjust the feed value in small steps, reprint, remeasure. Do not jump to a large correction based on one measurement, because a stretched vinyl will relax after it comes off the platen and read differently.
Three physical checks go with the calibration and take under ten minutes. Confirm the roll is loaded square against both guides, since a roll sitting a few millimeters off will steer and produce bands on one side only. Check that both edge pinch rollers are actually down on the media and not sitting on the liner edge or just outside it. Wipe the feed rollers with a lint free cloth and isopropyl alcohol, because a glazed roller slips unevenly and produces exactly the periodic banding people blame on the head. And on long unattended runs, look at the take up tension. A take up pulling harder than the feed motor expects will stretch the media progressively, which is why the defect often starts twenty meters in rather than at the first meter.
Ghosting, Soft Edges, and the Encoder Strip
If the defect is not horizontal banding but a soft double edge on vertical lines, or a slight blur that makes small text look like it was printed twice a hair apart, you are looking at bidirectional alignment rather than feed. In bidirectional mode the carriage fires on the way out and on the way back, and if those two passes are not landing in the same place, every edge gets a shadow. The test takes one print. Switch the job to unidirectional and print the same meter. If the ghosting vanishes, alignment is your cause and it needs to be recalibrated on the media and pass mode you actually run, because alignment values shift with media thickness and print speed.
The encoder strip is the other cheap suspect and it is the one most often missed. It is the thin transparent strip running the length of the carriage path, and the carriage reads its position from fine graduations on it. A film of dust, ink mist, or solvent haze on that strip corrupts the position reading, and the printer then places passes slightly wrong. The symptom is fine regular lines that change position between jobs, which is distinct from the rock steady interval of a feed error. Clean it with a lint free cloth and isopropyl alcohol, one direction along the length, not back and forth, and do not use acetone or aggressive solvent on it. Wipe the encoder sensor window on the carriage at the same time, and check that the strip is not bowed, kinked, or rubbing on a bracket at one end. A strip that touches anything will produce banding in one zone of the bed only.
Why It Only Shows Up Twenty Meters Into the Roll
The most expensive misdiagnosis in this whole list is a head replacement triggered by a defect that only appears on long runs. Short test strips look perfect, the operator assumes the machine is healthy, then the production job fails in the second half. That pattern almost never points to nozzles, which fail consistently from the first pass. It points to something that accumulates: tension, heat, ink supply, or static.
Here is how to tell them apart with one print. Run three meters of a solid 60 to 80 percent coverage block at production speed and mark every meter on the back. Note the exact distance where the first band appears. If that distance is the same on the next job, the cause accumulates with roll length, which means tension or media stretch, and you go to the take up and the feed rollers. If the distance varies job to job, the cause is intermittent, which means ink supply. On a long heavy coverage run the head pulls ink far faster than on a test strip, and a damper with a small air pocket, a weak ink pump, or a kinked tube that was fine at low demand will starve the nozzles partway through. Pull the damper, look for air in the ink chamber, reprime the line, and reprint. If the band position moves after that, you found it.
Two environmental factors are worth ruling out because both are free to check. First, humidity. Low humidity dries the meniscus in the nozzle plate faster between passes, so a head that is marginal on a dry afternoon will band on a long run and be fine the next morning. Second, heat. On solvent and latex machines the inline heater and dryer warm the media as it travels, and a media that stretches when warm will feed differently at meter twenty than at meter one. Put a thermometer on the platen area, or simply run the same three meter block with the heater set one step lower and compare. If the band moves later in the run, thermal stretch is part of your problem and the fix is tension, not hardware.
The Order to Work In, and When to Stop Spending
Print this order and tape it inside the carriage cover. It runs from free to expensive, and it exists because the most expensive component is also the least likely to be at fault.
One: measure the band interval and note light versus dark. Cost, zero. Two: nozzle check, then one normal clean, then nozzle check again. Cost, a few milliliters of ink. Three: reprint one meter at double the pass count. Cost, a meter of media and some time. Four: feed calibration on the actual media at the actual pass mode. Cost, twenty minutes and two meters. Five: unidirectional test, then bidirectional alignment. Cost, one meter. Six: clean the encoder strip and sensor, wipe the feed rollers, confirm pinch roller position and take up tension. Cost, ten minutes. Seven: replace dampers, cap top, and wiper as a set. Cost, a modest parts bill and an hour of labor. Eight: printhead. Cost, the largest single line item in the maintenance budget, and the one you reach only after steps one through seven have failed to move the defect.
Weigh that order against downtime, not against the parts price. An hour of stopped production on a wide format machine, plus thirty meters of wasted media, plus a missed delivery window, usually costs far more than the entire set of maintenance consumables in step seven. That is the argument for keeping dampers, a cap top, and a wiper on the shelf rather than ordering them after the machine fails. Replacing a four dollar wiper on a schedule is cheaper than the head it eventually scratches.
Can I mix OEM and compatible dampers on the same head?
Do not, at least not as a diagnostic. Dampers regulate the negative pressure seen by each channel, and mixing designs with different flow resistance means one channel behaves differently from its neighbors. If you are chasing a single channel dropout, run the same damper type across the whole head so the comparison is valid.
Do I have to recalibrate feed after switching ink brands?
Not for feed length, but do recheck your profile and ink limits. A different ink batch changes drop weight and how the media absorbs it, which shows up as tonal banding in dense solids rather than as a mechanical offset. Print one meter of the job before committing to a full roll.
Why does the banding appear in a different place on every print?
Because the defect is intermittent, which points to ink supply or a nozzle firing weakly rather than a nozzle that is dead. A dead nozzle produces a line in exactly the same place every time. A wandering band points to air in the damper, an unstable negative pressure, or a pump that cannot keep up at high coverage.
Is it worth replacing the printhead before the capping station?
Almost never. A cap top that does not seal lets the nozzle plate dry between jobs, so you get recurring dropouts that look like head failure and come back a week after any replacement. Reseal the maintenance station first, then judge the head.