The printhead you pulled out last month probably still had usable life in it. Most shops decide to replace on a feeling, or on the date somebody wrote on the box, and neither number has anything to do with how hard that head actually worked. The shop two streets over runs the same model on the same ink and gets far longer out of it, and that gap is rarely luck. It comes from writing three numbers down. How much ink has passed through the head. How completely it recovers after a cleaning. How long that recovery holds before the next defect appears. Once you track those three, the order date stops being a guess. Skip them and you stay in the expensive loop: buying heads early because a Monday nozzle check looked bad, or running a finished head through a deadline job until it quits mid roll. Setting the record up takes one afternoon.
Why Months Tell You Nothing
Ask two resellers how long a given wide format head lasts and you will get ranges that barely overlap. Same model, same ink family, different numbers, and both of them are reporting what they see in the field. The ranges are wide because the month figure hides the only variable that matters, which is how much work the head did. A head firing through two shifts of heavy coverage and a head doing one short run a day do not age at the same rate, even with identical maintenance. Selling both under one month range is an average, not a prediction.
Manufacturers who publish real numbers use a different unit. Some datasheets state a rated ink volume per head, counted in tens of liters across the life of the unit. Industrial head makers quote firing cycles instead, meaning the total number of drops each nozzle can be expected to deliver. Both units scale with production. Neither of them cares what the calendar says, and that is the clue worth copying.
Switch your own record to volume. Every time you change an ink bag, bottle, or bulk container, write the machine's square meter counter next to the liters you just loaded. Use a notebook taped to the side panel, or a phone note. Precision is not the point. The point is that after a few months you can say this head has pushed a real quantity of ink, instead of saying it is about a year old and hoping.
You also learn something about your own shop. Most owners underestimate their coverage. A banner with dense graphics burns several times the ink of a line drawing of the same length, and the square meter counter cannot tell them apart. If your liters per square meter climb while the counter looks normal, you are printing heavier than you did last quarter, and the head is aging faster than the calendar suggests.
One more reason to drop months: your head does not wear evenly. The nozzles that fire constantly reach their rated cycles first, and the ones at the edge of the pattern can still be fresh when the center is finished. If every job you take is laid out the same way, in the same position on the roll, you concentrate wear on the same group of nozzles month after month. Shift the image across the width when the layout allows it. Rotate which panel sits at the head of the roll, and stop feeding every banner through the same lane. It costs nothing and it spreads the load.
The Recovery Curve
Nobody can judge a head from one nozzle check. The test tells you what is wrong today, not how much life is left. What does tell you is the shape of the recovery. Run the check, clean, run it again, and write down two things: how much of the pattern came back, and how long it stayed back.
A healthy head in a shop that caps properly returns to a full pattern after one normal cleaning and holds it for days. A head entering its last quarter behaves differently. It still recovers, often completely, but each recovery gets shorter. It holds a week, then three days, then a single shift, then not even one long run. The pattern itself can look perfect after cleaning, and that is exactly what makes this signal easy to miss. The check passes, so the operator moves on to the next job.
Log the interval, not the pass. Write the date and time of each cleaning next to the nozzle check photo. When you can see the gap between cleanings shrinking across a month, you are looking at the end of the head, and you have weeks of warning instead of a failure halfway through a roll.
Before you accept that reading, rule out the cheap causes in order. Air in the dampers comes first, because a bubble starves a whole row of nozzles and looks identical to permanent dropout. Pull the damper, hold it against a light, and look for a bubble sitting in the ink chamber. Next comes the cap top, which has to seal flat against the nozzle plate; a hardened or warped lip lets the face dry between jobs. Then the wiper, because a frayed blade drags debris across the plate and creates deflection that cleaning cannot fix. Then the pump, which fails slowly and quietly and shows up as a channel that starves under load. Only after those four check out does the recovery curve belong to the head.
A shop that skips that order ends up replacing heads that were starved, not worn. It is the most common expensive mistake in this trade, and the nozzle check looks the same either way. Save the last three check photos in your phone, because a pattern that changes after a damper swap tells you the head was never the problem.
Three Numbers Worth Writing Down
You do not need software. You need three or four lines that take under a minute per job, and a place to keep them that will not get lost in a drawer.
- Ink volume plus counter reading. At every ink change, note the liters loaded and the machine's square meter count. This is the mileage of the head, and it is the only figure that compares fairly between two shops.
- Cleanings per square meter. Count cleaning cycles between two jobs and divide by the area printed. A rising number is the first honest sign that the head is fighting you.
- Hours of clean output between cleanings. After each recovery, note how long production ran before a defect came back. This is the curve from the previous section, in your own numbers.
- Wasted media blamed on the head. Keep a separate count of meters reprinted because of banding, dropout, or deflection. This is the number that turns a maintenance argument into a money conversation.
Put the sheet where the operator stands, not in an office folder. A printed nozzle check taped to the machine with the date in ballpoint pen beats a spreadsheet nobody opens. Photograph the check once a week and keep the images in one album, so flipping back thirty days takes seconds.
Assign the sheet to one person. Rotating responsibility is how these records die, because everyone assumes somebody else wrote Tuesday down. If the shop runs two shifts, each operator writes their own line and initials it. Two minutes per shift is cheaper than one wasted roll of media.
The sheet also catches problems that have nothing to do with wear. A sudden jump in cleanings per square meter in one color usually points at that channel's damper or its ink line, not at the whole head. A jump across all colors at once usually means the cap station stopped sealing or the room turned dry. Neither is fixed by ordering a printhead, and both are visible on paper before they reach a customer.
Keep the old sheets. When you finally swap the head, the record of the old one becomes the baseline for the new one, and it is the only way to know whether the replacement is behaving normally or starting badly. Shops that keep this habit stop arguing about whether a head was defective. They compare curves instead.
What Waiting Actually Costs
Waiting is not free, and neither is ordering early. The way to settle it is a number your accountant would accept, not a feeling about the job you almost missed last month.
Build the downtime figure in four parts. Take the machine's realistic hourly output value at your average selling price, multiply by the hours it will actually be down, then add the media scrapped in the failed run. Add freight for a rushed head and the technician call if you cannot fit the head yourself. Most owners land on a number several times larger than the head itself, which is the whole point of doing the arithmetic before the failure rather than after it.
The decision reads like this. If the head still holds a full shift between cleanings and the work is not close view, run it and keep recording. If it cannot hold one production run, order before the next big job, not after it fails. If it holds a shift but the wasted media count climbs every week, the head has already started charging you rent while you keep paying for ink.
When the new head arrives, change the parts around it at the same time. A new head dropped onto old dampers, an old cap top, and a worn wiper inherits the wear of all three from day one, and that is the most common reason a fresh head fails early. Flush the line, fit new dampers, fit a new cap top, fit a new wiper, then run a proper initial charge instead of a light clean. A dry ink path will not prime on a gentle cycle, and repeating light cleans on an empty head only adds heat.
Before you pay, confirm four things in writing with the seller: the exact head part number including any version suffix, the ink chemistry variant, the number of ink ports on your head versus the one in the photo, and what ships in the box. Some heads arrive with a manifold, some with a cable, some bare. Two suppliers can list the same part number with different contents, so ask for the packing list rather than assuming.
On stocking, one machine does not justify a spare head on the shelf. The money sits dead and the seals age while it waits. Two or more machines on the same head model is a different case. Track the recovery curve on the busiest one and order when it starts to shorten, so the head lands before the failure instead of after. DTF shops should plan white and color channels separately, because white channels age on their own schedule.
Questions from the counter
Can I run a head past its curve while the nozzle check stays clean? Yes, for work that is not close view, with nozzle compensation switched on. Know that you are trading quality margin for time, and keep the waste count running so the trade stays visible.
Does a flush bring a tired head back? It clears what is blocking the nozzles. It does not restore a scratched plate or channels that have lost their seal. If the curve does not change after a proper flush and fresh dampers, the head is your answer.