Here is the part nobody wants to hear after a color complaint: the new ink is almost never what moved your color. It gets blamed because it is the last thing that changed, and memory is very good at assigning cause to the most recent event. In most shops we walk through, the drift was already sitting there. The ink change simply pushed a print that was two points off into a print that was obviously wrong, and the invoice for the new ink took the blame.
That distinction matters because the fix is completely different. If the ink is at fault you spend weeks chasing batches, lot numbers, and suppliers. If the ink is not at fault, the real cause usually surfaces within one shift using the four steps and the seven question decision tree below. Start with the cheapest test and do not skip ahead because a later one feels more likely.
The Uncomfortable Part: A New Batch Rarely Shifts Color on Its Own
Ink manufacturers work to a color tolerance, not to one exact color. Two batches of the same product can differ slightly and both be within specification. That difference is normally small enough that a properly profiled workflow absorbs it without anyone noticing. What a batch change cannot do is turn a clean red into an orange, or make black print brown across an entire image. When the shift is that big, something else moved at the same time.
The things that habitually move together with an ink change are boring and predictable. A new cartridge lot arrives with a different date code. A different bag style changes how much ink reaches the damper. A flush gets skipped because the shop was busy. A profile gets rebuilt in a hurry by someone who had never built one. The machine sits overnight with a half filled line and the two inks separate. Any of those produces a shift far larger than batch variation ever will.
One habit pays for itself immediately. Keep a printed reference strip from every ink lot: a row of patches covering each channel, printed on the substrate you actually sell, labeled with the lot number and the date, stored flat in a drawer. When a complaint arrives you compare against your own strip instead of against memory. Without that strip, every color argument in the shop becomes a matter of opinion, and the person with the loudest voice wins.
Step One: Prove the Shift Is Real and Repeatable
Before changing a single setting, establish that the shift survives a reprint. Print the same file twice, on the same roll, at the same settings, ten minutes apart. If those two prints differ from each other, you are not looking at color drift at all. You are looking at instability, which usually means air in the line, a failing damper, or a temperature swing in the room.
- Print a target with known values: a gray ramp, a skin tone patch, and a solid mid red. Gray is the most sensitive of the three because a neutral tone turns green or magenta on a drift that a saturated color would hide completely.
- Print the target twice, back to back, and let both sheets dry fully before judging. Solvent and UV prints keep shifting while they cure, and judging wet is how good machines get condemned.
- Compare both prints under the same light. A print judged near a window in the morning and under a shop light at night will always look different, and that is not a machine fault.
- Write four lines on the back of each print: date, ink lot, profile name, substrate, and pass count. Four lines of handwriting end most arguments before they reach the counter.
If the two prints match each other but both differ from your reference strip, the shift is real and stable. That is the good outcome, because a stable shift is usually a profile problem, and profiles are cheap compared with hardware. If the two prints do not match each other, go back to the nozzle check and the supply path before reading another word of this article.
Step Two: Separate Ink From Substrate and From Profile
Three variables produce a color shift, and they are easy to confuse: the ink, the media, and the profile that translates between them. Change one and the other two take the blame. The way to separate them is to hold two of the three constant while you vary the third, which sounds like laboratory thinking but takes about twenty minutes in a real shop.
Print your target on a substrate from a known good roll. If the shift disappears, the media changed. Coated media has a shelf life, and a roll that has absorbed humidity prints darker and duller than the same product from a fresh box, especially in a shop with an open loading door. Then print the target with a profile you trust, one that produced acceptable work last month. If the shift disappears there, the new profile is the problem, not the ink.
Only if the shift survives both swaps is the ink worth investigating. Even then, the question is not whether the ink is bad. The question is whether the new ink needs a different ink limit. Pigment load varies between lots, and a profile built for a heavier load will over ink a lighter one. That shows up first as bronzing on dark colors, then as a flat, dull look across the whole image that no amount of saturation adjustment fixes.
- Keep one unopened bag of the previous lot as a control. It is the only honest way to test a batch complaint, and it costs one bag of inventory.
- Rebuild the profile last, never first. Profiling a machine with a failing damper bakes the fault into the profile, and then you own a profile that only works while the fault stays the same.
- Record the ink limit beside the profile name. If the limit value changed and nobody remembers changing it, someone changed it.
Step Three: Contamination and the Slow Mix in the Line
A line that was not flushed properly does not deliver new ink. It delivers a blend, and that blend changes over the first few meters as the old ink clears. This is why a color complaint often shows up on the first job after a change and then fades. Worse, it comes back after the machine sits overnight, because the two inks separate again in the tubing and the first print of the morning pulls the lighter fraction.
The signature is a shift that moves. If your red is orange on job one and closer to red by job three, you are watching a mix rather than a batch. The fix is mechanical and not negotiable: flush the line with the manufacturer's cleaning fluid until the return runs clear, replace the damper if it has held the old ink for a long time, and purge the first meter before production. Do not try to print your way through it, because the job where it finally clears is the job you deliver late.
Cross contamination between channels is the more serious version. A cap that does not seal lets ink from one channel creep into a neighbor during cleaning, and the result is a shift in a color nobody refilled. If exactly one channel drifted and it sits next to the channel you just changed, suspect the cap top and the wiper before you suspect the ink. That pattern is common enough that it deserves to be the first thing you look for when the drift is narrow rather than global.
Step Four: Mechanical Drift That Impersonates Ink Drift
Printheads do not deliver a constant drop volume forever. Piezo elements age, and the drive voltage that produced a full drop when the head was new produces a slightly smaller one a year later. The visible result is a print that looks lighter and less saturated. That reads as a color shift, it gets reported as a color shift, and it gets blamed on ink even though the ink is doing exactly what it was formulated to do.
Two mechanical factors are worth ruling out before any ink claim. Head temperature matters because ink viscosity changes with it, and a machine in a cold back room at seven in the morning prints differently from the same machine at two in the afternoon. Let the machine warm up and run a test print before judging, particularly in winter. The second factor is the nozzle check. A head that has lost a percentage of its nozzles still prints a recognizable image, but the image is lighter and the balance tilts toward whichever channel is healthiest, which looks uniform and therefore looks like ink.
Age is the third factor, and the easiest to miss. A head with a long service life behind it drifts slowly and in one direction. If your reference strips show a gradual move over months rather than a step change at the ink swap, the ink is innocent and the head is telling you something about its remaining life. That is useful information for planning a replacement, and it is not a reason to file a batch complaint.
The Decision Tree: Seven Questions, One Answer
Work down this list and stop at the first question that gives you a clear answer. The order is built so that each question costs less than the one below it, which is the only sensible way to troubleshoot anything that involves opening sealed containers.
- Does the shift repeat on a second print ten minutes later? If no, you have instability. Check for air, damper condition, and room temperature. If yes, continue.
- Does the shift disappear on a known good roll of media? If yes, the media changed. If no, continue.
- Does the shift disappear when you reload the previous profile? If yes, it is the profile or the ink limit. If no, continue.
- Did the shift appear only on the channel you refilled? If yes, flush the line and change the damper. If no, jump to question six.
- Does the shift fade over the first few meters of printing? If yes, the flush was incomplete. Purge and recheck. If no, continue.
- Is the nozzle check complete on every channel? If no, stop here. Color judgment on a head with missing nozzles is worthless and will send you in the wrong direction for a week.
- Does the previous lot still print correctly from your control bag? If yes, the new batch deserves a conversation with the supplier. If no, the machine moved and the ink change was a coincidence.
Seven questions, and only one of them requires opening a new bag. That ratio is the entire point of the exercise. The craft in this job is not knowing more about ink chemistry than the next shop. It is refusing to test the expensive variable before the cheap ones have been eliminated, and writing down the result each time so the next person does not start from zero.