A Dead CIJ Inkjet Printer and 36 Hours to Ship: What an Emergency Call Taught Me About Batch Coding

4:12 p.m., Tuesday

I was packing up when the phone rang. It was the plant manager of a contract packer we'd worked with for about four years — a mid-size operation filling glass bottles for several beverage brands.

"Our coder on Line 2 is dead. We've got 22,000 bottles to run and a truck booked Thursday at 6 a.m."

That was the whole call. No hello, no how-was-your-weekend (the shorter the call, the worse the problem — that's held true for the eleven years I've been handling rush work on coding and marking lines).

The batch coding printing machine in question was a CIJ unit that had been sitting on that line for six years. It wasn't throwing a warning. It was throwing a fault and refusing to run.

Here's what made it worse than it sounded. The bottles were bound for a grocery retailer's distribution center, and that retailer's compliance manual — the kind built around GS1 lot-coding conventions — treats a missing or illegible date code as a rejection trigger. Rejected freight doesn't come back Thursday. It comes back two weeks later with paperwork, and somebody eats the pallet, the freight, and the conversation with the brand owner.

So the clock wasn't really 36 hours. It was 36 hours minus however long it takes to find, install, and validate a replacement coder on a running filling line — which, if you've never done it, is not a small number.

My first instinct was wrong

When I first started taking these calls, I assumed the emergency was always the hardware. Find a machine, get it on a truck, done. Twenty or thirty rush calls later, I understood that the hardware is rarely the hard part. Shipping is easy. Commissioning is hard.

Normal lead time on a CIJ inkjet printer is typically three to five business days to ship, plus installation, plumbing, PLC integration, and code validation. Realistically that's another day or two, and longer if the plant's line has quirks nobody wrote down. That math doesn't work in 36 hours. Not with any vendor.

I still did the obvious thing first. I called our distributor network and started asking who had a loaner or a demo unit within driving distance. One distributor had a demo CIJ unit. It was configured for a different ink chemistry and a different substrate than what the client was running, and our window to prove it out was about four hours. I passed.

By 5:30, the straightforward answer — get another inkjet printing machine on site by Thursday morning — was dead. That's the point where you stop reading search results and start looking at the actual problem.

The spare unit nobody wanted

The plant had a UV inkjet printer. It was sitting in the maintenance room, half-retired, because UV ink is expensive.

That's usually the whole story with UV on a line like this. Somebody runs the UV inkjet printer price and the consumable numbers against their existing CIJ spend, sees ink at several times the per-litre cost, and the unit quietly becomes a backup nobody ever plugs in. Fair enough on paper.

The numbers said stick with the plan: get a CIJ back online, keep the cheap ink. My gut said run the UV unit Thursday and sort out the rest later. It took me another hour to articulate what my gut was actually picking up on, which was this: the printer dying wasn't the real failure.

The filling room runs humid. The solvent ink on that line had been slow to dry — not failing inspection, just borderline, a smeared digit here and there. They'd been catching it in QC for weeks (ugh, again) and treating it as a maintenance annoyance. The fault just made an existing problem impossible to ignore. Putting a fresh CIJ printer in the same spot would have reproduced the same marginal marks at higher capital cost.

UV ink cured on contact doesn't smear. That's the entire argument. On glass, in a humid room, with a case packer running right behind the coder, "doesn't smear" beats "cheaper per litre" most days of the week.

I have mixed feelings about how that conversation went. Part of me felt like I was talking a customer into the expensive consumable. Another part had been watching them fight smeared codes for two months and knew the ink price was never the real number.

The 36 hours

They ran the UV unit on the overnight shift. Two operators, one of our field techs on the phone until 1 a.m., and a QC person pulling bottles every 15 minutes for the first four hours.

22,000 bottles. Truck left at 6:04 a.m. Thursday.

What I remember isn't the relief. It's that the on-time delivery didn't actually fix anything — it bought them a quarter to make a real decision instead of a panicked one. Rush work is like that. Usually it isn't a solution; it's a delay you can afford.

The part I still kick myself for

I'd been on that plant floor in January. I watched a QC tech rub a bottle with her thumb and frown. I noted it, and moved on, because it wasn't my line and it wasn't my problem that day.

If I'd pushed a two-hour UV test in January, they'd have found this in February, on a Tuesday with nothing riding on it, instead of in March with a truck booked. That's the failure I own. Not the dead printer — the fact that the warning was legible six weeks earlier and I read past it.

What changed over the next two quarters

They didn't go back to the old setup. Line 2 kept the UV unit for short runs and label-heavy work, where instant cure matters. Then in Q3 they installed a CO2 laser marking machine for glass on the high-volume line.

Laser marking on glass works differently than ink — the beam creates a controlled matte mark in the surface rather than depositing anything on it. No ink, no solvent, no cure time, and no consumable to run out of at 2 a.m. The code is permanent in a way no ink is. We supply laser marking systems for exactly this category of application, and glass date coding is one of the places where the argument is fairly clean: high volume, humid environment, permanent mark required, and a product that can't tolerate much ink chemistry.

In my opinion, this is the part of the industry that has genuinely shifted. Five years ago, the default answer for batch coding on almost any substrate was a CIJ unit, and it stayed the default because it worked and because switching meant a capital request nobody wanted to write. That's changed — not everywhere, and not for every substrate, but for glass, washdown environments, and high-value product, laser and UV have moved from "interesting alternative" to the first thing a lot of plants evaluate.

The fundamentals haven't moved. The code still has to be legible, permanent, and compliant. The line still has to run at speed. Somebody still has to read it in a warehouse under bad lighting. What changed is which technology gets you there cheapest over three years instead of over one invoice.

The number everyone gets wrong

If you take one thing from this, take this: ink price per litre is not the cost of coding. Not even close.

The real comparison is cost per legible code that survives the rest of the line — the case packer, the pallet, the truck, and the receiving dock. That number includes:

• Consumable cost per code (ink, solvent, makeup, and the waste)
• Downtime — every nozzle cleaning, every purge, every fault
• Rework and rejected pallets, usually the biggest line item and the one nobody puts in the spreadsheet
• Labor, including the person standing there checking codes

On paper, a UV inkjet printer looks expensive next to CIJ. In our quote comparisons across four vendors in late 2024, UV ink ran several times the per-litre price of solvent CIJ ink, and UV systems generally came in well above comparable CIJ units in capital cost (based on quotes we collected Q3–Q4 2024; pricing moves fast in this segment, so verify current numbers before building a business case on them). What narrowed the gap was rework. Once you add rejected pallets and rework labor back in, the premium usually shrinks a lot — sometimes to nothing.

Laser flips the model entirely: high capital, near-zero consumable, and a maintenance rhythm measured in years rather than shifts. It's a bad fit for a plant changing substrates weekly. It's a very good fit for a plant running the same glass bottle 20 hours a day.

Prices above are for general reference only. Actual pricing varies by vendor, configuration, substrate, and integration scope.

What I'd tell the next person

The plant manager's first instinct — find an industrial inkjet printer supplier who can ship tomorrow — was reasonable, and it would have been wrong. It would have solved Thursday and hidden the problem until the following Thursday.

Three things I'd say to anyone in that position:

Separate the machine from the mark. Ask what's actually failing. A dead coder is a maintenance event. Codes that don't survive the case packer is a process problem wearing a maintenance event's clothes.

Keep a second coding path alive. Not a spare printer in a closet — an actual, tested, documented path. Their UV unit sat unused for eleven months and only worked Thursday because one operator remembered how to set it up.

Don't let a printer failure become a supply chain event. (Note to self: I should have put that on the whiteboard two years ago and I still haven't.) The cost of a rush fix is rarely the fix. It's the rejection, the freight, the retailer scorecard, and the two weeks of paperwork that follow.

We got lucky on this one. I'd rather the next plant not need the luck.