Nagoya Works · Japan | Americas · EMEA · APAC Service

60W Fiber Laser, a Cheap Chiller, and the Mitsubishi Electric LN25 Cooling Capacity (kW) Spec I Ignored

A real shop story: a 60W fiber laser failed 1,120 parts on a hot July week because the chiller was undersized for the actual shop conditions. The fix came down to reading the Mitsubishi Electric LN25 cooling capacity kW spec, wiring the alarm into a Mitsubishi Electric CNC, and making one boring checklist.

The second thermal alarm went off at 10:47 a.m. on a Tuesday in July 2022. I know the time because I wrote it in the logbook and circled it. First alarm that morning: 9:15. Second: 10:47. In between, the operator hit reset and kept going because the marks still looked fine from the other side of the safety screen.

We were running serial-number nameplates for a customer we had kept for four years. Four thousand pieces, due Friday. The cell doing the work was built around a used Mitsubishi Electric CNC, a rotary axis, and a 60W fiber laser marking head. It was supposed to be the boring, dependable part of the shop. It was not.

It also wasn't the laser's fault. It was the cooling system I chose nine months earlier because it was roughly $1,750 cheaper than the Mitsubishi Electric LN25 chiller the laser source manufacturer actually recommended.

The part of the spec sheet I skipped

I didn't start with a 60W fiber laser. I started in 2018 in a two-car garage, with a desktop diode laser and a stack of birch scraps. If you've ever searched how to use an Xtool laser engraver, you would have found me there—I watched three weeks of videos before I admitted the machine wouldn't set itself up.

That Xtool taught me the basics: focus height, material behavior, air assist. It could not do real production on stainless steel. When a contract for engraved stainless nameplates showed up, I had to move up to a 60W fiber laser. That was a game-changer. Cycle times dropped from minutes to seconds, and we could finally mark raw stainless without chemicals.

But a fiber laser source dumps heat into its coolant, and that heat has to go somewhere. The manufacturer's manual was blunt: recommended chiller class, 2.5 kW; coolant inlet temperature, keep it below 25°C. I read that line, nodded, and went shopping.

Two quotes came back. The Mitsubishi Electric LN25: roughly $3,200 with freight. The other option: $1,450. Both were described as 2.5 kW. Actually, the cheaper one claimed 2.8 kW on its nameplate, which made the decision look even easier.

Here is the part I'm not proud of. I stood in the shop, typed mitsubishi electric ln25 cooling capacity kw into my phone, saw that the answer was 2.5 kW, and told the Mitsubishi Electric distributor, that's less capacity for twice the price. I thought I was being smart. I wasn't being smart; I was comparing one number and ignoring the conditions attached to that number.

Even after I sent the purchase order for the cheaper chiller, I kept second-guessing. What if the spec sheet was hiding something? The Mitsubishi Electric unit came with a performance table. The budget unit came with one ambiguous chart. But the budget unit had the bigger number, and I wanted to be the guy who didn't overpay for a brand name.

For six months, I felt vindicated. In spring weather, the chiller held coolant temperature around 24°C, and the laser marked beautifully. I actually started to forget the warning in the small print. July reminded me.

What the cheap chiller did in July

The shop wasn't even in a severe heat wave. But our compressor exhaust pointed straight at the chiller's air intake, and I never thought about it. By mid-afternoon, the air entering that condenser was close to 38°C. A chiller rated at 2.8 kW under standard conditions is not producing 2.8 kW at 38°C. The manual didn't advertise that, and I didn't read deeply enough to find it.

During week two of the job, the coolant outlet temperature climbed from 25°C to 27°C to 30°C. The laser kept firing, but the marks got lighter and inconsistent. I didn't catch it. Our inspector did, after 1,120 pieces had already been run.

Then the high-temperature safety in the laser source tripped. The Mitsubishi Electric CNC did what we had programmed it to do: it stopped the axis, dropped the beam-enable signal, and sent the alarm to my phone. All of that worked exactly as designed. The problem was that by then, the coolant had been too warm for three hours, and the damage was done.

The laser source manufacturer's service tech arrived two days later, checked the coolant temperature at the inlet, and told me the source was fine but the chiller wasn't doing its job at our real ambient condition. Then he said the sentence I deserved:

You didn't buy a chiller. You bought a number. The number is only valid at the test condition.

He sent a quote for the replacement: Mitsubishi Electric LN25. I opened the PDF I had saved a year earlier and finally read the whole first page.

Mitsubishi Electric LN25 cooling capacity kW: the answer I ignored

Here is what that keyword search taught me, eventually. The LN25's nominal cooling capacity is 2.5 kW. That number appears on every distributor site, so if you shop by headline numbers, it looks like the smaller, more expensive option. The datasheet, though, shows the conditions behind the number: the ambient temperature, the water outlet temperature, and the performance curve. The budget chiller's datasheet had none of that detail. When our shop was hot, the budget chiller simply couldn't move the heat it needed to move.

An air-cooled chiller's condenser loses efficiency when the surrounding air is hot. It also loses capacity if the water temperature you need is lower than what the test used. The cheap chiller's manual gave me one nameplate number and almost no operating data. The Mitsubishi Electric documentation included the derating data. It just wasn't in the bullet point I wanted to read.

We bought the LN25, wired the chiller's alarm relay into the Mitsubishi Electric CNC, and set the process interlock to stop the cycle if coolant temperature crossed the laser's maximum. One quick note: laser safety interlocks are still separate—the machine is a Class 4 laser, and the enclosure interlocks follow IEC 60825-1, not the chiller circuit. The chiller alarm protects the process; it does not replace the safety circuit.

The bill for my 'value' decision

  • Budget chiller, delivered: $1,450. It's sitting in the back corner, unplugged.
  • Service visit and replacement pump head: $1,350.
  • 1,120 rejected nameplates—material plus setup and run labor: $3,700.
  • Weekend overtime to re-run before the customer deadline: $1,800.

That's $8,300, not counting the trust I lost with the customer, the gray hair on the operator, or the $3,200 for the LN25. I saved $1,750 and spent $8,300. The math was not complicated; I just did it in the wrong order.

What changed: boring paperwork

The real fix after the LN25 was installed wasn't another expensive part. It was a checklist. Every piece of equipment we add now gets a one-page spec sheet printed and hung next to the machine. Temperature limits, ambient limits, alarm settings, who installed it and when. It sounds like corporate nonsense for a shop our size, but this is where the Brother all-in-one laser printer in our office turned out to be the real MVP. That printer is nothing special—it's just been reliable for years, and it prints those spec sheets without drama.

Since we started that rule in late 2022, the checklist has caught eleven potential problems before they became expensive ones: a coolant flow rate set wrong, a focus height copied from the wrong material, an air compressor filter that everyone assumed someone else changed. Eleven. At our size, that's a huge deal.

If you're currently deciding between a 60W fiber laser, a Mitsubishi Electric CNC, or a chiller, don't take my word for the equipment. Take the habit: find the operating-condition table, not just the headline number. Then print it and stick it where the operators can see it.

Bottom line, after seven years and eleven documented buying mistakes worth roughly $26,000: the lowest price is only the lowest until the machine stops. Then it becomes an invoice with a schedule attached. Buy the system that works at your actual conditions. Trust me on this one—I've got the paperwork to prove why.

The Mitsubishi Electric CNC, the LN25, and the office Brother printer are all still running today. The cheap chiller is not. That's really the whole review.

Costs and specifications in this story come from our own 2022 invoices and equipment documentation. They are included for general reference only. Verify current specifications and prices with Mitsubishi Electric's official resources before making a purchase decision.

← Buying Mitsubishi Electric? What I Learned About CNC, Fiber Lasers, and the MSZ-HR35VF Why a Spec Sheet Is Not Enough: What I Check Before Buying AC, Printers, and 3D Printers →