Nobody calls me because their machine is running smoothly. The phone rings when a roller chain snaps, a coupling starts vibrating, or a line goes down at 2 p.m. on a Friday.
In my role coordinating emergency replacement parts for industrial plants, I've handled a lot of rush orders—47 last quarter alone, with 95% on-time delivery. And I've learned that there is no single answer to “how fast should this part get here?” The right move depends less on the part and more on your situation.
When I first started doing this, I assumed every urgent request deserved the fastest possible shipping. If someone was willing to pay for overnight air, why slow them down? Three returned orders—one wrong roller chain, one wrong shaft coupling, one bearing that had the right number but the wrong series—changed my mind. Speed doesn't help if the part isn't right.
Urgent replacement requests usually fall into one of three categories:
- Scenario 1: The machine is down right now. Production has stopped, and every hour costs money.
- Scenario 2: The part is failing, but the line is still running. You have a window—just maybe not a comfortable one.
- Scenario 3: It's planned maintenance. Nothing has failed yet. You want to keep it that way.
Here's how I'd handle each one. It's not the same answer, because the situations aren't the same.
Scenario 1: The machine is down right now
If this is you, stop reading and get the order moving. I'm serious. Twenty minutes spent reading while a line sits idle is not a good trade.
But do take five minutes to verify the part number before you confirm the order. That sounds like a contradiction, but it isn't. A chain that goes out on the next flight, arrives on time, and turns out to be the wrong pitch is the most expensive part you'll buy all year.
Check three things on roller chain:
- Pitch. Measure from the center of one pin to the center of the next. If you know the ANSI number, the pitch is usually printed on the side plate—for example, a 40 chain has a 1/2-inch pitch, and an 80 chain has a 1-inch pitch.
- Number of pitches. Count the links in the old strand if you can. A conveyor chain that's two links too short will ruin your afternoon.
- Single or multiple strand. A double-strand chain looks nothing like a single-strand chain once you try to install it.
If you're fairly sure the old chain was Tsubaki—many plants standardize on it—searching for “roller chain Tsubaki” on a distributor's site will quickly show the right series. The machine's manual or the plate on the drive frame will usually confirm the chain number faster than a worn side plate will.
Here's a real example. In March 2024, a food plant called at 2:15 p.m. with a conveyor down. Their normal three- to five-day turnaround wasn't going to cut it. We knew the machine, but nobody could confirm whether the chain had 96 or 98 pitches. We stayed on the phone while the maintenance lead counted pins—it took about eight minutes. The chain went out on a next-morning flight, cost roughly $310 more than ground freight, and was installed before the lunch break. That eight-minute count is what made the difference.
If you're in Scenario 1, rush shipping is usually justified. But it's not a substitute for knowing what you're ordering.
Scenario 2: It's still running, but something is wrong
This is where I see the most unnecessary rush fees. The line is running, but there's noise, vibration, or a coupling that's starting to fail. The temptation is to treat it like Scenario 1 and pay for next-day air before you've actually diagnosed the problem.
Don't do that yet. A failing part is often a symptom, not the root cause.
Take a shaft coupling, for example. Couplings usually fail because of misalignment, overload, or a bad bearing right next to them—not because the coupling itself was defective. If you replace the coupling but leave the misalignment, you'll be replacing it again in six months. Use the fact that the line is still running to check the alignment and the condition of the surrounding components first.
And speaking of bearings: what's a ball bearing? It's a shaft support with steel balls rolling between an inner and outer ring. It keeps the shaft spinning smoothly with less friction than sliding contact. When a ball bearing starts to fail, it makes noise that can easily be mistaken for a chain or coupling problem. I've seen plants order an expensive new chain when the real issue was a seized idler bearing. The chain wasn't bad; it was being dragged sideways by a bearing that had already given up.
So before you pay for expedited delivery in Scenario 2, ask yourself: have I found the actual cause? If the answer is yes, and the part is a chain or coupling that needs replacing, you usually have enough time for standard two- to three-day shipping. That can save a few hundred dollars in freight compared with next-day air—money better spent on alignment tools or spare parts.
This is also the right moment to use the Tsubaki chain catalog. If you can't read the worn markings on the old chain, the catalog has dimension drawings, part numbers, and load tables. In my experience, checking the catalog for ten minutes is faster than ordering the wrong part twice.
Five minutes of verification beats five days of correction. I've watched that play out more times than I can count.
There are exceptions. If the machine is making a grinding noise that's getting worse by the hour, don't wait for ground freight. But “it's running and we think it might fail next month” is not an emergency order—it's a planned replacement that you're doing early.
Scenario 3: Planned maintenance and preventive replacement
This is the scenario people undervalue, because there's no panic. Nothing is broken. You found a worn chain during an inspection, you're stocking spares before the busy season, or you're upgrading a drive before it causes problems.
Use that time. It's the cheapest insurance you'll ever buy.
If you're replacing a chain during a scheduled shutdown, order it well ahead, but also think about the parts around it: sprockets, bearings, and connecting links. A chain that's worn beyond its service life has probably worn the sprockets too. Replacing just the chain on badly worn sprockets means the new chain will wear quickly. That's not a sales pitch—it's basic mechanics.
This is also the right time for bigger changes. If you're considering replacing a pneumatic cylinder with an electric linear actuator, give yourself a real design window. Electric linear actuators need sizing for stroke, thrust, speed, and duty cycle, and they often require a controller and wiring changes. Nobody should be making those decisions in a rush. Lead time for a custom actuator can easily be six to eight weeks, so paying for speed doesn't even make sense if the design isn't finished.
Build a small stock of critical wear parts for each machine: one spare chain, one spare coupling element, a couple of bearings. Compare the cost of that stock to the cost of one emergency shipment, and you'll see why preventive spares make sense.
Which scenario are you in?
Still not sure? Ask yourself one question:
If I do nothing for the next eight hours, what happens?
- If the answer is “we stop producing”—you're in Scenario 1. Get the part moving and verify the specs on the way.
- If the answer is “it keeps running, but I don't like the noise”—you're in Scenario 2. Diagnose first, then decide how fast the part needs to travel.
- If the answer is “nothing happens, because it's scheduled maintenance next month”—you're in Scenario 3. Take the time to do it properly.
I still catch myself second-guessing orders after I hit confirm. Did I choose the right shipping? Was the part number right? What if the customer measured wrong? But after 12 years and 200+ rush orders, the one habit that has saved me the most pain is simple: check first, rush second. A fast order is only useful when it's also the right order.