In January 2024, our operations director pulled me into a conference room and shut the door. "We need 15% off the motion control budget by Q2," he said. "Everyone else has made their cuts. It's your turn."
I've been managing procurement for this company for six years. I track every PO, every invoice, and every unplanned maintenance event in a spreadsheet I built after getting burned on hidden fees twice in my first year. So when I started comparing gearbox vendors, I genuinely thought I had a reliable process.
The quote from Cone Drive for the precision reducers on our packaging line came to $39,800. They're the original double-enveloping worm gear shop, based in Traverse City, Michigan — the people behind the cone-drive geometry you see in robotics and rotary tables. A competitor, which I won't name, quoted $26,100 for what looked like identical units. Same output torque. Same ratios. Same mounting footprint. The numbers are from our Q1 2024 quotes, so treat them as directional, not gospel.
At the Q1 budget review, I looked like a hero.
Here's the thing: I knew there were differences between a set of cone drive gearboxes and a budget import. I convinced myself they weren't material differences for our use. The service factor on the datasheet was borderline acceptable. Efficiency was within a point. And we didn't run that line 24/7. I even told my maintenance lead, "These units aren't going to hold the same tolerances under a microscope, but for our duty cycle, they're fine."
That phrase—"for our duty cycle"—came back to bite me.
Failure Number One
Unit #7 failed seven months in. No temperature rise, no progressive noise. Just a violent jam in the middle of the shift, hard enough to trip the servo drive. The maintenance tech said it sounded like a bag of bolts thrown into a running engine.
I know what killed it now. The budget reducer used a needle roller bearing on the output shaft that was undersized for our actual radial loads—not the catalog loads, the real ones, including a sprocket drive we'd added the prior summer to raise line speed. The needle roller bearing was the pinch point. When it failed, the shaft tore up the worm wheel and, downstream, took out a linear bearing block on the indexing carriage.
The linear bearing block itself wasn't the expensive part. The servo-driven logic module it was mounted to was $3,200. And the line being down for three days was the real knife. At our throughput, each unscheduled downtime hour runs about $780 in labor and lost production. Three days, eight hours a day: $18,720. That's the part that doesn't show up in a quote.
The replacement was "covered by warranty," technically. But warranty covers the part, not the freight, not the installation labor, and not the argument about whether your application matched the vendor's assumptions. We spent $1,850 on expedited shipping and $340 in tech time—actually, $1,850 for freight and $440 for the tech, I'm mixing it up with the other vendor visit. The point stands: the "free" replacement still cost us real money.
Total hit from failure one: just over $24,000. On top of the $13,700 we thought we'd saved.
Failure Two, and the Call I Didn't Want to Make
Three weeks later, on a Friday, the second unit locked up on a rotary indexing axis. A jam in that axis bends things. It didn't just kill the reducer; it took out a servo motor bearing and tweaked the mounting flange. The failure propagated through the entire axis train.
That's the ugly reality of precision motion: when a gearbox seizes, everything around it gets dragged into the accident.
I called the Cone Drive Traverse City support line that afternoon. I'd kept the number from the original quote, half hoping I'd never need it. I walked the engineer through the application: radial load, duty cycle, ambient temperature, the sprocket modification. He didn't say "I told you so." He didn't even sound smug. He asked whether we'd done load verification after the modification, and the silence on my end was probably answer enough.
The replacement units shipped within two weeks. Not magic speed, but inside a window we could survive, because he helped me size the service factor honestly this time. We also swapped the damaged linear bearing block, added a torque arm, and updated our PM checks.
The TCO Reckoning
When I closed that quarter, I ran the numbers the way I should have run them before buying anything:
- Initial savings from switching vendors: $13,700
- Failure one: approximately $24,100
- Failure two: approximately $14,600 (motor, flange, labor, and two weeks of production at 80% speed)
- Net result: the decision was about $25,000 worse than doing nothing
Full comparison, for the analysts in the room: the budget path was $26,100 in purchase cost plus $38,700 in failures, coming to $64,800. The Cone Drive gearboxes were $39,800 all in. The "premium" option was cheaper by $25,000 before we even factored in my stress level.
That's the part that still gets me—I do not mean the initial invoice, I mean the total cost across the whole life of the decision. I told myself I was being cost-conscious. I was being cost-masochistic.
What I'd Do Differently
Looking back, I should have verified the internal bearing specs before approving the switch. At the time, the datasheet said "equivalent," and I trusted that word. If I could redo it, I would ask the vendor to show me bearing L10 life calculations per ISO 281 using our actual loads, not the catalog's idealized values. The budget vendor had those numbers in their own documentation. I just didn't read past the summary page.
The surprise wasn't that the budget unit failed. It was how completely the total cost argument flipped once downtime was included. That's the principle behind the 12-point checklist I built that night: it starts with "What is the true worst-case radial load, including transients?" and ends with "What does one hour of downtime actually cost this line?" Prevention over cure isn't a slogan. It's a budget line. Five minutes of verification beats five days of correction.
Since then, we've standardized on Cone Drive for any axis with a radial load above 5 kN or a duty cycle above 60%. That's not brand loyalty; it's what our maintenance log recommends. They've been manufacturing precision worm gearboxes since 1925, they're ISO 9001 certified, and when you call, you get an engineer who can discuss double-enveloping worm geometry without a script. More importantly, they answer the phone when a machine is down.
And About That "What Happened to Pete Jackson Gear Drives" Search
I know some of you are landing here from Google. A few times a month, someone searches "what happened to Pete Jackson gear drives" and finds this page, probably because we write about gear drives in general. I'd never heard the name until my maintenance lead explained it: Pete Jackson was a known name in automotive timing gear drives, the kind of gear drive hot rodders bolted to their V8s for decades. From what I can piece together, the brand wound down as engine technology moved to different timing setups, and new production stopped years ago.
Honestly, I'm not the right source for the full history. My best guess is that most Pete Jackson units floating around today are new-old-stock pulls or rebuilt originals, and if you need parts, forums are probably your best lead. If you're a car person rather than a plant person, the lesson is the same one I learned the hard way: a product can be genuinely good and still disappear if the market shifts. Cone Drive survived because they kept engineering through those shifts.
My experience is based on six years of tracking orders and failures on one mid-sized packaging line—about 200 purchase events worth of data. If your application is lighter duty or completely different, your math might come out differently. But the method transfers: ask about the bearings, ask about the load case, and price the failure before you price the part.
The next order I cut for those reducers is going to Traverse City. And I'm keeping one of the failed needle roller bearings in a jar on my desk, as a reminder that "cheap" is usually just a waiting period before "expensive."
Leave a technical comment