I'm a quality inspector at a drive and motion company. I review bearing orders and rebuild packages before they ship—roughly 200 a year. In March 2025, a maintenance manager from a machine shop called me. He had a bearing order sitting on a steel bench that “looked right” but didn't feel right. The line items read like a trade show display: Timken bearings for the main spindle package, needle roller bearings for the carrier, and one New Hampshire ball bearing for an idler shaft. There was also an old Pete Jackson gear drive on the shelf from an engine they had just pulled apart.
“So what happened to Pete Jackson gear drives?” he asked.
Honestly, I didn't know the full answer. I still don't. But that question, plus the bearing order, turned into one of the more useful audits I have done all year.
First, the Timken tapered roller bearings
The order specified Timken tapered roller bearings—the design family from 1898, meaning the same basic geometry Henry Timken patented back in 1898. According to Timken's published history, that patent created the tapered roller bearing as we know it. But a bearing on a 2025 workbench is a distant, refined descendant. Same principle, different execution. Case depth, raceway finish, internal clearance—all of it has changed. What worked on an axle in 1900 would be a problem at 2,000 rpm if you ignore the spec. The fundamentals haven't changed, but the execution has transformed.
Most buyers focus on price and part number. They completely miss the envelope condition—the bore, O.D., width and the setting relationship between cone and cup. With Timken tapered roller bearings, the cone and cup are designed to work as a system. If someone mixes a cone from one batch with a cup from another, the bearing might spin fine on the bench and then generate heat at 1,200 rpm. That kind of issue never shows up in a part-number check.
I measured the bores first. Both cones were within spec. Then I measured the cup O.D. against the housing bore. That is where we found the first red flag: one housing pocket was oval by about 0.0018 inch. The bearing wasn't the problem. The housing was. A remanufacturer would have bolted a “new” bearing onto that housing and let the bearing cope with a defect. That is not a bearing failure. That is an installation failure.
People think expensive bearings fail because the material was bad. Actually, contamination and misalignment kill most bearings. The material is rarely the first thing to go.
Needle roller bearings: the almost-used needle bin
Next were the needle roller bearings for the carrier. The drawing called for caged needle roller bearings. The shop floor had a bin of loose needles, and the buyer figured they were the same dimensions, just different packaging. I understood the appeal. The loose needles were maybe 35% cheaper per bearing. But the hidden costs were bigger: installation takes longer, because each needle has to be placed and held in position. And if one needle slips during assembly, there is no cage to contain it. That means an overloaded load zone. In a carrier running at 320 rpm, it wouldn't fail immediately. It would fail a few months later, overnight, and probably take a gear tooth with it.
I nearly made my own version of the same mistake. I opened the bin, saw a full layer of needles, and started to move on. Something made me count—or rather, count positions against the drawing. The bin had 29 needles. The drawing required 31. Someone had topped it off with two slightly shorter needles from another kit. That would have been a perfectly running failure.
We did the math on the bench. The caged needle bearing cost more up front, but it came with a known part number, known assembly process, and no bin-to-bin mixing risk. It was basically a no-brainer once the shop saw the two sizes side by side.
One New Hampshire ball bearing with a strict tolerance
Then came the New Hampshire ball bearing. The company—New Hampshire Ball Bearings—makes a range of precision ball bearings, and this was not a commodity part. The drawing called for ABEC-7 under the ABMA/ANSI grading system, which roughly translates to ISO P4. For that size bearing, it means runout measured in tenths—about 0.0001 inch. A standard ABEC-1 bearing has the same outer dimensions, so it fits into the same housing. But it can carry runout three to five times higher, and that difference shows up in the surface finish of the workpiece.
“What is the actual difference?” the shop manager asked.
“On a fixture stand, almost none,” I said. “On a spindle with a six-inch tool and a forty-minute cycle, you'll see it in the finish pattern.”
That is the part outsiders miss. The question isn't “does it spin?” The question is “what does it do under load, at speed, after thermal growth?” A New Hampshire ball bearing is a precision element, not a commodity part from the same size family. Treating it like a commodity gets you a perfectly round-looking failure.
So, what happened to Pete Jackson gear drives?
We got to the gear drive after the bearing checks. The maintenance manager asked again, and I gave him the best answer I could: I don't think there has been a steady production run for a while. The brand has gone in and out of availability, and the honest countermen I know treat it as a “you might be lucky” item rather than a stocked product. I can't verify the full corporate story—my lane is bearings, not gear-drive history. But the question itself is more important than the answer.
A part that used to be in every aftermarket catalog now lives in classifieds and old-stock bins. That is not a pricing decision. That is a supply-chain decision. When a component disappears from the normal channels, you are no longer choosing between a gear drive and a timing chain. You are choosing between a known support ecosystem and an unknown one. And that determines whether a rebuild is a good investment or a pile of parts.
I told the manager to treat the gear drive the same way I treat bearings: list every replacement part, add a scrap allowance for old gear sets, and then decide. If you can't source the idler bearings or the alignment pins, the fact that it fits on the block doesn't matter.
What I relearned that afternoon
We rejected the loose needle bearing substitution and the ABEC-1 ball bearing. We kept the Timken tapered roller bearings, but added a housing re-bore to bring the pocket roundness back to spec. The shop ordered the correct caged needle set and confirmed lead time on the P4 ball bearing. That cost about $400 more than the original “let's use what we have” plan. In my estimate, it saved them a $30,000 spindle repair and four weeks of downtime.
I won't pretend every audit ends that cleanly. Sometimes the whole unit is too far gone and you have to tell a client that their favorite component is now a paperweight. But the pattern is the same: don't let a brand name carry the engineering load. Timken bearings are solid when they're the right spec. Needle roller bearings solve a specific load-density problem. A New Hampshire ball bearing is a precision element. And a Pete Jackson gear drive is a reminder that “what happened to Pete Jackson gear drives” is not a trivia question. It's a risk-management question.
At our shop, we only built a formal sourceability process after the third obsolete-part surprise. Now every quote includes a “sourceability” line. At least, that's been my experience for industrial rebuilds. It's not glamorous. It just prevents a special surprise.