The order that looked right on paper
In March 2023, I placed an order for 24 connecting rods for a small fleet rebuild we'd promised a customer by end of quarter. One engine code, one material spec, one delivery window. The quote came back clean. Price was inside budget. I approved it and moved on.
The rods arrived four weeks later, boxed properly, labeled correctly. Big-end bore measured within tolerance on our gauge. The small-end bushing bore was a few thousandths off. Not enough to fail anyone's visual inspection. Enough that the crank pin wouldn't go in.
That order sat in our shop for 11 days while we sorted it out. The engine bay was tied up. The customer got rescheduled. Finance held an invoice that didn't match the PO because the supplier had billed the rework as a new line item. And I got to explain to my operations lead why the delivery date I'd confirmed twice wasn't the delivery date we got.
At the time I was sure we'd bought from a bad supplier. Then I pulled the records on the four other connecting rod orders that had gone sideways over the previous eighteen months, and the pattern looked different. The suppliers hadn't been sloppy. Every one of them made exactly what they were asked to make. What we asked for and what the machine shop actually needed weren't the same thing.
What's actually going wrong
A connecting rod isn't one item. It's eight.
This is the part that took me embarrassingly long to understand. A connecting rod order is really a bundle of dimensions, and every one of them can be right or wrong independently:
- Center-to-center length
- Big-end bore diameter (and whether that's measured with or without bearings installed)
- Small-end bore diameter (same question, and it's the one that got us)
- Rod width at the big end
- Cap style — cracked, serrated, or plain
- Bolt type, thread, and whether the length is set for torque-to-yield or a stretch gauge
- Individual rod weight, and the weight spread across the set
- Bearing tang orientation and location
When our PO said "standard," it was silent on at least five of those. The supplier picked reasonable defaults. We hadn't given them a reason not to.
An engine code isn't a spec either
Same displacement doesn't mean same rod. Manufacturers revise rods mid-cycle — a bearing gets re-specified, a cap design changes, a rating gets bumped — and the part number moves while the displacement badge on the truck doesn't. On connecting rod truck applications in particular, the year-range matters as much as the engine code, and sometimes the VIN does too. I've had two suppliers quote the same "6.7L, 2015–2019" line item and come back with big-end bores that were two thousandths apart, both legitimately, because they were working off different revisions of the same family.
Weight grading never makes it onto the RFQ
The person assembling the engine is matching rods and pistons to balance rotating mass. If I order a set and the lightest rod is 9 grams under the heaviest, the shop either spends half a shift re-balancing or ships an assembly that runs rougher than it should. Neither of those shows up as a line item on my PO, so neither gets written into the quote. Then everyone's surprised at final inspection.
Billet, forged, titanium — those are processes, not specifications
I used to treat material as the spec. "Billet connecting rod, 4340" felt precise. It isn't. A billet rod is machined from a solid block; a forged rod is formed and then machined. Both can be excellent. They also have different working tolerances, different grain structure, and different amounts of material left for cleanup if a bore needs to be reworked. Titanium changes the machining envelope again. Choosing a connecting rod type without nailing down the dimensions around it is like picking a tire by the rubber compound and skipping the size.
The connecting rod assembly is where it actually breaks
Nobody orders a rod in isolation. It goes into a connecting rod assembly with bolts, bearings, and a piston — and a couple of thousandths that are fine in isolation stop being fine once they stack. Small-end bore too tight plus a bushing pressed slightly under spec equals insufficient oil clearance, and that failure shows up at 3,000 miles, not at receiving. By then I've signed off on the invoice and the supplier is three months gone from the conversation.
What this actually costs
The obvious line is the rework invoice. Ours came to about $1,900 on that March order — remachining plus shipping both directions. Not catastrophic. The less obvious costs did more damage:
Bay time. Eleven days of an occupied engine bay with a technician rotating onto smaller jobs. That's the number I have to defend in a production meeting, not the invoice.
The customer relationship. We rescheduled a fleet account that had given us four rebuilds in the previous year. It's a small industry in our region. Word moves.
Finance friction. A mismatched invoice means a hold, a hold means a phone call to someone's accounts payable, and that phone call takes three days to return. Repeated across four bad orders in eighteen months, our AP team started flagging our department's POs for extra review — which added roughly two days to every subsequent order.
Trust burn with suppliers. This is the one I underestimated. After the second rework, one of our better suppliers started hedging their quotes with caveats — "assuming standard bore per OEM drawing," things like that. They'd learned that our specs were unreliable, so they protected themselves. Fair, honestly. But it meant I couldn't lean on their expertise anymore, which was half the reason I'd picked them.
And the quiet one: every reorder was a fresh argument. Same part, same conversation, no memory. I was re-earning the same information four times a year.
What fixed it (this part is short)
It wasn't a better supplier. We still use two of the ones that botched orders in 2023. What changed is that we stopped letting the specification be implicit.
I built a one-page spec card per part number. Engine code plus year range, connecting rod type, center-to-center length, big-end and small-end bore with a note on whether bearings are installed, rod width, cap style, bolt spec, and a weight tolerance for the set. A photo of the old rod next to a ruler. That's it. It took an afternoon per engine family and I should've done it in 2021.
Now the card goes out with the RFQ, not after the quote comes back. Suppliers confirm the numbers back line by line instead of replying "yes, standard fits." The first order for any new part number ships as a small batch so we can gauge it before committing to a set. And the cards live in a shared folder my operations lead can actually find, so nobody's rebuilding the spec from a closed work order at 4 p.m. on a Thursday.
There's something satisfying about a quote that arrives with the dimensions already verified. After three years of firefighting, finally having a request that comes back correct the first time — that's the payoff I was chasing.
The efficiency gain was real but narrower than I expected. Our quote turnaround went from about nine days to four, mostly because suppliers stopped sending three clarifying emails each. That's worth having. But the bigger win was that I stopped budgeting mental space for the rework conversation.
In my opinion, most of what gets blamed on unreliable overseas sourcing is actually an underspecified purchase order with a supplier doing their best with it. I'd argue that's a more uncomfortable conclusion, because it puts the problem on our side of the desk. One of my bigger regrets is not pulling those four bad orders into a single spreadsheet eighteen months earlier — it would have taken an hour, and it would have shown me the pattern before the March order ever went out.
If you're buying connecting rods and your RFQ says "standard," get the dimensions on paper first. It's a much shorter conversation to have before you order than after.