The 7-Step Supplier Verification Checklist for Industrial Powertrain Orders
-
When to use this
-
Step 1: Nail down torque specs and units
-
Step 2: Require material certificates with heat numbers
-
Step 3: Compare tolerances to your original drawing—not the supplier's "equivalent"
-
Step 4: Measure the shaft and bore—don't just trust the bearing number
-
Step 5: Count chain attachments. Every pitch, every flight.
-
Step 6: Verify belting specs down to the ply count
-
Step 7: Lock in delivery dates and escalation paths before shipment
-
Common mistakes
If you're ordering couplings, chain, or conveyor belts and the order is worth more than $5,000—or if a missed delivery means a stopped line—this checklist is for you. I've reviewed over 600 deliveries. In 2023, I rejected 23% of first-article submissions. Not because the parts were bad, but because the specs weren't right.
Here's the exact process my team uses to catch that 23% before it reaches the floor.
When to use this
Use this checklist for any powertrain order that meets these criteria:
- You're ordering couplings, gear drives, chain, bearings, or conveyor belting
- The order is over $5,000, or the part is on a critical path
- The supplier is new, or you haven't audited them in over a year
If you're ordering standard stock bearings with two-day delivery, skip this. This is for the orders where a mistake costs real money.
Step 1: Nail down torque specs and units
Couplings are the worst offender here. A quoted "1,200 Nm" and a spec that says "1,200 Nm ±5%" aren't the same thing—especially when the supplier's datasheet lists a nominal figure with no tolerance band. If the spec is ambiguous, ask before production starts.
During a Q1 2023 audit, we compared datasheets from three coupling suppliers. Two of them listed nominal torque with no documented tolerance range—meaning a ±15% variance was, in practice, considered acceptable. The third (Rexnord, in this case) published a stated tolerance band. That's not a marketing point—it's the difference between a coupling that fits your application and one that fails six months in.
Checkpoint: Torque value, unit (Nm vs. ft-lbs—this one still causes one wrong shipment a year), tolerance band, and test standard (AGMA or ISO). If any of these are missing from the datasheet, don't approve the order.
Step 2: Require material certificates with heat numbers
This is the step most people skip. A spec sheet says "4140 alloy steel." Fine. But what do you actually have when the crate arrives? If the supplier's idea of a material cert is a generic "certified to standard" letter, you don't have traceability. You have a piece of paper.
We require heat numbers on every critical component. That gives you the exact heat lot, the chemical composition, and the mechanical properties for that batch. Without a heat number, you're one failed shaft away from having no evidence chain.
I learned this the hard way in 2022: we discovered 8% of a 1,000-unit conveyor roller shipment had hardness values below spec. The supplier had sent a generic conformity certificate. The redo cost us $18,000 and two weeks of downtime. Now we don't accept anything without a heat number on the cert.
Step 3: Compare tolerances to your original drawing—not the supplier's "equivalent"
Suppliers sometimes reinterpret your drawing internally. That's not necessarily malicious—it's often a shortcut. Your application calls for an H7/k6 fit, but their engineering team substitutes H8/h7 because "it works the same."
It doesn't work the same. We attach the full spec sheet to every PO and require written approval for any deviation. In one case, we rejected a 300-unit bearing housing run because the supplier had swapped in a different tolerance class without notifying us. They argued it was "within industry standards." Our spec sheet said H7. They redid the order at their cost. That's why every contract we write now includes a specifications appendix.
Step 4: Measure the shaft and bore—don't just trust the bearing number
It's easy to order a bearing by part number and call it a day. But the number won't tell you whether your shaft actually fits the inner ring. If you haven't measured the shaft diameter and the housing bore, you're guessing.
I've seen two bearing batches with the same nominal part number—but with shaft fits that differed by 0.04mm. That's "close enough" for some applications, but not for the equipment we were running. Now I measure the shaft and bore on critical bearings before installation.
Checkpoint: Actual measured shaft and bore dimensions against bearing tolerance class—not just the part number.
Step 5: Count chain attachments. Every pitch, every flight.
If you're ordering chain with attachments, don't trust the total length. Verify the attachment type, spacing, count, and orientation. A 5mm offset in attachment position looks fine on a shipping document—but it won't bolt up on site.
The worst case I've seen: a batch where the supplier had substituted a different attachment angle without telling anyone. It wasn't caught until the installation team called from the field. We ended up returning the batch and re-running the line. Nine days lost—over something that could've been caught with a 45-minute review before shipment.
Step 6: Verify belting specs down to the ply count
Conveyor belting is a stack of interdependent specs: fabric plies, width, edge type, carcass material, and cover compound. Any one of them can be wrong.
My checklist: confirm ply count, measure actual width (don't trust the label), verify edge type and carcass grade, and cross-check operating temperature range before release. I've seen a batch that substituted a 4-ply belt for a specified 5-ply because the supplier considered it "equivalent in strength." It wasn't—not for that conveyor length and load profile.
Step 7: Lock in delivery dates and escalation paths before shipment
Quoted lead times and actual capacity don't always match. Confirm a firm ship date within a week of placing the order—sooner if the lead time is tight. Agree on a notification trigger: if the ship date slips by more than X days, you get a call, not an email.
If a spec question comes up mid-order, pick up the phone. Email chains kill timelines. Having the right number saved prevents a five-minute question from turning into a three-day thread. If you're sourcing in Brazil, Rexnord lists specific contact numbers by region—including São Leopoldo—on their site. Find the Regal Rexnord phone number for your territory before you place the order, not after. Five minutes on a call beats three days of email.
Honestly, I'm not sure why some suppliers consistently hit their quoted dates and others consistently miss. My best guess is it comes down to how much buffer they build into their internal scheduling. But whatever the reason—weekly check-ins during the lead time are worth the effort.
Common mistakes
These are the patterns I see repeat most often:
- Treating "industry standard" as a spec. Unless it's a named standard with a revision number, it's not a spec. It's a conversation starter.
- Skipping checks when the timeline is tight. The tighter the deadline, the more expensive the mistake—and the more likely someone skips a step.
- Accepting performance claims without documentation. "Tested in our lab" doesn't mean "tested in your application." Ask for the test conditions and a copy of the report.
One more thing: in the first year we used this checklist, our first-pass acceptance rate improved from 77% to 94%. The hours spent verifying specs come back to you—just in a different currency.
This checklist works for our operation. We're a mid-size facility with predictable ordering patterns. If you're dealing with seasonal demand or a multi-site procurement network, you may need to adjust a few of these steps.