Who This Checklist Is For
If you're a design engineer or procurement specialist who just received a sample run of Samtec connectors—maybe ERF8 or QSH headers—this checklist is for you. It's for the moment you open that first box, pull out the first unit, and need to decide: do I approve this for production, or do I send it back?
I've been on both sides. I'll walk through three steps that cover the mistakes I've seen most often—and a couple I've made myself. The goal is to spot issues before they become your problem, not after you've already assembled 500 boards.
Step 1: Check the Label and Packaging — This Is Where Costs Hide
You'd think the first step is measuring pin pitch or testing signal integrity. Not yet. Start with the box. I know, boring. But here's why:
I once rejected a batch of 1,200 Samtec TSW headers because the manufacturer's part number on the box didn't match the packing slip. The vendor said, 'it's the same part, just a different packaging revision.' I insisted we open ten units. Every single one was correct. But the paper trail was wrong. That mismatch cost us half a day in reconciliation and a $200 rush fee on the corrected paperwork.
What to check:
- Does the box label match the PO number and the packing slip? Not 'close enough.' Exact match.
- Are there any handwritten corrections or sticker overlays on the original label? In my experience, hand-corrected labels are a red flag. They suggest a last-minute substitution.
- Look at the date code. Is it consistent across all boxes in the shipment? Mixed date codes from the same line item often mean the vendor is shipping from multiple inventory bins—which can mean different manufacturing tolerances.
I still kick myself for not catching a mixed date code sooner. If I'd flagged it, we'd have had grounds to ask for a single-lot run. Instead, we found out during assembly that the 2022 stock and the 2024 stock had slightly different retention force. The difference was small—0.3 N per pin—but it was enough to cause intermittent connection issues in a high-vibration application.
Step 2: Verify Critical Dimensions Against Spec — Not by Eyeballing
Step two is where most engineers think they're being thorough—but they often skip the most important checks. Let me explain.
I've seen teams pull out a caliper, measure the overall length of a Samtec ERF8 connector, declare it 'within spec,' and move on. That's not enough. You need to measure the specific dimensions that matter for your application, not just the overall package size.
Here's what I recommend checking for a typical board-to-board connector like the ERF8 or QSH:
- Mating force sequence: Does the connector insert smoothly, or does it catch halfway? A 'catch' can mean the alignment pins are off by 0.05 mm. That doesn't fail a visual check, but it will cause field failures after thermal cycling.
- Pin co-planarity: Place the connector on a flat reference surface. Use a feeler gauge. Any pin that's more than 0.10 mm out of plane will likely fail solder joint reliability in wave soldering. You don't want to find this out after 1,000 units are assembled.
- Tail length after seating: This is the one most people miss. I once had a batch of Samtec TSW headers where the tail length was 0.15 mm shorter than spec. All ten units looked perfect. But when we populated the PCB and reflowed, the solder joints were 20% weaker on pull test. The vendor claimed 'it's within standard tolerance.' We rejected the batch. Now every contract I write includes the phrase: 'All critical dimensions per IPC-610 Class 2, with 100% sampling on first article.'
Look, I'm not saying you need to measure every single pin on every connector. But for the first 10 units from a new batch, take the time. That 45 minutes of measurement can save you from a $22,000 redo.
Step 3: Test a Real-World Mating Cycle — Not Just a Bench Test
This step sounds like common sense. It's not. Here's the trap: most people test a connector by inserting it once, wiggling it, and calling it good. That's a bench test. It doesn't tell you what happens after the 30th insertion or when the board is installed in its actual chassis.
A few years ago, I was reviewing a batch of Samtec QSH connectors for a ruggedized data link. The spec sheet said 'minimum 100 mating cycles.' We tested the first unit on the bench: inserted and removed it 50 times. This is what the vendor's data said it could handle. The connector worked fine. But then we mounted it in the actual enclosure and tested again. The alignment pin had a 0.7 mm interference with the enclosure wall—which wasn't in the spec because the enclosure was our design. When we cycled it in the real assembly, the plastic housing cracked on the 12th cycle.
What I now do:
- Test the connector in the exact assembly configuration—enclosure, brackets, cable routing. Not on a bare bench.
- Cycle it at least 20 times in that configuration. Record the insertion force curve. If it's not smooth, something is binding.
- Check the coaxial or differential pair signal integrity during the test if you have a TDR. A 2% increase in return loss after insertion can indicate a pin that's starting to deform.
The most frustrating part of this process? Vendors rarely capture this kind of data. You'd think a testing report from the manufacturer would include assembly-specific results, but it almost never does. That's why I now specify in every PO: 'Supplier must provide first article test report in the customer's specified assembly configuration, not just per standalone spec sheet.' It took me three rejections to finally get that language into our standard terms.
Two Quick Notes Before You Start
Note 1: Don't assume that 'Samtec' means 'perfect.' I'm not saying their product is bad—it's good, generally—but any high-volume manufacturer can have a bad batch. I've seen counterfeit Samtec-labeled parts show up from unauthorized distributors. I've seen 'new old stock' that had date codes from 2019 but were sold as current production. Don't let the brand name lower your guard.
Note 2: If you're comparing Samtec vs. a competitor (and I won't name names), don't trust spec sheets alone. I ran a blind test with my team once: same connector type from two vendors. Both spec sheets said '0.50 mm pitch, 20 mOhm contact resistance.' In practice, one vendor's connector required 40% more insertion force. The difference wasn't on the sheet. I'm not saying which vendor was which—I'm saying you have to test what matters to your assembly.
One more thing: if you're specifying a connector like the ERF8 for a 'ruggedized' application like the ones used in the Colorado Springs facility's defense programs, don't skip the sample test. I've seen 'ruggedized' connectors fail after 30 thermal cycles because the plating was slightly too thin. The only way to know for sure is to test your lot, not just the vendor's data.
A Final Thought
This checklist won't catch every problem. But it will catch the expensive ones—the ones that require a production stop, a recall, or a $400 rush reorder. And honestly? After the third rejection, you learn that the 45 minutes you spend on these checks is the cheapest quality insurance you can buy.