Some parts in an engineer's life are supposed to be boring. Connectors are supposed to be the boring part. Then a $200 platinum blood pressure monitor started failing because of a connector, and I stopped being bored.
This is a story about a 2019 project, a batch of 200 boards, and a week I don't want back. I write it down because the same mistake is sitting in a lot of design files. The problem wasn't the connector brand. It was how we treated the connector as an afterthought.
The Symptom: A Board That 'Intermittently' Worked
The big complaint came from a retailer: the product would sync for a week and then stop. One repair tech blamed the 3.5 mm audio jack on the side of the unit. It wasn't the audio jack.
When the boards came back to us, they still passed the functional test. But if you pressed on the display side of the unit, the signal dropped. The first time I saw it, I assumed a cold solder joint. I touched up every contact on every connector in a handful of units. Worse, I did it with a hand iron, which is the worst way to fix an SMT connector.
The truth was mechanical. The Hirose board to board connectors were half-mated. They were close enough to pass a bench test, but not tight enough to survive a customer's hand. The plug had a slight tilt because the specified stacking height was wrong for the distance between the boards.
If I had read the mechanical drawing instead of the electrical schematic, I would have seen the mismatch in five minutes. Instead, it took me a week and roughly $3,200 of re-test, rework, and expedited shipping.
Why I Kept Choosing the Wrong Half
The embarrassing part is that this was not my first connector mistake.
Earlier that year, I ordered a Hirose plug for a display board based on pin count and pitch. I didn't check whether the matching jack on the main board was the same 'stack height family.' It wasn't. The parts were individually valid; they just weren't a pair. They looked like they'd fit, and they did fit, but the second board couldn't close. The mechanical engineer gave me a look that I still remember.
That's the underrated thing about hirose board to board connectors: the product line is so broad that two parts with almost identical numbers can have different plastic bodies, different heights, or different mating alignment features. The electrical connection isn't the only spec you're buying.
The Deep Cause Was Our Own Process
The deeper problem—the problem behind the problem—was that connectors were viewed as 'off-the-shelf commodity items.' We picked them after the enclosure and PCB stackup were fixed. By then, we needed a 4 mm stack, found a 3 mm part that looked close, and called it done.
Connectors are not resistors. A resistor can be swapped with a similar part if you know the value. A board-to-board connector is a mechanical system that sits between two rigid boards. It has to absorb manufacturing tolerance, thermal expansion, and vibration. If the system isn't designed around the connector, the connector gets blamed for it.
There was also a naming problem in our internal CAD library. The jack side in this design was labeled vsrx_30—and yes, vsrx was a real label in our library. It was short for 'vertical surface mount receiver, 30 pins.' That label told us it was the receiver, but it didn't tell us which plug it was compatible with. It didn't tell us the stack height. It didn't tell us the alignment features. So every new engineer (including me) made the same guess.
We didn't have a formal 'connector check' step in the design review. The third time this type of mismatch happened, I finally created a checklist. The first time cost us a delay. The third time was the last time.
What This Actually Costs
Let me put it in concrete numbers, from the 2019 failure:
- 200 units shipped to a distribution partner
- 17 came back within a month
- Roughly $190 per field failure after shipping, logistics, and service time
- Plus a week of lab time to identify the mechanical issue
- Plus the awkward supplier conversation where I had to explain that the connector probably wasn't defective
The total out-of-pocket was near $3,200. That's not catastrophic in manufacturing, but it's a big number for a single afternoon of datasheet reading. The bigger cost was credibility. It's hard to tell a customer 'we used a slightly wrong plastics part but we can fix it' without sounding like you don't have a process.
The per-unit savings from picking a 'close enough' connector was probably zero, by the way. We didn't save anything. We just created hidden cost and put it on the final device.
The Checklist That Finally Fixed It
After the second incident, I made a one-page list. It lives in our team's shared drive now.
- Confirm stack height from the mechanical drawing. Not the part number, not the sales page, not the sample drawer. The mechanical drawing.
- Confirm plug/jack mating compatibility. Does the Hirose plug belong to the same family and same 'mated height' group as the jack?
- Check the reflow side. Board-to-board connectors often have features like vacuum pickup caps or bosses that are affected by which half goes to reflow first.
- Write down the internal library abbreviation. If you call the jack
vsrx, call the plugvsplor something clearer. Don't let a vague label lead to a guessing game. - Build one electrical and one mechanical sample. If the team can't press two board halves together by hand without stress, the design is wrong.
The biggest change was putting the connector requirement in the schematic itself: '30-pin plug, 5.0 mm stack, gold plating, DF12 series' instead of just 'Hirose.' People still make mistakes, but now they make them during review, not on the production line.
Where My Experience Ends
To be fair, my experience is based on about 200 consumer and medical electronics board designs. I've only worked with Hirose DF12 and FX10 families in any serious depth. I can't speak to how these principles apply to high-vibration automotive environments or hermetic sealed connectors. Those applications have different rules, and I'd probably be the one making the mistake there.
What I do know is that the 'boring' parts are often the ones that teach you the most. The connector wasn't the enemy. It was just the messenger.
The connector is not the problem until you make it one by ignoring the mechanical drawing.
If you're staring at a failing board, don't blame the plug, the jack, or the factory. Check the height. Check the orientation. Check the label. The answer is usually written down somewhere—you just need to read it before the boards come back.
Ask an engineer about this topic