10 Components You Should NEVER Use in a Product
These 10 components show up in almost every tutorial, dev kit, and prototype out there.
But if you design any of them into a product you plan to manufacture and sell, you could be looking at failed certifications, unreliable products, sourcing nightmares, dead-on-arrival units, and expensive redesigns that could’ve been avoided.
Every one of these works fine on a bench, and that’s exactly what makes them so dangerous, because they trick you into thinking they’ll work in production too.
So in this video, I’m going to go through each one, explain why it doesn’t belong in a real product, and tell you what to use instead.
Component #1 – USB Micro-B Connectors
I’ll be honest, I hate USB-B connectors.
I went through multiple Amazon Kindles over the years, and the Micro-B port was always the first thing to fail.
Then we recently bought the newest Amazon Fire tablet, and I was pretty annoyed to see it still had Micro-B instead of USB-C.
And that frustration isn’t just mine, it’s a real engineering problem.
USB Micro-B is the most notorious mechanical failure point in consumer electronics.
The connector is mechanically fragile, and the solder joints and internal structure can’t handle the stress users put on it.
It fails after far fewer insertion cycles compared to USB-C, and EU regulations already require USB-C for most consumer electronics.
Beyond reliability, customers increasingly expect USB-C on new products, and a Micro-B port signals that your product is outdated before it even ships.
There is almost no reason to start a new product design with USB Micro-B in 2026.
Component #2 – Through-Hole Components (When SMD Alternatives Exist)
Through-hole parts are bigger, harder to place with automated assembly equipment, and they drive up manufacturing cost.
People coming from the hobbyist world default to through-hole because that’s what they learned to solder by hand.
But if an SMD equivalent exists, and it almost always does, you should use it.
Through-hole components require wave soldering or hand soldering as a separate step after SMD reflow, and that adds both cost and complexity to every single production run.
There are legitimate exceptions like connectors that need mechanical anchor strength, some large power components, and certain high-reliability applications.
But for the vast majority of standard resistors, capacitors, and ICs, SMD is the right choice.
Component #3 – Barrel Jack Power Connectors
Barrel jacks are frustrating from an engineering standpoint, but they’re even worse from a customer experience standpoint.
With USB-C, your customer already has a dozen chargers lying around the house that will work, but with a barrel jack, they have to keep track of that one specific adapter that came in the box.
I’ve owned products where I lost the barrel jack charger, and trying to find a replacement with the right size, right polarity, and right current rating is such a hassle that I just ended up buying a whole new product instead.
That’s the kind of experience you do not want your customers to have.
On the engineering side, there are too many similar but incompatible sizes, so a 5.5×2.1 mm plug looks almost identical to a 5.5×2.5 mm plug, but one won’t fit right in the other’s jack.
There’s no built-in polarity protection either, which means if a customer plugs in the wrong adapter, they can fry the board instantly.
For most low-to-moderate power consumer products, USB-C is the better default, and if you need more current than USB-C Power Delivery can handle, use a locking connector rated for your specific current requirements.
Component #4 – Unshielded DC-DC Converter Inductors
Switching regulators are everywhere in product designs, but pairing one with an unshielded inductor is asking for EMI trouble.
The magnetic field from an unshielded inductor radiates freely, couples into nearby traces, and contaminates sensitive analog or RF sections of your board.
This often shows up as noise in a pre-compliance scan, or worse, as an outright failure at the test lab when you’re paying thousands of dollars for certification testing.
Shielded or semi-shielded inductors from manufacturers like Wurth, Coilcraft, or TDK cost only slightly more per unit, and they reduce radiated EMI significantly, though good layout around your switching regulator is still critical.
So that tiny savings on your BOM can easily turn into a $5,000 or more certification failure if your design doesn’t pass EMC testing.
Component #5 – Cheap No-Name Electrolytic Capacitors
The “capacitor plague” isn’t ancient history, it’s still happening with bottom-tier capacitor brands.
Cheap electrolytics dry out early, drift out of spec, and slowly destroy your product’s reliability over one to three years, long after the warranty period is the last thing on your mind.
This is one component category where brand matters, so stick with Panasonic, Nichicon, or Rubycon.
Even better, switch to ceramic or polymer aluminum capacitors where your design allows it.
And if your contract manufacturer is substituting mystery-brand caps into your BOM without your approval, that’s a BOM control problem you need to fix immediately.
Component #6 – Bare 2.54 mm Pin Headers as Production Connectors
Bare pin headers are perfect for prototyping, but they’re terrible for anything that actually ships to a customer.
They have no locking mechanism and no strain relief, which means they vibrate loose in anything that gets moved, shipped, or even just handled roughly.
They’re also easy to plug in offset by one pin since there’s no keying to prevent it, and that can damage your board or connected components instantly.
So instead, use connectors rated for your application with proper retention and keying, brands like JST, Molex Pico-Blade, or Hirose.
The cost increase is usually justified by the reliability and usability improvement.
Component #7 – Mechanical Relays (When Solid-State Works)
Mechanical relays wear out after a limited number of cycles, they’re loud, they’re physically large, and they draw significant coil current.
For many products, especially battery-powered or compact designs, a solid-state alternative is the better choice since there are no moving parts, no audible click, a fraction of the size, and dramatically longer life, and for DC loads, a MOSFET is often the first thing to consider.
Mechanical relays do have their place for things like high-current switching, galvanic isolation requirements, and certain safety certifications.
But founders default to them way too often simply because they’re simple to understand.
So before locking in a mechanical relay, ask yourself whether you actually need the specific properties it provides, because most of the time you don’t.
Component #8 – Single-Source or End-of-Life Components
If only one factory on earth makes your critical component, you’re one supply disruption away from a dead product.
The COVID chip shortage made this painfully obvious, and it will happen again.
End-of-life components are even worse because you’re designing on a countdown timer that you might not know about.
So always check lifecycle status before committing any component to your BOM, and look for second-source alternatives.
If no second source exists, that’s a risk you need to consciously accept up front, not discover later when you’re trying to place a production order.
This is especially dangerous for startups with long development timelines, because your key part can go end-of-life before you even ship your first unit.
Component #9 – Counterfeit or Clone ICs
Fake STM32s, cloned FTDI chips, and suspiciously cheap “original” components from unauthorized distributors are more common than most people realize.
They pass bench testing just often enough to make it into a production run, and then they fail in the field at scale when it’s far too late and far too expensive to fix.
Common signs include pricing that seems too good to be true, packaging inconsistencies, and unusual behavior under edge conditions.
The fix is simple: buy from authorized distributors like Mouser, DigiKey, Arrow, or Farnell.
If you’re sourcing ICs from Alibaba or Taobao to save a few cents per chip, you’re gambling your entire production run.
Component #10 – Hobby-Grade Sensors
Product creators love carrying over the sensors they used in their first Arduino or ESP32 prototype, and I get the appeal because they’re cheap, easy to wire up, and everywhere online.
But sensors like the DHT11 or the MQ-series gas sensors were never designed for production products, and they show it through low accuracy, poor long-term stability, and inconsistent performance across units.
The common thread with all hobby-grade sensors is that they work well enough on a bench to make you think they’ll work in production, but they lack the consistency and reliability your customers will expect.
So when you’re choosing sensors for a real product, look for parts from reputable suppliers that offer factory calibration, standard digital interfaces like I2C or SPI, and solid documentation.
