Frankenswitches: Mixing Stems and Housings
Build a switch that doesn't exist by combining parts: recipes that actually work.
Every mechanical switch is three interchangeable parts pretending to be one product: a stem, a housing, and a spring. The moment you realize those parts come apart, a question follows almost immediately: what happens if I put the stem from one switch into the housing of another?
That question is the whole hobby of frankenswitching. You take the best traits of two or three switches and build a hybrid that no manufacturer sells. Done well, it produces something genuinely better than its parts. Done carelessly, it produces a scratchy mess. This guide explains how the parts interact and gives you recipes that actually work.
Why mixing parts changes the feel
A switch’s character comes from how its three components interact. Swap one and you shift the whole balance.
- The stem controls the slider profile, the legs that touch the contact leaf, and the bump on a tactile switch. It has the biggest single effect on feel.
- The housing (top and bottom) controls the sound signature and how the stem is guided. POM bottoms sound deep and clacky; nylon sounds fuller; polycarbonate tops sharpen the top-out.
- The spring controls weight and the return behavior, independent of everything else.
Because these are loosely standardized across the MX ecosystem, a stem from one brand usually drops into another brand’s housing. “Usually” is the operative word: tolerances vary, and a too-tight or too-loose fit is the main reason a franken recipe fails.
The three things that go wrong
Before the recipes, learn the failure modes so you can diagnose your own builds.
- Wobble. If the stem is narrower than the housing’s stem chamber, you get side-to-side play. Annoying, but usually livable.
- Scratch and binding. If the stem legs or rails don’t match the housing’s contours, the switch drags. This is the dealbreaker: no amount of lube fully hides it.
- Inconsistent actuation. Mismatched stem legs can sit at the wrong height against the contact leaf, causing chattering or a heavier-than-expected press.
A quick test press of one assembled switch before you build forty of them will save you an evening.
Recipes that work
These combinations have circulated in the community for years because the geometry happens to cooperate. Treat them as starting points, not gospel.
| Recipe | Stem | Housing | Result |
|---|---|---|---|
| The classic deep linear | Gateron Ink Black | JWK/Durock housing | Deep, smooth, clacky |
| Tactile with a clean top | Boba U4T stem | Gazzew housing (matched) | Strong bump, rounded sound |
| Poppy linear | Cherry MX Black stem | C³ Tangerine housing | Snappy, bright |
| Budget glow-up | KTT stem | Gateron housing | Smoother than either alone |
The general principle: pair a stem you like the feel of with a housing you like the sound of, keeping both within the same broad tolerance family. JWK/Durock and Gateron housings are popular bases because they’re consistent and widely available.
A sensible process
Don’t franken a whole board on a hunch. Work in stages.
- Build one. Assemble a single test switch and press it against a known reference.
- Lube deliberately. Frankenswitches almost always benefit from a thin, even coat of 205g0 on the rails and legs, since mismatched parts create friction that lube tames.
- Add the spring last. Pick weight independently. A 62-67g spring suits most linears; tactiles often feel best a touch lighter so the bump dominates.
- Batch only after you’re sure. Once one switch feels right, replicate it and listen for outliers.
Where to find parts
You don’t need to buy full switches twice. Many vendors sell loose stems, housings, and springs specifically for this, and a tube of mixed “switch openers” plus a parts organizer turns the whole thing into an afternoon project rather than a chore.
Is it worth it?
Sometimes the answer is no. The market is now so deep that a stock switch often nails what you’d franken toward, for less effort. Frankenswitching earns its place when you want something specific that nothing sells: a particular bump in a particular housing, or a sound that no off-the-shelf switch quite delivers.
It’s also simply fun, and it teaches you more about how switches work than any amount of reading. After building a few hybrids, you’ll read a spec sheet completely differently.
The takeaway
A frankenswitch is just a stem, a housing, and a spring chosen on purpose instead of by a manufacturer. Start from a proven recipe, build one before you build forty, and lube to tame the friction that mismatched parts introduce. The payoff is a switch tuned exactly to your ears and fingers, plus a much deeper understanding of why your favorite switches feel the way they do.