The 1.5 mm Key Travel Cut That Split a Keyboard Maker's Own Team
Archi5 min read
Contents
The Tradeoff Hiding Under the Space Bar#
The prototype arrived in a foam case with no branding and a note that said "deck B." It was the third variant of a low profile mechanical keyboard from a company I'll call Halyard, a mid sized peripheral maker that sells direct and through two regional distributors. The company asked me to test three decks and give feedback, but the real decision was already framed inside the box. Deck A used a standard 3.5 mm switch. Deck C used a custom 3.0 mm switch. Deck B split the difference at 2.5 mm and had been the engineering team's favorite until the marketing team saw the CAD renders.
I typed on all three for six days, roughly 4,000 words per deck across drafts, chat, and two long code review sessions. The notes below are about a single decision: how much key travel to give up to make the case thinner. That decision reached into switch sourcing, tooling costs, battery placement, and the company's return rate.
What the three decks felt like#
Deck A was the known quantity. The 3.5 mm travel had a clear bottom out and a spring curve that most typists recognize immediately. Activation landed around 2.0 mm, and accidental presses were rare even during fast bursts. The case measured 31 mm at the thickest point, which is thick for a keyboard that people carry in a backpack.
Deck B was the compromise. At 2.5 mm of travel, activation came at roughly 1.2 mm. My first hour was messy. I triggered keys while resting my fingers on the home row, mostly the A and S keys. By day three, my error rate on a 60 second typing test had dropped to within two words per minute of Deck A. But the bottom out felt blunt, like typing on a firm rubber mat.
Deck C was the slimmest at 22 mm and the most divisive. The 3.0 mm custom switch had a crisp feel, but the shorter spring made the force curve steep near the bottom. Two testers loved it. One described it as "snappy." Three others, including me, found long sessions fatiguing in the last knuckles.
The decision the company actually faced#
Halyard's engineers wanted Deck B because it let them reuse an existing switch housing with a shorter stem and a modified spring. That meant no new tooling for the switch itself, only a new plate and a revised PCB. The marketing team wanted Deck C because the 22 mm deck photographed better and matched the thin profile of the laptops the keyboard was meant to sit beside. The finance team had a third view: Deck C required a custom switch from a supplier in Taiwan, a minimum order of 30,000 units, and a 14 week lead time. Deck B could ship in nine weeks using parts already in the warehouse.
The tiebreaker was not feel. It was the return rate on a previous low profile model, which had run at 8 percent in the first 90 days. Most returns cited "keys too sensitive" or "typing feels off." The support team traced many of those returns to users coming from full height keyboards who rested their palms on the deck. Deck C's 1.0 mm activation point made that habit costly. Deck B's 1.2 mm activation point was barely better on paper, but the stiffer spring near the top of the press gave a small amount of resistance before activation, and that resistance mattered.
Field notes from the bench#
I measured actuation force with a small spring scale on ten switches from each deck. Deck A averaged 52 grams at activation. Deck B averaged 48 grams. Deck C averaged 44 grams, with a sharper rise after activation. The numbers are close, but the feel at the fingertip was not. Deck C's light initial press invited resting fingers to trigger keys.
Battery placement was the hidden constraint. The 22 mm deck left room for a 1,800 mAh cell. The 2.5 mm deck fit a 2,400 mAh cell in the same footprint because the switch cavity was deeper. In my wireless test with backlighting at 50 percent, Deck B ran 11 days between charges. Deck C ran seven. For a keyboard marketed to people who carry it between offices, that gap is a feature, not a footnote.
I also opened the decks to look at the stabilizers. Deck C used a new wire bend to clear the shorter cavity, and the space bar had a faint rattle that the other two did not. The engineering lead told me the rattle was a tooling issue they could fix, but the fix would add six weeks. That estimate alone pushed the schedule past the holiday sales window.
What shipped#
Halyard shipped Deck B. The final case measured 27 mm, not the 22 mm in the early renders. The product page calls it "low profile" and lists 2.5 mm of travel in the spec table. The company kept the custom switch program alive for a future model and negotiated the minimum order down to 15,000 units with a smaller supplier.
I asked the engineering lead whether he thought the company made the right call. He said the 22 mm deck would have been the better product in two years, once the switch supply chain matured and the support team had data on how people actually typed on it. For this cycle, the return rate and the nine week lead time decided it. The keyboard that reached customers is thicker than the one in the renders. It is also the one that most of the testers, including me, kept using after the trial ended.
What to watch in the next round#
The custom switch program is the thing to track. If Halyard can get the 3.0 mm switch into a 15,000 unit run with a fixed stabilizer and a 1.3 mm activation point, the thin deck becomes a real option. Until then, the 2.5 mm compromise is the safer bet, and the spec sheet is honest about it. Buyers who want the thinnest deck should read the travel number, not the marketing photo. The difference between 2.5 mm and 3.0 mm is small on a ruler and large under your fingers.
More articles