Unpacking The Kailh Box Switch Debacle
Do Kailh Box switches break keycaps? The answer is not so simple.
It all started with a crack and a post on GeekHack. User Kavik discovered that their Kailh Box switches were cracking the stems on the expensive keycaps they were putting on them. What gives? It seemed that the X-axis of the cross stems on the Kailh Box switches were possibly too large, and when people pushed their keycaps onto them, the posts cracked.
Numerous individuals in the keyboard community immediately started measuring the Box switches they had on hand and comparing notes. Meanwhile, Kaihua began an internal evaluation to see what could have gone wrong at the factory, and they worked with vendors like Novelkeys to identify, scrutinize, and abate the problem.
The Thing That Happened
It didn’t take long to drill down to the issue at hand: The original Kailh Box switch specification had the X-axis thickness of the cross stems at 1.3mm with a manufacturing tolerance of +/-0.02mm. However, that specification had changed. Back when Kaihua was working to get its (then new) Box switches onto shipping products, Dareu was its first major customer. Dareu wanted a little more girth from the Box switch stems for its own keycaps and asked Kaihua to alter them. The switch maker obliged, adding tiny nibs to the tips of the cross stems to increase the thickness from 1.3mm +/- 0.02mm to 1.32mm +/- 0.02mm. Since then, all Kailh Box switches have been manufactured to the new thicker specification.
Therefore, in the original specification, Kailh Box switches would be within spec if they measured between 1.28-1.32mm. In the new specification, that changed to 1.3-1.34mm. Once this information came to light, the consensus was that the cross stems on some Box switches were a wee bit too thick, and that was enough to crack caps. If that was the case, then Kaihua screwed up. The company certainly seemed to admit that it made an error, and they immediately agreed to change the Box switch spec back to the original 1.3mm +/-0.02mm, by simply removing the nibs.
This is the new specification:
Of course, that’s little consolation to the many people (and vendors) who were stuck with what appeared to be keycap-cracking Box switches. But is it all much ado about nothing?
The central question here is whether or not that additional 0.02mm Kaihua added to the Box switch spec was too thick.
First, let’s assume that all reports of cracked keycaps that were placed on Box switches are accurate. It’s reasonable to assume that this is the case, given the number of reports from community members.
Second, let’s assume (for now) that all Box switches in the wild are within spec. Kaihua maintains that this is the case, and at least one vendor has stated that all of the stock he measured appears to be within spec as well. It is true that many community members have performed their own measurements, using calipers on any Box switches they have on hand. However, those measurements cannot be used as definitive proof that any switch stems are or are not within spec. Trying to accurately measure to the hundredths of millimeters with any handheld tool is extraordinarily difficult. Complicating matters is that the plastic of the cross stems is soft enough that it can be slightly compressed fairly easily, thereby throwing off a caliper reading. This is not to mention the fact that any miniscule misplacement or over-tightening of the caliper tips can produce dramatically different results. (In this case, a difference of 0.02mm would be “dramatic,” and 0.02mm is extremely thin.)
Third, it’s important to bear in mind that cross stem thickness is not the only variable here; the keycaps themselves present multiple additional variables. Like switches, keycaps have a manufacturing tolerance. For example, Signature Plastics (SP) told Manofinterests that their keycap stem tolerances (after shrinkage) are +/-0.005 inches, which is 0.127mm. Further, plastics are not homogeneous. Some plastics are softer or harder than others, and some can become more brittle with age. The colors used in various plastics affect those characteristics, too.
In sum, if the reports of damage are all true, and all Box switches are within spec, and all keycaps involved in the cracking issues are also within spec, then it stands to reason that we have identified a problem: Switches with cross stems at the thickest edge of their manufacturing tolerance (1.34mm) may cause damage to some keycaps, which are presumably also on the edge of their own manufacturing tolerances.
But even that needs to be qualified.
Comparing Specifications: Is 1.34mm The Enemy?
To better understand whether or not that 1.34mm thickness can legitimately be considered a primary cause of the keycap cracks, we need to examine the specifications of other switches. Here is an illustrative but not exhaustive list:
The X-axis cross stem on standard Kailh desktop switches (PG151101) is 1.3mm +/-0.03mm, so up to 1.33mm would be within spec. (Kaihua also produced a document for the keyboard community listing many of its the cross stem measurements.) TTC’s is thinner, at 1.3mm +/-0.02mm. Greetech’s X-axis cross stems are 1.3mm +/-0.04mm, meaning 1.34mm is within their spec.
But the most notable measurement comes from Cherry. The X-axis cross stem is 1.31mm +/-0.02mm. That means Cherry MX switches that are within spec have cross stems that range from 1.29-1.33mm.
We can draw some conclusions from this data. Of the above list, only Kailh Box switches and Greetech switches can be up to 1.34mm thick and be considered within spec. However, Kaihua’s non-Box desktop switches, and Cherry MX switches, are considered within spec up to a thickness of 1.33mm.
|Kailh Box (old)||1.32mm||+/-0.02mm||1.3mm||1.34mm|
|Kailh Box (new)||1.3mm||+/-0.02mm||1.28mm||1.32mm|
We can reasonably conclude that X-axis cross stems measuring up to 1.33mm do not pose a cracking threat to keycaps. If they did, Cherry MX and Kaihua switches would have been wreaking havoc on keycap stems for years and years.
It’s also rather thin (if you’ll forgive the pun) to state that X-axis switch stems up to 1.34mm pose a clear cracking threat to keycaps. If the thickness was truly and entirely at fault, then Greetech switches would be cracking keycap stems as well. Granted, the sample size for Greetech switches is much smaller by dint of the fact that they likely aren’t employed as widely as Kailh Box switches are at this point (at least in certain circles), and also that only Greetech switches at the edge of the tolerance achieve the 1.34mm thickness. Of course, by the same token, only a certain number of Kailh Box switch cross stems will actually be 1.34mm. But more to the point, the difference between 1.33mm (safe) and 1.34mm (possibly not safe) is extremely miniscule.
Regardless, given the above chart, if Kaihua changes the Box stem X-axis to 1.3mm +/-0.02mm, which they have stated they will do, the problem will disappear with all new batches of Kailh Box switches and branded Box variants.
My conclusion is that although some Kailh Box switches caused some keycap stems to crack, 1) the problem is not widespread; 2) is not necessarily the result of a faulty specification; and 3) likely has more (or at least as much) to do with the tolerances, materials, and age of the keycaps than the thickness of the switch’s cross stems.
However, there are two unanswered questions that warrant more discussion and should compel Kaihua to investigate further.
The first question is whether any Kailh Box switches in the wild are out of spec. For example, some photos that have circulated in the community appear to show a bulge along the x-axis. Given that the additional thickness on Kailh Box switches is from the nibs on the ends, a bulge in the middle of the axis arm indicates a problem, especially if it protrudes further than the nibs.
The second (and in my opinion more interesting) question is whether the nib design itself is problematic. (This is where I make a very loud disclaimer that I am not a materials scientist nor an engineer. These are speculations based on the observations and common sense reasoning of a non-expert, and they should be taken as the beginning of a conversation, not hard and fast conclusions.) Other cross stem switch designs have uniform thicknesses across their axes that ostensibly produce a snug and even fit between stem and keycap. There is no single point of pressure. However, the nibs on the tips of the Box switch cross stems look like they create four distinct pressure points.
It’s like how you can lay on a bed of nails without injury. Your weight is evenly and widely distributed across all of the nails. But if you step on one nail and put all your weight on one foot, you’ll have a puncture wound.
We may never have a complete answer to the first question above. My hope is that we can at some point definitively answer the second question. Meanwhile, it appears likely that, given what we know, only in a perfect storm will Kailh Box switches break keycap stems. You should be able to use any old Box switches you have on hand, although if you have expensive or older keycap sets, you may want to proceed with caution before pressing them onto Box switches. Meanwhile, Kaihua is obviating any semblance of the problem on the manufacturing side, so all future Box switches should be safe for keycaps.