Measured, not designed-to
The number is the weight of a production sample without cable, with the standard skates on, on a 0.1 g scale — not the CAD figure and not the pre-production shell.
Home / About
We are a measurement company that happens to sell hardware.
CEMMORONX makes ten input products and three chairs. What makes them worth buying is not that the numbers are the biggest — several of them are not — but that every number was produced by a method we are willing to print next to it.
What we make
Ten peripherals, three chairs, one platform underneath each.
Four mice, four keyboards and two headsets share three platforms: the CX-30K sensor, the CX-HE magnetic switch and the CX-LINK radio. That is deliberate. A spec quoted for one product is true of every product that names the same platform, and a revision is a revision of the range rather than a quiet substitution in one model that nobody notices for eighteen months.
The seating line is three chairs on one frame. It exists because the same customers asked for it and because an eight-hour session is a real engineering problem — but it is a line of three, it is not planned to be four, and the homepage gives it one section against the range's five. If that ratio ever changes, the site has to be rewritten rather than extended.
10Input products
3Chairs, and no fourth planned
8,000Hz, every input device
6Test rigs behind the figures
What we will not claim
The list that decides what this company is.
Every one of these is a sentence a competitor prints and we do not. They are not modesty; they are the reason the numbers we do print are worth reading.
- We will not say "zero latency". There is no such thing. A radio link has a floor, a USB stack has a floor, and our own published median is 1.09 ms — a number we would rather defend than beat with an adjective.
- We will not tell you a mouse improves your aim. Hardware removes excuses; it does not add skill. Anyone selling you the second thing is selling you something they cannot deliver.
- We will not print a DPI figure as a headline. The CX-30K reaches 30,000 DPI and nobody plays there. The 50 DPI step size is the specification that changes anything, so that is the one we lead with.
- We will not claim a chair is good for your back. A chair is furniture. Ours is well made, adjustable and rated to a load; it is not a medical device and it does not correct anything.
- We will not ship SOCD cleaning. Rapid trigger yes, adjustable actuation yes — but not a feature that several tournament organisers now disqualify players for using.
- We will not print a testimonial we did not receive. There are no reviews, no star ratings and no review markup anywhere on this site, because a rating on a shop with no customers is the fastest way to look like a fake one.
1.09 msPublished median, all four stages
0Telemetry endpoints in the software
0Star ratings on this site
100 MActuations, magnetic switch
The naming rule
A mouse is called what it weighs, and that binds us.
AXIS 54 weighs 54 grams. If the next revision comes in at 56, it is not an AXIS 54 with a footnote — it is a different product with a different name. Most specification changes on most brands arrive as a silently edited web page; this one cannot, because the name is on the shell.
Keyboards are named for their size, chairs for nothing at all
RAIL 60 HE is 60 %, RAIL 80 HE is tenkeyless. SESSION, SESSION XL and SESSION MESH carry no number, because the figure that matters on a chair is a load rating and putting 150 on a backrest would be absurd.
The wired products are not the cheap ones
BASE 68 is the least expensive mouse we make and the lowest-latency one, because it has no radio stage. We would rather explain that than hide it behind the price.
Spares outlive the product
Skates, keycaps, earpads, cables, castors and gas lifts stay orderable for three years after a product leaves production. A brand that stops selling a $6 part is telling you what it thinks of the $149 one.
Everything above is checkable, and we would rather you checked it.
If a figure on this site does not survive your own testing, tell us and we will either show the method or change the figure.