Kit Signal Specs Team Blog FAQ Reserve a kit
All posts

Five hundredths of a millimetre

The arm is a spring, the shell is one printed piece, and the fit comes out of a 0.05 mm tolerance.

Five hundredths of a millimetre

A headset is a fit problem wearing a signal problem's clothes. Eight contacts have to hold against a curved, moving, uneven surface for an hour at a time, on heads that differ by centimetres — and it has to do that without a technician deciding how tight is tight enough.

The arm in the Research Kit is a spring. The shell is one printed piece. The parts that matter are held to a tolerance of 0.05 mm, and this piece is about why that number is on the page at all.

scroll to flex

A spring, not a strap

A conventional headset asks you to tighten something until the impedance is acceptable. That works, and it makes every session depend on whoever put it on. The arm here is a sprung element: pressed onto the head, each of the eight contacts finds its own pressure, and the shell settles from S to XL without a fitting step.

The behaviour that matters is not the pressure itself but its consistency — the arm has to behave the same way on the hundredth unit as it did on the first, or the montage stops being comparable between sessions, and the whole argument for a 10–20 layout collapses.

One piece

The shell is printed as a single piece rather than assembled from a carrier, a frame and a set of fasteners. There are no screws to lose, no clips to snap, and no seam where a housing normally splits after a year of being taken on and off.

In the film, the industrial designer's line about it is blunter than anything we would write: "it's really nearly impossible to break it." That is a claim about how it is made, and the honest way to test it is to hold one.

Why 0.05 mm

Laser printing is not the cheap way to make a shell. It is the way that gives us the material behaviour the spring needs, and it holds a dimensional tolerance of 0.05 mm — five hundredths, or about a twentieth of the thickness of a sheet of paper.

At that tolerance, two things stop being variable: the geometry of the sprung arm, so contact pressure repeats across units, and the interface between the arm and the head, so the same montage lands in the same place.

What 0.05 mm is not: it is not a signal-quality figure. It does not make the noise floor lower and it does not improve the model. It is the manufacturing claim underneath a fitting claim, and it belongs on the specifications table rather than in a headline.

What we would want you to check

Tolerance claims are easy to write and hard to audit, so the useful question is not "is it 0.05 mm" but "on what, measured how". Ours refers to the printed shell and its arm geometry; if you want it broken down per feature, ask us and we will publish the drawing with the number on it.

The rest of the numbers — eight channels, 250 Hz, 0.16 µVrms, 24-bit — are on the specifications table, and the kit is open for reservation.

More posts See the specifications