Test every USB cable in the drawer, then label it

A bench device that measures what a USB-C cable can actually carry — power, data, orientation — and prints a sticker so you never guess again.

The idea

A small bench box with two USB-C receptacles. Plug both ends of a cable into it, press the button, and eight seconds later a screen tells you what that cable really is: how many conductors are wired, what voltage drop it shows under a 3 A load, whether the data pairs are present or it is charge-only, and what power profile it negotiates. Then it prints a sticker — 100W · USB2 · 1.8m — that you wrap around the cable. The drawer stops being a lottery.

Why build this

USB-C made every cable look identical and behave completely differently. A 240 W e-marked cable and a charge-only freebie from a hotel are visually indistinguishable, and the failure modes are miserable: a phone that charges at 5 W instead of 45 W, a monitor that won't carry video, a dock that drops out under load, a firmware flash that fails halfway. People debug the device for an hour and it was the cable.

Commercial answers exist at both extremes and nothing in between. A $12 USB tester dongle reads voltage and current while the cable is in use, which tells you what happened but not what the cable is capable of. A proper USB-IF compliance rig costs five figures. Meanwhile anyone with a workshop, a repair shop, a hardware lab, or a 3D printer and a soldering iron has the same drawer and the same problem, and the parts to solve it are now boring and cheap.

Stack sketch

MCU: RP2350 on a Pico 2, chosen for the PIO blocks and plenty of GPIO. Firmware in C with the pico-sdk.
Continuity and orientation: an analog mux (74HC4051) scanning each contact against each contact at low voltage, so you learn exactly which of the 24 pins are landed — CC1/CC2, SBU, both data pairs, the shield.
Power measurement: an INA260 on the supply side, plus a MOSFET-switched load bank of power resistors to pull 1 A / 3 A / 5 A steps. Voltage drop across the cable at a known current gives you the conductor resistance directly.
PD negotiation: an FUSB302B as the port controller so you can talk real USB-PD, read advertised source capabilities, and issue a Discover Identity VDM to pull the cable's e-marker.
Data path: a second RP2350 running PIO-USB as a host, enumerating a known cheap USB 2.0 device through the cable under test. Enumeration success at full speed is a pass; SuperSpeed pairs get a continuity check only.
Output: a 2.42" SSD1309 OLED for the verdict, and a Niimbot D11 over BLE for the sticker.
Log: CBOR records over USB CDC into a SQLite file on a laptop, with a tiny Tauri viewer for history and re-tests.

Scope for v1

In:

Pin-level continuity and orientation map.
Conductor resistance and voltage drop at 3 A, with a pass/warn/fail threshold.
USB 2.0 data enumeration pass/fail.
E-marker read over PD where the cable has one; "no e-marker" is a valid, useful result.
Cable length estimate from resistance, printed sticker, CSV export.

Deliberately out:

Any SuperSpeed or USB4 signal-integrity measurement. You cannot do eye diagrams with a $40 BOM and should not pretend to.
Alt-mode / DisplayPort testing.
Charger and power-bank testing — different fixture, later.
An enclosure you are proud of. Print a box, ship the board.

Where it could go

The obvious extension is the other half of the problem: chargers, power banks, and docks lie about their output the same way cables lie about their throughput. The same PD front end plus a beefier electronic load becomes a bench supply tester that logs a charger's actual curve across its advertised profiles, which is a genuinely useful thing to hand a reviewer or a procurement team.

The second path is batch mode for people who do this at volume. A repair shop or an IT team refreshing a floor of desks wants to run forty cables, print forty stickers, and get a CSV that flags the three that are marginal rather than dead — the ones that will generate a support ticket in four months. That is a firmware mode and a queue UI, not new hardware.

The third is selling it as a kit. This is a well-behaved open hardware product: small BOM, no RF certification burden, an audience that will happily solder, and a clear reason to buy the assembled version instead.

Watch out for

You cannot use USB-IF marks or the word "certified" anywhere near this, and claiming a cable is good for 240 W when you only loaded it to 5 A at 5 V is how you end up responsible for someone's melted connector — state the measurement, not the rating. Also treat the load bank seriously: 100 W of resistors gets hot enough to matter, so thermal cutoff and a fused supply belong in v1, not v2.