TROUBLESHOOTING.md — two load-cell checks from the support channel #44

docs/TROUBLESHOOTING.md covers "The scale reads 0 g, or negative, no matter what" (a swapped
load-cell wire, check the sign of the raw reading) and "Weight drifts, or the reading is jumpy"
(recalibrate, no mechanical binding). On 27/08 a builder hit a negative reading during assembly,
and the team's answers on Discord added two checks the file doesn't carry yet:

1. Test with the load cell mounted in its printed slot, the right way round. The builder had
   fixed it to a makeshift support for testing; the answer was that a load cell lying on a table
   or on a test bracket gives meaningless values, and to mount it in the enclosure, oriented as in
   the reference photo, before judging the reading.
2. Check that DT and SCK are not swapped between the HX711 and the board (and that the
   Dupont leads are fully seated) — suggested alongside the load-cell orientation.

The thread doesn't say which one fixed it, so I'd rather ask than write it up:

• Should (1) go under Weight drifts (mechanical binding is already there) or under 0 g, or
  negative?
• What does a DT/SCK swap actually look like on this board — [HX711] not ready forever and no
  weight, or a wrong reading? If it's the former, it belongs under the [HX711] not ready
  section rather than the negative one.

Happy to open the PR on this file once you say where each goes. The wiki is about to route builders here — TigerTag-Project/TigerSystem-Docs#48 adds a link to
this file from products/tigerscale.md → Troubleshooting — so this is where they'll look.