Therion in general seems to work at the 1 cm level when exporting certain formats, but I think (someone correct me if I am wrong) that internally, it works on much higher resolution, and therefore can work even with tiny legs.

I tested this, and can confirm that Therion can indeed work at tiny, tiny leg lengths. I produced a survey of 100'000 legs, of 1 mm length. (I tried 1 million, but Loch just crashed!). And then at 0.1 mm length It is below the resolution of the Loch graphics handling to render each of the tiny lines correctly at that scale, but the overall rendering is correct. When fixing the points at the start and end of the survey, the overall rendering is still correct, and the error is distributed correctly. So basically, Therion can work at crazy small leg lengths, it does not cause strange distortions. But that can take it beyond the capabilities of the app being used to render the output. Your error probably lies elsewhere.

On 25/08/2022 20:16, Tarquin Wilton-Jones--- via Therion wrote:

...

I am only suggesting removing one of the radiolocation points because your siphon survey sounds like it's fairly accurate, and as tarquin says not likely to be causing the error. I would think the accuracy of the radiolocation could be the weak link, but i'm no expert with radiolocation.

Depends. I would say something that uses gyroscopes to detect its position, is likely to be quite inaccurate, but that really depends on the quality of the device. I know Andrew has worked with some supremely accurate devices, but in this case, I would consider it quite suspect.

Did some more looking into it. The manufacturers do not state the accuracy of their inertial sensors (gyroscopes and accelerometers). This is a summary of what I found: This particular one works using GPS when in contact with the surface, then gyroscopes, accelerometers, compass and depth gauge underground or underwater. Its design seems to be aimed at the use of scooters, rather than being handheld. The compass has only 3-5° accuracy (unless specifically calibrated), and the depth gauge has 0.3m accuracy. Part of its operation uses a speed sensor which detects water flow, so its results vary depending on the current within the cave, and its accuracy is 2% when moving at scooter speed (and likely far worse at slow speeds). The main accuracy limitation is the quality of the inertial sensors, which will become progressively worse over time, so the accuracy decays over longer dives. Even a high quality device is expected to lose 50 m of accuracy within 15 minutes, and a poor one within just 10 seconds [taken from a forum thread about this exact problem with this device]. The manufacturer does not state the quality of the sensors or their use of them. The equivalent cave surveying grade is therefore very hard to estimate, and probably equates to grade 2 at best. It would be interesting to know how well the device actually performs in a test without GPS, comparing a disto survey with an underwater dive survey done using an ENC3.