Spartan Ultra accelerometer/fusedspeed issue
I try to describe the wrong operation (in my opinion) of the fusedspeed of a spartan ultra (but I could say the same of a spartan trainer that I used previously).
The “malfunction” I find is the following: during a movement activity (trekking or walking) when I stop the speed shown by the clock is reset only after a long (too) time, sometimes after more than 10 seconds. This implies that the tracks will have wrong lengths with deviations of about 10-12% (comparison made using OruxMaps and internal GPS of an Oneplus 5 in parallel).
The tests I have made to understand the problem:
1 - reset the clock by reinstalling the firmware
2 - activate/deactivate the Glonass option
3 - use Best and Good in energy saving
4 - disable GPS
The test conditions have always been in the open field (no trees, no houses or other constructions).
I began to suspect that the problem was related to the accelerometer.
So I did this test:
- sitting on a chair I activated a trekking session (but also “walking” gives the same result)
- I simulated the walk by moving my arm for several seconds so that the speed reached around 4-5 km/h
- when locking the arm in front of me to see the speed values, it did not reset until after many (too many) seconds.
- Stabilization on speed 0 has less latency if I put the clock on the table for example.
My conclusions: the accelerometer has a sensitivity too high in order to understand when I’m really still.
Is there anyone who has encountered this strange behavior (I think the whole Spartan series) and maybe has some suggestions to give me?
Thank you very much!
PS: I beg your pardon for my bad English
From what I know the watch calibrates the accelerometer after each best gps activity.
If you are with gps off the accelerometer based speed should be responsive. If you are on gps then it bases on gps not on accelerometer so much
So my try without GPS inside my house tells me that the accelerometer is defective? Because in that case only information from accelerometer are taken into account by the watch in order to calculate speed. Is it normal? am I too demanding?
PS: I also called suunto support in Italy and the operator tells me that the accelerometer wins over GPS when both are used!
FWIW, I just did this quick test with an A3P (GPS on, Best accuracy) and an SSU (GPS Off). Modified Walking sport mode. Cadence On, on both, as a control to test if they’re both registering bumps similarly. One on each wrist. Middle portion is me using the SSU hand to do cash register stuff.
The distance was about 2km on the flat in low city canyons. The A3P registered 2.49km, the SSU 301m. Same duration.
As you can see (if I’m reading the results correctly), the modal (average) SSU speed, as calculated by the only thing vaguely capable of it (the accelerometer) was basically 0km/h. My typical walking speed is about 7km/h: 6.9 today, according to the A3P. Which makes sense for my understanding of accelerometers as measuring delta-v. (I have no idea where 77km/h came from.)
Once my basic cadence is established (perhaps linked to arm swing to-and-fro), the accelerometer uses that as a baseline and spikes only when I accelerate/decelerate veer noticeably left, right (crossing roads abruptly), up or down (stairs). Or used the cash register: G-force on 2/3 axes, rather than one.
So the SSU only considered me moving in space (for a distance/time speed calculation) for that 300m of non-straight-line movement.
@Simone-Marchi What does your in-house 4-5km/h movement look like in graph form? A collection of spikes or a consistent, level reading?
TLDR: I don’t know whether yours is wonky but don’t trust accelerometers to measure speed: that’s not what they do well. I would think that the accelerometer overrides the GPS on cornering/ascents/descents/speed changes when the GPS accuracy/fix rate is going to provide a less-granular result. Measuring instantaneous force+direction input vs ex post-facto track calculation. (Bet I’m doing something wrong somehow, tho’.)
P.S. Left to its own devices (sorry), the SSU’s altimeter did better than the A3P’s FusedAlti (Auto), in consistency terms: 6m up/down vs 9m up/6m down. However, the GPS conditions are lousy so any GPS input will be for the worse.
@Simone-Marchi As if the above wasn’t long enough, do you have the following slow response issue with the SSU?
The route/map screen is way too slow to stabilise on raising wrist to view. I don’t know whether it’s the accelerometer, compass or GPS squabbling, or screen drawing speed but, when wanting a quick orientation reference, I find myself waiting for my Spartan’s route screen to stabilise, find North, doubt itself, move my location 50m, reset its aspect, move me again and demand compass recalibration.