"Acceleration with g" is not drawn after clearing

[bodqhrohro]

1. Open "Acceleration with g".
2. Start measurements.
3. Wait for 40 seconds until the graphs stop growing.
4. Pause the measurement.
5. Clear the data.
6. Start measurement again.

[Staacks]

Could not reproduce on an Pixel 6 with the current version on the Play Store.

But I think that "Wait for 40 seconds until the graphs stop growing." is already a problem. It should not stop growing, but continue to measure indefinitely. (In practice until the phone runs out of memory, which sometimes comes sooner than one thinks, but certainly not after 40 seconds).

Could you post some details about your device, so I can try to reproduce with the same Android version or even the same device (if I can find one)? The simplest way would be to post the "device info" (the "i"-icon in the main menu of phyphox).

Also, did this work before? Do the other sensors work without that problem (maybe after 200s as for example the accelerometer "without g" produces data at about a fifth of the rate of the one "with g" on some devices).

[bodqhrohro]

phyphox
Version: 1.1.10
Build: 1011002
File format: 1.14

Permissions
HIGH_SAMPLING_RATE_SENSORS: no
INTERNET: yes
RECORD_AUDIO: yes
CAMERA: no
READ_EXTERNAL_STORAGE: no
BLUETOOTH: yes
BLUETOOTH_ADMIN: yes
BLUETOOTH_SCAN: no
BLUETOOTH_CONNECT: no
ACCESS_COARSE_LOCATION: no
ACCESS_FINE_LOCATION: no
ACCESS_NETWORK_STATE: yes
WAKE_LOCK: yes
SCHEDULE_EXACT_ALARM: no

Device
Model: HRY-LX1
Brand: HONOR
Board: HWHRY-H
Manufacturer: HUAWEI
ABIS: arm64-v8a, armeabi-v7a, armeabi
Base OS: HONOR/HRY-LX1/HWHRY-H:9/HONORHRY-L21/262C461:user/release-keys
Codename: REL
Release: 9
Patch: 2020-01-01

Sensors

Accelerometer (type 1)
- Name: rohm-kx023
- Range: 78.4532 m/s²
- Resolution: 9.576806E-6 m/s²
- Min delay: 5000 µs
- Max delay: 200000 µs
- Power: 0.23 mA
- Vendor: KIONIX
- Version: 1

Magnetic Field (type 2)
- Name: mag-akm09918
- Range: 2000.0 µT
- Resolution: 0.0625 µT
- Min delay: 10000 µs
- Max delay: 200000 µs
- Power: 6.8 mA
- Vendor: akm
- Version: 1

unknown (type 3)
- Name: orientation
- Range: 360.0 
- Resolution: 0.1 
- Min delay: 10000 µs
- Max delay: 20000 µs
- Power: 13.0 mA
- Vendor: huawei
- Version: 1

Light (type 5)
- Name: als-B
- Range: 10000.0 lx
- Resolution: 1.0 lx
- Min delay: 0 µs
- Max delay: 0 µs
- Power: 0.75 mA
- Vendor: B
- Version: 1

Proximity (type 8)
- Name: proximity-tmd3702
- Range: 5.0 cm
- Resolution: 5.0 cm
- Min delay: 0 µs
- Max delay: 200000 µs
- Power: 0.75 mA
- Vendor: ams
- Version: 1

Gyroscope (type 4)
- Name: gyroscope
- Range: 34.906586 rad/s
- Resolution: 1.7453292E-5 rad/s
- Min delay: 5000 µs
- Max delay: 200000 µs
- Power: 6.1 mA
- Vendor: STMicroelectronics
- Version: 1

unknown (type 9)
- Name: gravity
- Range: 9.80665 
- Resolution: 0.15328126 
- Min delay: 10000 µs
- Max delay: 20000 µs
- Power: 0.2 mA
- Vendor: huawei
- Version: 1

Linear Acceleration (type 10)
- Name: linear Acceleration
- Range: 78.4532 m/s²
- Resolution: 0.009576807 m/s²
- Min delay: 10000 µs
- Max delay: 20000 µs
- Power: 0.2 mA
- Vendor: huawei
- Version: 1

Attitude (type 11)
- Name: rotation Vector
- Range: 1.0 
- Resolution: 5.9604645E-8 
- Min delay: 10000 µs
- Max delay: 20000 µs
- Power: 6.1 mA
- Vendor: huawei
- Version: 1

Vendor specific (type 65538)
- Name: HALL sensor
- Range: 1.0 
- Resolution: 1.0 
- Min delay: 0 µs
- Max delay: 0 µs
- Power: 0.75 mA
- Vendor: huawei
- Version: 1

Magnetic Field (type 14)
- Name: uncalibrated Magnetic Field
- Range: 2000.0 µT
- Resolution: 0.0625 µT
- Min delay: 16667 µs
- Max delay: 200000 µs
- Power: 6.8 mA
- Vendor: Asahi Kasei Microdevices
- Version: 1

unknown (type 17)
- Name: significant Motion
- Range: 2.14748365E9 
- Resolution: 1.0 
- Min delay: -1 µs
- Max delay: 0 µs
- Power: 0.23 mA
- Vendor: huawei
- Version: 1

unknown (type 19)
- Name: step counter
- Range: 2.14748365E9 
- Resolution: 1.0 
- Min delay: 0 µs
- Max delay: 0 µs
- Power: 0.23 mA
- Vendor: huawei
- Version: 1

unknown (type 20)
- Name: geomagnetic Rotation Vector
- Range: 1.0 
- Resolution: 5.9604645E-8 
- Min delay: 10000 µs
- Max delay: 200000 µs
- Power: 6.1 mA
- Vendor: huawei
- Version: 1

Vendor specific (type 65552)
- Name: RPC sensor
- Range: 65535.0 
- Resolution: 1.0 
- Min delay: 500000 µs
- Max delay: 1000000 µs
- Power: 5.0 mA
- Vendor: huawei
- Version: 1

unknown (type 15)
- Name: Game Rotation Vector Sensor
- Range: 1.0 
- Resolution: 5.9604645E-8 
- Min delay: 5000 µs
- Max delay: 0 µs
- Power: 13.13 mA
- Vendor: AOSP
- Version: 3



Cameras

Depth sensors
- Depth sensors front: 0
- Max resolution front: -
- Max frame rate front: -
- Depth sensors back: 0
- Max resolution back: -
- Max frame rate back: -


Camera 2 API
[{"id":"0","facing":"LENS_FACING_BACK","hardwareLevel":"HARDWARE_LEVEL_LIMITED","capabilities":["CAPABILITIES_BACKWARD_COMPATIBLE","CAPABILITIES_MANUAL_POST_PROCESSING","CAPABILITIES_RAW","CAPABILITIES_LOGICAL_MULTI_CAMERA"]},{"id":"1","facing":"LENS_FACING_FRONT","hardwareLevel":"HARDWARE_LEVEL_LIMITED","capabilities":["CAPABILITIES_BACKWARD_COMPATIBLE","CAPABILITIES_MANUAL_POST_PROCESSING"]},{"id":"2","facing":"LENS_FACING_BACK","hardwareLevel":"HARDWARE_LEVEL_LIMITED","capabilities":["CAPABILITIES_BACKWARD_COMPATIBLE","CAPABILITIES_MANUAL_POST_PROCESSING","CAPABILITIES_RAW"]},{"id":"3","facing":"LENS_FACING_BACK","hardwareLevel":"HARDWARE_LEVEL_LIMITED","capabilities":["CAPABILITIES_BACKWARD_COMPATIBLE","CAPABILITIES_MANUAL_POST_PROCESSING","CAPABILITIES_RAW"]}]

In practice until the phone runs out of memory, which sometimes comes sooner than one thinks, but certainly not after 40 seconds

Yeah, this time it got stuck after 80 seconds. And I couldn't reproduce it neither in "Acceleration (without g)", nor in other sections I tried.

did this work before?

I don't know, I try it first time.

[Staacks]

Thanks for all the details. Unfortunately, I could not yet find a colleague with a Huawei that reproduces the problem. If you are willing to try a few more things, you could help us narrowing it down:

1. Using the "+" on the main screen and then "create simple experiment", could you please create an experiment with a sample rate of only 50 Hz and the accelerometer and see if it happens in this case, too?
2. Could you try a different app to rule out that this is a bug by the phone itself? For example you could try "Physics Toolbox" (another nice app to access the sensors). In case of Physics Toolbox, please make sure that in the settings menu (gear symbol in top right corner) the sensor rate is set to maximum (further down in the menu).

Thanks a lot!

[bodqhrohro]

Using the "+" on the main screen and then "create simple experiment", could you please create an experiment with a sample rate of only 50 Hz and the accelerometer and see if it happens in this case, too?

Waited for 300 seconds, works well.

The bug in "Acceleration with g" experiment now behaves differently: after a few seconds, the graphs start over from zero and paint over existing ones, with a retrace line like on a CRT:

After that, it doesn't stop even after 200 seconds. When I clear the data and start the experiment again, it still doesn't paint though, like usual.

Could you try a different app to rule out that this is a bug by the phone itself? For example you could try "Physics Toolbox" (another nice app to access the sensors). In case of Physics Toolbox, please make sure that in the settings menu (gear symbol in top right corner) the sensor rate is set to maximum (further down in the menu).

Tried "g-Force Meter" in "Physics Toolbox Accelerometer", Sensor Collection Rate is set to Fastest and oscillates near 200 Hz, couldn't reproduce anything like that.

[Staacks]

That's even weirder. Could you export one of these measurements with the horizontal line for me?

The strange thing is that some things point to a problem with the device and some to a problem with phyphox. The fact that it works smoothly if you set a fixed sensor rate ("+" menu) points to a problem with the device, because the only difference here is that phyphox requests a different rate. But the fact that Physics Toolbox does not have the same problem, of course, strongly suggests that it's a phyphox issue (I don't know their source code, but the "max" setting should make it request the maximum rate like phyphox does by default).

Finally, this weird horizontal line confuses me even more: The device generates sensor events with a timestamp relative to the time the device was turned on. Phyphox subtracts the time at which you press the start button to get a time that starts at zero. From the exported data I would like to see if it jumps back to zero (indicating that phyphox for some reason created a new point of reference, although I do not see how and why) or if it jumps further back and just eventually comes back into view.

This gives me another idea: Does the problem also exist if you use the "+" menu and set the frequency to zero? This should be the same as the default of getting the maximum frequency, so I would expect that this also happens then. If so, could you set the frequency to zero and select two sensors, for example the accelerometer and the gyroscope and see what happens then? Would be very interesting to see if one runs normally at the same time as the other one jumps or stops. (Also, exported data might help).

Best wishes and thanks for helping to narrow this down, although admittedly I am not to sure if I can fix it without a device at hand that reproduces the problem.

[bodqhrohro]

Acceleration with g 2022-08-08 19-28-20.zip
My Experiment 2022-08-08 19-29-27.zip