Joulescope UI might display settings that are not actually in device

[atsju]

Joulescope model

JS220

UI version

other

What OS are you seeing the problem on?

Windows 11

What happened?

UI1.0.42

I had the sampling rate at 50kHz but I'm pretty sure it was actually sampling at 1MHz (I was evaluating the noise level)

What was expected?

UI should always ensure the actual setting are same between display and device

How to reproduce?

1. Open UI and set sampling to 50kHz
2. unplug JS220
3. replug JS220
4. UI still displays 50kHz but JS is 1Mhz

Extra information

No response

[mliberty1]

I was able to successfully duplicate this behavior. The sampling_frequency setting is restored correctly when the device first connects. However, it is not restored correctly for subsequent disconnect-reconnects.

I captured the following log snippet:

INFO:2023-11-16 10:34:03,104:jsdrv_stream_buffer.py:223:joulescope_ui.devices.jsdrv.jsdrv_stream_buffer.001:device_close JS220-001707
...
INFO:2023-11-16 10:34:09,962:backend.c:737:jsdrv:device_add_msg u/js220/001707 : \\?\usb#vid_16d0&pid_10ba#001707#{e27188c8-98ff-41de-be50-653324c6b19c}
INFO:2023-11-16 10:34:09,962:jsdrv.c:457:jsdrv:device add sn=u/js220/001707
...
INFO:2023-11-16 10:34:10,105:jsdrv_stream_buffer.py:211:joulescope_ui.devices.jsdrv.jsdrv_stream_buffer.001:device_open JS220-001707
INFO:2023-11-16 10:34:10,105:jsdrv_stream_buffer.py:353:joulescope_ui.devices.jsdrv.jsdrv_stream_buffer.001:add JS220-001707.i
INFO:2023-11-16 10:34:10,105:js220.py:736:joulescope_ui.devices.jsdrv.device.u.js220.001707:firmware_update_available JS220-001707: up to date
INFO:2023-11-16 10:34:10,105:js220.py:691:joulescope_ui.devices.jsdrv.device.u.js220.001707:current_range auto
INFO:2023-11-16 10:34:10,105:js220.py:609:joulescope_ui.devices.jsdrv.device.u.js220.001707:setting(Device(u/js220/001707)): signal_frequency <= 50000
INFO:2023-11-16 10:34:10,105:js220.py:609:joulescope_ui.devices.jsdrv.device.u.js220.001707:setting(Device(u/js220/001707)): statistics_frequency <= 2
INFO:2023-11-16 10:34:10,105:js220.py:609:joulescope_ui.devices.jsdrv.device.u.js220.001707:setting(Device(u/js220/001707)): target_power <= True

The UI is doing the "right thing" in attempting to restore the signal_frequency parameter value.

The Joulescope UI uses self._driver.open(self._path, 'restore') to open the pyjoulescope_driver device. 'restore' gets converted to open mode 1 by pyjoulescope_driver.binding.Driver.open which is JSDRV_DEVICE_OPEN_MODE_RESUME. This gets all settings values from the instrument, which it does here.

However, sampling_frequency is "special" for now in that the joulescope_driver knows it as h/fs, which is implemented on the host, not the instrument. The long-term goal is to implement downsampling on the instrument. With the existing host-based implementation, restoring the value from the instrument does nothing, so the pubsub retained value remains but without being reprocessed. The driver needs to be reconfigured correctly using the retained value. The joulescope_driver.js220_usb.d_open should call on_sampling_frequency with the retained pubsub value for h/fs.

[mliberty1]

Fixed in jetperch/joulescope_driver@40e6342 and will be included in the next release.

[mliberty1]

Fixed in 1.0.45 (currently alpha).