Initial review of Owon virtual (USB) oscilloscopes

I purchased two Owon virtual (USB) oscilloscopes recently:

  • RDS1021i; and
  • VDS1022i.

RDS1021i

The RDS1021i is a single channel ‘pen scope' with 25MHz bandwidth, and the i suffix denotes isolation of the USB ground from the instrument ground.

The supplied application software is fairly good, but has some usability issues:

  • the application opens with a fixed (ie non configurable) window size that may be larger than the physical screen size;
  • the application does not persist the window size to new application sessions;
  • the supplied driver (libusbK) and application combination are very flaky, the driver often disconnects and the scope has to be unplugged and reinserted to restart operation (we are warned that libusbK is not tolerant of poor usb implementations);
  • USB operation was less reliable on a cascaded USB3 hub;
  • on two of three hosts tested, the RDS1021i is identified as “Unknown device #1” in Windows;
  • True triggered operation is not possible with the timebase set below 50ms/div (1s total sweep), it forces ‘auto' trigger mode disabling true triggered display, and as a result, disabling single sweep. The ‘auto' mode is actually unsynchronised above 50ms/div. Though the menu permits single mode at up to 50ms/div, it isn't reliable above 2ms/div.
  • The ‘update online' facility does not work.

A good time base trigger facility is essential to effective utilisation of any DSO, virtual or not.

Experimentation with libusb v1.2.6 seemed to resolve the USB driver flakiness, this can easily be installed using zadig.

The roller ball device is rather dicky to use.

VDS1022i

The VDS1022i is a two channel scope with 25MHz bandwidth, and the i suffix denotes isolation of the USB ground from the instrument ground.

The supplied application software is fairly good, but has some usability issues:

  • the application opens with a fixed (ie non configurable) window size that may be larger than the physical screen size;
  • the application does not persist the window size to new application sessions;
  • the supplied driver (libusbK) and application combination are very flaky, the driver often disconnects and the scope has to be unplugged and reinserted to restart operation (we are warned that libusbK is not tolerant of poor usb implementations);
  • USB operation was less reliable on a cascaded USB3 hub;
  • on two of hosts tested, the VDS1022i is identified as “Unknown device #1” in Windows;
  • True triggered operation is not possible with the timebase set below 50ms/div (1s total sweep), it forces ‘auto' trigger mode disabling true triggered display, and as a result, disabling single sweep. The ‘auto' mode is actually unsynchronised above 50ms/div.
  • The ‘update online' facility does not work.

A good time base trigger facility is essential to effective utilisation of any DSO, virtual or not.

Experimentation with libusb v1.2.6 seemed to resolve the USB driver flakiness, this can easily be installed using zadig.

RDS and VDS together

A test was conducted to see if both devices and their respective applications are compatible for concurrent operation.

On one of three hosts tested, the RDS1021i and VDS1022i both identified as “Oscilloscope” in Windows, they did not identify by model number. The applications did not automatically select the compatible device type, and present a list of identically named devices for the user to choose for each one.

It would be nicer if a single application supported both hardware types and effectively mapped the devices to different channels, but that is not so. Diagnostic messages suggest that each application, although they look the same, load their respective device specific FPGA images to the attached device (rather than recognising the device type and loading the relevant image).

Conclusions

Assuming that the issues that exist with these two products exist also with the VDS 100MHz 4 channel devices which are priced at $800+, I would not be even thinking of them unless there was some specific need for the PC integration.

A good time base trigger facility is essential to effective utilisation of any DSO, virtual or not… and both devices fall well short of the mark. Lack of performance in the timebase area makes channel performance somewhat irrelevant.

The application software and firmware on the tested device is a year or more old which hints a lack of interest by the manufacturer in fixing significant problems. Their web site hints a lack of support resources.

The fact that it works better with libusb0 than libusbK might indicate a non-compliant usb implementation.

There is no doubt that they are both much better than commonly available sound card oscilloscopes without an interface unit, and these low end models at modest prices might be good value for some applications, but they fall a long way short of a good quality DSO.