Cannot set "ai_conv_rate" on NI DAQ 9209 due to missing active device modifier for timing attributes

[CleSpi]

To avoid ghosting during switching, we need to increase the ai_conv_rate. However, the standard command "task.timing.ai_conv_rate = 1.1 * sampling_rate * n_channels" does not work on the DAQ model 9209.

AB#3064732

[zhindes]

This is likely because setting the convert rate on a cDAQ module requires specifying an active device on the timing property, which the Python API does not support.

Quoting some internal information:

As a workaround, it is possible to use the internal "interpreter" interface to call the set_timing_attribute_ex_double method, but this is not a stable API and future versions of nidaqmx-python may break compatibility with it: https://github.com/search?q=repo%3Ani/nidaqmx-python%20set_timing_attribute_ex_double&type=code

# Unsupported workaround, may break in future releases
task._interpreter.set_timing_attribute_ex_double(task._handle, "cDAQ1Mod1", 0x1848, 100.0)

[CleSpi]

Yes, I am aware of this workaround. However, as you mention, it may not work in future versions.

[bkeryan]

If we follow the design of the DAQmx .NET API, implementing this feature will involve making the Timing class indexable with a string:

task.timing["cDAQ1Mod1"].ai_conv_rate = 1.1 * sampling_rate * n_channels

We would add an _active_devs field to the Timing class and a Timing.__getitem__ method that returns a new instance of Timing with _active_devs set to the index. Some of the get/set/reset methods in the Timing class would have a condition to call the get|set_reset_timing_attribute_ex method if _active_devs is set. We wouldn't want to use the ex methods in the normal case because they support only a very small number of attributes (basically the AI convert clock and delay from sample clock attributes). Get/set/reset methods that don't support active devs would error if it's set.

[davitdan]

@bkeryan And if I am not mistaken, the getter for this func is not implemented either right? (get_timing_attribute_ex_double)

[bkeryan]

@davitdan It's implemented for LibraryInterpreter: https://github.com/ni/nidaqmx-python/blob/master/generated/nidaqmx/_library_interpreter.py#L3572
and GrpcInterpreter: https://github.com/ni/nidaqmx-python/blob/master/generated/nidaqmx/_grpc_interpreter.py#L2135

BaseInterpreter is an abstract base class, so all of its methods are stubs that raise NotImplementedError.

[zhindes]

Still in work.