QA40X Mask File REST API

1

I’m using the REST API to apply some masks with the QA403. When the JSON returns it gives a left/right pass or fail. Is there a way to check what frequency(ies) the waveform failed at?

Also occasionally it will fail to find the file even if the path I’m using in the code and in my files hasn’t changed.

Here’s the error viewed in the console. It only happens rarely. And I can’t think of any operating conditions that would cause it to behave differently.

image
image975×328 127 KB

Here’s a successful call viewed in the console of a similar file.

image
image975×301 119 KB

2

Hi @Isaac, if you keep the setup the same and just run the same request over and over a hundred times, how many errors do you see? Would the situation change if you added a 200mS delay between calls? If you stop your code when the error occurs, and you do a control+space (to run a single cycle manually) does it complete OK? Are you absolutely sure the measurement has completed before you ask for the mask test to be run?

There’s no way to to learn automatically where the frequency failed. But you could pull over the data via code and do the test manually when you get a failure, and learn that way.

3

It’s probably close to around 1 in 100. I can try adding timing to it, and trying the single acquisition after the error.

The acquisition is the synchronous type and from my understanding blocks execution while getting the measurement, so it shouldn’t be possible to request the measurement before its available. There’s also some code the executes between the acquisition and the request for the mask results that get the correct absolute path, though it wouldn’t be close to 200mS.

4

so it shouldn’t be possible to request the measurement before its available.

Technically yes, but there’s a whole bunch of stuff that happens asynchronously inside the software, with many threads coming together to make a measurement and expose it via HTTP.

So, if you add a delay of 100 mS between the acq and the mask processing request and it goes from 1 in 100 to 1 in 1000, then that a good clue on where to start looking and how to repro on this end!

If you add in the delay and nothing changes, then that’s a completely different vector.