The I2S generator of the QA403 is messed up after using PUT /Settings/Default.
The way to solve t is to switch off I2S wait for 250msec, switch on I2S and wait again 250msec.
I found this issue when using a python scrypt to do THD=f(Pout) measurement.
Use v 1.2.2.
Behavior with the python script fix:
Result is OK:
Now without the I2S gen odff/on and the delays.
Result is not good.
Is this a known issue?
Hi @JP-Huijser I think the expectation should be that the I2S is off once you issue a /Settings/Default
Sounds like you are saying it’s still on but broken/nonsensical?
I noted in another post there’s a lot of re-work happening with the front panel I2S (on the next product, the I2S moves to a DB9 on the rear panel, but the operation will be the same). The I2S will mirror the main generator for everything. The only question is whether it’s on and the bit depth. So, if you have a multitone sine selected for the main generator, then the I2S will be multitone sine too. Now, this doesn’t solve your immediate problem, but I point this out to say a lot of time won’t be invested in the current I2S operation because what is coming is very different.
But I will check to see about your current problem as it shouldn’t be working this way. Thanks very much for reporting this!
Hi Matt,
yes indeed. It remains on but the data out is messed up. To reset behavior I need to program I2S off wait a little and program I2S on again, wait a little and start testing. Then ok. See also code snippet.
So for me it works now.
By the way, some other questions:
is there also a put command for setting the fft averaging on and the number of averaging?
also would be good to have an automatic function to trigger the QA403 to find/get the input signal for reaching 1% thd. Since the supply is weak this needs to be an itterative process. Run this now in python myself. Goal is to make a graph Pout1% =f(Vsupply) which is a common graph. I basically loop in python the supply (using pyvisa) and search for 1% power.
Hi @JP-Huijser
I think the smarts that could be built into the various REST commands is enormous. But the thinking right now is since you are writing code the controlling everything, things like averaging and iterating are very easy to do–just a few lines of code. I think that as soon as very complex routines were added to REST, there would be requests to allow each of these routines to be tweaked. Everything from averaging (lin/log, number of average) to graph colors, to line thicknesses….it would be never ending. And so, the best way to approach IMO is to provide the basics, and then the user can unleash modern numeric and graphing libraries on their particular problems.
Thanks for the additional details on the I2S bug!
Hi @JP-Huijser, in release 1.221 I think when you issue a REST ‘defaults’ now the I2S should revert to its off state instead of that indeterminant state. Please confirm at your convenience.
Thanks, Matt
Hi Matt, Thank. Will test tomorrow and will let you know. Bye.
Hi Matt,
Yep problem solved with v1.221.
Thanks for fixing it.
BR Jan-Paul
