After some frustration with the Logic Analyzer implementation in the Rigol MSO2072A ( see other thread on the forum), I dared to write a negative review in the Tequipment Review section.
Apparently that got Rigol's attention and to my surprise I actually got an answer from Rigol. I sent numerous reports to Rigol directly with these issues and never got much info back except as "the developers are working on it."
Here is what I wrote in my review:
"The decoding fails if the memory setting goes beyond 1.4MPoints or if the sampling rate is set to low values which renders this function and the deep memory entirely useless.
The timing interface does not allow a setting of the sampling frequency which is bizarre
When downloading the LA data, no command is provided to receive the Logic Analyzer sampling rate and it cannot be determined from the Analog Signal sampling rate
Over all the logic Analyzer is very poorly implemented and not worth the additional $400.-. Rigol is aware of the decoding issues since over a year and hasn't fixed them."
Rigol Response:
The decoding is done entirely on the displayed data. You only decode what you see. The deep memory is not used for decoding.
The sampling rate is set primarily by the time per division and the number of active channels. In the context of decoding you simply want to make sure you can view the serial data stream clearly on the screen.
You can use the :WAVEFORM:XINCREMENT? command to query the time between samples. Simply set :WAVEFORM:SOURCE to D0 (or other Digital channel) and query the Xinc value in seconds. The value here also depends on the waveform mode. So, you get the time between samples you are about to receive. That time is higher if you are about to view the displayed data than if you ask for the deep memory.
Rigol is constantly making improvements and has released a number of recent firmware versions that improve decoding usability and features, but architecturally we only decode the data on the screen. The deep memory can be pulled from the scope for further analysis offline, but real time analysis is about what you can view on the screen. That is the best way to use the digital capabilities of the MSO2000A series."
Unfortunately you cannot answer to this response. So I thought I bring this here for everybody's enlightenment.
For the first time I hear decoding is not intended for the deep memory!!!! However, the specifications or web page never mention this anywhere. At least I couldn't find any information or warning. So decoding only works when the MemDepth is set to Auto and when you can see the data on the screen and when it doesn’t use too much memory. I haven't tested this yet fully, but whatever... I wonder for what reason Rigol then implemented the EventTable which lets you scroll through all the data. Even in Auto-mode at a certain memory depth the decoding will fail. Also, you cannot chose the memory depth directly as it depends in Auto-mode on the time scale setting. Considering this, it is highly questionable whether the $400 price difference is worth it as I stated above.
Now to the second issue regarding the timing interface: Rigol partially stepped over this issue and showed that there is a work around solution for the query issue which is certainly appreciated. However, my first complaint was related to the interface that was implemented for setting the sampling frequency. While the scope displays both values SaRate and LASaRate, both values are shaded and cannot manually be altered. That might be ok for manual selection, but using SCPI you are lost and need a really complicated series of commands. It is nice to hear that you can get the LASaRate back into the scope with the :WAVEFORM:XINCREMENT? command. (BTW, don’t bother using ":WAVE:SOURCE D0“ as explained as this command doesn’t work with "D0". It needs to be set to LA and the scope needs to be in STOP mode or you won’t get the sample rate but just time scale setting. The programming guide does not tell you this)
I really wonder why I didn’t hear any of this from Rigol directly and I asked for this multiple times. But if you complain on a seller website then that get’s their attention.
It is however true that Rigol brings out updates and fixes certain things from time to time. But it took a full year to get the download SCPI command to work.
So now you know it officially, there is no decoding with deep memory on your MSO2k. The optional (not inexpensive) decoding option can't handle that and it was apparently never intended to do so. I guess we can write a fix for this off (which is on the other hand probably not such a big deal as long as there is no search function).
atx