Assembler Direct Plot add to outputs

One of my colleagues also noticed this behavior regarding spm('sp 1 1) last year. CCR 2538096 has been filed for this, its current state is "Characterized".

We have also observed the update problem in Assembler but apparently never bothered to report it because the workaround was so easy. According to https://support.cadence.com/apex/ArticleAttachmentPortal?id=a1O0V0000091BsWUAU this may have been fixed in IC618 ISR30 by CCR 2727624.

It seems CCR 2727624 'ADE Assembler does not refresh the Outputs Setup tab' is not fixed in IC618 ISR31 which I am using.

Btw I agree the workaround for the update problem is not difficult (although I did waste a bit of time figuring it out initially) but the necessity of using a bunch of work arounds has a cumulative effect which negatively impacts a designer's efficiency and "user experience" with the tools. As an seasoned designer I already know how to setup Assembler for a wide variety of tests, so the next valuable metric is UX and efficiency using the tools and interacting with data.

Brad RFeng said:

It seems CCR 2727624 'ADE Assembler does not refresh the Outputs Setup tab' is not fixed in IC618 ISR31 which I am using.

The CCR might have been for a different problem, we only know the title.

Brad,

The duplicate output names looks a bit odd, and the refresh issue is definitely a bug - please report that to customer support; it's probably a different corner case than what was fixed already (there have also been some refresh problems related to the detection of invalid output node names which have been fixed in recent ISRs, but anything outstanding I would recommend creating a new case).

When you use the "add to outputs" with the s-parameter plots, if I remember rightly there is some additional "hint" information data that is saved in the output structures which are not user-accessible (that's what Frank's CCR is related to). Rather than over-complicate the output UI to allow setting of how exactly you want each output to be plotted, the recommendation is to use the plotting templates capability in ADE Explorer/Assembler as this allows you to set up the outputs graphically and get them the same way next time.

Regards,

Andrew

Andrew,

Unfortunately templates won't work in Append mode and seem to not work across tests in many cases. Fixing the Append mode to work with templates is desperately needed.

Also, a simple but valuable enhancement would be to include a check box on the Assembler Plotting/Printing Options form to select plot type (R, Z,Y, ZY, P) for s-parameters (or any complex expression) when not using a template. This would result in all complex expressions going using the same type of plot but at least the user gets to decide which type. When using a template its settings would override the plotting/printing options form. Long ago when the Assembler plotting scheme was coded up, someone made a poor decision to have complex values default to a rectangular magnitude plot type. I'm willing to bet that person was not an RF designer.

Regards
Brad

I agree with Brad RFEng. I gave up using plotting templates since they are not usable for my applications. That's unfortunate.

What Brad RFeng asks for is very specific to his particular needs. The trouble is that to support all the different possible ways of plotting stuff, you'd end up with an enormoform (should I trademark that?) with lots of unwieldy options to control what you'd want. Then people would want to be able to specify this per-output so that they could plot both Z and Y parameters (say) for different outputs. The whole point with plotting templates is that it should give you the flexibility to do what you want.

Of course, there may be gaps, and those should be reported to us.

There is a CCR (with a number of duplicates) asking for support for plotting templates in append mode (CCR 1948714). A lot of work has gone into that CCR but it's not been released yet because a few issues were run into as part of the development (some of the development is quite recent). It's actually a complicate problem - trying to make sure that what you're appending into is compatible with the plotting template - after all, you may have rearranged the subwindows and removed stuff, and so what is the correct thing to do if the previous plot has been changed with respect to the plotting template. Replace and New Window are much easier of course, because they don't have to add to existing data/arrangements.

Anyway, I'd suggest that you contact customer support and request a duplicate for append mode explaining the need - that all helps with prioritisation.

I'm not sure what issues you had TommasoF, but I suggest you contact customer support as otherwise we won't know why they don't work for you.

Andrew

Hi Andrew ,
what I recall is something similar to what Brad mentioned, that across Tests the plotting template does not work anymore. To be specific in order to open a ticket I would need to go back and track down the issues I had, but I recall that were "basic" limitations, so I was not doing anything advanced with the tool.

If I can make an example based on my memories:
I have a maestro assembler view with 6 Tests (LNA1, LNA2..LNA6) which they all produce the same input/output, just the DUT is different.
Then for Test "LNA1" I create a template where I plot some formulas like S11, S22, S21, Gp and so on... 
 If I desire to reuse the template for Tests "LNA2" and the others then it does not work, since the template refers somehow to "LNA1". 

Ideally I would like to have a template which is Test, History (and so on...) agnostic. At the end of the day the S11 of a simulation result is the same regardless of which test/history is used, instead it seems to make a difference which eventually cause an evaluation error.

Sorry for not being too specific.
Tommaso

There's a request (I think there might be others) for plotting templates which are per-test (or rather don't have the test coded in) - CCR 2022003. I understand that - it makes sense. They already are history-agnostic (unless you are including signals from an older history as a reference). Of course, it gets complicated - because of potential ambiguities as to what you want to do.

This is the general challenge with plotting templates as an idea - they work well in general, but the complexity comes when you change something. A lot of work was done to handle changes in corners and so on, but altering tests is another level of complexity. It's not that trivial to adapt the current architecture to handle that, but I do see why you would want this. 

Andrew

To avoid the enormoform(tm) but still make an improvement to the poor choice of sending complex expressions to a rectangular mag plot by default, a simple enhancement to the Assembler Plotting/Printing form could be made. Just need a check box for Plot Type which would apply only to complex expressions. At an absolute minimum, skip the Plot Type enhancement and just send complex expressions to a Z-Smith chart.