Never miss a story from Analog/Custom Design. Subscribe for in-depth analysis and articles.
Spectre EMIR, the simulation engine inside Voltus-XFi, provides the IR drop and EM current analyses. In reviews of the reported customer problems, it turns out that many Spectre EMIR problems can be avoided by proper preparation and setup. The most common problem Spectre EMIR customers report is iterated method IR drop and EM current inaccuracies when compared against the golden direct EMIR method. Often, these accuracy problems are caused by improper preparation or wrong setup. This blog discusses two typical problems and explains how to address them.
The DSPF netlist is the post-layout netlist used in Spectre EMIR. It is included in the Spectre netlist with the dspf_include statement. In many cases, the DSPF file content has different types of problems, such as wrong connectivity, wrong net or element names, missing content, or a mismatch between RC and instance section.
Cadence provides the spfchecker utility as part of the Spectre installation. It is recommended to always run spfchecker on a new DSPF file.
After spfchecker is run, check the Message Statistics section in the *.chklog file. If this section contains any error message, then the DSPF netlist has problems that must be fixed in extraction.
There is no sense in running a Spectre EMIR simulation when the DSPF file has problems being reported by spfchecker. Spectre will create garbage results when a bad DSPF file is used as input. If you have questions on how to address DSPF netlist problems, contact Cadence Customer Support.
Many of the EMIR cases evaluated by the Spectre R&D team have multiple DC solutions. That means the selected nodes have multiple valid DC solutions at time=0. Often, we observe that these different DC solutions cause different transient behavior, which then leads to different IR drop and EM results when comparing iterated with direct method.
If you want to compare the DC solution differences between direct and iterated EMIR method, then use the spw script shipped with the Spectre installation.
spw ic direct.ic iterated.ic -d
The report appears, as shown below.
We can observe that the nodes a, b, and c have a significant value difference between the direct and iterated methods. Most likely, it means that these nodes have multiple DC solutions and that the user needs to define which solution should be used.
The recommended way is to discuss multiple DC solutions problems with the designer, and to address them in the setup by setting an IC statement for the relevant nodes. Nevertheless, if the designer is not available, then we can just use the same IC file for both direct and iterated EMIR simulation. Commonly, we create the IC file for the iterated method simulation.
tr1 tran stop=100n write=″golden.ic″
Afterwards, we use this IC file for any of the subsequent direct or iterated method simulations.
tr1 tran stop=100n readic=″golden.ic″
This approach helps to void an EMIR inaccuracy caused by the multiple DC solutions problem.
Besides addressing these two common setup problems, there are a few further recommendations for Spectre EMIR users:
More details about the above recommendations can be found in the recently published application note Addressing Common Spectre EMIR Problems.
Following the guidance and avoiding the discussed problems will enable you to quickly qualify and use the Spectre EMIR solution in Voltus-XFi.
You may also contact your Cadence support team for guidance.
For more information on Cadence products and services, visit www.cadence.com.
Spectre Tech Tips is a blog series aimed at exploring the capabilities and potential of Spectre® circuit simulator. In addition to providing insight into the useful features and enhancements in Spectre, this series broadcasts the voice of different bloggers and experts who share their knowledge and experience on all things related to Spectre. Click Subscribe in the Subscriptions box to receive notifications about our latest Spectre Tech Tips posts.