Stimulus File Issues

There's a few issues here. First of all, stimulus files are not directly included into the simulator - they go through a process of mapping schematic names to simulator names (mostly this was useful in the days of flat netlisters when there was no clear relationship between the name on the schematic and the name in the netlist). That's achieved by using [#...] to map nets or [$...] to map instance names. For more info on this search in cdnshelp (which you can launch via the help menus in Virtuoso) for "stimulus file" or "Template Control File" (include the quotation marks in the second) to find the section which describes the syntax.

In your case you've escaped the open square parenthesis, which means that the name translation doesn't happen, and the simulator gets the the literal square brackets - ([#LIGHT<0>] 0) - which is invalid spectre syntax. That's not what you wanted. So (in theory) you should have just omitted the backslash, but unfortunately there's a bug (CCR 752498 - has been there for some time) where the default mapping mechanism doesn't take into the account that it needs to add backslashes before the < and > in the resulting name, so it still fails, but in a different way. The older mapping mechanism that was used in the distance past worked OK, but since it was changed many years ago, it's been broken. The fact that it hasn't been fixed yet is an indication that not many people rely on this mapping mechanism. It is possible to tell ADE to revert to the older name mapping in the netlist, but I wouldn't recommend that in general just to fix this issue.

By the way, you can click on the hyperlinks in the spectre log file and see the mapped file - you will see that it doesn't directly include your file, but instead includes a copy which has been copied into the netlist directory after mapping has been performed.

So, the solution is just to use the names that are in the netlist - and then you can either specify the stimulus via the stimulus file or a model library or definition file (since you don't need it to do any mapping any more, you can also use model library or definition file). If you use model library or definition file to include it, it's a bit simpler:

simulator lang=spectre
LIGHT1 (LIGHT\<0\> 0) vsource dc=0.882353n
LIGHT2 (LIGHT\<1\> 0) vsource dc=1.217647n
LIGHT3 (LIGHT\<2\> 0) vsource dc=1.105882n
LIGHT4 (LIGHT\<3\> 0) vsource dc=0.988235n
LIGHT5 (LIGHT\<4\> 0) vsource dc=0.205882n
LIGHT6 (LIGHT\<5\> 0) vsource dc=0.270588n
LIGHT7 (LIGHT\<6\> 0) vsource dc=0.276471n
LIGHT8 (LIGHT\<7\> 0) vsource dc=0.258824n

If you use stimulus file still, you have to escape the escape, because the mapping process removes the first backslashes:

simulator lang=spectre
LIGHT1 (LIGHT\\<0\\> 0) vsource dc=0.882353n
LIGHT2 (LIGHT\\<1\\> 0) vsource dc=1.217647n
LIGHT3 (LIGHT\\<2\\> 0) vsource dc=1.105882n
LIGHT4 (LIGHT\\<3\\> 0) vsource dc=0.988235n
LIGHT5 (LIGHT\\<4\\> 0) vsource dc=0.205882n
LIGHT6 (LIGHT\\<5\\> 0) vsource dc=0.270588n
LIGHT7 (LIGHT\\<6\\> 0) vsource dc=0.276471n
LIGHT8 (LIGHT\\<7\\> 0) vsource dc=0.258824n

If  you want to get the real mapping issue fixed, I suggest you contact customer support and ask for a duplicate of CCR 752498 to be filed on your behalf.

Regards,

Andrew.