The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> Circuit Simulators >> Viewing innards of Spectre netlists/primitives
https://designers-guide.org/forum/YaBB.pl?num=1091574316

Message started by aznaragorn on Aug 3^rd, 2004, 4:05pm

Title: Viewing innards of Spectre netlists/primitives
Post by aznaragorn on Aug 3^rd, 2004, 4:05pm

I am running mixed level simulation combing behavioral Verilog AMS code with Spectre Netlists (using modelpath variable) and the only problem I have is the spectre netlist subcircuits behave like a black box; i can't view any of the voltages and currents within it. For example, say I was co-simulating a gain amplifier with a filter and the amplifier was a subcircuit of a spectre netlist. Say i wanted to view the voltage on a particular transistor within an instantion of amplifier, how would I do this?
Basically i cannot look at anything within the spectre netlist (subcircuits), I can only look at the inputs and outputs of the subcircuits I instantiate. Any ideas?

Title: Re: Viewing innards of Spectre netlists/primitives
Post by Andrew Beckett on Aug 3^rd, 2004, 9:48pm

I think this should work... I've not done it for a while, but I think it should be OK:

In the analog control file, you can add lines:

Code:

saveNodes options rawfmt=sst2 rawfile="trans1.tran"
save top.block1.block2.sigName1 top.block1.block2.sigName2

You can then save (into another file) hierarchical paths into subckt blocks.

However, why are you including spectre netlists? If those spectre netlists came from Composer schematics, would it not be better to get them in Verilog-AMS format (the AMS Designer interface will do this for you)? That way you can see everything in simVision.

Regards,

Andrew.

Title: Re: Viewing innards of Spectre netlists/primitives
Post by aznaragorn on Aug 4^th, 2004, 3:04pm

Unfortunately much of our schematics use 3 terminal devices and because progn isn't supported by the spectre netlister our spectre and verilog ams netlisters do not work properly (the fourth terminal of all the three terminal devices is being mapped to something like vdd! or vss! instead of Bn). Our company uses our own netlister. I've worked out a fix (based on changing the spectre stopping view and changing the parameter Bn to a netexpression) but we have legacy issues with it that we are working out. So right now the best solution is to use our own netlister and use modelpath to utilize the netlists.

Title: Re: Viewing innards of Spectre netlists/primitives
Post by Andrew Beckett on Aug 4^th, 2004, 3:22pm

If you contact customer support, they should be able to tell you about solution 11004934 (an internal solution that I wrote a while ago), which describes how you can write a custom netlist procedure for 3 terminal devices in the spectre interface which emulates the progn(bn) behaviour. I advise the use of this for compatibility with existing designs, but that all new designs should use inherited connections instead.

The same principles in that solution could also be used to write a custom AMS netlisting procedure.

Regards,

Andrew.

Title: Re: Viewing innards of Spectre netlists/primitives
Post by aznaragorn on Aug 10^th, 2004, 1:53pm

We have used your solution to imitate progn behavior for Spectre netlisting. However, for AMS netlisting there is no field in the Siminfo for ams view for netlistProcedure so I don't think a custom netlist procedure can be used. The ams simulator also seems to say that this is impossible. Taken from the AMS Simulator manual:

For traditional Cadence netlisters, which are usually written in Cadence SKILL language, a SKILL procedure that formats instances can be named in the netlistProcedure field for a device. The AMS netlister is not written in SKILL and cannot use a customized procedure for devices. Specifying a procedure does not affect the formatting of instances because the AMS
netlister does not recognize the procedure. New fields provided in the ams section provide the mechanism for specializing device instantiation formatting.

Title: Re: Viewing innards of Spectre netlists/primitives
Post by Andrew Beckett on Aug 19^th, 2004, 10:24am

Support for netlist procedures for AMS was added in IC5033.

So you need to use IC5033 or later.

Andrew.