deepprobe node name for extracted netlist

Hi Kevin,

The right names would be:

• Itop.Q2\:1
• Itop.I113\/MMP0\:G

The forward slashes and colons need to be escaped (if you leave the colon unescaped then this is illegal syntax in Spectre). Unfortunately it seems that one of the options that ADE passes to spectre interferes with the handling of colons, and that is a bit of a problem... it sees the escaped names as above as unconnected.

There are two choices here:

1. Substitute all the colons in the included file to be something else (e.g.underscore).
2. Add -escchars to the User Command Line Options field on Setup->Environment in ADE

The first is probably my preference; whilst the second works, I'm not entirely sure what it might break by turning "escchars".

Andrew

Hi Andrew,

As you can see from my original post, I already tried substituting colon with underscore in extracted netlist and it doesn't seem to work in my case:

I also try to replace colon with underscore in the netlist file (Q2:1 ->  Q2_1), but below still doesn't work

Itop.Q2_1

thanks,

Kevin

ps. just for the information, if I just probe internal nodes, below still works :

save below to a file:

save Itop.Q2\:1
save Itop.I112\/MMN0\:D

and include that file to "Definition FIles" .  I need deepprobe here because I'd like to apply signals to the nodes

Kevin,

Sorry, I misinterpreted your original post as meaning that you'd just tried replacing the : with _ in the deepprobe property, not having edited the netlist.

I cannot see any reason why it shouldn't work with underscore if you've edited the netlist (I don't know how you're including the netlist, but it ought to work). The deepprobe is quite straightforward (I created it originally) - it simply creates an iprobe instance between the net connected to the pin, and the property passed in. Provided they are both valid hierarchical spectre names, then all should be OK. 

I think to progress this further, an application engineer would need to see what you're doing - maybe you've made a small mistake somewhere along the way. So I'd suggest contacting customer support as I assume you can't provide the complete example here (input.scs plus extracted netlist that is being included).

Andrew.

Hi Andrew,

I'm thinking of alternative solution (before our CAD is back to work on deepprobe support). i.e  for node net111  in the extracted netlist, can I just add pin A and add a 1nOhm resistor between net111 in the netlist, and add A to the end of pinlist of .subckt definition

then what would be the cleanest way to update the main cell ? there needs to be this added new pin. and our schematics are currently locked

thanks,

Kevin

Hi Andrew,

I'm thinking of alternative solution (before our CAD is back to work on deepprobe support). i.e  for node net111  in the extracted netlist, can I just add pin A and add a 1nOhm resistor between net111 in the netlist, and add A to the end of pinlist of .subckt definition

then what would be the cleanest way to update the main cell ? there needs to be this added new pin. and our schematics are currently locked

thanks,

Kevin

Adding a 1nOhm resistor is a very bad idea. Such tiny resistors can cause convergence difficulties. You'd be better off to make it 0 ohms or put a 0 volt source.

That said, if you can't edit the schematic, I don't see how this would help. I don't know how you're including this extracted netlist, but presumably you have a symbol with pins that match the subckt header.

A better hack might be to add:

.GLOBAL myProbe!

in the included file, and put a zero-volt source between net111 in the netlist and myProbe!. You can then add a wire in your test bench called myProbe! and that would connect to it...

Andrew.

global variable rocks!  this works just as good as deepprobe. thanks  Andrew!

Kevin