Cell is not defined error (maybe an auCdl issue)

Which LVS tool? Which version? It appears that the error is coming from an LVS tool, not the auCdl netlister.

What does the instance look like in the CDL netlist (assuming you are using a CDL netlist)?

Regards,

Andrew.

It's PVS.

But the tool requires a auCDL view to get that far. I can get a successful LVS by telling the LVS to use auLVS, but that only works when LVSing the layouts/schematics with the structures I've modified the rule deck to extract (And created a symbol for).

It's something to do with extracting a netlist from the schematic it is having a problem with, but I can find nothing in the documentation about this.

I need to have the LVS options for schematic as create CDL or it doesn't seem to be able to match transistors/diodes/resistors etc. from layout to schematic. It seems to see the devices in the layout and schematic but not recognise they are the same device.

It may not be related to auCDL, but it is some sort of CDL issue, but I'm not sure where to even look in the documentation to find out what's going on.

I've been trying to view a netlist to see what's going on with that, but I can't see how to get one extracted. I assume there isn't one, I haven't created one, I've made a symbol and an auCdl (as lvs moaned about there not being one.), but I haven't physically written a netlist, I'm not sure where I'd do that.

OK I've managed to extract a CDL netlist from a schematic I've instantiated a cell in.************************************************************************
* auCdl Netlist:
*
* Library Name: test
* Top Cell Name: CCD_pixel_9
* View Name: schematic
* Netlisted on: Feb 2 14:54:35 2015
************************************************************************

*.BIPOLAR
*.RESI = 2000
*.RESVAL
*.CAPVAL
*.DIOPERI
*.DIOAREA
*.EQUATION
*.SCALE METER
*.MEGA
.PARAM



************************************************************************
* Library Name: test
* Cell Name: CCD_pixel_9
* View Name: schematic
************************************************************************

.SUBCKT CCD_pixel_9 p1 p2 p3 p4
*.PININFO p1:B p2:B p3:B p4:B
XI4 p1 p2 p3 p4 / CCD
.ENDS

Can you contact customer support? It's going to be much easier to figure this all out if we can see your data.

Kind Regards,

Andrew.

I work for an academic institute and we get the Cadence software through the microelectronic centre at the stfc. For some reason that bars us from using Cadence's customer support, I emailed stfc last week but they haven't replied yet. So no, unfortunately I can't contact customer support.

Because there are a very large number of academic users within the Europractice scheme, who receive access to the software for very low access rates, the agreement is that Europractice provides the first level support and then they escalate to us any issue that they cannot handle themselves.

Trying to debug this via a forum is going to be quite hard when I can't see your rule deck, can't see your layout, and can't see all the files from the LVS to see what is going on - which is why it really needs to be through a support channel.

I'll email my contacts at Europractice to see if I can get things moving.

Kind Regards,

Andrew.

Cheers,

I guess if students were contacting you that the numbers would be large. I work for the largest academic institution in the country I and I am the sole user of our single licence. :(

I don't necessarily think that we're at the file/text level debugging stage yet. I suspect I've just not followed the correct steps when generating a symbol which I wan't LVS to associate with a device that I've specified in the lvs rule deck.

I'm fairly sure the rule deck has been modified correctly. I think the problem is my creation of a symbol. I had no idea if I had to create this in a a certain way or add a line of text to some definition file etc. So I just created a symbol and named it the same as the device that the LVS should pick up in layout.

The documentation for this was really poor. The PVS ref manual has so much stuff/commands missing. I pretty much had to reverse engineer the method of modification from the recognition layers of an NMOS and apply the same technique to the structure I wanted to be recognised.

I can't find anything relating to the generation of a symbol or similar for the purpose I require. Do you know which document would have this information?

I'd have to do some experiments and checking through the documentation to work through this. Whilst I have a lot of knowledge of Physical Verification in general (dating back over 25 years), I've not done that much PVS rule generation or technology setup so I'd have to do some revision and research...

Anyway, I've spoken to the Europractice folks, and they've assured me that your issue has been assigned an owner and they will be dealing with it as soon as possible. Rather than double working (and I do this forum in my "spare" time), I'll let them deal with this .

Kind Regards,

Andrew.