Cadence IC 5 parts brower still broken!!! NCSU mess's it up

My parts brower is still showing 1 level and 1 cant select any components for that I require the 3 levels as it works fine without the NCSU design kit, what is going on here, can someone actually get this kit working for gods sake.

KLEOS

Re: Cadence IC 5 parts brower still broken!!! NCSU mess's i

Can someone provide a detailed installation procedure for the kit, step by step to try and solce this stuipid thing as iam not getting anything done.

KLEOS

Re: Cadence IC 5 parts brower still broken!!! NCSU mess's i

Does anyone have a fix for this? Iam waiting patiently. Iam also trying to simulate nmos/pmos transister schematics but iam unable to because of the models. With the NCSU kit the component brower shows only one level and I cant select instances AT ALL.

KLEOS

Re: Cadence IC 5 parts brower still broken!!! NCSU mess's i

Are you not able to descend in the component browser?

Try unsetting the LANG environment variable before firing icfb...

Re: Cadence IC 5 parts brower still broken!!! NCSU mess's i

Yes iam not able to decend exactly right!!
Where is this LANG variable?
Its always the same nomater what i run,
Is it possible to install the kit EXACTLY AS the installation instructions say, though it requires compiling some files for some reason, I bet doing it this way the KIT will behave flawlessly.

KLEOS

Re: Cadence IC 5 parts brower still broken!!! NCSU mess's i

thanks m8 i sent the lang variable to 0 and all is fine yeah, I can now browse.

Now when I for example start icms, then new library I attach to a library but when instancing nmos/pmos (ffrom NCSU analog parts after attaching to a techfile) transisters it gives me and error saying some parrameters cannot be found. I assuming that the models are not being found therefor its probably missing form the Library path editer.

I have added

NCSU_Analog_Parts
NCSU_Digital_Parts
NCSU_Sheets_8ths
basic


to the library path editor, is there anything missing, what about the model files etc...
Looking at the tutorials there seems to be several extra library lines being


cmosx_Basic
cmosx_Cells
cmosxLPCells
cmosx_Pads
cmosx_dodBorders
cmosx_genericBasic
cmosx_mosisLocal


Can anyone please include all thier libray names and paths,

Thanks
KLEOS

Re: Cadence IC 5 parts brower still broken!!! NCSU mess's i

OK here are the errors


Set Mosfet Properties Window

Library Propertie gridResolution was not found

Library Propertie minlength was not found

Library Propertie minwidth was not found

are these suppose to be in the actuall tech file?

KLEOS

Re: Cadence IC 5 parts brower still broken!!! NCSU mess's i

Ok all is fine I setup the library files and ran the install.pl script. The only problem that remains is when simulating schematics using SPECTRE, I have to manually specify the model files but when simulating it gives an error saying that transistor M1's bulk voltage has exceeded imelt, whatever that means, the output waveform of an invertor allthough correct only reaches 0.67 volts with a 5V input. The circuit is from the NCSU turorials.

Any ideas on this, also in virtuso layout when doing instancing parametised cells from the included TSMC librarys, all I see is the outline I cant see the poly, metal and diffusion etc...


KLEOS

Re: Cadence IC 5 parts brower still broken!!! NCSU mess's i

It seems that I've finally found a reason of that pmos imelt blah blah problem. :sm16: When you look in the netlist, there's wrong bulk node connection assigned (gnd! (or 0) instead vdd!). Then there is a positive biased pn junction and therefore huge current flowing through the pmos's bulk node.
idea: The point is to redefine the bulk node of pmos in the netlist or in directly in the pmos symbol cellview to vdd!. Due some incompatibility between IC4** and IC5, the NCSU connection inheritance fails. There is also a conversion utility in CIW tools (analog bulk conversion), but that doesn't seem to work for me. :? so now im going to edit the part symbol view manualy.

Until now I had to get through this with the stupid workaround by using the standard parts from analoglib with NCSU CDK models, which is not that good for all the cellview dependencies in NCSU toolkit, especially by parasitic extraction etc., now i can finally enjoy the NCSU students' good job, it is surely worth it.


best regards

Lempour