how to debug geomStamp() errors

[DReynolds]

I am trying to clean up the lvs on my mixed signal chip, but I get
psub_stamp_floating and psub_stamp_multi error messages which come
from geomStamp() commands in the foundry provided rule deck. I
understand that the idea of the checks is to look for floating
substate areas and for places where I have both digital and analog
gnd shorted.

My problem is that the output appears to just be a huge bBox, with no
info about locations of problems or even what nets are involved....
does anyone have any suggestion on how to go about debugging?

Is there a way to get more detail out of geomStamp() about why it
thinks there are errors??


David Reynolds

 

[Ed "Mr. Diva" Kalenda]

On Wed, 26 Oct 2005 11:10:10 -0500,
david.reynolds@ieee-dot-org.no-spam.invalid (DReynolds) wrote:

I am trying to clean up the lvs on my mixed signal chip, but I get
psub_stamp_floating and psub_stamp_multi error messages which come
from geomStamp() commands in the foundry provided rule deck. I
understand that the idea of the checks is to look for floating
substate areas and for places where I have both digital and analog
gnd shorted.

My problem is that the output appears to just be a huge bBox, with no
info about locations of problems or even what nets are involved....
does anyone have any suggestion on how to go about debugging?

Is there a way to get more detail out of geomStamp() about why it
thinks there are errors??


David Reynolds

My technique is poor, but it is all I have been able to come up with
over the years. If you think of something that would help, let me know
and I'll see if it can be added to Diva.

You should have two sets of errors from geomStamp. One will be the shape
which is stamped by multiple nets. The other is all the shapes which
stamped it. Correct all other errors, then make only the metal1 and
error layers visible. You can then find the metal over the stamping
shapes and see how it is in different nets.

Ed "Mr. Diva" Kalenda
Cadence Design Systems

This is just me blathering, not the company, since they don't let talk for them.

 

[Tim]

Hi David,

It depends on how the geomStamp rule was written. I was writing a rule
and only the shape with the stamp error was showing as the error. With
a little rewriting I got it to show the shape and the offending shapes
that were stamping it.

They can be confusing to debug mostly when there are many layers that
are defined with connectivity that could be causing the problem.
Substrate is the worst one.

Good Luck,
Tim

 

[Gerry Vandevalk]

Try looking at how many layer1 polygons intersect layer2

The error is that 2 (or more ) layer2 polygons (with node information) are
tring to stamp layer1
(or no layer is ... )

The problem is that "substrate" is usually the size of the design.

On trick that I have used is to make a layer that can cut the substrate and
do so where you think the different areas are

-- Gerry


"Tim" <troy1@irf.com> wrote in message
news:1130419118.498484.164810@g49g2000cwa.googlegroups.com...

Hi David,

My geomStamp rule looks like this.

geomStamp( substrate subcon error )

I believe it is the options (multStamp or floating) that might change
the output. Mine only has the error option. I got this from the manual
from taking the Diva training class. The options in your rule aren't
list in my manual but it does say that the "error" option must be used.

Note: The second layer in the geomStamp rule must be in a geomConnect
statement and the geomStamp must be located after the geomConnect in
the rule file. The first layer is not in a geomConnect statement.

Tim

 

[DReynolds]

Tim, can you explain more how you coded the stamp

How do you direct geomstamp() to only output the connecting shapes a
errors

from my extract.rul.....
geomStamp(psub ptap multStamp floating

these commands look pretty basic.... I don't see anything abou
directing ouput in them...

since it is stamping all of psub, it just plots all of psub as th
error (I think) how would I recode to get the shapes

it would be great if I could get the net names when it does a mul
... anybody know how to get that

Davi

 

[Tim]

Hi David,

My geomStamp rule looks like this.

geomStamp( substrate subcon error )

I believe it is the options (multStamp or floating) that might change
the output. Mine only has the error option. I got this from the manual
from taking the Diva training class. The options in your rule aren't
list in my manual but it does say that the "error" option must be used.

Note: The second layer in the geomStamp rule must be in a geomConnect
statement and the geomStamp must be located after the geomConnect in
the rule file. The first layer is not in a geomConnect statement.

Tim