Forum Replies Created
-
Posts
-
OK there is a problem with the code for this type of structure. I'll see if I can get the fix into 4.5.31.
Again its hard to make a comment without seeing an example. The code takes the area of any shape, and subtracts the area of any hole(s) in that shape. I'll see if I can create an example to reproduce it.
Hi Riyaz,
So you have 2 parallel lines spaced 0.6um and that's ok, but 2 perpendicular lines of the same geometry spaced 0.6um and you get an error? Just trying to understand what the cases look like.
Keith
No, at the moment only MOS devices are reduced.
Edit: but in 4.5.30, series/parallel resistors/capacitor reduction will be supported…
Hi Riyaz,
If Gemini is reporting nets as mismatched, they most probably are mismatched.
It's possible for some symmetrical circuits you will get mismatches reported which are false, in this case Gemini will have reported that symmetry was found. If this is the case then sometimes using an equivalence file can help resolve them.Check if after reduction you are getting different numbers of devices, if so then the circuits ARE different. If the number of nets is different then there may also be a problem.
Also check your bulk node nets for mos devices. Missing well taps may not be immediately obvious but will cause mismatch.
Gemini does not match by net names except if you use an equivalence file, so different generated names in schematics and layout are not a problem for it.
Keith
You can use e.g.
Code:ptlist = [[0,0],[1000,0],[1000,1000],[0,1000],[0,0]]
num = len(ptlist)
width=100
style = TRUNCATE
begExt=0
endExt=0
p = cv.dbCreatePath(ptlist, num, layer, width, style, begExt, endExt)which will create a path with the same start and end points…
What do you mean by a closed path? A path is just a series of connected vertices. Do you mean a path where the start and end vertices are the same?
Polygons are just a set of vertices with an implicit closing edge between the first and last vertex, at least as far as the external interface goes.
The build for Windows has been updated to the latest v141 C runtime libraries. Please run the vcredist executable to install them.
It will read via definitions of the form:
viaDefs(
standardViaDefs(
;( viaDefName layer1 layer2 (cutLayer cutWidth cutHeight [resistancePerCut])
; (cutRows cutCol (cutSpace))
; (layer1Enc) (layer2Enc) (layer1Offset) (layer2Offset) (origOffset)
; [implant1 (implant1Enc) [implant2 (implant2Enc) [well/substrate]]])
;(
)
( M1_PO PO M1 ("CO" 0.04 0.04)
(1 1 (0.085 0.085))
(0.02 0.02) (0.02 0.02) (0.0 0.0) (0.0 0.0) (0.0 0.0)
)…etc.
2 options:
1) in the NLPDeviceFormat property, use M[@elementNum] instead of [@instName]
2) Use [@instName] but set a string property 'type' to 'mos'. This should force creation of instances of this symbol with names M0, M1 etc. rather than I0, I1 if you have the schematic display option Instance Names set to 'Force Spice compatible'.
