gabi71ro
Junior Member level 2
- Joined
- Sep 3, 2003
- Messages
- 21
- Helped
- 0
- Reputation
- 0
- Reaction score
- 0
- Trophy points
- 1,281
- Location
- Iasi Romania
- Activity points
- 259
Hello All,
After I used many years Calibre and from time to time Dracula for layout parasitic extractions, now I have to do this with Assura.
I find few things that put me in a difficult position, and make to untrust Assura results. (Like a conffirmation that Assura don't work properly)
Description case 1:
I have two wide metals (metal 1 and metal 2), both with 0.12 Ohms resistace. This two metals are connected trough a 2x2 via 1 matrix. Supposing that Rvia1=5 Ohms, the resulted resistance of this matrix should be 1.25 Ohms. Now the total resistance of metal1 + via1 matrix + metal2 must be equal with:
0.12 + 1.25 + 0.12 = 1.49 Ohms.
Assura extractions show me a value about 20.24 Ohms, which suggest that Assura doesen't calculate paralell resistor values and it add like a serial resistor structure.
Description case 2:
in case of three short paralel metal lines (agressor-victim-agressor), with the same width, placed at minimum space, Assura calculate a total capacitance for the victim with the formula Ct=2cc + cba + 2cbf + cta + 2ctf = 1 fF (cc= coupled or lateral cap, cba =area cap to bottom layer, cbf = fringe capto bottom, cta = area cap to top layer, ctf = fringe cap to top layer).
If I increase only the space between lines (double space), the total capacitancies increase!!! for short lines.
I am agree that increasing the space, a part of cc will be transformed in cbf and ctf, but both cbf and ctf have let say a small contribution in the Ct. On the other hand, with a hand calculation, the second value (for double space) is smaller than the first one, normal I say.
In case of long lines, Assura extract show me Ct(single space) < Ct (double space) but the difference is too small. (comparing with the hand calculation).
Do you know why Assura give me these results?
Is an Assura bug, or a mistake in the rule file?
It is possible that all this to come from Vampire?
But Vampire have nothing to do with paralell resistors.....
Can I trust Assura? (no hard feelings for Cadence's engineers)
Best Regards,
Gabriel.
After I used many years Calibre and from time to time Dracula for layout parasitic extractions, now I have to do this with Assura.
I find few things that put me in a difficult position, and make to untrust Assura results. (Like a conffirmation that Assura don't work properly)
Description case 1:
I have two wide metals (metal 1 and metal 2), both with 0.12 Ohms resistace. This two metals are connected trough a 2x2 via 1 matrix. Supposing that Rvia1=5 Ohms, the resulted resistance of this matrix should be 1.25 Ohms. Now the total resistance of metal1 + via1 matrix + metal2 must be equal with:
0.12 + 1.25 + 0.12 = 1.49 Ohms.
Assura extractions show me a value about 20.24 Ohms, which suggest that Assura doesen't calculate paralell resistor values and it add like a serial resistor structure.
Description case 2:
in case of three short paralel metal lines (agressor-victim-agressor), with the same width, placed at minimum space, Assura calculate a total capacitance for the victim with the formula Ct=2cc + cba + 2cbf + cta + 2ctf = 1 fF (cc= coupled or lateral cap, cba =area cap to bottom layer, cbf = fringe capto bottom, cta = area cap to top layer, ctf = fringe cap to top layer).
If I increase only the space between lines (double space), the total capacitancies increase!!! for short lines.
I am agree that increasing the space, a part of cc will be transformed in cbf and ctf, but both cbf and ctf have let say a small contribution in the Ct. On the other hand, with a hand calculation, the second value (for double space) is smaller than the first one, normal I say.
In case of long lines, Assura extract show me Ct(single space) < Ct (double space) but the difference is too small. (comparing with the hand calculation).
Do you know why Assura give me these results?
Is an Assura bug, or a mistake in the rule file?
It is possible that all this to come from Vampire?
But Vampire have nothing to do with paralell resistors.....
Can I trust Assura? (no hard feelings for Cadence's engineers)
Best Regards,
Gabriel.