help for Momentum setting VIA

[steven_l]

​

Hi,everyone. I am using momentum to design transformer. But two problem appears when I have define sub.
1、 how can I know the conductance of VIA.
Here is one segment of PDK parameter :
----------------------------------------------------------------
Film | Valid Width | Rsh Mean/Range | Unit
----------------------------------------------------------------
VIA1| W/S=0.26/0.26 | 6.4(<20) | Ω/ct
----------------------------------------------------------------
A、 what does W/S mean？
B、 What does (<20) mean？
C、 What doesΩ/ct mean？ I mean “ct”
D、 Can I calculate the conductance of VIA1 from the data above？ If I can， how to calculate？ If I cannot， which data need I to know and how to calculate the conductance？
2、 PDK said : the dielectric layers outside the metal are overetched XXX um。
When setting metal in momentum， I can set metal as “intrude into substrate”， but can not set XXX um。 In fact, there is not option for intrusion distance。
Whicn of the following setting in the momentum is better for VIA1 of PDK : sheet，intrude into substrate or expand the substrate?
Thank you !

 

[volker@muehlhaus]

A、 what does W/S mean？

Via size (W) and spacing between vias (S)
Vias are square, so the via size is W*W

B、 What does (<20) mean？

Typical resistance is 6.4 Ohm per single via, but worst case it can be more (but not more than 20 Ohm per single via)

C、 What doesΩ/ct mean？ I mean “ct”

Contact. In a more general way, it means one single via.

D、 Can I calculate the conductance of VIA1 from the data above？ If I can， how to calculate？ If I cannot， which data need I to know and how to calculate the conductance？

You will need the via length also. You get this from the distance between the metals that are connected by this via -> process description. And then it's simply R=rho*l/(w*w).

2、 PDK said : the dielectric layers outside the metal are overetched XXX um。

Is this some oversize of the via in the x-y plane (horizontal)?

When setting metal in momentum， I can set metal as “intrude into substrate”， but can not set XXX um。 In fact, there is not option for intrusion distance。
Whicn of the following setting in the momentum is better for VIA1 of PDK : sheet，intrude into substrate or expand the substrate?

That's well described in the ADS online help. Sheet is for thin metal simulation -> not recommended for RFIC work. Intrude or expand both create a thick metal model for the conductors, and the only difference is how the dielectric thickness is included/excluded with the surrounding dielectric. See ADS documentation.

To simplify building EM stackups for Momentum or Sonnet, I have created this tool:
https://muehlhaus.com/products/material-file-utility
It creates the *.subst and materials.matdb files for Momentum. Free to use and manuals included, but I don't have time to give any support.

 

[steven_l]

Is this some oversize of the via in the x-y plane (horizontal)?

Thank you very much ~~~~~~~~~
About “PDK said : the dielectric layers outside the metal are overetched XXX um”， I think it means it is about the relation between the dielectric layers and the metal layers， because this sentence appears where the PDK describes the dielectric layers。
About the sentence, I confuses if the metal layers intrudes into the dielectric layers. And it determines how can I set the metal model in the momentum: intrude or expand?

 

[volker@muehlhaus]

Let's say you want to create this stackup shown below:



There are two ways you can do this:

(1) If you define M1 and M2 as "intrude" then dielectric layer thickness must be defined as 2µm, which is the distance from the bottom side of M1 to the bottom side of M2.

(1) If you define M1 and M2 as "expand" then dielectric layer thickness must be defined as 1µm because each metal will automatically expand (grow) the dieelectric layer by the metal thickness. If the metals are set to "expand" then you define the dielectric layer thickness as the distance between the two metals (here: 1µm between top of M1 to bottom of M2)

 

[steven_l]

Let's say you want to create this stackup shown below:

View attachment 118378

There are two ways you can do this:

(1) If you define M1 and M2 as "intrude" then dielectric layer thickness must be defined as 2µm, which is the distance from the bottom side of M1 to the bottom side of M2.

(1) If you define M1 and M2 as "expand" then dielectric layer thickness must be defined as 1µm because each metal will automatically expand (grow) the dieelectric layer by the metal thickness. If the metals are set to "expand" then you define the dielectric layer thickness as the distance between the two metals (here: 1µm between top of M1 to bottom of M2)

Thanks a lot. Your explain is more understandable than help file. And your mmatl.exe is more convenient than momentum itself.
In your mmatl_user_guide_v3.pdf, there is a Air dielectric layer that its thickness = 1000um, epsilon =1, Resistivity=INF Ohm*cm.
However, when exporting the .son file, it also exists“TMET “Free Space” 0 FRESSSPACE 376.7303136 0 0 0 ” and “BMET “Lossless” 0 SUP 0 0 0 0”.
It seems include a Air layer already. So need I define the Air layer (thickness =100um) in the Dielectrics column？I mean a Air layer on top and a conduct layer at bottom?
Actually I do not have PDK of ADS version, but have Cadence version. So I want to design a transformer in momentum. After finishing, I import the transformer from momentum to cadence.

 

[volker@muehlhaus]

In your mmatl_user_guide_v3.pdf, there is a Air dielectric layer that its thickness = 1000um, epsilon =1, Resistivity=INF Ohm*cm.
However, when exporting the .son file, it also exists“TMET “Free Space” 0 FRESSSPACE 376.7303136 0 0 0 ” and “BMET “Lossless” 0 SUP 0 0 0 0”.

m/matl was initially designed for the Sonnet EM solver which does not have true open boundary condition. An approximate open boundary in Sonnet is created by a finite air layer with absorber above. But you are correct that the finite air layer is not needed for Momentum, because Momentum has a real open boundary condition. Yes, you can delete that air layer in Momentum.