How to create a wide route with coaxial shielding in Innovus?

Hi,

you can create coaxial shielding on nets as described here : How to apply coaxial shielding on clock nets

Note that the shielding is applied opportunistically. For best results I would suggest prerouting the net before any other nets are routed.
You can route only selected nets with the following commands
selectNet <netname>
setNanoRouteMode -routeSelectedNetOnly true
routeDesign

In my understanding EGR routes should not have an effect on real routes, but you can also try to delete them first just in case:
editDelete -status unknown

Then you can also fix the net wire and via shapes so that it is not further modified by NanoRoute: How to change the status of a net in legacy or common UI ?

- Dimo

Dear Dimo

Thank you very much for your answer. I have some follow-up questions.

1. Can you elaborate a bit more on when a shield is created? I noticed that NanoRoute refuses to create any shield unless the core pins are already created. For example, if I only create a power grid on layer 5 and 6 then no shield is ever created. If I however first run

sroute -connect {corePin}  -corePinTarget {stripe ring blockring}

and then start NanoRoute, the coaxial shields are created. Why is that?

2. How can I tell the tool to place as many vias as possible during layer change? For now, I use

setNanoRouteMode -drouteUseMultiCutViaEffort high

This generates 2 vias per layer change, but I would like to place more. How to do that?

Here is the complete mini-script I use so far:

add_ndr -width {MET1 1.46 MET2 1.46 MET3 1.46 MET4 1.46 MET5 1.46 METTP 1.46} -name 1_46w1s
setAttribute -net myNet vssd
setAttribute -net myNet -coaxial_shielded_layers {1:4}
setAttribute -net myNet -top_preferred_routing_layer 3
setAttribute -net myNet -bottom_preferred_routing_layer 2
setAttribute -net myNet -non_default_rule 1_46w1s
setNanoRouteMode -routeSelectedNetOnly true -drouteUseMultiCutViaEffort high
selectNet myNet
routeDesign
editChangeStatus -to FIXED
deselectAll

Thank you.

Hi,
I have not used coaxial routing a lot, so I am not very familiar with it. As I am not with Cadence anymore, I only have access to the publicly available data.
I assume this behavior has to do with the connection of the shielding to the vssd net. As stated, the coaxial shielding is done opportunistically, so I guess it does not try too hard to create connections, if it does not find an easy way to ground the shielding it just gives up.

Regarding dropping multiple vias, I suggest you create a via definition with multiple cuts using add_via_definition and add it to the NDR for the router to use on this net.

Btw you can use setAttribute with multiple attributes at once and merge your commands into one. E.g.
setAttribute -net myNet -top_preferred_routing_layer 3 -bottom_preferred_routing_layer 2 -non_default_rule 1_46w1s

- Dimo

Dear Dimo

I can confirm that using add_via_definition is the way to go. Here is one example where I am happy with the outcome:

Unfortunately, there are still situations where Innovus decides to yield another output. For example, I have another such critical net that is 3 microns wide. I use the same commands as before except that the non-default rule now uses even wider traces. So I use

add_via_definition -via_rule via1Array -row_col {4 4} -name via1_4x4
add_via_definition -via_rule via2Array -row_col {4 4} -name via2_4x4
add_via_definition -via_rule via3Array -row_col {4 4} -name via3_4x4
add_via_definition -via_rule via4Array -row_col {4 4} -name via4_4x4
add_ndr -name 3_00w1s -width {MET1 3.00 MET2 3.00 MET3 3.00 MET4 3.00 MET5 3.00 METTP 3.00} -via {via1_4x4 via2_4x4 via3_4x4 via4_4x4} -generate_via
setAttribute -net myNet -shield_net vssd -coaxial_shielded_layers {1:4} -top_preferred_routing_layer 3 -bottom_preferred_routing_layer 2 -preferred_routing_layer_effort hard -non_default_rule 3_00w1s
setNanoRouteMode -routeSelectedNetOnly true
selectNet myNet
routeDesign

which yields the following result:

Any idea why this happens? I tried playing around with the settings like using other widths in the NDR or using other via definitions - nothing seems to work. Innovus always places those slim traces with at most two vias.

Dear Dimo

I can confirm that using add_via_definition is the way to go. Here is one example where I am happy with the outcome:

Unfortunately, there are still situations where Innovus decides to yield another output. For example, I have another such critical net that is 3 microns wide. I use the same commands as before except that the non-default rule now uses even wider traces. So I use

add_via_definition -via_rule via1Array -row_col {4 4} -name via1_4x4
add_via_definition -via_rule via2Array -row_col {4 4} -name via2_4x4
add_via_definition -via_rule via3Array -row_col {4 4} -name via3_4x4
add_via_definition -via_rule via4Array -row_col {4 4} -name via4_4x4
add_ndr -name 3_00w1s -width {MET1 3.00 MET2 3.00 MET3 3.00 MET4 3.00 MET5 3.00 METTP 3.00} -via {via1_4x4 via2_4x4 via3_4x4 via4_4x4} -generate_via
setAttribute -net myNet -shield_net vssd -coaxial_shielded_layers {1:4} -top_preferred_routing_layer 3 -bottom_preferred_routing_layer 2 -preferred_routing_layer_effort hard -non_default_rule 3_00w1s
setNanoRouteMode -routeSelectedNetOnly true
selectNet myNet
routeDesign

which yields the following result:

Any idea why this happens? I tried playing around with the settings like using other widths in the NDR or using other via definitions - nothing seems to work. Innovus always places those slim traces with at most two vias.