Power polygons & shielding vias
mulfycrowh , 03-06-2023, 11:34 AM
Hi,
When you add shielding vias to your design, you then get many holes on the power polygons.
Is it an issue?
Thanks a lot.
When you add shielding vias to your design, you then get many holes on the power polygons.
Is it an issue?
Thanks a lot.
qdrives , 03-06-2023, 03:01 PM
The narrower polygon in the (vertical) middle is almost cut in half by vias. Does it need to go anywhere to the left?
There are only vias connected to it in the bottom right.
Whether this is an issue mainly depends on the current it needs to carry.
There are only vias connected to it in the bottom right.
Whether this is an issue mainly depends on the current it needs to carry.
mulfycrowh , 03-06-2023, 03:06 PM
@qdrivesI am not done. I have to optimize between current and shielding.
But if you say that the matter only involves current, that's OK.
Thank you.
But if you say that the matter only involves current, that's OK.
Thank you.
qdrives , 03-07-2023, 01:24 PM
No, I am nit stating that it only involves current.
I am working on a design where the inner power 'tracks' are 14mm wide, 3oz (105um) copper and still the temperature would rise 10°C @25Arms. To many via's reducing the width would have an even more negative impact.
The word issue in the sentence "Whether this is an issue mainly depends on the current it needs to carry." is on the via's 'cutting' the track.
One of the thing you could do is to vary the position of the via so that the polygons can 'flow' between them.
In the picture below - pick the right style and avoid the left style.
My expectation is that as long as the (large amount) of vias are Gnd (signal return), then it does not matter so much how many you place (other than it may cost a bit more), as long as you do not prevent the flow.
I am working on a design where the inner power 'tracks' are 14mm wide, 3oz (105um) copper and still the temperature would rise 10°C @25Arms. To many via's reducing the width would have an even more negative impact.
The word issue in the sentence "Whether this is an issue mainly depends on the current it needs to carry." is on the via's 'cutting' the track.
One of the thing you could do is to vary the position of the via so that the polygons can 'flow' between them.
In the picture below - pick the right style and avoid the left style.
My expectation is that as long as the (large amount) of vias are Gnd (signal return), then it does not matter so much how many you place (other than it may cost a bit more), as long as you do not prevent the flow.
mulfycrowh , 03-07-2023, 01:42 PM
@qdrives Thanks for this information.
In my case, the current needed on the power polygons is low.
High current is on power planes.
I had this idea to move some vias to prevent from having a perpendicular wall effect.
In my case, the current needed on the power polygons is low.
High current is on power planes.
I had this idea to move some vias to prevent from having a perpendicular wall effect.
mulfycrowh , 03-07-2023, 03:38 PM
I am getting the following idea: is there a simulation tool, let's say in Altium, that computes the current and rise of temperature on a power polygon, taking account of the vias on it?
robertferanec , 03-08-2023, 07:44 AM
I have never used it, but I remember Altium had a PDN plugin - but it was not free. I am not sure if something has changed about it.
qdrives , 03-08-2023, 02:19 PM
I did use an evaluation version of the (original) PDN extension. It confirmed the difference in resistance, as I also measured experimentally, between the 2 Gnd paths of a DC brush motor driver.
Because I learned that, I currently am making my designs more symmetrically.
I hope to get an update from Lukas Henkel on his PDN analysis with Elmer - https://www.linkedin.com/posts/lukas...m_source=share
Perhaps @robertferanec can make a video about it with him.
Just today I watched a webinar on thermal simulation and current density (with Cadence FlowCad Celcius). Recording is not yet available, but here is link to some details: https://pages.ema-eda.com/EMA-2023-W...B+Applications
Because I learned that, I currently am making my designs more symmetrically.
I hope to get an update from Lukas Henkel on his PDN analysis with Elmer - https://www.linkedin.com/posts/lukas...m_source=share
Perhaps @robertferanec can make a video about it with him.
Just today I watched a webinar on thermal simulation and current density (with Cadence FlowCad Celcius). Recording is not yet available, but here is link to some details: https://pages.ema-eda.com/EMA-2023-W...B+Applications
robertferanec , 03-08-2023, 11:25 PM
@qdrives I am talking to Lukas about making a video together - PDN was one of the suggestions.
Use our interactive Discord forum to reply or ask new questions.