merge two different gnd from a single point (net tie) issue

[electronx]

I am having problems when I want to merge two different gnd from a single point with net tie and polygoun pouring.I want to combine them into one piece. Any help?

[hwasti]

Two thoughts:

1) Having multiple grounds is a dumb idea. It leads to far more problems than it solves. Think about placement and return paths and use a single ground plane.

2) I am not seeing what the issue is. The net tie is properly connecting your separate PGND and GND nets. What exactly do you want to happen that is not happening?

[ajb]

It looks like you need to change the polygon pour settings (in the polygon properties) from 'pour over same net polygons only' to 'pour over all same net objects'.  The polygon is currently pouring over the pads, but not the track that links them in the net tie.  Alternatively, you could change the net tie to use a polygon instead of a track, I think. 

1) Having multiple grounds is a dumb idea. It leads to far more problems than it solves. Think about placement and return paths and use a single ground plane.

Generally agreed, eg in the sense of splitting analog and digital grounds, but there are definitely situations where you want to control the ground connection point between circuit segments, and net ties are great for that -- especially if you need to control return paths within a board that otherwise has a single solid ground plane. Those situations aren't terribly common, but based on the routing, maybe that's what OP is dealing with. 

[electronx]

actually its a remake project  ,and this project has passed of emc issues.CE and fcc certificed . So I dont hestitate split the smps grounds

[hwasti]

Getting back to basics of your original post: What exactly are you expecting to see, that you are not getting? As far as I can tell, the software is doing exactly what you have asked it to do.

If you can explain what you are expecting to see, I or someone else, can help command the software to create that.

And one final thought on split planes: Many products work and pass EMI with split planes, guard traces and other counter-productive features. They work/pass not because of these, but despite these.

If your design works and you are tasked to just replicating it, by all means replicate what already exists. That is the best option, especially if you are able to make a case that the new design is equivalent to the old design and re-testing is not required. That mitigates both risk and cost.

But based on this question, I suspect that the layout is not identical and will need to be re-tested. In that case you are taking on unnecessary risks by perpetuating the problems.

Good luck.

[Pseudobyte]

What primitive is the actual bridge of the net-tie?

From what I remember this can't be a track and needs to be a region.

Then draw it so the actual bridge barely overlaps with the pads.



Also make sure you design the net tie for your specific use. Meaning that the overall length and length of the bridge comply with your design rules. ie the pads are long enough to make connections without causing the connecting features to be in violation with the bridge of the net-tie.



The alternative to this is to scope a rule to ignore the primitives within your net-tie for the specific nets you are shorting.