Sorry if I seem to be on a roll.
I have several airwires that seem to be an unwillingness by Fusion to accept the connections I manually routed.
Other than just Approving because I know it is a false positive, should I be doing something different to allow Fusion to understand?
This does look like a false positive from what I can see. One thing you can try if you haven't done it already is to run the ratsnest; and see if that's going to heal this stray air-wires. I assume that any copper pour polygons you have will be re-poured when the ratsnets command is run.
If this still doesn't fix the issue then if you think it is worth it you can send me your design I will have a quick look. My email address is: [email protected].
Please let me know if there is anything else I can help you with.
Kind Regards,
@constantin.popescuXD3CL , thanks for the response! I did try the RATSNEST command without any parameters, and it did list the airwires I have, but maybe I needed to add parameters to get it to act on them?
I hid the other layers to really show the airwires, all of which are False Positives.
A couple are the odd results of the editing of wires. Often a new segment is created when an attempt to move a wire is performed.
I guess it would be too dangerous to allow a user to select an airwire and delete it. However, as this is a design block, even if I Approve the 13 errors, I get 13 NEW errors for every one of the over 60 copies I place into the 500mm strip, LOL
Hello @chris_cummingsDT,
I hope you're doing well. Would you be willing to share the designblock for review? I want to check if there's maybe something in the file that leads to these extra airwires.
Let me know if there's anything else I can do for you.
Best Regards,
Thanks a lot for sending the design, I will have a look and let you know about my findings.
Kind Regards,
I'm interested in hearing what @constantin.popescuXD3CL but just want to share a couple of things that may help.
One of the issues is that use of wires for the GND and +5V rail on the bottom, you'll get a cleaner layout and fewer airwires if you instead use a polygon for each of those on the bottom layers. Polygons are designed to have vias connected to them and will handle them better than using wires which are designed for point to point connections. That will go a long way to cleaning up these airwires.
How did you make the pads on the ends? In the library you would want to follow the arbitrary pad shape workflow for this. The way the nets are connecting to them feels like it could be cleaner but I would have to see the component to be sure.
Addressing these two points might be enough to clean up the airwires.
Let me know if theres anything else I can do for you.
Best Regards,
@jorge_garcia , You are on a roll today, with lots of great advice! I will make those wires polygons and see how that turns out. As to the pads / fingers, they are a library item. I was not finding a way to assign a net name or pin definition to a polygon, so I added SMD pads in the middle, which is likely driving the airwires:
Do you know of a way to assign a pin or net name to a polygon in the library editor? Maybe I just missed it?
You did it the correct way. You have pads on top and bottom correct? This is what I was referring to by arbitrary pad shape functionality.
Here's more in depth documentation on it
Let me know if there's anything else I can do for you.
Best Regards,
Aha! So, What I need to do is make the pad itself the shape I need, in another part of the system? Is that the Footprint Editor?
@jorge_garcia @constantin.popescuXD3CL ,
I guess I am wondering why polygons in a PCB design can be assigned to a Net but polygons in a component cannot be. By converting the wires to polygons, the polygons are not blending / overlapping.
I had a look into your design block document and I agree with @jorge_garcia's assessment. Most of the air-wires are caused by the wires not being connected point to point especially the thick ones on bottom signal layer.
Regarding your comment: "I guess I am wondering why polygons in a PCB design can be assigned to a Net but polygons in a component cannot be." - the answer is simple the nets (signals) are only present in the design and they don't have any meaning in the Library, the signals are containers of connected objects (wires, pads, vias, polygons, etc.). The polygons (polygonal regions) in the Library are polygon shapes that cannot be poured and they don't have any connectivity properties. The polygon pours are only available in board design and can be poured around any other objects.
Regarding the copper pour not blending / overlapping this is caused by the fact that you have allowed polygon thermal relief connection for all pads / smds in the same signal as the polygon. When drawing the polygon please turn off the thermals, like below:
The Thermals option needs to be set as shown above. Now because you have already placed the polygon pours if you click on the dotted polygon outline you can change the Thermals option using the Inspector like it is shown below:
Uncheck the Thermals check-box and this will give you direct connection of the polygon pour to all pads / smds of the same signal. After changing the thick botom layer wires to polygon pours the number of air-wire errors has dropped from 13 to 6. The +5V thru-hole via in the middle is also not connected point to point and after routing an extra small wire from the via center the air-wire error for this via is removed. All remaining air-wires have the same cause.
I hope this helps.
Please let me know if there is anything else I can help you with.
Kind Regards,
Here is the result, after conversion to polygons and turning off thermals (duh! Good catch @constantin.popescuXD3CL !!)
Hard to see, but still 7 airwires in my design, that I will try to weed out.
In these types of scenarios, it's preferable to re-route the last segment than to try to move it onto the connecting point of the pad. This DB block is simple so it shouldn't be much work to do that.
Another option, which I like, is a practical application of the autorouter. Since these stubs are short, the end result is that the complete connection will be made and nothing will look different. I like this solution so give it a try.
Let me know if you continue to run into problems.
Best Regards,
After ripping the one trace (looks like I did not even have to) I ran the Autorouter and while it made a bunch of odd little wires, it eliminated all 7 airwires in .4 seconds. Gotta Love it, right?
