Snapping bug in PCBNEW

I while ago I did a post about 2 kicad bugs. I promised to make a video but could not do so at the time.

One of the problems is that Snapping a trace does not work well. I have to hold the mouse perfectly still if I want to add a waypoint for a trace. This fails in 50% of the time, the other 50% I double or triple click and I leave the routing mode…

I also have this annoyance… continously
image

I will also make a video about the other bug. That bug is about the routing being incredibly slow and often refuses to draw traces to your mouse even without obstacles in between. This simple PCB could not trigger it. But now I can make video’s it shouldn’t take me to long

This video is shot in kicad 6.0.4. My work PC has 6.0.9 and there this is still present.

Is there something I can do to fix the snapping issue?

Regards,

Bas

1 Like

For the routing issue, I encourage you to post the layout (or PM it to me). This will help us fixing the bug much faster.

About snapping: do you have the option “fix all segments on click” in the routing settings turned on or off? Does changing it help?

Tom

Tom, from what I’ve seen, the snapping issue requires “Fix all segments on click” disabled and “Use mouse path to set track posture” enabled.

1 Like

It was off. Now I turned it on. And now it finally works again!!!

So apparantly this ‘bug’ was a ‘feature’. Is it perhaps an idea to enable this by default? I cannot imagine that anybody would like to have this turned off, but I suppose some will need it for something? This has plagued me for months…

Kind regards,

Bas

Many PCB programs work this way by default and quite frankly I myself find the “fix all segments on click” behaviour less productive than the default one… But it’s my very subjective feeling that others may share or not.

As for turning it on by default - I can bet there would be users who would suddenly think KiCad has a new bug :slight_smile:

2 Likes

In V 6.0.9 I have ‘Fix all segments on click’ off and never had it on. I have not problems like you describe.
Do you have in Preferences-PCB Editing-Editing Options - Magnetic Points - Snap to pads (and Snap to tracks) set (as I have) to Always? Or may be Never?
The problem for me is opposite (and I reported it few years ago) that snapping works too hard and when I want being already inside a pad to change the track width for the last short segment ending at pad center even I have pad occupying my entire screen the snapping takes my cursor and jumps half the screen to pad center. I suggested to limit the snap action to not take place if jump is longer than some specified screen pixels. Hijacking my cursor and suddenly moving it half the screen bothers me a lot, but when I (tried now) switch off magnetic then I have problems like you.

For information:

“Fix all segments on click” is enabled by default in 6.0 and 6.99, unless the configuration was migrated from versions 5.x or older - then it’ll be disabled.

Some aspects of differential pair routing work better with the option disabled currently.

This setting shouldn’t have any effect in this case. The “Use mouse path to set track posture” option does.

Holding Shift key temporarily disables snapping to items. That doesn’t work in your case?

We need more details here.

Now I think I don’t understand what is the problem (even previously I thought that I understood).
I read:

I looked at:
track-track

and was sure that track started or finished at another track not snap to track end. I can get such results only when I switch Magnetic off so I supposed it can be the reason. Now I see that I shouldn’t write anything. Sorry.

It works. When I was reporting that (I think it was V4 those time) I didn’t know that I can use Shift. But in any case I still think that snapping should be always limited by a screen distance as snapping is only (in my opinion) to correct human precision and jumping too far distance has nothing to do with it.
If it would work that way I would use mouse wheel to stop with cursor inside pad but not at its center.
I can imagine also (but not sure if it can happen to be real problem) that it could be useful if two magnetic points are too close to each other. Zooming in you could simpler to snap to one of them.

Now I see this my sentence was simply not true.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.