John (and other experts),
I was waiting to see if NX 11 was going to address the concerns I have in the attached PDF, but alas, they did not. I would be interested in your (collective you all) comments regarding the workflow speed limitations I have experienced in sketcher as described in the following attached PDF.
Rensselaer Polytechnic Institute
Unfortunately I'll never get to use NX 11.0 (at least not the finished version) as I retired better than a year ago.
Thank you for this feedback. I am a bit surprised by your comments being labelled "Not fixed in NX11.0.1mp1". Should we have read your mind? We are smart, but not that advanced. :-)
Maybe I just missed thme in the ER listing.
It is perfectly fine to ask these questions, or even log IR/ER/PR's for this.
I am going to take your comment and look at them in more detail. You are on to something.
Here are my initial reactions:
1. Constraining to other sketches.
Here you got two things fighting. Selection of features and selection of sketch geometry. I do like the suggestion where we change the selection filter on the second selection once a sketch curve has been selected.
We are advising users to think twice before making a reference outside the sketch. But I agree it could be easier.
2. Selecting a dimension when a curve has been selected
My proposed solution for #1 would also address this one. The filter could adapt after the first selection. This will really help when the dimension extension line is over the curve.
3. Selection of constraints is prioritized over curves
In most cases curves are easier to pick because they are bigger (cover more pixels on screen). We made constraints be on top for a reason, they are often so small it is hard to select and delete them. This was done at the request of customers. Now you hit the unfortunate case of an extension line over a curve... There is no simple solution for this one. Addressing #1 will help on the second pick.
4. The projected datum line is very long and get in they way of fit
I would assume the line is inferred from the length of the datum. We could investigate excluding these lines from fit. Please note that if you use Orient View to Sketch that these lines are ignore. Press SHIFT+F8 or use the view radial to access it quickly.
All of these are all perfect Enhancement Requests. I suggest you file them.
PS. by now you probably discovered that one of the areas I work in is the sketcher area
PS2. I believe John should have been given a free NX license for life when he retired.
Well John, your retirement is deserved, to say the least. Did miss you at the last PACE forum. Now get off these forums and relax!
Best wishes to you!
I have mixed emotions about not being offered (and not having asked for) an NX license when I retired as I know its been done in the past (for years I signed for the annual renewal of a courtesy license for someone who's still a frequent contributor to this forum). But if I did have a license I would probably be spending even MORE time on these forums which might have been counter-productive, at least personally (I won't go into the details here but I have talked about this in one of the Eng-Tips forums: http://www.eng-tips.com/viewthread.cfm?qid=419979 ).
Anyway, I'll continue to follow these forums at least for a while yet, until either NX gets so advanced that I can't recognize it anymore, or I start to forget more of the details
Thanks for the in-depth response. My "not fixed.." comment was just letting you know that I'm trying test these on the latest version possible, not that they should have been fixed by then -, but we could always hope. I haven't logged them yet, wanted to come here and see if I was way off base before that.
Short answer, I'll log the ER's as mentioned through GTAC.
1) Part of my reason was to get feedback on methods from users/developers like yourself. Are my methods out-dated? When you caution users to "think twice" before referencing curves outside of sketcher (to other sketch curves) - are you cautioning against that method? Is there a better, quicker way to reference previous geometry? I'm open to thoughts here. Being able to quickly reference existing edges and curves, with minimal clicks is, to my knowledge, an advantage of NX.
3) I agree - most cases, the constraints are smaller, and you've made it easier to select them (for deletion). But by adding priorty to them, the selection of the curves is made more sluggish. I think I'm speaking of only constraints that can be applied at distances and could potentially 'fully cover' a sketch line (point-on-curve, collinear, midpoint) - these dashed lines should not be priority over actual segements, in my view.
4) Thank you. "Orient view to sketch" does ignore the dashes, my feeling is regular fit should do the same thing, or resize your dashed lines to the geometry extents, or delete them. Since they are not selectable anyway, why do they exist? I don't see any information they are providing other than "geometry has been constrained to the axis/datum".
Now I've investigated this deeper and came across something else:
5) I've mentioned about that Point-On-Curve has priority over segments - well EXCEPT for [at least] this case: (see attached picture). On the left instance PIC "1": I've attached the right end-point to be Point-On-Curve [POC] to the X-sketch axis. This POC constriant is now ONLY selectable if you QUICKPICK LIST the end-point (red arrow location). The black arrow location is not selectable. There is no dotted line back to the origin, there is nothing selectable other than the line itself. Contrast to PIC "2" where the line is dragged away from the axis to the left, so the dashed line is now visible and selectable. This seems inconsistent.
So what should be the behavior? IMO - the system should operate like PIC "2" for both instances, but the POC dotted line should not have priority over the line. The longer, blue, unselectable dashed lines, could be removed or resized and made selectable as the constaint that created them. I'm of the opinion that the software should never automatically create an entity on-screen that cannot be selected/edited by the user.
Thanks for the feedback!
On #1 constraining to objects outside the sketch. Yes, this is an advantage but it can also make your life difficult. Let me give you two examples:
1. When you want to reattach then the 'other' external references are not reparentable. You do see them when you copy and paste.
2. When external references are modeled away, then your constraints and dimensions to these objects will fail to solve. Please note in order releases they are blindly deleted.
As soon as you blend away an edge that is referenced by a constraint, there is no way to repair a geometric constraint. You can only delete it and recreate another one. At least you are warned, so we are making progress.
There is an NX 11.0.1 enhancement that retains projected curves in sketcher when their inputs are modelled away. If you constrain to a projected curve, then you can repair the constraint by repairing the projected curve. Since reference curves are hidden when the sketch is inactive (also NX 11.0.1) you part is not polluted with extra curves.
If you work like this, then sketch scope set to within active sketch only is suddenly a nice thing.
Now I am not telling you all to change your workflows overnight. I am just making you aware of how you can take advantage of the enhancement we made in NX 11.0.1.
Now on topic #5
The line does not go back to the origin because the constraint has nothing to do with the origin. You probably constrained the end point of the line to be on the sketch X axis. In picture 1, it looks like that. So in picture 2, the line endpoint is extended to the projected sketch axis.
Regards, **bleep** (trying to type my first name but for some childish reason the forum assumes I named after a bad word :-))
Thank you - please note in the both pictures of point #5 - it is the same: POC (endpoint to axis). I merely dragged the endpiont to the left for pic2. I only mention the 'origin' becuase in pic2, that is where the dotted line ceases. In my view, if the dotted line is going to terminate at the 'origin' (according to you: the start of x-axis) when the point is on the left of the origin, it should exist and terminate at the same location when it is on the right of the origin. For consistency.
From your video example and explanation, I believe I understand.
In the video, the user 're-orders' the blend before the sketch. Not sure why this is done, I would not teach my students this type of behavior, for exactly the reasons you mention (it would break parent-child). I see that by adding the project curve, you are trying to make it more 'robust' for potential re-ordering. But I still abide by chronoligical modeling, so I try to teach those guidelines.
I did try to duplicate your steps in the video, but after I re-associate the projected curve to the rear edge, finish sketch, I was still getting the Warning Information window: Parent of the point is either misisng or modeled away..then I realized, in your video, your CSYS was placed at the bottom left corner, while mine defaulted to the upper right (the one that was removed with the blend re-order)...so alas, re-ordering like you have done may come with more baggage than I'm afraid the project curve enhancement can alleviate, so I remain extremely cautious upon most feature re-orders and continue to teach my students the same.
Thank you though, I love to see and hear about other workflows.