Search found 237 matches

by project4
Wed Jul 25, 2018 9:49 am
Forum: Assembly
Topic: Another approach to assembly solver (A2plus)
Replies: 1842
Views: 492756

Re: Another approach to assembly solver (A2plus)

There is a small bug in your recent code. Somewhere you forgot to replace rig.(a_member) by self.(_member). Look for it. The mistake is in a try/except section, therefore you might not have it realized at moment. In my solver-stabi branch, it is fixed. Can you copy&paste some code around it? I ...
by project4
Wed Jul 25, 2018 9:38 am
Forum: Assembly
Topic: Another approach to assembly solver (A2plus)
Replies: 1842
Views: 492756

Re: Another approach to assembly solver (A2plus)

I think you are looking for a single center of rotation that will rotate both holes into place. Graphically, draw a line connecting the centers of one hole, a second line connecting the centers of the second hole. Make perpendicular bisectors to both of these lines. The intersection point is the lo...
by project4
Wed Jul 25, 2018 9:37 am
Forum: Assembly
Topic: Another approach to assembly solver (A2plus)
Replies: 1842
Views: 492756

Re: Another approach to assembly solver (A2plus)

There is a small bug in your recent code. Somewhere you forgot to replace rig.(a_member) by self.(_member). Look for it. The mistake is in a try/except section, therefore you might not have it realized at moment. In my solver-stabi branch, it is fixed. Possible. I've found one place as well before ...
by project4
Wed Jul 25, 2018 8:05 am
Forum: Assembly
Topic: Another approach to assembly solver (A2plus)
Replies: 1842
Views: 492756

Re: Another approach to assembly solver (A2plus)

I couldn't sleep well with that information... :? Me to, more than one night ! It is one of the biggest problems. . What I think we should do is finding the rotation point as a weighted average between the rotation points. Nothing else i did when introducing the "refPointsBoundBoxCenter" ...
by project4
Wed Jul 25, 2018 6:46 am
Forum: Assembly
Topic: Another approach to assembly solver (A2plus)
Replies: 1842
Views: 492756

Re: Another approach to assembly solver (A2plus)

The top bar will rotate a little bit arounds it refpointsBoundBoxCenter, then it is moved again, and so on. At last, everything should fit. (if there were not the problems with slow rotation) I couldn't sleep well with that information... :? Let's see the following examples. Those are simple 1 rigi...
by project4
Tue Jul 24, 2018 7:31 pm
Forum: Assembly
Topic: Another approach to assembly solver (A2plus)
Replies: 1842
Views: 492756

Re: Another approach to assembly solver (A2plus)

BTW, maybe the "assembly_hierarchy.html" file should be placed, where the assembly lives (path and assembly-filename added) and not blindly into home folder. That's what I wanted as well, but I don't know (and didn't had time to look for) how to get the path to the assembly file. Please r...
by project4
Tue Jul 24, 2018 5:56 pm
Forum: Assembly
Topic: Another approach to assembly solver (A2plus)
Replies: 1842
Views: 492756

Re: Another approach to assembly solver (A2plus)

@project4, i just merged your branch with branch solver-stabilization. We can add the flag later. Thank you for all your work ! :D I will PR the button to your solver-stabilization branch soon. Here are some nice numbers for the crankshaft assembly (it's hierarchy allows simple breakdown to parts):...
by project4
Tue Jul 24, 2018 4:17 pm
Forum: Assembly
Topic: Another approach to assembly solver (A2plus)
Replies: 1842
Views: 492756

Re: Another approach to assembly solver (A2plus)

@project4, i just merged your branch with branch solver-stabilization. We can add the flag later. Thank you for all your work ! :D Thanks! Now try to keep the logic in relevant places. Everything that is done on the variables or related to dependency should be done in the dependency classes. Everyt...
by project4
Tue Jul 24, 2018 3:28 pm
Forum: Assembly
Topic: Another approach to assembly solver (A2plus)
Replies: 1842
Views: 492756

Re: Another approach to assembly solver (A2plus)

my 2 cents idea taken from ffmpeg used on xvid encoding: 1st pass at very rough accuracy i.e. 0.3mm and 0.3deg with big rotation steps calibrated according to the accuracy set i.e 0.1mm steps and 0.1deg rotation, this should rapidly find a valid solution even on complex assemblies. 2nd pass, improv...
by project4
Tue Jul 24, 2018 2:06 pm
Forum: Assembly
Topic: Another approach to assembly solver (A2plus)
Replies: 1842
Views: 492756

Re: Another approach to assembly solver (A2plus)

I think that it's better to calculate the refPoints movement that are affected by the rotation and apply it (mathematically) in the same round. That will probably decrease the wobbling around the needed point (back and forth). Since every movement and rotation apply forces in different directions, ...