SWX does my !"£$%^& head in

I do this quite often....laying out a pattern of machines in a layout assembly...then change config in each distance in the pattern. Attching 3D sketches to each machine and rarely no problems. Not using flex configs though.

// Krister

"Krister_L" wrote

Which version are you using?

I think this "bug" is related to a problem that one of my colleagues is also having, whereby in order for a 2nd tier sub-assy to appear correctly in the

1st tier assy, you have to have first opened up a 3rd tier sub-assy (which is used by the 2nd tier one). If you don't open up one of these lower level assys first, it all gets screwed up. The problem my colleague is having is that SWX seems to remember how the subassy was at some time in the past, and it keeps reverting back to this - in his case, dimensions and supression states keep changing.

I've seen this "remembering" problem in other areas - eg. it frequently changes the colour of assy components back to what they were some time in the past.

John H

I'm on Sp4.0 now but we've been doing this since maybe 2000.

// Krister

John, Welcome to Solidquirks world!

Ok this is the answer to your problems with your "flexible" assemblies, (clear description so the drunks can understand :-))

Say a rose joint, consisting of two parts, Rose, Body, with the assembly called, Rose assy. IF you have two, Rose assy's, in an assembly and have both flexible, you will have problems, period! That's Solidquirks for you.

Now a way around this is to make a copy of the Rose part, (save as copy, box ticked) and save as Rose1, repeat for the Body and also the Rose assy,( remember the save as copy tick box!!!!)

you should have:- Rose1 Body1 Rose assy1

Read the following paragraph fully and not line by line before proceeding

Now use File /open, Find the assembly Rose1, But do not open yet!, click on the references box and change the parts to the new ones, remembering to tick the boxes to show the green ticks!

How to do this:- File/open single LMB on the Rose1 assy Lmb on the references In the new path name window double LMB on the Rose part, a new window opens, find and double LMB on Rose 1 Then single LMB in the box next to this new path and you will see an green tick in the box. Repeat for the Body part

Now single LMB on the open box. Ctrl C, to rebuild the assy and save.

In you original assembly with the two Rose assy's, replace one of the Rose assy's with Rose1 assy, ( your mates will stay intact).

You will not get your mate errors this way.

I follow what you're saying, though it sounds like the problems you have experienced may be different to mine. I don't actually get mate errors (though I do get mates flipping). I experience 2 distinct problems:-

1) Higher level assys that contain flexible assys show as being permanently in need of an update

2) Higher level assys that contain 2 or of the same subassy in different configs causes the incorrect config to be shown when you reopen the assy, UNLESS (jury's still out on this one) you first open the sub-assy. This would be fixed by your suggestion of copying the assy.

One thing I don't understand about your suggestion, is why is it necessary to copy the parts and not just the assy?

John H

I've found SWX just can't reliably deal with more that one Flexible SubAssy's... Sure you can make them work with a lot of 'work-arounds' but I don't have time for that malarkey, and have pretty much given up on them. We're still on sw06, I'm hoping for the day it this feature works better, but not holding my breath.

Happy Friday

John H wrote:

Because the mates are refereeing to the faces of the parts in the assembly, so effectively, you have two identical mates, but "different placed" mates, on the same face of the same part. This is not be a problem with different parts in different sub-assemblies. Solidquirks can not work out which mate belongs to which assembly, hence Solidquirks wants to rebuild the main assembly, Mad!

I think!, that this confuses Solidquirks, which then causes the flipped mate errors you are seeing.

So, just copying the sub-assemblies is not enough, you have to make copies of the parts too.

A pain, I know, but it is the guaranteed way to sort your problems out.

I chose a rose joint, especially because the "rose" can be constrained by just a concentric mate, allowing the rose to swivel.

The rose joint assembly also shows the mate errors more easily.

I hope this has been fixed in 2008.

Other people, may have found a better way to solve this problem, I hope they have and pass on their knowledge here.

not just talking CAD software.

assembly, and each sub-assy instance might be

configs round completely screwing up mates

config, but this may be a coincidence.

assemblies", so that a hinged mechanism

answer on any topic from them (the only

software"). However, the last time I got that

config issues". hahahahah

lamentable performance? It makes Microsoft

John,

I've had the exact same problems and have found no good solution. The only thing that I've found that works (and it is a pain in the arss) is to P&G the assembly and then rename the assembly and parts. Then I have to manually adjust the BOM to reflect the proper quantity of sub-assemblies for the main assembly.

Please someone tell us there is a better way.........

Joseph

"Joseph" wrote >

I used to use I-DEAS and the VAR I had back then said they got a lot of sales because SWX, Inventor, Edge etc couldn't do flexible assemblies (I'm talking 7-8 years ago) and so their BOM's were buggered up by having to have copies of assemblies. SWX Corp obviously addressed this in their seemingly typical fashion of introducing half-baked features to get them over the problem during the sales pitch - it just is crap in practice.

John H

John,

BTW, are there any incontext relations in any of the parts or subs?

Pete has it, imho. You'll need to break down what common faces are used and then apply either the replace part(s) Pete is talking about or trick it using some other construction geometry to eliminate the similar mate instances in the configs.

I've seen this myself and, as it is now, you will NEVER get to stop rebuilding,.. even when you think you have it, it will rebuild!? The flipping of the faces issues Is something SW deals with poorly, the face or direction with mates, inside or outside, have been around for awhile, between sp's, releases....

Sorry man, imho, this program was never meant to do many of the things it is doing, it's so mosh-posh (I had to use that word!).

..

I use flexible subs only when really needed because of the overhead and possible issues. Usually our solution is to make 2 configs of the sub, each constrained in the proper position, such as forks out, forks in, etc.

WT

I'm with Wayne on this. Just this week I was working on an assembly from a coworker and he used a flexible subassembly so he could test out different positions of the top level machine. That flexible subassembly cause the top level asm to puke (not in lost mates, just in wild-ass solutions that could not happen in the real world)

I went to the sub asm and added configs for a sampling of position conditions. Since it was rotating around a shaft I added 12:00, 2:00,

4:00, 6:00, 8:00, 10:00 configs to the subassembly to cycle the thing.

Then I got rid of the flexible option in the properties of the subasm and simply cycled through the positional configs for the subasm to check out design issues and it worked like a champ (no more wild-ass solutions that couldn't happen in the real world).

Is this reversion to older techniques righteous? No. Is it practical? Yes.

Swx is very consistent at introducing promising new stuff in a release then not revisiting it to work out the kinks.

I have worked in SWx for ten years and have some old, hard-won habits that are probably not completely justified, but they always work. When more recent adoptees at work try newer techniques, they fail, and I find myself fixing their issues with the old workarounds from the late

90's or early 00's because those workarounds root principles have been worked out and debugged for so long. It aint right, but it gets the job done.

Enter rant mode, just cuz: example from last week... sketch blocks don't work with library features. Why not? nobody at SWx tried. That is wrong. Sketch blocks could be a neat powerful additon to our design arsenal, but this lack of curioisty on the part of SWx developers on how they might be applied throughout the software mean they have lingering unidentified limitations, and when it comes to crunch time its just safer to resort to the old tried-and-true worrkarounds. Exit rant mode.

The lesson is that whenever something new and cool comes out, be very skeptical of it and don't hinge your design on it. I will put money down that most new features will not work consistently, for multiple releases, across all the ways we want to apply it (and win - anyone want to test me on this? I can use some extra cash) .

I wish it were a harder bet to win. I hope that some day it will be.

Ed

Just curious but have you looked at your configuration settings in the sub assembly and the properties for the sub assembly in the top level assembly's feature tree? There are some settings there that might impact your problem.

TOP

Edward,

I would also tend to use a specific rigid sub-assy config, but in this particular case it's really inconvenient.

We have a custom hinge assembly that is built in LH and RH variants, and gets used as a pair of hinges on LOADS of different jobs.. The LH/RH variants are done as 2 configs of the same hinge assy, with one degree of freedom left unconstrained so that the hinge "works".

For each job, the hinge angle is different, and frequently changes many times at the layout stage. The parts that bolt to the hinge have to align with other geometry, and its very difficult to calculate the angles by hand, so I want to use SWX to work it out for me.

I therefore have an assy with a pair of hinges (plus a few other essential components for determining the angle) which are used "flexible".

The method suggested of copying parts and sub-assys would result ,over time, in countless copies of the same thing. I really don't like to go down that route for this particular real-world component.

The alternative would require multiple configs in the hinge sub-assy at maybe 0.2 degree increments between 40-70 degrees, for both handed variants. I would then have to keep switching between configs to select the closest approximation.

This might hopefully solve some of the issues related to using flexible sub-assys, but I'm not yet sure whether that is the only source of problems. I still feel that SWX has a problem with using more than one instance of an assembly, where each instance uses a different config.

John H