I have the same problem, but with SW 2010 and it happens with small assemblies. As soon as you use a lot of configuration, solidworks has trouble finding the constraints.
On the other hand, your backup solution is interesting, because I redo the constraints and it's quite long in general.
We regularly work on large assemblies: 6500 components
It is common to have many dysfunctions in the constraints. However, we don't use configurations much. Especially because of this kind of problem. Despite everything, the problems are still present, for example with free components in sub-assemblies...
It is often enough to open the subassemblies and set the components to solved...
In 2011 SolidWorks was equipped with the "SpeedPak" function. It makes it easy to insert large assemblies into higher-level assemblies.
I don't use it much. Because it seems tedious to me. It is necessary to determine the reference elements that we wish to keep. When you forget some, you waste a lot of time adding them. On top of that, SolidWorks frequently asks us to update the speed pak because they find it outdated and we are still wasting time.
To try then...
I would also be interested to have your feedback on this function. In order to know if I don't use it as I should or if it doesn't correspond to my use.
I have the same problems. I tried the SpeedPak mode on the sub-assemblies. It saves time when opening and it bugs less. You can also lock the crafting tree for certain rooms by scrolling down the yellow bar.
I also noticed that we had less of a problem by assembling the parts isostatically. There is too much of a tendency to make the 3 planes coincidental, while one point is enough for the last constraint.
I am of the opinion of RMorel, we commonly handle ASMs of 6000 to 10000 pieces under an average of 6 levels and the configurations are nests of annoyances; so almost forbidden in our country because the slightest problem of constraint spreads throughout the tree.
Without these constraints, our assemblies perform quite well
I signed up after reading this post and in particular following rmorel's response.
Before I will answer s4kuj0
I found using a sub-assembly (a kind of jack to put it simply) with 3 configurations:
-free (with a limit distance constraint)
-stem outlet
-retracted stem
if I use this cylinder 3 times using 1 different config it works
On the other hand, if I use the free configuration for my 3 cylinders in flexible mode and I use constraints to output the rod of different values, I get errors. If I remove the limit distance constraint in the cylinder it works
In order to avoid this problem and not to use the flexible mode, I created a configuration by position (not practical if it's a real cylinder used in different machines). It is also possible to create a new file each time it is used (jack1, jack2, jack3) so that you can keep the limit distance constraint and the flexible mode (useful for simulating movements)
This is a continuation of my previous essay which was an answer to the question asked.
As I said, rmorel's answer made me react
It is common to have many dysfunctions in the constraints [...] It is often enough to open the subassemblies and set the components to solved
I admit I use this method when I open a machine (about 300 unique pieces, 2500 with repetitions, opening, waiting) I open the sub-assembly(s) that are causing problems (opening, waiting), I correct the problem and go back to my machine (waiting, rebuilding, waiting). The worst thing is that I even pay more attention to these problems, it's normal. With this method, it is very difficult to convince new users of the advantage of 3D
Do you find it normal to spend more time trying to have an assembly without software errors rather than checking if there are any design errors?
The speedpack is very powerful and considerably reduces the opening times, but then if you want to modify the whole thing you have to load the complete model and then update the speedpack in each configuration.
Today I feel like I spend more time correcting constraint errors than designing new assemblies, yet we try to make assemblies as clean as possible:
- Stresses on a single fixed base part -no loops in constraints - no constraint on parts from repeat functions - no advanced stress (symmetry, gearing...) -No flexible set -decomposition into sub-assemblies to reduce the number of first-level constraints (no set with 300 1st level pieces but a set with 10 sub-sets of 30 pieces)
- no component chain (3 or 4 max)
- Never a part with sketches or functions with external references
-one part is constrained in relation to only one other part (the basic one if possible) and not two
In short, we follow the recommendations: http://help.solidworks.com/2012/French/SolidWorks/sldworks/Best_Practices_for_Mates.htm Despite this, the opening times are long >5min for a complete machine and there are often errors.
On the other hand, we do not totally constrain all the parts, there is often a degree of rotation for the
cylindrical parts (entratoise, pin, pin, screws, washers...)
First of all, as far as the cylinders are concerned, I use exactly the same solution with the 3 constraints. And it is virtually impossible to use two cylinders simultaneously to move the same workpiece.
No, I don't think it's normal to spend more time fixing the incessant problems of stability than designing. Just like you, these problems are routine and daily to such an extent that they become almost "normal".
In your second message you mentioned the "SpeedPak". I would like to know if you use it frequently, if it saves you time and stability and how do you use it? Because as I wrote before, I have tested it several times on recurring assemblies but without any real benefits.
Can you please also tell me which version you are using and the type of your pc?
Personally I have been using the 2013 SP4 EV version for a short time. (I signed up for the Early Visibility program to provide feedback to developers.) My pc is running Win7 64 bits, I7 processor 2.8Ghz, 8GB of ram and a Quadro FX1800...
What to do: don't forget the basics, don't get into bad habits related to what SW allows and keep simple constraints with simple configurations
I've never done big assemblies on speedpak but I'd be curious to see the difference in loading time between a "speedpak" assembly and a "light" assembly just in removing elements.
Has anyone already been able to do this test?
I've never seen how it works with other software?
Can someone give their opinion?
For my configuration, I use SW2012 SP05 with Win7SP1 64Bits
I take the liberty of relaunching the subject because, 4 years later and with Solidworks 2017 (SP3) for my part, the problem is still the same!!
The "large design, complex assembly" modes do not change the problem since the constraints are not rebuilt if the solved mode is not activated.
Even if Speedpack has improved a little, it's still a method where you have to update and therefore wait in front of your machine.
The only plus is the "lock rotation" option in coaxial stresses, so assemblies can be easily fully constrained.
My method is the same as y.pacquelet, with system settings that prohibit the light mode.
More and more customers are asking us for ifc files to do BIM, great, solidworks knows how to convert its files to ifc (2x3 or 4) on the other hand, you have to be patient, very patient. Like several hours, or even days depending on the crashes to create an ifc file of several GB. The assembly in question has "only" 3500 parts...
My PC:
HP Z440 with
Intel Xeon E5-1650 v3@3.5GHz
16GB RAM
Nvidia Quadro K5200
Windows 7 64-bit SP1
Software limit (tjs no multithread my proc never exceeds 12% usage)
Machine limit? (RAM saturates when exporting/importing large files)
Thanks to the SW developers for thinking about the guys who spend a lot of time waiting in front of PCs wondering if the machine is well sized, why we only find PCs with 8-12-16 processors when only one is used at a time...
Personally I would put the connecting rod for the piston directly on the big wheel, then increase the height of the cylinder to increase the piston stroke and have more inertia towards the end of the stroke. But it will probably be necessary to provide grooves in the cylinder to pass the connecting rod.
Or simply create an inertia wheel that would have to be driven when empty, and with a cam fixed on the big wheel drive the piston that comes back to it thanks to a spring (a bit like a valve).
