how to apply mate to parts in the same sub assembly in an upper level assembly?

[bnemec]

Due to assembly structure requirements, we have things that are modeled in a sub assembly that is not fully defined, then used as flexible in the upper-level assemblies they go to. It is normal to need to position components in the same sub assembly in an upper-level assembly. How can we do this?

Did not have this problem in Solid Edge, this is new to me.

image.png (5.48 KiB) Viewed 5598 times

Thank you

[berg_lauritz]

Can you be a little more specific?

The error you get is due to the fact that you are trying to mate from Assembly A but you only select components/mating entities from Subassembly B within A.
Is that the intention? Why are you not able to mate those earlier? Can you choose different entities not from Subassembly B for mating?

Can you set up your assemblies in a different way?
Example:
We "split" our hinges into 3 configurations.
1 full hinge (part a + b)
1/2 hinge part a
1/2 hinge part b
"1/2 hinge part a" does not show up on a BOM. So if we ever have to put the movement in, we can just put the config "1/2 hinge part a" in and on the top level place the configuration "1/2 hinge part b". The movement can be simulated like this without having to have any flexible sub-assemblies.
I know that this is not always possible - esp. regarding more complex movements/assemblies but maybe this helps.

[DanPihlaja]

Position the sub assembly as needed in the upper level assembly.

Then edit the sub assembly in the context of the main assembly.

Add the mates inside the sub assembly

Now your sub assembly is in the position you want.

[bnemec]

Position the sub assembly as needed in the upper level assembly.

Then edit the sub assembly in the context of the main assembly.

Add the mates inside the sub assembly

Now your sub assembly is in the position you want.

For one where used, but not all the others. This is the point of a "flexible assembly"

[bnemec]

Can you be a little more specific?

The error you get is due to the fact that you are trying to mate from Assembly A but you only select components/mating entities from Subassembly B within A.
Is that the intention? Why are you not able to mate those earlier? Can you choose different entities not from Subassembly B for mating?

Can you set up your assemblies in a different way?
Example:
We "split" our hinges into 3 configurations.
1 full hinge (part a + b)
1/2 hinge part a
1/2 hinge part b
"1/2 hinge part a" does not show up on a BOM. So if we ever have to put the movement in, we can just put the config "1/2 hinge part a" in and on the top level place the configuration "1/2 hinge part b". The movement can be simulated like this without having to have any flexible sub-assemblies.
I know that this is not always possible - esp. regarding more complex movements/assemblies but maybe this helps.

I was hoping the process could be defined in more generic terms. Such as assembly A has parts 1 and 2 that are not fully constrained. Assembly B has components A, 3 and 4; component A is set to be flexible. In assembly B I can mate part 1 to part 4. How can I mate part 1 to part 2 in assembly B and leave them loose in assembly A?

Why do this? Because we have subassemblies in CAD to match the BOM structure in ERP, the hardware and other bits are positioned differently in the various where used. It does not work very well to make them each a config in the subassembly.

[Frederick_Law]

You can't add mate to components in the Flex assembly.
You need to mate them to components in the Upper assembly.
Upper assembly should have features to locate the flex assembly.

I use flex on cylinders. To show ext, ret and free.
I use "Positional Rep" (IV) or config.

Need to set the assembly Flexible in the Upper assembly.

[Frederick_Law]

How can I mate part 1 to part 2 in assembly B and leave them loose in assembly A?

You cannot mate Part 1 and Part 2 in B
You need to mate Part 1 to B and Part 2 to B.

Door hinge assembly. Upper and Lower. Free to rotate.
Put it in cabinet assembly.
Upper hinge mate to door, Lower mate to cabinet.
Add angle mate between door and cabinet.

[bnemec]

You cannot mate Part 1 and Part 2 in B
You need to mate Part 1 to B and Part 2 to B.

We've learned a lot of new ways to accomplish the same end while migrating to SW and much of it's ok. But I've been blind sighted (or blindsided) by a few missing functions that ended up with crappy work arounds. This is one of them.

[jcapriotti]

@bnemec I think I see what you want although I don't know why without seeing an example. We have lots of kits that we manufacture that get assembled in the field. So we create "bolt" kits as assemblies that contain for example:

10x Hex Bolts
10x Flat washers

Then what you want is to insert that kit "as flexible" into an upper assembly and mate them together at that level, right? I see the error, but we've never encountered it as we would mate the bolt to the washer in the kit sub-assembly. Then just mate the "group" to the location in the upper assembly. You must have a different use case I'm not visualizing.

[Frederick_Law]

He try to add mate on bolt and washer in upper assembly.

[SPerman]

I typically create configurations for critical positions in the sub assembly. (For example, an actuator will have fully retracted, fully extended, and any other critical positions needed.) These configurations have to travel all the way up the assembly to the level where you want to control the position. It isn't the quickest way to solve the problem, and it can be a bit of a mess if the actuator is used across different designs and needs several different positions. Despite the time investment, it has been rock solid.

[bnemec]

@bnemec I think I see what you want although I don't know why without seeing an example. We have lots of kits that we manufacture that get assembled in the field. So we create "bolt" kits as assemblies that contain for example:

10x Hex Bolts
10x Flat washers

Then what you want is to insert that kit "as flexible" into an upper assembly and mate them together at that level, right? I see the error, but we've never encountered it as we would mate the bolt to the washer in the kit sub-assembly. Then just mate the "group" to the location in the upper assembly. You must have a different use case I'm not visualizing.

As Fred said, mating bolt to washer in upper level. Not exactly that simple but it paints the picture. Sometimes we can do as you state; mate the washer on the bolt in the subassembly then position the bolt in the upper level. But when there's more hardware with various combinations it must be done in the upper assembly.

So now when there's a stack of parts in the subassembly the only option we've come up with is to position them with distance using "magic numbers" for the distance or aligning them with parts that are far from adjacent. Then some just gave up and locked/fixed the parts.

[Glenn Schroeder]

I feel like I'm missing something here, but for the bolt and nut assembly you can leave it flexible, then mate the bolt and nut to their appropriate locations in the upper level Assembly. That shouldn't throw any errors.

Or is this example too generic?

[bnemec]

I typically create configurations for critical positions in the sub assembly. (For example, an actuator will have fully retracted, fully extended, and any other critical positions needed.) These configurations have to travel all the way up the assembly to the level where you want to control the position. It isn't the quickest way to solve the problem, and it can be a bit of a mess if the actuator is used across different designs and needs several different positions. Despite the time investment, it has been rock solid.

For what we're doing, we've avoided configs for things like this. Couple of reasons.
The sub kit should not have an awareness of how it will be used; we shouldn't need to edit the sub assembly file when using it in a new way. That is cemented with states in PDM as the sub assembly is in a released state with rev control so editing it to add a new config would require PDM admin to put the file in WIP with a no rev change; something we're trying to keep out of our normal process.
Second reason you already mentioned, it's a bit of a mess when the where used tree fans out to dozens of files, with unknown number of people maintaining the models for the next couple decades or until we change CAD again.

But as usual it's a choice of the least bad options.

[Frederick_Law]

A flex robot arm will go crazy. Too many DoF.
You'll need some way to limit it's movement/travel which usually is not possible in CAD.
Sometime you'll need to add plane, axis, point, sketch to help locate it in upper assembly.

[bnemec]

I feel like I'm missing something here, but for the bolt and nut assembly you can leave it flexible, then mate the bolt and nut to their appropriate locations in the upper level Assembly. That shouldn't throw any errors.

Or is this example too generic?

Hi Glenn, add a few custom spacers. Keep in mind there's probably 4 screws and let's say 8 spacers. Some will have 4 spacers on 2 screws, some will have 2 spacers on four screws. The problem goes back to the MFG BOM only cares about the pn and count (well along with where it's consumed and such but...) it does not define the position. So to actually work in all the where used we need to mate spacers to spacers sometimes.

But as Fred plainly stated SW does not support that. I'm a fool for assuming it could.

[SPerman]

But as usual it's a choice of the least bad options.

This is the Solidworks way.

[Frederick_Law]

Sometime its process, purchasing and sales want their BOM a certain way.
I've encountered a recursive part list.

[bnemec]

Sometime its process, purchasing and sales want their BOM a certain way.
I've encountered a recursive part list.

We haven't done that yet. LOL. I can't imagine what would happen when MRP rolls that night after someone tries that. There might be some locks that prevent it.

[jcapriotti]

This is why our BOMs are often manually done. We design a complete product but it's assembled in the field on the job site. Manufacturing builds sub-components of the products and does a lot of kitting and combining that makes little sense from an "Engineers" view. But they are forced to work with MFG and build the structure their way so often our models don't reflect the BOM.......meaning no kits usually in the model. Or if we do, its a flexible kit sub-assy. Wiring is worse than kits with harnesses and the different ways tot route the same thing. Honestly it's a bit of mess and I envy those places that get to completely work on a model and BOM that match.

As we move to Windchill, we now have a global standard design engineering view. The model and BOM have the same structure. Windchill allows you to create a "Mfg" BOM view where you can rip the Eng-BOM apart, insert sub-assys and kits, etc. We still have the challenge that the MFG BOM and engineering 3d model/drawing don't jive. Now we are creating a near duplicate Mfg drawing and in some cases a near duplicate Mfg model because they deviate so much from the E-BOM. I'm not sure which way is better but I can't imagine that any CAD system handles this sort of thing that well.

[Frederick_Law]

Add upper management that keep changing "kits".
One way is just parts in CAD, let ERP do the assembly.

[berg_lauritz]

Here is a strategy for you that you might be able to use, @bnemec (although I know you're opposed to virtual parts):

Use virtual components in your assembly.
At one point we were discussing the possibility of having a library part in our assembly that we would drag in, make virtual & modify to our gusto depending on the situation.

You can set up assemblies/parts with the correct custom properties so that the only thing you have to do, is to modify them.
It comes with all the salt that virtual parts come with though...

[bnemec]

Here is a strategy for you that you might be able to use, @bnemec (although I know you're opposed to virtual parts):

Use virtual components in your assembly.
At one point we were discussing the possibility of having a library part in our assembly that we would drag in, make virtual & modify to our gusto depending on the situation.

You can set up assemblies/parts with the correct custom properties so that the only thing you have to do, is to modify them.
It comes with all the salt that virtual parts come with though...

We've considered VCs in places similar to this, we thought there was merit but end up shooting it full of holes whenever we model real parts that exist in ERP as VCs. The final straw is maintaining the data. When the part is revised and the model needs updated, finding all the where used in CAD can become difficult. I have a SQL querry to find the assemblies that contain VCs who's name contains the specified keyword (the part number for example) but there's no control over the VC names. So we wind up without a where used search on the CAD side, we would find all the assemblies to update based on the ERP where used. That would be acceptable (assuming we could train users how to corrolate the ERP BOM structure to CAD structure) but the edit would still need to be done to every place that component is represented by a VC vs just updating one file that goes to a bunch of assemblies.

[Frederick_Law]

I've made a few "VC" just to fill the BOM.
Not really virtual, a part without model, just the required properties.
ie a hardware pack, ERP can change whatever in the pack and it won't change the model.

[AlexLachance]

How about this:

A promoted assembly with 2 configurations

One for your sub-assembly
One for your upper level assembly

Mated origin on origin.

The promoted assembly configs would serve the purpose you require for the sub-assembly and your upper level assembly, each calling out whatever's required in the BOM. Any future modification would update itself across your products without having to reassign mates or whatevs

[berg_lauritz]

I've made a few "VC" just to fill the BOM.
Not really virtual, a part without model, just the required properties.
ie a hardware pack, ERP can change whatever in the pack and it won't change the model.

Oh, that might work combined with what I proposed with VCs:

You can make an empty config for you real part that you put in there (has all the custom properties you have).
You put the geometry for the parts into that assembly as a virtual parts that are excluded from the BOM.
You can modify the geometry as needed and you would not loose the custom properties.
I might look weird and have a bit of overhead in it, but you could modify the geometry as needed and have the part data, too.
Yes, geometry changes might be a pain in this case, but you would at least FIND it in PDM (because the part is in the same assembly...).

Neat ideas there, @Frederick_Law .

[Dwight]

We "split" our hinges into 3 configurations.
1 full hinge (part a + b)
1/2 hinge part a
1/2 hinge part b
"1/2 hinge part a" does not show up on a BOM. So if we ever have to put the movement in, we can just put the config "1/2 hinge part a" in and on the top level place the configuration "1/2 hinge part b". The movement can be simulated like this without having to have any flexible sub-assemblies.
I know that this is not always possible - esp. regarding more complex movements/assemblies but maybe this helps.

This is what we do, for hinges, slides, and such. Works best for us.

Dwight

[jcapriotti]

We "split" our hinges into 3 configurations.
1 full hinge (part a + b)
1/2 hinge part a
1/2 hinge part b
"1/2 hinge part a" does not show up on a BOM. So if we ever have to put the movement in, we can just put the config "1/2 hinge part a" in and on the top level place the configuration "1/2 hinge part b". The movement can be simulated like this without having to have any flexible sub-assemblies.

We used to do what you're doing in the early days of SolidWorks (90's-2000's). We now build our hinges as an assembly with virtual parts for the pieces, then set to flexible. We've been using them nearly 10 years now and while buggy in the beginning, its a lot better today.

[Dwight]

We now build our hinges as an assembly with virtual parts for the pieces, then set to flexible.

We've never been able to apply flexible more than one level down from the top assembly. Is that still the case with 2022?

[TTevolve]

I went to straight open/closed configurations on hinges, they are tough to do as "flexible" assemblies if you have more than one in an assembly, if you move the first one the others do not move since they are flexible as well. Not sure how you could get around this.

[Frederick_Law]

Sliding door wheel and stop kit.
This is a flexible assembly in IV.
The 4 "sub-assemblies" are mate so they can slide on same plane.
The wheels will mount on different width door. Stop on different length rail.

The assembly is set to "Purchased" in BOM. It will only show as one part in BOM, cannot get parts inside the assembly in BOM.
"Reference" will not show in BOM at all.
All the setting will show differently in BOM. This allow BOM manipulation in part and assembly level.

[bnemec]

I went to straight open/closed configurations on hinges, they are tough to do as "flexible" assemblies if you have more than one in an assembly, if you move the first one the others do not move since they are flexible as well. Not sure how you could get around this.

Use Solid Edge? Inventor maybe?

[Frank_Oostendorp]

We've never been able to apply flexible more than one level down from the top assembly. Is that still the case with 2022?

Simple flexible subs go at least three levels down in SolidWorks 2019. And more.

[SPerman]

We've never been able to apply flexible more than one level down from the top assembly. Is that still the case with 2022?

I've got a flexible assembly 3 levels deep and it is working fine. (2022 SP4.) Every sub in the chain has to be flexible.

I am a GIF. Click me!

[AlexB]

I've got a flexible assembly 3 levels deep and it is working fine. (2022 SP4.) Every sub in the chain has to be flexible.

I also use these quite frequently. I've noticed that flexibility can get turned off several levels deep in one document and not affect another document so it's best to make sure everything is flexible on every level in every document if that makes sense.

I had a case where my top level model had something non-flexible 2 levels deep, but when I opened that sub-assembly directly, the flexibility settings were all correctly set to be flexible. Very odd but worth noting.

[jcapriotti]

BOM-01.jpg
The assembly is set to "Purchased" in BOM. It will only show as one part in BOM, cannot get parts inside the assembly in BOM.
"Reference" will not show in BOM at all.
All the setting will show differently in BOM. This allow BOM manipulation in part and assembly level.

Nifty option names. The SolidWorks equivalent...I think:

Envelope would be the same as reference, however it's set at the assembly tree as a component property rather than at the part/assy level...more of an override. I do like the IV names for them, makes it easier on the end users to pick the right setting.

[berg_lauritz]

We used to do what you're doing in the early days of SolidWorks (90's-2000's). We now build our hinges as an assembly with virtual parts for the pieces, then set to flexible. We've been using them nearly 10 years now and while buggy in the beginning, its a lot better today.

Virtual parts show the "edited but not saved" pencil of death in PDM. Do you use PDM? How do you get around that?

[bnemec]

Virtual parts show the "edited but not saved" pencil of death in PDM. Do you use PDM? How do you get around that?

Is this "pencil of death" to which you are referring?

[jcapriotti]

Virtual parts show the "edited but not saved" pencil of death in PDM. Do you use PDM? How do you get around that?

We've had PDM since 2008. We mostly just ignore the edit pencil as it can be futile in large assemblies to have all components fully rebuilt to the point it never shows it. As for virtual components, they share the edit status as the assembly they belong to.

Before saving assembly:

After saving assembly: