Chasis Modeling w.r.t. the Veh.Co.Sys.

[salil]

Hi All,



Working on theComplete veh. chasis modeling & all comps. are to be modelled w.r.t. the Veh. Co. Sys. here my quest. is do I need to use only the default co sys. in the prt as V.C.S. or I can also use a seperate co. sys. which will be at some dist from the V.C.S. as per the comp. location(Part will be having 2 co.sys one will represent the V.C.S. & other wll be at the the intersection of the Datums of the individual comp)



Regards,

Salil.

 

[Nose Bleed]

I wouldn't see any problem using the default as a standard for all the components. That is, unless you are using intralink where other designers have used different locations...

 

[dougr]

Working on theComplete veh. chasis modeling & all comps. are to be modelled w.r.t. the Veh. Co. Sys.



Sounds like it's up to you as long as have the V.C.S placed ACCURATELY somewhere in your part or assembly.



The basic premise behind this is that you can assemble sub-assemblies and parts simply by using a coord-sys to coord-sys mating relation.



These leads to the holy grail of automating assemblies.

 

[miked]

I create and place an empty part in the assy using the default coord sys of the part and the vehicle assy coord sys to mate.

Once you have assembled the part you can begin building features and copying surfaces from the vehicle as references and they will share the same coord sys.



There are two problems with this:



You will need another csys for the part in tooling position, which you can solve by creating an assy strictly for placing the part in tooling position and exporting from there.



The other problem is rotating the view of the part.

If you use anything other than wildfire, the part will rotate about the vehicle assy coord sys which is suboptimal if you are working on something at the back of the vehicle because your spin centre will be more than ten feet from your part ,depending on the length of the car.



* you can remedy this slightly with pre wildfire versions if you create a mapkey that changes the spin center to the middle of the screen but pro has trouble remembering this setting and you will have to conitnually use your mapkey to reset it.

Mike

 

[dr_gallup]

I think you would be better off modeling your part with respect to the default Coor Sys and putting in a sepperate VCS Coor Sys. This way you can suppress the VCS when you don't need it and not have all the problems associated with having the geometry 10 feet away fom your spin center. Might want to make a family table of the part and put the VCS only in the instance that goes in the assembly.



Personally, I think assembling everything to a VCS is very poor modeling practice. It is easy to implement and maintain because you don't have a lot of external dependancies but it also means that your assemblies are not parametric. If you change the suspension the wheels stay in the same place so you might never notice that now they interfere with the fendors until the car gets built!

 

[miked]

Its the standard in automotive.

If they cant take your part and drop it into their assembly you do not get paid.



They will not interpret anything nor will you be able to specify which coord sys they use when they assemble your parts.



I actually think its a good system and it prevents cumulative errors that you wold get from positioning parts relative to each other.



For wheels they use instances for each position and they provide a jounce file that is basically a single surface to represent every position of the wheel.

Mike

 

[dr_gallup]

Well, the cumulative errors will show up in the finished vehicle cuz the parts are going to be bolted to each other, not some hypothetical coordinate system. I understand the reasoning, I just think it is flawed.

 

[miked]

Cumulative errors are pretty common.

I have seen 3/4 inch build variances from one side of the vehicle to the other.

Is it possible that using the vehicle coord system makes it easier to track the errors?

There must be some compelling reason for every vehicle manufacturer to use the same system?

Does anyone know why they do it this way?

Mike

 

[Brian_Adkins]

I think it all just goes back to the pre-3D days for both automotive and for aerospace (stations and waterlines). It was important to reference everything from a known common location.



With the advent of solid modeling, knowing the precise x,y,z locations for a rivet became slightly less important because you could just make it follow the hole it fills.



In industries where part (and model) re-use is critical, it started making sense to switch away from a system where every part had to 'know' where the global vehicle origin was. In fact, that system can make part re-use down right difficult because any given part would have to contain the Veh-Co-Sys of every vehicle it gets assembled to. If the part is used on 20 vehicles, it's got to contain 20 coordinate systems.



In other industries (e.g., automotive and aerospace), the practice seems to have stuck... especially for models without much re-use (e.g., interior & exterior styled components).



Another factor that makes the global-origin system appealing is when you are farming out large portions of the design (again... automotive and aerospace). Making your suppliers reference a single csys makes it easier to reintegrate everything as data comes back.



Just my thoughts on the subject...



We still use both methods here and they both have advantages and disadvantages. We tend to use the common-origin system for the top 2-3 levels of our assembly structures and then mate-align from there on down.



-Brian Adkins

 

[dougr]

Think the real issue here is the technique of using the csys-to-csys constraint for assembly:



1) It's much easier, especially on the higher level assemblies, as only one assy constraint is required instead of 2 or 3



2) Assembly can be done blindly by using select-by-menu. This eliminates the need to rotate big assemblies in order to select mating faces.



3) Mates and aligns only work when you have something to mate and align to which is not always the case.



4) Csys assembly also eliminates a bunch of parent-child relationships so tend to be significantly more stable.



I do believe having a VCS in a part model is nfg as the only real advantage comes at the higher assemblies where mates and aligns can become really hard to do.



Also the VCS only needs to be in the top assembly - other section or module csys can be referenced off the VCS in this assembly too.



Anyone with experience in large assembly modeling sure appreciates the virtues of assembling by csys.



both automotive and for aerospace (stations and waterlines).



Actually stations, waterlines and buttock lines stem from the ship building industry and like any other convention are very useful.

 

[Brian_Adkins]

>>stem from the ship building industry



Yeah... I never worked in shipbuilding, but they are still used in aircraft design as well (although I've been out of aerospace since '92).



what is nfg? ... never mind... just looked it up



-Brian

 

[dougr]

Brian,



Just in case:



nfg - no f****** good.



If the part is used on 20 vehicles, it's got to contain 20 coordinate systems.



Just leave the VCS out of parts and restrict them to upper level assemblies where they are useful and really needed.

 

[Brian_Adkins]

nfg... got it.



However, VCS within parts definitely has some useful applications as I mentioned...



With parts such as styled surfaces that must be continuous all the way around a vehicle, using a common csys within the part is still the best approach in my experience.



We currently do this with parts such as hoods, molded exterior tanks, fenders, cowlings, etc. On a given vehicle, these part surfaces should be continuous (c2 would be nice) all around. Therefore, the visible surfaces are usually modeled using the master-model approach which means they all have a common csys (usually vcs). These parts are almost never used on other products, so part re-use is usually not an issue.



I haven't worked in automotive, but I can see the exterior sheetmetal and interior styled components being handled the same way with the same benefits.



-Brian

 

[miked]

In automotive the entire exterior starts off as a surface file then it gets passed down the line for each division to part out (body in white, which consists of the sheet metal surfs but no offsets)

For anything that fits onto the vehicle I put this BIW file in my assy first then copy surfs from it as required.

It can be time consuming because the quality of the files varies considerably between the different manufacturers and you have to stitch them up before oyu can work on them.

Mike