I worked at one company that did some avionics related work. One of the biggest single problems they had was swapping out clamps per customer design reviews at the end of projects. They had roughly 6 different family tables of clamps, and they were all modeled differently and in different orientations. Some of them had the default coordinate system at the center of where the routing of cables went through and some were facing left or others right, I think there may have been one or two that were modeled so they were extruded in the Y direction.or X direction, rather than all being in in the Z direction. To top it off there was also a glitch with some of them, they would often "flip" when you set them to fixed, I believe this was due to someone defining the coordinate system so it conformed to Catia's orientation, where I believe the Z direction is up rather than toward the user.
Anyway, due to this difficulty they didn't really do their harnessing the way it was supposed to be done. I don't know harnessing that well and don't know all the particular, but they were just routing a basic harness to represent the routing to customers.
I developed aclamp model that I believed was the most practical solution. It was difficult to develop, and utilized many of the things I had learned over the years. One model containing all of the clamps that would be used. It would definitely streamline the process of swapping out clamps, AND preserve routing references as well. I no longer work there, as far as I know it's possible they they are using them now, but at the time they weren't utilized although the head of engineering originally said we should. I think the main problem was actually that if they worked as well as I expected them to, that others would be in trouble for all of the difficulties they had with routings over the years. Most of the "concerns" voiced were pretty silly to me, one being how to tell what generic an instance was from.
So basically, I'm wondering if I was right or wrong, I can't see how they wouldn't improve the situation, and potentially totally resolve difficulties with swapping out clamps at the last minute, or even earlier in the process. I thought it was an acheivement to be able to get all of the clamp variations in one generic model.
This is a Creo 2.0 model.
I don't agree, they're not equal, I've done design automation using pro/program, and interchange assemblies were always a second choice for me, with family tables being the first choice.
I'm well aware of the difficulties many users have had with family tables, expecially in a pdm system, but I haven't had those problems because I know what I'm doing and don't just plow through it like some do. It's really a matter of hiring more qualified and better users, but a lot of companies don't really consider CAD skills to be very important in the hiring process, or even later on.
I've seen a lot of issues with interchange assemblies, individual users having different ones, I've seen a lot of situations where pro-e knows there's an interchange assembly but it isn't available becuase it's some individual user's or has been lost. A lot of them haven't been done that well, and getting every single referenceevery user might select can be difficult.
The actual replacement can probably be done, but it can be so much cleaner and simpler with family tables.
PS: I hit post to quickly, there area lot of other references to be considered, such as attachments in drawings, exploded views and simplified reps which canall gomuch moresmoothly when components are replacedwith family tables.
I dislike that interchange assemblies need to embed a link to the assembly in each part, which complicates things when an item might be part of more than one group or that some item might have been released and now needs to be revised prior to being included, even though zilch has changed except for the embeded reference.
I agree on the training part. I came across an assembly with a family table with about 130 columns of essentially gibberish. By adding named groups, this was reduced by nearly half by placing Y/N on the groups rather than Y/N on the individual components and features. How long would it last back in the original hands? Not 10 seconds. I exaggerate a bit.
One thing that would help is if PTC allowed for in-tree per-feature and per-component comments. Based on my view of parametric modeling as an interactive, GUI based software development environment for an intepreted program that happens to produce shaded images and drawings of parts, I have to ask - where can comments be placed? Other software places comments in-line with the code so one can see the comment and the code at the same time, but PTC doesn't, at least for parametric modeling. There are comments for some things, but they are hidden and so you can't know they are there unless you ask to see them. Mostly they aren't there, like for columns in family tables.
Dan, a family table is one of the better solutions. The only thing that occurs to me worth trying, and also just as much a start-over as the whole family table, is creating a prototypical clamp routing model (datum plane, points to guide the cable, maybe an axis) and use inheritance to duplicate it in clamp geometry models so that swapping unrelated will have the identical basis. That way, each clamp model can be managed independently, but still offer interchangeability.
Now to stop users from polluting the pool with free STEP models that are built badly and have no redeeming characteristics, but can arrive with hundreds of layers. Yay.
I had used an existing part from a company's database which wasn't very accurate, for a few reasons, I was told to use the part number and that's what was in their database, so using something else would create issues with multiple models, and although it was modeled poorly, if I discussed it with my supervisor he would have gotten ticked off. Anyway it ended up being used in quite a few places, when all of a sudden some guy who liked to inject himself into things and cause problems decided to replace one or two of them with a step file, which caused problems with other assemblies not being able to find the part or if it did, place it. It was a totally ridiculous and foolish act on his part. It took a bit of time to resolve the problems he caused. It was just a purchased handle and it really didn't matter or affect anything, in any way, at least on a design or engineering basis.
They also had a very emphasized rule there, that if they weren't manufacturing it, or profiting in any way from it, they didn't want to invest time making parts look prettysuch asputting cosmetic rounds on them. Since he made the accuracy of the partan issue, I spent about the same amount of time that he took todownload the step file, to modify the part so it looked more like the actual part and got rid of his stupid step file part that screwed up everything. If he would have put a little effort into modifying the part in the database, he would have prevented a good four to five hours of work fixing his foolishness.
But none of this has much to do with all of the problems related to changing out clamps, unless of course you've had users bringing in step files of clamps and have been forced to use them when swapping out clamps.