Best Practice and CAD Snobbery
This is one of those things I’m wondering about. In SolidWorks, I’ve written pages and pages talking about potential best practice approaches. It really gets tiring. And then I argue that in Solid Edge, you don’t have to worry about best practice. But then I heard Matt Johnson was presenting at his Ohio user group on Solid Edge best practice. It’s really hard not being an expert.
I can understand that ordered modeling in Edge requires best practice in the same way that Works does. But what happens in Synchronous modeling? My feeling is that the model is its own best practice. As long as the geometry is correct, and free of “funk”, it doesn’t matter how you got there. The only things that are going to hinder ST from working are “funk” in the model, like small vestigial faces, unintended non-tangencies, or faces that might have been knocked out of square accidentally.
Making the jump from a system where the rules are so rigid, and you spend literally years understanding them and writing them, it’s hard to just totally give up on best practice and suddenly say its ok to do “hack and whack” modeling. I’ve spent years developing senses as a “CAD snob”, and now I have to give up that sense of superiority.
I’ve spent most of last week talking with a guy who just can’t let go of all that over on the Dezignstuff blog. I get the sense that this is going to be a common argument. Guys who used to go to work every day in a starched white shirt, red tie, and blue blazer are going to have difficulty just going in jeans and a t-shirt. Whatever gets the job done is correct. That’s a hard pill for CAD snobs to accept. I know, it’s one I’ve been wrestling with myself, and I can see I’m not the only one.
Hi Billy
I will do my devil advocate 🙂
“..Keep features to a minimum for regeneration time..”
In term of best practice, not sure this one should be include. Maybe this can find it’s root in people’s being afraid of management, but keeping feature to a minimum if probably one of the best way to get yourself in trouble.
People get in vicious circle by saying i have problem i should reduce my feature to reduce problem. They keep having problem and keep reduce feature….
We should build with the maximum feature as possible keeping the GBS in mine ( Gros Bon Sense in french CG Common Sense in english)
If you need 100 features don’t try to make 10 that just not make sense.
And this less feature is in contradiction with
“…Figure out method of modeling part such that you do not build yourself into a corner, and have to start over due to reorder and change model hierarchy…”
Check this link and browse the topics on the left
http://www.soliddna.com/SEcommunity/page/help/part.html
Matt,
Best Practice’s pertain to both worlds, for similar and different reasons.
In the SW and or history world, BP main function is to also a creation of a model that delivers design content, that is changeable, without blowing up part. BP is SW and or history modeling is there for “style” of modeling, not really to benefit the final part geometry. In history modeling of complex parts much of design intent of the final part geometry is lost all of the dimensioned 2d sketches that a lot of times have nothing to do with the final part design. BP are there in History to keep models efficient, stable and to keep part blowouts to a minimum.
In History Only Some Are: (there are many only a few listed for example):
Keep feature tree clean and orderly
Keep features to a minimum for regeneration time
Fully Constrained Sketches
Figure out method of modeling part such that you do not build yourself into a corner, and have to start over due to reorder and change model hierarchy
Filets, chamfers last if possible
In Synchronous Tech there are Best Practices too. I’m sure there are many that I have not thought of yet. But here are a few things that we are doing. But the BP in ST are not there to keep features blowing up. From what I believe BP in ST allow you to harness the technology of ST.
A few things we do in ST, follow:
Easy one – Create part geometry about origin in such a manner that best utilize Live Rules
Side note: >> If you don’t do this step properly initially, power of ST allows moving entire part to more effective location to take advantage of live rules (a no-no in History cause it just adds another feature for clean part u have to start over)
Easy one – Fillets, chamfers later in part if possible, (we use both sync and ordered)
Easy one – So on every part we GROUND 1 face and make all others RIDGID with persistent relationships to lock all part geometry down.
We are not fully constraining parts with dimensions, {{{something that gives you an uneasy feeling for a while, but before parametric solid modeling we did not constrain everything in a 2D sketch did we}}} (locking every detail with a red dimension).
This is because Live Rules controls a lot of part behavior. Understanding part behavior that is controlled by the Live Rules is the biggest challenge. With ST5 and the implementation of the Live Rules solution manager greatly helps understand the dynamics of control that is being undertaken by Live Rules.
Easy one – All sheet metal done in Sync Tech
Easy one – Use Assembly Groups
These are a few things that we are doing off the top of my head I’m sure that as time goes along we will adhere to more Best Practices as we learn more about Solid Edge.
After all when you boil it down, it’s only about “Designing Better” and a few Best Practices thrown in there only helps.
Larry brings up a point here about locking things down. When I mention locking dimensions down I don’t mean the ones you are using to make the edits by. However there are times where unrelated things you don’t want to edit will if they are not locked down. For instance sometimes I create a set of holes that are related to each other center to center. At one end though I want to create a dimension from edge to a point on the edge of a circle. If the dimension is not locked down on the circle the diameter may change as an unintended consequence of the xy edits. There are design features I want to preserve when I create parts and not have them edit by accident and locking things down will prevent this. I can still change everything by unlocking but I want it to be a deliberate choice to do this and just to the things I want. Like many other users I find live rules are a bit tedious at times and I just turn them off and control it all with dimensions.
Dave, you are right about the way some things need to be locked. I still do hole patterns in ordered!
Larry, I am with you that Sync Patterns needed some work. They are in concept a beautiful thing (no massive recompute required), but in practice had some issues. We’ve learned a lot from users about this and ST6 I believe has all these issues addressed. I urge you to give patterns a “re-try” with ST6. I think you will be very happy.
Leaving dimensions not locked down is the hardest thing to get used to. My old habit was to fully constrain each part and make any changes in the variable table when editing in an assembly. Then I went to the meeting in Massachusetts and saw that you can grab an unconstrained face in assembly and simply move it where you need it. No need to open a variable table or open the part. You could almost actually see the light bulb flash above my head. Remember control+spacebar switches from part select to face select mode.
If the part is created symmetricaly about the base plane and you have it constrained to one end in the assembly it will change by double the distance when you move the other end. That introduces a small problem, because if you move the face to a picked point it will end up moving twice as far as you wanted it to. I have to determine the distance I want it to move and then, after initiating the move, enter it in the input box with “/2” behind it to divide it in half.
Without seeing Matt’s presentation the only thing that comes to my mind is dimensions. If you don’t dimension everything there can be problems with edits at times. If you dimension everything but don’t lock the dimensions down there can be problems in this case too with unintended consequences. It used to be before ST5 that you could dimension the same thing twice and lock it down and quite often this would trip people up as they would only catch the one and wonder why the edit would not work not realizing that the other locked dimension still prevented edits.
Matt,
Not sure what Matt Johnson presented, but I suspect it was good and well organized.
I think Synchronous Technology does still require some form of “Best Practices” but not necessarily to the same level as History based modeling. I model almost completely in Synch now, and once was one of those “CAD snobs” that required a lot of persuasion to unbind myself from my old habits. But I find the lessons learned in History Based Modeling still have a lot of value and importance in Synch as well. Like placing Radii as a separate feature instead of adding this detail to the base body. Although Synch handles these radii well in either scenario, I have had occasions where the results were not as intended with Sync… and not because of anything Synch did. It’s just that Sync can’t anticipate every possible scenario and deliver the desired results 100% of the time… although I’m finding this less and less of a problem as I become more familiar with it’s Sync’s mindset.
Bottom line is there still needs to be Best Practices, it’s just that synch allows you to manipulate the geometry later with less clean up. Cause we all know that even with the most robust set of “Best Practices” in History, you never seem to make a perfect model when going back later… something is always lurking and looking for trouble.
Now as far as me giving up my Blue Blazer, White Shirt, and the Red Bow Tie… that ain’t happening. Although my wife would be happy if I could at least put on some pants!
Bob