Sync variables problem in ST3

[beachcomber]

I'm trying to model a plate with a staggered hole pattern in ST3.
As i've discovered there is no stagger option in sync mode but thats not a problem, I can put 2 holes and then pattern each one.
What I want to do is have the plate size increase/decrease according to how many rows/columns in the pattern.
To achieve this I've created variables for for all the required parameters and linked them as required but the plate will not change size.
The error message I get is -

"The edit of the feature variable cannot be applied to the model
The original value will be restored"

Any ideas?

bc.
2.4GHz Core2 Quad, 4GB RAM,
Quadro FX4600.

Where would we be without sat-nav?

 

[teebarr]

BC, there is a workflow I noticed a while back being used by some trainers and that is to do patterns and rounds in ordered. I have subsequently adopted this approcah and found that sync edits work much better as the ordered features are not considered during the edit. Typically I will only use sync pattern tool for small patterns - all others are done in ordered.
Tony

 

[beachcomber]

But does that work if I want to link the variables?
Basically the length and width of the plate is determined by how many rows/columns in the pattern. The distance of first hole from bottom and LH side of plate is constant, as is distance of last hole from top and RH side.
It isn't essential that I do it in sync, I was just experimenting to see IF I could do it.
If the holes have to be put in in ordered there is no point doing it in sync at all as they are more or less the only features.
What disappoints me is the difference between patterning in sync and oredered.

bc.
2.4GHz Core2 Quad, 4GB RAM,
Quadro FX4600.

Where would we be without sat-nav?

 

[teebarr]

If it is just a simple plate then I would say ordered is better on this occassion. The advantage of ordered pattern is that it is defined by a sketch which can be constrained at BOTH corners and move with the plate size so no formulas are required. Sync pattern can only be constrained at the seed feature so a formula would be required to drive the plate size.

For me there are some things sync is very good at but I have definitely had less predictable results when driving geometry via the variables table. Horses for courses I guess.

 

[beachcomber]

Sync pattern can only be constrained at the seed feature so a formula would be required to drive the plate size.

That's the problem - the formula doesn't work.
The plate size doesn't change with the pattern.

bc.
2.4GHz Core2 Quad, 4GB RAM,
Quadro FX4600.

Where would we be without sat-nav?

 

[teebarr]

It's working fine here. Are you sure you are selecting the variable name to construct the formula as opposed to the value?

 

[beachcomber]

I'm definately using the variable names.
I've created my own variables, with appropriate names and vales then linked the actual driving dimensions to those, rather than just renaming the dimensions.
No matter what I do I can't drive the plate size from the variables that define number and spacing of holes, and the distances to the edge of the plate.
For example I have something like:
FIRST_HOLE1_XDIST = 55
FIRST_HOLE1_YDIST = 53
ROWS = 6
ROW_SPACING = 30
COLUMNS = 10
COLUMN_SPACING = 60
LAST_HOLE1_XDIST = 85
LAST_HOLE1_YDIST = 53

Then the plate length variable is defined as
PLATE_LENGTH = FIRST_HOLE1_XDIST + ((COLUMNS-1)* COLUMN_SPACING))+ LAST_HOLE1_XDIST.

I then set the plate length dim eg. V1234 = PLATE_LENGTH
Plate width is done in a similar way.

When I change the number of columns the plate length will not change in the sync environment.

bc.
2.4GHz Core2 Quad, 4GB RAM,
Quadro FX4600.

Where would we be without sat-nav?

 

[teebarr]

I think the original assertion is correct - don't use sync patterns for this type of thing.
From what I can see, while a patterned instance can drive the plate size, it only seems to work when the number of instances remains the same. Once you change the number of instances it fails.

I've come to realize that adopting a mixed modelling approach is the route to getting the best from SE.

 

[beachcomber]

I've done a bit of experimenting and while the plate length and pattern are independant it works fine, but as soon as you link them it fails.
Very poor really.

bc.
2.4GHz Core2 Quad, 4GB RAM,
Quadro FX4600.

Where would we be without sat-nav?

 

[teebarr]

BC
Yes it is poor IF you confine yourself to sync mode only - but why would you do this when you can use the strengths of both?
I don't know of ANY direct modeller that can handle the sort of parametric operation you want to perform.

That said, I understand where you are coming from. I used to try everything in sync because it looked like the tools were there to do it, but I hit too many dead ends and limitations.

This is quite a good article as to why the mixed approach is valid

http://www.dezignstuff.com/blog/?p=6103

 

[beachcomber]

Teebar,
I was doing this to compare the functionality of ordered and sync modes, and for such a simple problem you should really be able to do it equally well in either. Not being able to drive the 2 parameters from a common variable is poor.
I did try the mixed approach and it works - I can create a variable and use that to drive the plate size in sync mode, and the hole pattern in ordered.
I can also do it all in ordered - which is what I would probably do anyway. I've almost given up now with sync mode as it always seems to fail when I come to do real-world problems like this. I also find it far too complex, but that's possibly an age thing. It seems OK for imported parts that you might want to modify.
I also see the mixing of the 2 modes as a bit of a failure really - a sort of compromise because sync won't work for every case, although the link you provided does explain reasons why it might be necessary.
Thanks for your input.

bc.
2.4GHz Core2 Quad, 4GB RAM,
Quadro FX4600.

Where would we be without sat-nav?