1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Part Error Why?

Discussion in 'Using Alibre Design' started by GIOV, Dec 28, 2019.

  1. GIOV

    GIOV Alibre Super User

    Hello,
    I like to know what's wrong with the piece. Something I'm wrong about or is a problem in the program. Some faces on the surface are hidden. How to correct this.
    Thanks in advance for your help,
    I think is a AD issue regarding to normal surface vector direction or I am wrong:
    https://en.wikipedia.org/wiki/Normal_(geometry)

    GIOV
     

    Attached Files:

  2. bigseb

    bigseb Alibre Super User

    I too think this is an Alibre issue. Perhaps one could get it to work with a ton of tweaking but I don't have the time or patience. Alibre's loft tool is still pretty poor.
     
  3. idslk

    idslk Alibre Super User

    Hello giov,
    you can use simplfy surface in your loft.
    Then you get this:
    upload_2019-12-29_0-24-7.png

    The italic Loft<4> comes up evry time the loft gets edited, maybe there is something wrong with the sketches...
    Regards
    Stefan
     
    Eddy So likes this.
  4. GIOV

    GIOV Alibre Super User

    Sebastian: In general I agree with you because I have several issues with loft and the support are studying.
    Stefan: Yes, I've erased sketch 3 or 5 and now it's fine.
    Thanks,
    GIOV
     
    Last edited: Dec 29, 2019
  5. JST

    JST Alibre Super User

    Concur.... Loft has many issues.

    For that matter, fillets have problems.
     
  6. Ralf

    Ralf Alibre Super User

    The loft feature is not the best, not fault-tolerant, but good to use
    if you have a concept and if you can create clean 2D sketches, but impatience poisons the creation process.
    The same goes for filets, not the best in the world, but useful.
     
    Last edited: Jan 1, 2020
  7. JST

    JST Alibre Super User

    Fillets have the worst problem, actually. There are fewer settings, so less that can be done wrong. Yet the fillet tool has refused to fillet a cubical form, with the usual "capping error" or similar error that does not allow the fillet to "close" onto itself to put a fillet around a face. Not much to be done wrong there, so long as the fillet is appropriately sized for the cube. It points to a problem which may be down in the "dust", the very last "decimal places".

    Lofts..... there are more variables, and so more ways to do something "wrong". But the path of "rightness" is much harder to find. So many things that can and will be "automatically wrong" unless actively blocked by geometry-dependent actions.

    And so much of the time it does "work right", then the next time some minor looking factor is the thing that ruins the entire operation.

    I believe that much of the problem lies in the rather clumsy way that the lofts must be defined. The loft tool would benefit from a rewrite that assists the user in creating a correct loft of the form wanted.
     
  8. Ralf

    Ralf Alibre Super User

    As always, general statements and/or vague statements do not lead to a result but poison the atmosphere.
    Sending facts and sample files here or better to technical support is the way to go.
    If Gmail and/or lost passwords for dropbox are a problem, support sends a link to upload, no problem.
    This is the way to go, for all who are interested in solutions.

    By the way, a question with facts was asked by GIOV above, he provided a file and Stefan showed the solution.
    This is also a way to go, for all who are interested in solutions.
     
    Last edited: Jan 1, 2020
  9. oldfox

    oldfox Alibre Super User

    If I understand your explanation of the failure correctly, then this cube filleted just fine using no decimals at all. Just integers. You could try it out on one of your own parts with your dimensions. Personally, if I needed a sphere, then I would just rotate a 2D hemisphere.

    Or a half-filleted cube. Which naturally turns the cubic into a circular. Grandma can't make
    an apple pie using apple slices if she tries it by slicing up an orange.
     

    Attached Files:

    Last edited: Jan 1, 2020
  10. JST

    JST Alibre Super User

    I do not understand either your or Ralf's statements.

    The assumption seems to be either that I am stupid or obstinate. If that is your opinion, then please just stay out of the discussion.

    Ralf:

    If I knew what the problem is, than I would tell support what it is. I do not know what the issue is, and neither do you.

    If I simply stay silent because I cannot supply the solution, that does nothing helpful. So if I say that I find the fillet tool fails to close fillets around an object in many cases, and mention the error, then at least that is known.

    I have supplied examples of the problem in the past. Started so long ago that 3DSystems owned the program. I have no idea what was done, I was told that the new ACIS would fix everything.

    It did not "fix everything".

    If we have come down to the place where it is "poisoning the atmosphere" to say "there is a problem which I cannot explain the cause of", then it is clear that there is no intention of fixing anything, and we are all supposed to shut up unless we say we think the program is perfect.

    Somehow I do not think that is what the development team wants.

    Old fox:

    No, you are off the mark, unfortunately.

    I have mentioned it before, but have nothing reasonable that can be given to support. Tried that, got back nothing useful, just a "cannot duplicate".

    The failure to fillet is random. There seems to be no reason for it. The failure to create a good loft is also random. A fillet that fails, may work fine the next attempt, even if it is done the same way, but something else must usually be done in between, even if it does not affect the dimensions at all.

    In other cases, changing a dimension by the smallest amount possible may make a fillet work perfectly every time.

    A Loft that works may fail the next time the model is opened. Or a loft that fails may work another time.

    The only explanation I can come up with is that differences in numbers, which are so far down "in the noise" that they are almost random, have an effect on the result. One time, they may differ by some amount that causes a "fail", but making any change at all, will affect these tiny numbers enough to make it work.. I have suspected this in other cases, where parts that actually measure the same exact dimension, cannot be aligned, for no good reason.

    This is probably not the true explanation, but it at least illustrates the type of problem. Naturally, exporting a package and sending to support probably changes things enough to erase the problem at least temporarily.
     
    Last edited: Jan 1, 2020
  11. idslk

    idslk Alibre Super User

    Hello JST,

    the fillet function is not the best and could be more stable, but it is not to worse for me...
    upload_2020-1-2_9-59-4.png

    Regards
    Stefan
     
  12. DavidJ

    DavidJ Alibre Super User Staff Member

    If the response from support was 'cannot duplicate' - that suggests that either the problem is specific to your installation or that you haven't described the precise workflow in enough detail to allow it to be replicated.

    Until the problem can be replicated, it is unlikely to be analysed (and then hopefully addressed).

    Please send files/examples/log files anyway - even if you don't understand the problem/cause. Even if problems can't be replicated, being made aware of them does help to build a picture - if only one person reports it, that tends to suggest a problematic install or hardware. If several people report the same/similar issue, it might suggest possible areas to explore.

    If you just moan and don't provide any examples, then any fix will only come about by dumb luck.

    Be aware that written posts don't always get interpreted by the reader as the originator intended - written communication is very different from face to face interaction. JST - a lot of your posts do sound aggressive or confrontational as I read them, that might explain some of the responses. Where I don't understand a response (happens a few times each week), I either ask for clarification, or more often just ignore it.

    Do remember that English is not the first language for everyone here, and that 'American' is not the same as UK English - so it isn't surprising that we don't always clearly understand each others postings.

    Can we all bear that in mind, and try to make allowances for each other?
     
    markporter and simonb65 like this.
  13. JST

    JST Alibre Super User

    I disagree about "cannot duplicate".

    First, the workflow may not be KNOWN. A random problem also may occur or not occur with the same basic workflow, due to internal program issues.

    Without a keylogger, I do not remember what went on a bit before a problem, I am working along thinking about the design. All I know is that suddenly I have a problem, which I can describe. If I had known I was going to have a problem, I would have made notes, but unfortunately that degree of prescience is not a skill or capability of mine, even though it could be very useful (or annoying).

    The problem may persist, but then may go away if some other unrelated operation is performed, and then the (presumed) previous operation is done again. Or perhaps the operation is done in a different order, and then the problem is gone.

    I ask you, how can this be documented sufficiently? The problem has disappeared, but it really WAS there. And not due to operator error invoking the operation, because it persisted.

    The problem may easily be (and I think it IS) due to the wrong combination of numbers somewhere down in the decimal places that leads to a failure. That is clearly the issue when two items, that measure identical dimensions with the measure tool, cannot be aligned. Obviously there really IS a difference at some invisible decimal place (binary place!). Seemingly unrelated work can actually affect the result of a different operation if some location for temporary results is not initialized properly, for instance, and still contains a previous result. That location may not be used often, perhaps only when more than a certain number of calculations is needed, or perhaps that location is usually initialized, but for certain sequences is not, in which case the problem will appear "random".

    In such cases, where the problem occurs, but then fails to occur if some unrelated operation is performed in between, the problem is not even replicable ON THE SAME SYSTEM, yet it may occur again "randomly".

    That sort of error is very hard to find, but causes problems, likely more problems with operations which require the largest number of calculations. I suspect that lofts and fillets are operations that involve many calculations, and so are going to be more susceptible to such problems.

    I have now used the program extensively on three different computers, with three different OS versions, and very different memory, processors, etc. Since the problems continue to occur, I am forced to discount the idea that they are system specific, and I suggest that the support folks may want to avoid using that "excuse", no matter that it is easy, and that it may even be true sometimes. Whenever I have discounted problems in that way during product development, it has usually become clear that I was wrong, and have lost valuable time that might have been used to solve the issue.

    I would say "I" was the problem, since "I" am obviously a "common factor", but all I can do is ask for a fillet and do what the program demands to get it. Not much to do wrong, yet two ends of a 1/4" fillet may fail to "connect" at a corner unless I select the sections in the reverse order..... Or maybe only if I delete the operation, and re-do it. Go figure.

    I have tended to use the program very extensively, and ask it to do things that others may not. That may be why I see things that others do not.

    If the problems do not recur easily, then perhaps they can be borne in mind and addressed later. There may come a moment when you say "Ah, THAT is what that blasted idiot (meaning me) was talking about".

    Yes, "intermittent" problems are hard to find. Sometimes very hard to find. Welcome to the world of engineering and debugging......

    Confrontational? No.

    I simply put the issue out there, as it is or appears, and assume we are all secure adults here. That may appear blunt. I note that there are some other regular commenters, both users and company folks, who share that characteristic. It comes with the territory among technical folks who must deal with facts as they are.
     
    Last edited: Jan 2, 2020
  14. DavidJ

    DavidJ Alibre Super User Staff Member

    OK - I did say 'suggests' , there may also be other possible causes. The software, combined with any PC, is a 'complex system' and may exhibit unexpected behaviours.

    For your 'won't align, even though dimensions identical' cases - surely you can send those to support ? That does sound like something fairly concrete that can either be diagnosed and explained, or elevated to Development for deeper study.
     
  15. JST

    JST Alibre Super User

    I have, but no idea what the resolution is. Been a while.

    Unfortunately, that is also not consistent. Happens once in a while.

    But it IS persistent once it occurs in a model, so yes, should be identifiable.
     

Share This Page