Complex Profile Beams "Flipping" after Back-Saving

[Jeff Griffin]

I've noticed that certain complex profile beams end up in the wrong orientation (mirrored 180 degrees about the reference line) after back-saving to previous version. Specifically, this issue seems to show up at the point I save from AC22-to-AC21. (AC24-to-AC23 and AC23-to-AC22 seem to work fine.)

These particular complex profile beams are ones that I've made to show fascias. To work as efficiently as possible, as I place these complex profile beams, I mirror them about the ridges, since most of the time the roof slopes are symmetrical on both sides a ridge. It's been several months since these complex profiles have been originally placed so there's no way to know for sure, but I suspect that the ones that are now in the wrong orientation (after back-saving) were ones that were mirrored across ridges when I had initially placed them. Approximately half of the rest of these beams are in the correct orientation.

Interestingly, when I open the AC21 version using AC22, the issue can be seen. When I open the AC22 version in AC22, everything looks fine.

AC22 version of PLN, opened in AC22: red line added to show approx. location of reference line
https://www.dropbox.com/home/Archicad%20issues/AC22%20to%20AC21?preview=22.jpg

AC21 version of PLN, opened in AC22: red line added to show approx. location of reference line
https://www.dropbox.com/home/Archicad%20issues/AC22%20to%20AC21?preview=21.jpg

AC21 version of PLN, opened in AC22: beam selected to show inversion.
https://www.dropbox.com/home/Archicad%20issues/AC22%20to%20AC21?preview=21+(CP+Selected).jpg

Here's the whole progression, from AC24-to-AC21 if it helps:
https://www.dropbox.com/sh/zy85y0pbu8hzkl9/AAABFMMMuwJeuChfxYP4Xg8ua?dl=0

I've already tried opening & repairing the PLNs along the way to see if that would help, but it didn't make a difference.

Any ideas what may be causing this defect during the AC22-to-AC21 back-saving step, or how I can resolve it?

Thank you,
Jeff

[Barry Kelly]

I could understand the problem occurring when you save from 23 to 22, as 23 was when the new segmented beams were introduced.
But I have no idea why it would happen from 22 to 21.
I have done this a few times myself and not noticed a problem, but generally I would not have used complex profile beams.

You say only with 'certain' complex profiles.
So you have some complex profiles that are not affected?
If that is the case it will be a matter of determining what is different about those profiles.
Origin location, stretch modifiers, etc.


Barry.

[Jeff Griffin]

Hi Barry,

Thanks for your reply. What do you use to model fascias if not complex profile beams? I'm not aware of any other tool that can be used for modeling fascias, since I need something that can stretched, sloped, rotated & easily updated after placed. If you use another tool for fascias & frieze boards, I'd love to know what that is.

I suspect that the ones that are now in the wrong orientation (after back-saving) were ones that were mirrored across ridges when I had initially placed them. Approximately half of the rest of these beams are in the correct orientation (guessing the ones that weren't mirrored when placing them). I now notice that some of the complex profile beams that I use for frieze boards are also inverted. Approximately half of these were also likely mirrored while I was placing them in order to work efficiently.

We're only taking about two different complex profiles here (one is the fascia and the other is the frieze board and both are relatively simple) and I don't think that there is anything wrong with either of them. The origin is set correctly and I'm not using any modifiers on these. Half of them are inverted after saving from AC22 to AC21 (guessing that these were the ones that were mirrored while I was placing them initially). All are fine in AC22, AC23 & AC24. It's the AC22-to-AC21 step that is introducing this defect.

[Jeff Griffin]

I should probably mention that this isn't the first time I've experienced this particular issue. I had something similar happen in 2018 when saving from AC22 to AC21 and uploaded my PLA to GSNA for eval on 11/30/2018, but didn't hear anything back. I should probably have been more persistent regarding finding a resolution to this issue then. On that project (in addition to half of my fascias flipping), custom timber brackets made from complex profile beams (which one had been mirrored after placing the first) flipped orientation upon back-saving from AC22 to AC21. Here's a link to a screenshot with all of the inverted beams highlighted.

https://www.dropbox.com/s/bkpls8qkwx63m32/2018%20Issue.jpg?dl=0

Again, this defect was introduced when saving a AC22 file to AC21. The AC22 file is perfectly fine.

[Barry Kelly]

jffgrffn wrote:
Thanks for your reply. What do you use to model fascias if not complex profile beams? I'm not aware of any other tool that can be used for modeling fascias, since I need something that can stretched, sloped, rotated & easily updated after placed. If you use another tool for fascias & frieze boards, I'd love to know what that is.

For the horizontal I have a custom scripted gutter/fascia object that has different profiles for the gutter and fascia.
For the raking fascia I again have a custom scripted object that has preset profiles and a few different end options.
Construction methods for the type of work I do is probably different to what you do as well.

Barry.

[Jeff Griffin]

Thanks for your reply Barry. Interesting approach. If I had to script all of the trim profiles I find myself using, I would probably chose another occupation! So, you don't use complex profile beams for anything? If so, why not? Is it due to some limitation or some other shortcoming with CP beams that you are aware of? A CP beam should be bread and butter to Archicad, a basic building block.

Has anyone else noticed this defect (complex profile beams flipping orientation after saving from AC22-to-AC21) and if so, were you able to overcome it? To recreate this issue, all I think one would need to do is open AC22, place a (non-symmetrical) complex profile beam in plan, mirror a copy of it (end for end) and then save the PLN as AC21. Next, open the AC21 PLN in AC22 (or AC21 for that matter) and see if the defect (the mirrored one flipping orientation 180 degrees) happens for you. I'm thinking it probably will. I can't be the only one. I was able to recreate the exact issue using the OOTB USA AC22 template with a standard OOTB CP beam called "Trim-Wood-Base". See links below.

New PLN created from AC22 USA OOTB Template:
https://www.dropbox.com/s/r5sz9232n89qoog/Test_22_pln.jpg?dl=0

Saved AC22-to-AC21 (Defect present when reopened in AC22 or AC21):
https://www.dropbox.com/s/t6hnlb43qld5fi0/Test_21_pln.jpg?dl=0

Any ideas on if and/or how this can be resolved?

[Barry Kelly]

jffgrffn wrote:
So, you don't use complex profile beams for anything? If so, why not? Is it due to some limitation or some other shortcoming with CP beams that you are aware of? A CP beam should be bread and butter to Archicad, a basic building block.

We do use complex profile beams, for things like 'I' beams (universal beams) and channels and concrete strip footings occasionally.
But rarely would we use them for profiled fascias because we are happy with the objects we have.

I can confirm you problem though.
Mirror a copy of a complex profile beam end to end, and then save as a 21 file.
Open in 21 and the mirrored copy is in the correct location but the beam is now flipped along the axis.
Unfortunately because it is in version 21, I doubt there will ever be a fix for this as it is no longer updated any more.

In fact even if you mirror it along the axis, it will flip when saved to 21.
It must be in the conversion because if you open the 21 file back in 22 the beams are still flipped.


Barry.

[Jeff Griffin]

Thanks for confirming this defect Barry. I guess this is something to stay aware of, moving forward one should be wary of back-saving beyond AC22. Or, spend more time during the placement of CP beams (no mirroring at all during placement), or convert all CP beams to morphs prior to save as 21 (ugh), just in case there's ever the possibility of needing to work with anyone using AC21 (or earlier). Unfortunate, since back-saving has been pretty seamless up until this.

[Minh Nguyen]

Dear All,

Thank you very much for your reports and testings for this issue!

Thanks to Barry's comment about the mirroring operation, I was able to replicate this issue quite easily. Unfortunately, since this issue hasn't been raised by the market when AC21 and AC22 were still supported, we were not aware of the issue and issue a fix in time. As these versions won't receive any further updates, I am very sorry to say that there won't be any fix in the future for this issue.

Luckily, in AC23 we found this issue in-house, thus backward migration from AC24 to AC23 was correct.

I am very sorry for the trouble that you had with this one! Please let me know if you have any further questions!

Best regards,
MInh

[Jeff Griffin]

mnguyen wrote:
...
Unfortunately, since this issue hasn't been raised by the market when AC21 and AC22 were still supported, we were not aware of the issue and issue a fix in time.
...

Thanks Minh. Well, as I had mentioned, the issue was reported to GSNA on 11/30/2018 and my PLA was uploaded as directed on that same date, which was well within the AC22 support dates. I just wasn't an ArchiPLUS customer at the time, so perhaps that's why it didn't go anywhere. I do understand that the ship has sailed. You can be certain that I'll be more persistent with follow ups next time around with issues like this.