This is really bad - ad21.0.9/20 fails to detect duplicate designators

[ajawamnet]

[ Specified attachment is not available ]A person on the official forum just posted a project that exhibits this problem - see attached vid. They have lost all control of the code base it seems. In the almost 30 years and 3,000+ designs I've done using Protel/Altium, this is the worst I've ever seen as to bugs.  I recall the Chinese coders had some issues and got them fairly well sorted out at AD16/17, but these Ukrainian guys have really hosed the code base. They seem to have no clue as to what the software did/is supposed to do, and they have no handle on dependencies when adding features, and so trash critical operations.

I think I know what happened - see this vid:

s

You shouldn't need a facecrack account to watch it...

Vid that compares ad17 to ad21(ad20 is just as flaky) - http://www.ajawamnet.com/nodupe.mp4

[ANTALIFE]

Is this them opening a AD17 project in AD21, or the other way around?
Also did they show what the project error rules are?

I guess one thing I have learned with Altium is to always stay one revision behind if you want stability/reliability

[ajawamnet]

This appears to be that was initially done in AD20. When I open the PCB in AD17 I get a warning HTML about later version and prop delays.  Using the tool that John Williams wrote it states AD20 - see attached screenshot. It's possible that it was from an earlier version tho.   But there's have been recent other complaints about it failing to catch duplicates. 

And it seems that regardless how it was generated, the ERC should flag it. I did some experimentation, and if I select one of the parts with the dupe designator, and Copy/Paste (with the Reset Designators on Paste set to OFF) it gladly pastes a THIRD copy so now there's three parts with hte same designator. A verify still does NOT catch this error

  This failure to detect dupe designators occurs in both AD20 and AD21.

The only way to catch them is to run Annotation > Reset Duplicate... . Weird thing is that you get and incessant PCB update of it trying to add the original designator parts to the sheet class after correcting the duplicates.

As to staying a major version behind, until AD20 I was at AD17. It just works. And before I jumped into AD15, I stayed at AD6.9; probably did a thousand or so PCB designs with it.

For me, AD21 fixes a bunch of stuff including an issue with the Layer Stack Manager as well as the ability to separate arc's in the snaps (as was in AD17's Snap dialog). We lost that in AD20 (maybe before that.)

BTW - John Williams is the guy that wrote QualeCAD that was a plugin for Altium back in the day. I recall asking him in the mid 2000's to make it so that it'd generate PDF build sheets.  Just a badass tool for doing prototype builds.

[Elasia]

I've never trusted altium or any other for duplicates..  full bom review cross checking number of parts and full designator listing.. there better be one to one or if a designator number somehow has multiple entries.. well there you go..  I usually do this in excel, active bom is handy but not an excel replacement. I dont silk screen designators usually less its hand parts so it better damn well be right

[VK3DRB]

I am using 21.1.0 (the latest). It changed R14 to R13 so there are two R13's. I ran the validation check and it detected the duplicate designators, no problem.

The bigger issue I have, is Altium hanging, showing exception errors or immediately shutting down for no reason. No other package does this, including Autocad which I also use.  Very bloody frustrating, especially when you are under the hammer. They need to focus on software quality assurance and remove focus from features few people use. For the money, they can do better.

[VK3DRB]

...I dont silk screen designators usually less its hand parts so it better damn well be right

In contrast, I silkscreen every designator. This helps save time for staff in production debug, R & D, the field, and helps manufacturing documentation (Altium Draftsman can show the DNF's clearly and provide a designator search facility). In the end, a good design is about it working robustly, meeting regulatory requirements, and communicating effectively to the next poor bunny that has to use my work. My designators are generally oriented so the bottom left of the first character is closest to the component to greatly reduce the chance of ambiguity. So the designator can be upside down even. It works a treat. Furthermore, my pins 1 markers are never those crappy circles that look like fly vomit. I don't know why people still use these, and ever so inconsistently. That are substandard and look aweful. I use small triangles equilateral traingles (filled region which works a treat in removing pin 1 ambiguity and greatly improving readibility.

I generally do not use "TP" for test point designators either. There is absolutely no reason to use "TP" on high density boards, other than to just make life difficult. Just use a number near the testpoint. For example, instead for TP23, just use 23. Altium handles it perfectly and there is no ambiguity with any other designator preamble. After doing many high density boards for industry, no-one else has an issue with this, not even an extreme QA manager I once had to work with.

[PlainName]

Why is a triangle better than a circle? The only thing I can quickly think of off-hand is that a corner can point to the pin, but is there something else?

[Ghydda]

Hi, quick question:

Is anyone else having trouble with Altium 21.1.1 (or any other version) not updating the designators between schematic and layout?

It goes like this;

• first import into layout loads all components, using whatever designators contained in the schematic
• any subsequent changes to designators or adding new parts in schematic simply goes into the bin when importing changes again
• inspecting each component reveals that the schematic-given RefDes indeed is linked to every part in the layout, but Altium gives me the middle finger anyway and complains endlesly of duplicate components and not being able to make a rats nest until I fix it

.

Attached screenshot shows a row of recently added and imported components. Everyone of them seems undesignated.
However, inspecting any one of them - please see attached screenshot: a resistor - reveals it to be known as R307 in schematic-land, but = R? in layout-land Bravo. Well done. Good job. 
No amount of tickling or thrashing seems to make any difference - once I got my first import, then forget it.
Reset of all designators, re-annotating, re-import, even importing into a new clean layout file: none of it makes the sligtets difference. I'm stuck with a bunch outdated RefDes values dating back to the initial import, and equally many un-annotated.

And yes, the 21.x.x releases does seems to be the worst/flakiest/wonkiest releases in years.

But come on, this is core functionality - not some nitty-gritty edgecase Tomfoolery nobody uses - how can this break and slip through QA?