DipTrace Version 2.3 is out today!

[reagle]

Nice mix of updates:

Online DRC - dynamically checks design rules while you add objects, route traces or edit anything in layout. Static DRC check was improved and optimized.
3D VRML 2.0 Export of PC board. Allows you to export solid body of your board with models to mechanical CAD.
Custom Non-Signal layers.
Arranging PCB components by hierarchical blocks. Copying routing/placement between similar blocks.
Automatic showing/hiding ratlines for copper pours depending on pouring areas and traces ("Hide Net Ratlines" option has been removed).
Pour priority for copper pour.
Separate "Place Ratline" mode, no more ratlines in "default" and "edit traces" modes.
"Manual Route" mode hotkey (~).
Configuring and saving filenames for Gerber export by layers (name and extension for each layer).
Exporting dimensions into Gerber and DXF.
Checking pin/wire superimposing without connection in Schematic ERC.
Saving connections without wires while copying/pasting parts in Schematic.
Moving objects between pages in Schematic.
Rearranging sheets in Schematic.
Saving "open/save dialog" initial folders by file type (globally for all programs).
Blocking file by the program while editing it. Overwriting by second/network appearance is prohibited.
Backward compatibility of binary file formats starting from this version (it can open files of further versions).
Windows 8 compatibility.
Ordering engine has been added (see File/Order PCB).
Online check for new versions and DipTrace news.
Library update: 8000+ new components, euro symbols, new patterns.

[reagle]

Hmm, and now I can complain

1. Even though I am on their newsletter list, no notification came in

2. It would be nice to see a list of details to go with the release notes. Almost like (gasp) Altium does on their live blog when a new release is posted.
Here is where the new feature is, here is how it works, etc

3. Can't' seem to make online DRC to do anything, or maybe it needs enabling. It will let me short traces with no complaints for example or put parts on top of each other

[novarm44]

1. Already sent. Seems you noticed it earlier, than we updated all installers and web-site before sending emails.

2. We will think on making video blog about new features or something like that in the future.

3. You should enable it in Verification / Design Rules.

[reagle]

Hey, thanks for answering!

I did receive an email later, but it was " We have published new build of DipTrace (2.2.0.7) today"

"3. You should enable it in Verification / Design Rules."
Duh, didn't think to look there! Works like a charm now.

Igor

[novarm44]

I did receive an email later, but it was " We have published new build of DipTrace (2.2.0.7) today"

That was a bug of phplist system - I noticed that when sent it already (several hundreds of subscribers including me received that email for some reason). You should also receive second correct email (I queued the new email only, and seems old email was queued before and not removed - didn't notice that).

[shebu18]

In PCB layout you still can not hide the cooper pour ratlines, or am I wrong?

[novarm44]

In PCB layout you still can not hide the cooper pour ratlines, or am I wrong?

It should hide them automatically when you pour copper pour if both ends of the ratline are electrically connected by copper pours and nets. In previous version it hides ratline if "Hide Net Ratlines" box is checked and both ends are inside copper pour border polygon.

[Mechatrommer]

Arranging PCB components by hierarchical blocks. Copying routing/placement between similar blocks.

my voice was heard! now i hope the Novarm team can consider this...
1) hiding the "silkscreen" when in 3d preview (currently only components and open vias option available). sometime i want to see electrical connection only with no fancy texting and bounding boxes silk
2) bug: wrong silkscreen placement when converting to gerbers. not sure if its fixed in this version (or full paid version)
3) bug: 3d components sometime only showing ghost shadow (ie grayscaled pcb behind the components), but reloading will render correctly. not sure if its fixed in this version (or full paid version)
4) when will be the "built in" more robust spice simulation?
thats so far i can spot. i hope its not too much, except #4, i know that can be a hard decision

[novarm44]

2) bug: wrong silkscreen placement when converting to gerbers. not sure if its fixed in this version (or full paid version)
3) bug: 3d components sometime only showing ghost shadow (ie grayscaled pcb behind the components), but reloading will render correctly. not sure if its fixed in this version (or full paid version)

Please send examples to us or point me to mantis bug numbers if you reported them.
Especially the first bug - gerber export should be ALWAYS correct. We do not experience any problems with this feature ourselves

[Mechatrommer]

2) bug: wrong silkscreen placement when converting to gerbers. not sure if its fixed in this version (or full paid version)
3) bug: 3d components sometime only showing ghost shadow (ie grayscaled pcb behind the components), but reloading will render correctly. not sure if its fixed in this version (or full paid version)

Please send examples to us or point me to mantis bug numbers if you reported them.
Especially the first bug - gerber export should be ALWAYS correct. We do not experience any problems with this feature ourselves

i think i'll just report here in case someone else could reproduce it. if you need my original file, please ask i'll PM (do you need my whole library as well? ). i'm running in ver2.3 now. i cannot reproduce bug #3, i guess its corrected. for bug #2, its still there in ver2.3. please refer to picture below.

1e: my bottom pcb. red arrow indicating the correct silk, please note letter A, I, and O.
2e: when i view in 3D, the letters start to dislocated (i believe i've mistaken it with gerber export, i think its your application translating the silk, not gerber export module.
3e: preview in "export gerbers" showing bottom silk. dislocated A, I, O (if analyzed closely, that silk will sit on top of a solder mask which can cause problem when the order arrived from fab house)
4e: after the process, whether i export or cancel the preview, the pcb layout program now showing dislocated A, I, O. to correct this problem i have to "update component" from library. i hope thats clear and may help you in the next bug fix.

sorry i'm not sure i understand "mantis bug numbers", i'm a newbie, sorry.

[novarm44]

Can you send this board to support [at] diptrace [ dot ] com for verification.
We do not need any libraries.

Or you can just send wrong component (as library or PCB file).

[poorchava]

There's some problem (at least was in the beta) with dxf export of copper. I wanted to mill a simple board on my cnc router, and it turns out that copper pours seem to be omitted when exporting copper outline.

http://www.diptrace.com/forum/viewtopic.php?f=9&t=6591

Other than that still no way to change footprints in multiple components (eg. i have different part in the library for each resistor value,and i have to update them all one by one if i change some minor detail in the 0805 footprint). Also the footprint has to be re-selected from the library. I would kil for a possibility to just select some parts and click 'update footprints'.

There's still the annoying feature that when you select multiple parts and rotate them without grouping, they rotate each around it's own center rather than rotating whole group around its center. I know that for example CadStar works like that too, but that seems really counter intuitive. Maybe a button to choose the rotation mode would be a good idea?

[Mechatrommer]

1) occasionally my objects/parts is off-grided. when i remove them they keep snapping into the grid, and then later at another time when i re-select the same part its off-grided again maybe because i renewed pcb from updated schematic? not sure. sorry too busy to go into diptrace forum hope novarm guy can re-transmit the message to Mr Michael Jackson? or something in support team.

2) adding vias to a net was a charm, now i have to manually click the edit net (place ratline) toolbar up above.

[shebu18]

When will diptrace be native on linux?

Sent from my GT-I5800 using Tapatalk

[novarm44]

For copper pours it doesn't make outlines while exporting outline for G-code - please use polygons in signal layers instead.

Other than that still no way to change footprints in multiple components (eg. i have different part in the library for each resistor value,and i have to update them all one by one if i change some minor detail in the 0805 footprint). Also the footprint has to be re-selected from the library. I would kil for a possibility to just select some parts and click 'update footprints'.

We have this in to-do, but it is not implemented yet.

There's still the annoying feature that when you select multiple parts and rotate them without grouping, they rotate each around it's own center rather than rotating whole group around its center. I know that for example CadStar works like that too, but that seems really counter intuitive. Maybe a button to choose the rotation mode would be a good idea?

Ctrl+G to group, then rotate and Ctrl+Alt+G to ungroup. In PCB there is also feature to "rotate selected" (see in Edit submenu).

1) occasionally my objects/parts is off-grided. when i remove them they keep snapping into the grid, and then later at another time when i re-select the same part its off-grided again  maybe because i renewed pcb from updated schematic? not sure. sorry too busy to go into diptrace forum hope novarm guy can re-transmit the message to Mr Michael Jackson? or something in support team.

Maybe the real grid is different from visible - you should zoom in to check that. If you will have time - better provide us with the video of the bug.

2) adding vias to a net was a charm, now i have to manually click the edit net (place ratline) toolbar up above.

Actually this is improvement: "separate mode to place ratline" requested by many of customers. In previous versions ratlines were placed in Default and Edit Traces modes, that was a problem while working with the board (you was able to make ratline occasionally).

Can anyone point out how the new 'moving objects between pages in schematic' feature is implemented? I am perhaps looking in the wrong place but I can't see to find anything different in this respect from version 2.2, and the help file doesn't seem to clarify it either. I am a new user of the software so perhaps am looking in the wrong place.

I did try to ask this on the official diptrace forum but my post was auto-deleted for some reason.

Right click on the object and see "Change Sheet" item in sub-menu. I will also review help-files as some new features are not there yet.
We moderate first message from new user (remove 10-20 spammers daily), probably your message is on-hold yet.

When will diptrace be native on linux?

I don't know. We plan to move to Delphi XE3 with VCL (currently DipTrace is developed with Delphi 5 - made in 1999 ), then switch to FireMonkey and make native Mac version. Maybe they will also support Linux anytime soon.

[Mechatrommer]

Maybe the real grid is different from visible - you should zoom in to check that. If you will have time - better provide us with the video of the bug.

i dont have time for video sorry, but its getting in my nerve when i had to realign a route keepout 3 times! maybe it was due to i keep changing grid between 0.32 and 0.16 but i'm pretty sure they are all aligned to 0.32, i know which one that aligned to 0.16. i also frequently change origin that might be the cause as well? not sure. i'll live by it until someone else make another report sigh. it was not there in ver 2.2

[Mechatrommer]

1) copy and paste route keepout will not keeping copper pour out unless we open property box, make sure its route keepout (which it already is) and click ok.

[Mechatrommer]

here specially made for the thread... out of alignment bug when we change origin. please note the dotted grid and see the "out of alignment" in origin 2 (below), i'm newbie in eda but i think this should not be happening. changing origin will result in multiple coordinate system parts alignment. this is very bad i want to punch or bite something biological after keeping aligning my board outline for so many times

i dont mind if the bug is brought from previous version but i recall i never had this problem previously. upgrade something please upgrade to better or at least the same, but certainly not worst like this, this is diptrace version 1.7 darwinian movement!

[poorchava]

For copper pours it doesn't make outlines while exporting outline for G-code - please use polygons in signal layers instead.

But the polygon will not avoid other signals or go around keepouts, am I right? So I would have to draw outline of every polygon completly manually, whereas using copper pours this can be automated to some extent. I mean with the design where i stumbled upon this, it would be doable - it's not that complicated. But on some more complicated smps design drawing all polygons completly by hand would elongate the process alot.

[novarm44]

here specially made for the thread... out of alignment bug when we change origin. please note the dotted grid and see the "out of alignment" in origin 2 (below), i'm newbie in eda but i think this should not be happening. changing origin will result in multiple coordinate system parts alignment. this is very bad i want to punch or bite something biological after keeping aligning my board outline for so many times

Any chance to see coordinates of mounting hole center and how it is aligned to the grid on the first image?
The grid is always aligned to origin, so if you have, for example, .5 grid and 1.2 between board outline and hole center, when you move origin from board outline to hole center - the grid will be shifted to 0.2 as distance is not multiple to the grid step.

But the polygon will not avoid other signals or go around keepouts, am I right? So I would have to draw outline of every polygon completly manually, whereas using copper pours this can be automated to some extent. I mean with the design where i stumbled upon this, it would be doable - it's not that complicated. But on some more complicated smps design drawing all polygons completly by hand would elongate the process alot.

For milling method copper pour is actually external copper area of the board, which is separated from pads/traces while milling. If you want to get solid area - just use shape. How to mill copper pour, isolated from pads/traces with specified gap, as both copper pour and pads/traces are parts of layout? If we have copper pour clearance bigger, than mill diameter - it's possible, but I don't see any sense in it as you already have external copper area.

[Mechatrommer]

Any chance to see coordinates of mounting hole center and how it is aligned to the grid on the first image?

and this become weirder and weirder (1st image). the coordinate is not multiple of 0.16mm! i never touch those coordinates manually. i only rely on the diptrace grid system. editing coordinate manually is a good feature but lets dont use it as grid reference or origin? if ever we want to limit or snap our design to 0.16 (or 0.32 or else for this matter) diptrace should have it's own grid reference, not relying on what user input xy manually, it will screw the rest of component placement, but as i said, i never touch those manual editing XY coordinate, so its weird. FWIW, the mounting holes and board outline was copy pasted from another file, previous file created in diptrace 2.2XX.

ok to make thing simpler, i started a whole brand new pcb editing from zero, first thing i did? mounting hole placement, using 0.32 mm grid, here's the coordinate result (2nd pic), can you see it? WTF? i think i get it, "pentium floating point operation rounding off" error? except its worse. it should be 0.32 not 0.318! if i'm the programmer, i treat this as serious bug that need immideate fix! duh i should be a beta tester and get 95% discount on unlimited version sorry i dont mean to put diptrace down by putting the report here infact i very like it its an evolution! its just i cant be in many forums at one time and i have many other things to do. diptrace representative presence here is actually a blessing. thanks to diptrace team, full thumbs up to you!

[Mechatrommer]

i tried to reproduce it in diptrace 2.2X i also notice this 0.318 error (or is it intentional? i dont know) but i cant trace it down to my pattern library its now "format unsupported", its not a good practice to keep changing file format frequently it should be the last programmer's effort. so i'm doomed to 2.3 now sigh. btw i edited and attached my library to this particular pattern and please ask your programmer why i cant align this through hole pad to 0.16mm grid? or why it became nonaligned in the first place? maybe i manually edited the pad spacing? deleted the other pad etc i dont remember what happended. if so as i mentioned, diptrace should stick to its own grid reference to avoid this mess. this may cause the mentioned bug i'm not sure this the least i can do, hope it helps.

[Kremmen]

I got intrigued about this and tried to reproduce the error. But i cannot. I have the latest DipTrace and i cannot see any of the errors you have mentioned. All libraries i have created using 2.2 work flawlessly in 2.3, and i also cannot cause the misalignment problem, no matter what i do. I'm not saying that you are seeing things, but could it be something in your particular setup or the way you use Diptrace?
Just in case the rest of us have missed something, could you outline step by step what you do to cause the misalignment. I can then repeat the steps and test on my machine whether the same happens here and report back.

[Mechatrommer]

the problem was i didnt know what happened and was not aware sorry i thought diptrace programmer would know better. btw, now a bit organized... i managed to produce it from scratch.

1) make a pattern library (22.lib) with 2mm spacing and pad size (i think this is where the problem begins)
2) make another pattern 11.lib just for testing scratch from new
3) load both library into diptrace pcb and place both parts on the pcb area (0.16mm grid)
4) make rectangle board outline and try changing origin between part 11's and 22's pads, there will be 5 pads, try all you'll see the grid shifting.

just for example i attached the sample files, try aligning the board outline straight in 22.dip by moving top left corner alone if you can. if you get the same result as i am you need to move all corner to realign them straight rectangle, and then changing the origin again to another pad until the grid shifts, you'll have another headache realigning board outline.

please tell me if you get what i get. or its me thats mistaken i should have never made the 2mm spacing and 2mm sized pad?

[Kremmen]

OK, i followed your steps. I didn't actually create the pads since the included file was clear enough. I could see how the pads were done in the lib files.
I also believe i know what your problem is basically; different nonaligning grids. I am all the time pissed off with the imperial grid used in most of Diptrace lib components. Doesn't match the metric grid i use exclusively and as a consequence the pins in schematics seldom align with my metric components. But no big deal as the symbols are all wrong anyway, i use IEC symbols and those only, so have had to redraw everything anyway to a metric grid.
But back to your issue. The four pad footprint was created using a 2x2 mm grid. The pad centers are aligned at grid points and so far so good. But now, when you insert these footprints on a grid that is not evenly divisible with the creation grid ( 2mm vs 0.16mm), some pads will necessarily fall on points not on the grid. In itself that is not a problem; you will encounter all kinds of footprints, both metric and imperial dimensioned that don't fall on any grid you care to use.
Now, wherever your coordinate origin is, if you define the board outline using the grid, the outline will naturally be aligned on the points of the current grid. However, if you next switch the grid origin to a point not currently coinciding with the grid (such as between pads created on the 2 mm grid, while using a 0.16 mm grid) none of your new grid points will coincide with the previous points either. Features like the board outline are not affected *provided you do not touch them*. But move a corner and it will naturally snap to the current grid while all the other vertices stay on the points of the previous grid. Now you lost straight angles and will have to move all the other corners as well.

But let me suggest that this is not the proper workflow how you should define the board outline (or many other features as well). Sometimes you have a size constraint, sometimes you can freely use whatever real estate you need, and sometimes the board is a curious shape mandated by some quirk of the enclosure where it has to fit including various holes and cutouts in specific locations. The workflow should reflect this to make life easiest. There are many approaches you can take but such mechanical features are nearly always best created by using specific coordinates instead of relying on a grid. Can be done of course if the board is a simple rectangle and in such a case i do use the grid.
Easy solution: Just lock the board outline when you are happy, then it will not move whatever you do with the grids.

This is just me but the grids i work with in both capture and layout are 0.05 - 0.1 - 0.2 - 0.5 - 1 - 2 - 5 - 10 mm where all bigger grids are evenly divisible into the smaller grids. Everything must coincide with one of those grids - only individual footprints are created with exact dimensions.

P.S. Why 0.16 / 0.32 grid? Where does that come from?

P.P.S. OK for some specific things i have also the 1.27, 2.54 and 5.08 grids to properly align things like connector pins etc.