KiCad 5.1.5 released, plus interview with KiCad lead programmer Wayne Stambaugh

[I wanted a rude username]

KiCad 5.1.5 has just been released will be released on 27th/28th November.

While we wait for them to build binaries/installers, here's an interview with KiCad lead programmer Wayne Stambaugh on Open Hardware, commercial KiCad adoption, and future directions for the project.

[poeschlr]

The source code of KiCad has been tagged for 5.1.5. Now the other asset maintainers (libraries, documentation, ...) need to tag their things. After all of these have been tagged will start the work for the packagers.

Meaning 5.1.5 has not been released!

See https://lists.launchpad.net/kicad-developers/msg42597.html

[I wanted a rude username]

OK, that's an interesting way to do it. Makes sense if they don't have a unified repository and/or don't have a person designated as release master.

Release expected on 27th/28th November then. Thanks for the clarification.

[chickenHeadKnob]

There was nothing in the quoted interview which mentioned sturm und drang with respect to Wayne's employment, unless I missed it. Wasn't there a rumour that he was quitting or being let go?

[jaromir]

The "rumor" was spread by Wayne himself https://lists.launchpad.net/kicad-developers/msg42269.html

[Yansi]

Hopefully there will be at least some progress on improvement. Currently KiCAD is still useless tool for me. Waiting to adopt it.

[Warhawk]

There has been no rumors, there's an official statement on the kicad webpage:
http://kicad-pcb.org/blog/2019/10/KiCad-Lead-Developer-Announcement/

Unfortunately I never understood who WIT is, why they fund it and why it was such a big deal. I googled yesterday but did not get a clue what was going on.
Any cpt. obvious here?

[voltsandjolts]

http://kicad-pcb.org/blog/2019/04/Full-Time-Project-Leader-Announcement/

[nfmax]

Hopefully there will be at least some progress on improvement. Currently KiCAD is still useless tool for me. Waiting to adopt it.

Why do you say this? What capabilities does it lack for you? (serious question)

[Yansi]

First thing I needed to design in KiCAD was a flex PCB. That was my first task, when I was put in front of KiCAD.

And let me tell you - it SUCKED.  It can't make anything past a sharp 45° corner. No smooth curves. At all.  I had to make to PCB in a different tool, export as DXF and thought I'd import it. No. That thing can't import DXF to a copper layer.  Well OK, thinking I just change layer of the artwork. No, can't do it. No way to swap layer of multiple objects, unless willing to sped tens of hours learning with the python console (that crashes the thing right after it is opened anyway). Or to say specifically - there is absolutely no capability to change any property of multiple objects. So you can cry in anger when wanting to change even such basic thing as a font style of component identifiers. Yes, there is a global setting. However changing that does not apply backwards to the already placed components.
I had to change the layer of hundreds of partial lines, one after another. Manually. At least 5 mouse clicks per piece of line.
No polygon-selection capability. Just Altium-like left-right and right-left swept rectangle. Quite annoying, for a user used to use the polygon selection in Eagle a lot. Also, the rectangular selection malfunctioned repeatedly in a few places where couple of curved or sloped!=45° traces in different layers intersected.
Moving a line or trace breaks it in half. No way around it - or I couldn't figure out how to do it (maybe is it just me?)
Made myself a library with a FPC connector. Thought I'd use the filter, to find it - but no, I can't because letter B and V can't be typed in there. A year or more old bug that still isn't fixed (or the bugfix still isn't pushed into the stable release).
The outline of the PCB must be a closed curve. Like closed to the last nanometer, that it does not even show you the number for so much precision. So both objects start and end on the same XY point, but it still bitches the curve is not closed there. Because likely someone compares two double precision float variables to each other. Or I don't know wtf.
Lastly, what kind of person invented the fourth quadrant operation? Every single CAD ever is operating in a first quadrant. But no, KiCAD must be special and with a flipped Y axis, with a load of excuses of it affecting performance badly, hence it must be like this.

So, my first experience in KiCAD is pisspoor.  Or, the PCBNEW software is pisspoor to be specific. I quite like the schematic editor. It works like a charm and I never stumbled upon any kind of bug or problem in the EESCHEMA. I am quite impressed by the ease of use of EESCHEMA.

It seems to me, that the PCB editor is full of modern features, like 3D and such, but it lacks basic functionality, that every other PCB cad should have from the first time it is supposed to be used for actual PCB work. Such as the multiple item property change.

Maybe, I've had too much expectations from KiCAD. I knew a lot of quite complex PCBs have been already made in it successfully. But as I've said - it seem, that it only works for simple rectangular PCBs, with just simple 45° traces. And don't ever think about changing multiple things later.

I got quite pissed, when I've seen that the smooth curved traces were already there as a functionality - but was ditched recently due to whatever reasons. I've seen numerous tutorials how to do them - yet I couldn't, the feature wasn't there. I thought I must be stupid or what.

Also please note I used the latest stable build for Windows, at the time (say a couple of months back that was, don't remember the exact version number).  So using the nightly build that may contain much more features and actually might be workable, is not the option for a windows user. I think there are none nightly builds for windows, if I am not mistaken.

I hope KiCAD gets better and usable in the future. Hopefully before I decide to invest and buy another professional tool license, be it Altium, the new Autodesk Eagle, OrCAD or whatever.

Note that I am not bashing KiCAD as being "just unusable". As above, I found numerous (even major) fuckups that are the reasons why I think it is not ready for serious work yet.  Or it is just me that I can't use it for the stuff I do.

[Monkeh]

So using the nightly build that may contain much more features and actually might be workable, is not the option for a windows user. I think there are none nightly builds for windows, if I am not mistaken.

You're very much mistaken by virtue of not even looking, seeing as they're right there on the download page.

[Yansi]

I haven't looked, I was told so. Hence even didn't bother to look. I stay being corrected. Unfortunatelly, repeated nightly install on a PC at work is problematic at least to say. 

[MitjaN]

@Yansi

I can agree. The UX can be rough. It took me more than a year to find out there is a filter selection tool. So when you select a square, you can filter down the selection to include only tracks, or zones, or drawings or ...

You still can not change the properties of multiple selected items besides tracks.

Regarding the MCAD collaboration I agree that using DXF can result in some pain, but if you do this using FreeCAD's stepup, the stepup workbench knows the limitations of KiCad so you can get custom board shape and you can change it. I am aware that this means learning and using a new tool so this might not be an option.

And as for the nightly releases I can concur if there is some feature available in nightly release you really can not say it has been released. Even if nightly builds are available for your platform. The nightly builds are just not sane to use in production and I'd rather not see people suggesting to use them. Once you are using KiCad for more than a year, then, when you know the community and the release cycle and what comes with using nightlies, only then maybe it might be sense to use them.

And lack of round tracks and having footprints bound to top and bottom layer exclusively you can not really use KiCad for rigid-flex design.

But opposite to your experience, I really appreciate KiCad scripting ability. You can really do a lot. I think this can be mostly seen in the official libraries. With V5, the quality of the libraries has increased significantly also due to the automation of checks on the symbols and footprint contributed by the community. This is the first time I've seen CI pipeline used for something other than code.

So I accept it. Warts and all. But I can understand that some people have different workflow and KiCad is not suitable for them.

[poeschlr]

@Yansi

I find it strange that people use a release notification as the place to complain about missing features. Is there a way for moderators to split discussions?

---

KiCad can already do a lot of things. Working with requirements for flex pcb is not really one of them.

There were always ways to get curved traces via external tools. The best one i know of right now is the toolset collected by the forum user maui (github easyw) https://forum.kicad.info/t/rounded-tracks-reloaded-again-rf-tools-for-kicad/19190
It is expected that the next version will include support for curved traces but that is still at least year away.

Another currently missing feature is creating hatched zones. Right now you would need to use either python or a footprint to create such a thing. The next version will certainly have native support for this (is already in the current development snapshot)

---

Changing multiple objects (other than traces, vias and text) is indeed not straight forward. I would however argue that this is a minor issue at most (when do you need to globally change anything other than traces, vias or text.

---

> Made myself a library with a FPC connector. Thought I'd use the filter, to find it - but no, I can't because letter B and V can't be typed in there. A year or more old bug that still isn't fixed (or the bugfix still isn't pushed into the stable release).

This is a bug on windows. https://bugs.launchpad.net/kicad/+bug/1843161 (Not yet fixed) Workaround: do not assign single letter hotkeys for relevant tools in windows.

---

> The outline of the PCB must be a closed curve. Like closed to the last nanometer, that it does not even show you the number for so much precision.

The tolerance is 0.01mm. This is sadly not a user setable variable. Source https://www.mail-archive.com/kicad-developers@lists.launchpad.net/msg30084.html (discussion about courtyard but the same algo is used for the edge cut)
How did you come up with nanometer?

Also v5.1 offers a DRC check that points to the problem area (and snap to object which makes it easy to correct such a problem)

---

> I got quite pissed, when I've seen that the smooth curved traces were already there as a functionality - but was ditched recently due to whatever reasons. I've seen numerous tutorials how to do them - yet I couldn't, the feature wasn't there. I thought I must be stupid or what.

Disclaimer: i was not involved in this. I can only inform you about what was going on on the public mailing list (there might have been private messages which were never published.)

There was indeed at least one patch published shortly before the v4 release. Sadly at that point in time there was already a policy of not accepting patches that only implement a feature using the old backend that was kept alive for legacy reasons (It was always the plan to drop that backend by v5 which happend for at least some platforms. The reason is that this backend used an outdated library which is no longer supported by a lot of modern operating systems.)

There was another patch where the contributor was not willing to take any input from the main devs. Both of these kind of show that in a large project sometimes hard decissions need to be made. In this case a nice feature needed to be dropped (twice) because the way it was implemented would have hampered the future development of kicad.

[Yansi]

@Yansi

I find it strange that people use a release notification as the place to complain about missing features. Is there a way for moderators to split discussions?

I was asked by a member, hence responded, I wasn't going to say anything else otherwise.

[JoseR]

Hopefully there will be at least some progress on improvement. Currently KiCAD is still useless tool for me. Waiting to adopt it.

Why do you say this? What capabilities does it lack for you? (serious question)

If I remember correctly since my last try, things as advanced as:

- File - Save As
- Cntrl+C / Cntrl+V from one project to another (copy - paste)
- Add sheet to the project (without plain hierarchy or stuff like that)
- Simple BOM generation where you can select the properties shown

(I know you can do it following some 10 page tutorial or manipulating in the console a couple of environment variables)

[nfmax]

@Yansi

I find it strange that people use a release notification as the place to complain about missing features. Is there a way for moderators to split discussions?

I was asked by a member, hence responded, I wasn't going to say anything else otherwise.

And you answered my question: thank you! I was interested to know if you had UI issues or capability issues - clearly the latter!

[awallin]

5.1.5 package seems available for Ubuntu now, from the js-renaud PPA (https://launchpad.net/~js-reynaud/+archive/ubuntu/kicad-5.1)

[SiliconWizard]

Available on Arch now as well.

Note that this is only, AFAIK, a bug fix release (although a sizeable 96 bugs fixed), so don't expect any change except if you were running into one of these bugs.

[Docara]

………….. I had to make to PCB in a different tool, export as DXF and thought I'd import it. No. That thing can't import DXF to a copper layer.  Well OK, thinking I just change layer of the artwork. No, can't do it.

Don't even get me started on that one!! I recently had a bitch on the KiCad forum concerning exactly this issue, and was slapped down via a PM from a developer. I am working on a  CapSense layout. All I wanted to do was produce a hatched layer and 5 custom circles of copper with offset holes in for LEDs. The native 'shapes' contained within KiCad are also not able to be used with copper layers - F**ckin' madness and F**ckin' annoying!
OK there is no hatch copper fill command fair enough, so what I did was make my Top Copper 'shape' in Autocad and import the DXF into KiCad only to find out there was no option to import graphics onto copper etc.

I made the point that someone in development must have specifically excluded this functionality from the available import onto layer dialogue box. If this simple functionality was just enabled it would massively increase the productivity and usability of Kicad and not to mention overcoming the limitations of the software. We accept that it is not a £7000 Altium or whatever but don't force limitations on us. If you have a list of layers to import onto, the basic functionality is in place within the software,  but to then specifically exclude the copper layers it just blows your mind over their short sightedness - what if I wanted a custom logo in copper  aarrgghhhh!!!!!!!!

For us to have a custom copper shape in Kicad you must download (and learn) Inkscape, download and learn a pluggin. Import this in to Kicad. Then repeat as necessary because the position and sizing's are probably wrong because Inkspace isn't an engineering CAD package but a drawing prog. All because someone has specifically removed a function.

Then the old favourite response from the developers when they know they are wrong and backed into a corner     "...... if you don't like it learn to program and change it yourself" makes your blood boil. It reminds me of the silly flat earthers rhetoric.  I have been made redundant, trying to relearn electronics from over thirty years ago and don't have the ability to spend hundreds if not thousands of pounds on software just to try and get ahead

Anyway rant over (for now LOL)

[2N3055]

This is Kicad roadmap. Pay attention to bold sentence:


************************************************************
Version 6 Road Map
This document is the KiCad version 6 Developer's road map document. It is living document that should be maintained during the version 6 development cycle. The goal of this document is to provide an overview for developers of the goals for the project for the version 6 release of KiCad. It is broken into sections for each major component of the KiCad source code and documentation. It defines tasks that developers an use to contribute to the project and provides updated status information. Tasks should define clear objectives and avoid vague generalizations so that a new developer can complete the task. It is not a place for developers to add their own personal wish. list. It should only be updated with approval of the project manager after discussion with the lead developers.

Each entry in the road map is made up of four sections. The goal should be a brief description of the what the road map entry will accomplish. The task section should be a list of deliverable items that are specific enough hat they can be documented as completed. The dependencies sections is a list of requirements that must be completed before work can begin on any of the tasks. The status section should include a list of completed tasks or marked as complete as when the goal is met.

General
This section defines the tasks that affect all or most of KiCad or do not fit under as specific part of the code such as the board editor or the schematic editor.

User Interface Modernization
Goal:

Give KiCad a more modern user interface with dockable tool bars and windows. Create perspectives to allow users to arrange dockable windows as they prefer.

Task:

Take advantage of the advanced UI features in wxAui such as detaching and hiding.
Study ergonomics of various commercial/proprietary PCB applications (when in doubt about any particular UI solution, check how it has been done in a certain proprietary app that is very popular among OSHW folks and do exactly opposite).

************************************************************

Seriously, can't make that shit up....
This, right there, is what is wrong with what otherwise could be best thing since the invention of fire ....

If attitude wasn't like that, ALL of UI problems wouldn't even be there. Most of crappy stuff in Kicad is not bad programming, but deliberately bad (and obviously spiteful) choices how to implement certain functionalities...

Don't get me wrong, current version came a LONG way in last year alone... But it could have been perfect by now if it weren't for apparently deliberately bad implementation choices, and misguided priorities (Spice integration wasn't more important that implementing proper PCB editor)......
At this point, it is not fully featured, and deficiencies are not obvious. On paper it looks nice and all, but when you start project, than you start hitting roadblocks on things that should be basic functionality.
So, you cannot make a full switch to it, and I cannot afford a learning curve for something that I will have only limited use for...
So no switch for me at this moment...
 

[poeschlr]

@Docara i can not comment on what happened in a private conversation but i know what public conversation you reference (I could however not find it right now so i work from memory).

You were told that yes the current version of kicad has problems with the thing you want to import from dxf (direct import to copper is not possible).
You were also told that the current development version would already support what you want to achieve. You did not like that answer so you were given the suggestion to use freecad with stepup. Which you also did not like (because why learn yet another tool). Then you were given the option to use python scripting. Again something you did not like.
In the end your topic felt more like a place for you to rant than a place were you wanted to be helped.

---

@2N3055 As far as i am aware the "certain other pcb tool" that this sentence references is eagle version 6 (Yes this part of the roadmap is from quite some time ago. It got copied from at least the  version 5 roadmap possibly even earlier) I am convinced it is not fully serious as discussions on the mailing list indicate that the developers take the user experience of other tools into account when making decisions about the graphical interface (this includes modern eagle).  I will ask for clarification on the mailing list as you are right this type of humour has nothing to do in the roadmap of software that aims to be a professional tool.

[iMo]

Where will KiCad be in 5 to 10 years?

Wayne: I’d like to see us start moving to more simulation. One of the other places where I think the open hardware movement is just starting to branch out into is silicon. Maybe we could add a set of tools for chip design. I think we are pushing the whole open ecosystem down the stack.

The RISC-V is really interesting. But even that is still not completely open because manufacturing processes are not open. You still do not know completely what’s inside your silicon. I would like to see that opened as well. Anything that KiCad can do to help support that would be great. That’s a long way off in the future but I think that’s what I’d like to see 5 or 10 years out.

[SiliconWizard]

The RISC-V is really interesting. But even that is still not completely open because manufacturing processes are not open. You still do not know completely what’s inside your silicon. I would like to see that opened as well. Anything that KiCad can do to help support that would be great. That’s a long way off in the future but I think that’s what I’d like to see 5 or 10 years out.

Yeah, I read that as well. I'm not sure what kind of crack he was on though. How adding IC design capabilities in KiCad (and I wish them already good luck with that if they want something usable) would make manufacturing processes suddenly open? I fail to see the direct link. There are certainly already other IC design open CAD tools, and that magic (no pun intended, I think 'magic' is one of them) hasn't happened. But sure, mine is always bigger than yours. So there.

[SiliconWizard]

Official binaries are now available.