Why Choose KiCad over Eagle?


#1

I know this question has been asked in the past. I want opinions now that KiCad is at version 4.x and Eagle is now part of Autodesk.

I will be using the replies as a means to convince ~10 others that KiCad is worth taking the time to learn.

Thanks!


#2

You will likely only convince people if they are already looking to escape Eagle. Some might be too afraid to leave. There are many good reasons to use KiCad rather than eagle, but a few I can think of at the moment are:

  1. kicad is free and there is no limitation on the purpose of your use - no one cares if it is for hobby or for a commercial production environment
  2. format specifications are public, so even in the future if kicad would somehow disappear, your data is not hostage
  3. kicad does not have artificial limitations on number of pins or board size and even supports 32 copper layers at the moment
  4. MCAD: for eagle you add a ULP and pay $$$ for an external package to provide STEP export. With kicad 4.x you can use FreeCAD (free) + Maurice’s (free) StepUp tools to create STEP models. In the dev version of kicad (hopefully soon KiCad 5.X) you can export directly to STEP.
  5. The push and shove router is very convenient to use
  6. If you need length tuning, KiCad supports that
  7. KiCad has some support for differential pairs
  8. KiCad is continually improving and you get all the latest features for free. Eagle has been stagnant for years and even now the biggest change made by Autodesk seems to be to diddle the licensing - I don’t know any engineer who would consider that progress. Expect more of the same games from Autodesk. If you really want to pay money, use Altium - the cost ends up about the same and it leaves Eagle for dead.

The transition will not be so easy though; we still don’t have a complete set of tools to translate Eagle XML design files to KiCad. So KiCad is definitely useful for future projects, but at the moment porting older projects is not trivial.


#3

THANK YOU!

These are people that are completely new to PCB EDA software.

Your list is VERY good. I do thank you for your time.


#4

Is a critical point.
I have several old projects in my office that were created on Protel, Orcad and Cadstar.
None of these are accessible due to problems with licence keys and dongles and Windows versions.
Maintaining access to an old delivered project requires significant amounts of money every year, something the accountants don’t like at all for work that doesn’t obviously earn revenue


#5

One current qurik with KiCAD are the libraries for symbols and footprints.
Due to KiCAD being in continuous development with limited resources the way symbols are handled is more cumbersome and irritating than the way the footprints are handled.
At some point the symbol libraries will behave like the footprints do and have the same kind of features (github integration, etc. pp.) but until that point it is as it is.
If you can wrap your head around it though and don’t expect any magic in that area you should be fine.
We even have someone around that posted a KiCAD classroom setup for students and a couple of other people use it for collaborative work - so a lot is possible if one is willing.

This is the essence of it:


#6
  1. Kicad is free
  2. Kicad have interactive routing ,difference pair routing, length match, basic auto routing. If other package have same feature of these price is $$$$
  3. Eagle is $$$$ now.
  4. Most of thing can do on kicad and same eagle why pay?

#7
  1. KiCad is free and opensource, not just free to use. This means it can never suddenly stop being free due to a takeover or change in policy
  2. Eagle has a strange user interface, If you want something with saner UI and more features, you are going to have to pay a lot of money for Altium or Orcad

#8

Ahh, yes, “the very first time”!

I will affirm that the first time I did a circuit design, schematic capture, and board layout “from scratch” - with “PADS Power PCB” - I experienced plenty of frustration, re-work, and time with my nose buried in the manual. I suspect the same would have been true with either KiCAD, or Eagle.

In prior incarnations over the last 20 years or so I have had superficial acquaintance with several commercial EDA programs. In most cases it was just for a board or two; perhaps just to revise a previously designed board. As best I recall, moving between those programs took less effort than moving to KiCAD. Perhaps this is because I’m older and less agile now than I was then. Perhaps it’s because I developed too many expectations of how EDA software “should” be organized and how it’s “supposed to respond”. Somebody coming to KiCAD without those expectations may find it easier to learn.

“Beauty is in the eye of the beholder.” Every EDA program I’ve used has it’s share of quirks and idiosyncracies. Or should I say, unique personalities and charming character? “Different” isn’t necessarily "better, or “worse”.

With KiCAD, I can turn out boards that are the technical equal of anything I did with any of the other programs - and it doesn’t cost $K’s per copy. The library system is a weakness, maybe a little worse than average, but then I’ve never really liked any of the other library systems I’ve encountered.

The open, natural-language file formats are an advantage! After you figure out the organization and syntax - which isn’t terribly difficult - you can often find ways to do things (using a text editor) that can’t be done in the program itself. Or, maybe you decide that doing it wasn’t all that important after all.

KiCAD’s libraries are NOT integrated into a corporate MRP or Configuration Control system like I experienced at several employers. It seems like the capability is present in KiCAD but it hasn’t been exploited. While I appreciate the power and usefulness of such systems when they are done well, I also know they can waste time and money when not done well. I hope the KiCAD development team will not get sidetracked into developing an MRP system until after KiCAD’s core program is a recognized, world-class schematic and layout tool. (And, until they genuinely have the resources to take on development of an MRP program.) This topic surfaces on this Forum every month or two.

Dale


#9

While everything @cbernardo said are valid argument, to convince people not aquainted with EDA software, the #1 argument is easy: “Don’t expect my support unless you use KiCad”. :wink:


#10

I want to THANK EVERYONE for their replies tonight.

Had a small beginners class tonight at a maker space called Olymega. The space is trying to minimise the different EDA programs that are expected to be supported in the space. The space has computers, 3-D printers, and a CNC machine. Learing all the different EDAs for all of these task is likely impossible with the current membership.

The only way to “make” things happen is to have the space “adopt” certain software for each task. This thread was presented at the beginning of the class and the brutally honest comments about some things not being perfect was a great way for them to know that I wasn’t trying to be a “fan-boy” of any particular product.

I am fairly certain that the space will recognize that the space will only guarantee support to members if they used KiCad for their PCB EDA. Those that replied here tonight helped make that happen.


#11

Well, if they need support you can always send them here, if there is a solution we’ll find it after a couple of hours or days - someone always knows something. :nerd:
If not, then we either have something new to strive for (maybe some in that makerspace even turn into devs as the source code is open) or need to live with it :wink:

If they’re all new to this I suggest this thread for a nice cheatsheet by @Foalyy :

There are some basic quirks/how-do-I-do-this things mostly anyone stumbles over within the first couple days of using the program, which also means a lot of people know the answers to those.


#12

HAH!

I already printed those out and provided a sheet to everyone that attended! I need to thank Foalyy for it. For info, I printed the Landscape version, as I think for the class that I did that version visually puts emphasis on the HotKeys. With Wide Screen monitors now common at this time, moving the mouse from center to the right EVERY time is an awful lot of wasted motion when the other hand could hit a key.

Thanks for adding helpful information to this thread!


#13

Others have provided lists of features/licence issues that of course come out very positive for KiCad, but for me the decision to start using KiCad was about how I felt that the program worked and interacted with me. I had Eagle on my computer for almost 10 years, telling myself that I should learn it properly and start using it, but I never got to the point where I designed a board, and just kept hard wiring my one of a kind devices, using a generic drawing program (Designer) for my schematics. I just was not able to like Eagle enough to invest time in it. The user interface felt outdated and the way it incorporated footprint information into the symbols just felt cumbersome.

Then I had some extra time, 3 weeks that I spent evaluating different EDA software, discovered KiCad, and spent the rest of that time learning it by making a board for something I had already designed on a breadboard (a dsPIC based signal generator). I have used KiCad ever since, including for design of several scientific instrumentation projects. (I am a physiologist by profession.)

A contributing factor for my choice was the nice integration of the 3D rendering; perhaps the design process felt closer to the hard wiring that I was used to. :relaxed: I also discovered Walter Lain’s excellent collection of symbols, footprints and 3D models ( http://smisioto.no-ip.org/elettronica/kicad/kicad-en.htm ), which I still use as my base library. When there were some changes to the way KiCad treated the 3D models, I helped relaying some information from the Yahoo forum that made him solve the problems. (I just wish those models could make it over to the STEP model world…) I also signed up on the KiCad bugtracker, used the development version until V. 4 was released, contributed several bug reports instead of spending much time in fora, and was amazed at how responsive the developers were, fixing reported bugs within days or sometimes even hours. I doubt that would happen with a commercial product.

Another feature of KiCad that I like very much (even though it feels very different from a regular drawing program) is the efficient use of the left hand for keyboard commands in combination with the right hand for mouse moves. It saves a lot of mouse clicks and feels good ergonomically. So in my mind, the choice is a no brainer, and with the new licencing scheme I would not even look at Eagle if I was a new user. Investing some time to learn KiCad is highly rewarding.


#14

Yeah, we can be a cynical, grumpy, lot of wet-blankets at times. :unamused: Not afraid to call a “spade”, a spade - but neither afraid to put a little muscle into using the spade to till a little soil.

In an environment like that I’d find it difficult to justify anything but KiCAD.

KiCAD removes all cost barriers to everybody - even the most casual or infrequent user - having a personal copy of the software on as many machines as they like. You probably have an incentive to create some “standard” libraries, page layouts, templates, filename conventions (for Gerbers), the boiler-plate fab notes, etc, so that projects can be easily shared among members. (Of course, such standardization must have provisions for creating personal extensions - with full awareness that using personal extensions may disrupt project portability.)

Glad to hear your presentation went well and was well received!

Dale


#15

When I started looking for a PCB CAD tool in the late 1990s I looked at gEDA, KiCad, and Eagle, among others. I felt gEDA was a collection of unrelated tools that people somehow made work together by force of will. KiCad looked promising but was immature in the extreme. Eagle met my requirements, including running natively on Linux. I’ve been using Eagle ever since, and have paid, perpetual “Hobbyist” licenses for V5, V6, and V7. Like many, I’m unwilling to switch to a subscription license for a tool I use sporadically. Thus I’ve taken a new look at KiCad.

I see that it’s improved tremendously since I last looked. Its routing capabilities seem superior to Eagle’s. But as I start to use it for a simple project I find it still has some serious user interface issues.

  • The Y-axis increases as you move down, rather than up. Every engineering drawing I’ve ever seen has the Y-axis increase as you move up the page. Having it in the upper left requires transposing positive Y positions for negative.
  • The coordinate origin is fixed in the corner of some arbitrary paper page (A4, US Letter, etc.) rather than a point on the board. This is utterly absurd. If I want to place a component at a particular location on a board based on a standard engineering drawing, I have to check where on the page the drawing origin is, then add the component’s X-axis and subtract the component’s Y-axis from that location.
  • If I move a board layout on the page all my component locations change.
  • I can select a dynamic origin by tapping the space bar which shows at the bottom of the editor window, but it can’t be used in the dialog boxes. It’s also far too easy to move by accident.
  • There is a “Set Grid Origin” option, apparently for manufacturing exports, but that doesn’t help me while doing the layout.
  • The “Move Footprint Exactly” is a relative movement from its current position, thus requiring even more math to figure where it should go if you want to move it relative to another component or the origin of the board. I understand a patch has been submitted recently to address some of this feature.
  • There seems to be no way to place a via except as part of routing a signal.
  • There seems to be no way to change the length of a graphic line; I have to delete and redraw it. There appears to be no “edit” capability, nor any way to see its characteristics.
  • I haven’t found any way to highlight a particular net in the layout editor other than trying to route it.
  • There seems to be no way to highlight all segments of a net in the schematic editor. The best suggestion I’ve found was to export the netlist and check the connections using a text editor.(!!)
  • The Default canvas leaves artifacts all over the screen.
  • Changing the rendering colors for things like ratsnest airwires doesn’t seem to work at all in the OpenGL canvas. I saw some comment that it was because the OpenGL support was new, but that report was almost two years ago.

I see many people have complained about the bizarre origin issues. Most of the answers are, “Can’t fix it; it’d break compatibility.” Poppycock. The addition of a simple PCB attribute specifying the origin location and which way is positive would be sufficient to maintain compatibility. Board files lacking this attribute work the way they’ve always worked.

So why choose KiCad over Eagle? Cost, a general preference for open-source products, and the recently-added push-and-shove router (which Eagle devs claim to be working on). Other than that, Eagle is a far superior product.

Frankly I’d love to ditch Eagle for all these reasons. But I’m not sure I can, yet.


#16

The top to bottom Y axis is due to computer memory mapped graphics always starting top left, a hangover from television based CRT displays

Moving footprints to an exact location can be done directly by entering in the properties dialog X and Y positions

Changing length of graphic lines is again done by editing properties - I always set the board outline on Edge Cuts exactly this way


#17

That’s an explanation for this brokenness, but it’s not a justification. It’s utterly irrelevant to a person doing PCB layout. Even worse is measuring from the corner of a page frame rather than a designated point on the PCB.

Both could be fixed without changing the layout file format except for the addition of one attribute specifying the origin and the directions in which the X and Y axes increase. Existing layouts could continue as-is, while the user could specify the attribute characteristics. The current manufacturing grid origin could be used as the user-designated (0,0) point.


#18

All valid points there - most get used to this quirk after a couple of weeks though and don’t think about it anymore.
Might also be a reason why no one is tackling it really, cause as soon as one is firm enough with the software to actually be able to poke around in the source code to remedy this situation he’s got used to it and does not care anymore (= been with KiCAD for more than 6 months, probably longer, would be my best guess).
Catch-22.

PS: I’m aware a lot of KiCAD starters are turned off by this and it would help retention rates a lot if it was tackled… but it’s not a high priority item as there are bigger fish in the water that affect a lot more users of the software right now.


#19

If KiCad wants to successfully play in the field of professional CAD, it simply has to tackle the issue.

But then, does anybody really want to be in this field in the developer community?

I feel that KiCad will be facing the chasm (refer to Geoffrey A. Moore’s marketing book) really soon now and will have to decide whether it wants to cross it, or not. Marketing mechanisms also apply to KiCad, that’s not a matter of free/paid for product, it is about market penetration. Sorry to mention marketing in a tech environment, but that’s what it is.

So: what is the “market”? This has to be decided. But then: how and by whom?

There are a lot of issues currently in discussion - at least in this forum - which point in the same direction: the layer manager discussion, customer configurable menus, the coordinate issue, library structure and management, to name only a few of them.

It will be interesting to see what KiCad wants to be, where it will be headed.


#20

Having used Protel, Visula, Orcad, Cadstar, Eagle and DesignSpark, I haven’t come across any CAD package with a properly logical structure without serious quirks