OK, thank you for clarifying the position. Just to be clear, I have absolutely no interest in rubbishing KiCAD or demanding any changes from it - for me, this is just about identifying the best way into PCB design. I have paid for the KiCAD training course that Peter Dalmaris offers, and KiCAD is currently my favourite choice. Autotrax DEX looks superficially attractive, but I’ve already raised that for discussion in the past.
Thanks again, everyone.
The problem with comparing applications is that there are 3 ways to compare:
Beginner in X, beginner in Y compares software
Problem with this is that the results are entirely random. It depends on what they used first and what they happened to spend more time on. Their experience is also random since it depends heavily on where they encountered a problem first.
Expert in X compares Y which he has only a cursory understanding.
Problem with this is that this is not a honest compassion. Its going to come out as i can do A,B and C in X but Y is weird that does not allow for this workflow. But here is the thing workflow and software are very intertwined, changing software automatically alters your entire workflow. So this is hardly useful.
Expert in many software of category X and Y and expert in both X and Y compares applications.
This is extremely rare to find. Why? Because people of this kind are rare. Worse its unlikely that they have time to do a comparison.
Its also less ideal to most users. People like this are really pragmatic to begin with. And they don’t see things as being better or worse due to necessity of using several software. Their opinions also dont resonate well with beginners since they have a really weird take on things that contradicts with what many other sources are saying.
Worse they are not immune to “Expert in X compares Y which he has only a cursory understanding”, since they use something more.
Comparing software is less useful than most think.
I see your point, but following your approach to it’s logical conclusion would mean that all software reviews are essentially meaningless, because it’s unrealistic to review a product without comparing it with its competitors.
Having said that, I have quite often found that a reviewer will make no mention of something that I subsequently find really annoying. For example, ‘double-click-and-hold-then-drag’ to select a word and extend the selection word-by-word, is an essential feature in any word processor I use, yet almost none of them implement it. And it never gets mentioned in reviews.
Despite that, I still find comparative reviews helpful.
Back in the world of PCBs, one of the things I like about that DEX product I mentioned is that it feels very tightly integrated and coherent, whereas KiCAD still feels like separate applications with a layer on top through which they interwork. I don’t want KiCAD to change - I’m just giving it as an example of a comparison that is meaningful to me.
One thing to be wary of… this is a kicad forum and thus will have kicad advocates here.
if you want to two different pubs and asks people whether football or rugby was better… the outcome will be based upon the type of pub not whether rugby or football is better
Anything deemed to improve Kicad is fair game. But in general, the more that benefit the better. In general this is a forum of engineers or engineering minded people so the signal to noise is good.
Since I have just spent a lot of hours trying to compare Klipper and Marlin via online reviews and research (I’m trying to decide whether or not to update my 3D printer), this really hits home…
John
My $0.02 worth.
Most of my PCB’s are mixed signal. Don’t really push the boundaries of size, layers, complexity, etc. But I design commerical products, and have been for 30+ years.
My PCB design experience spans back to the AutoTrax days in the late 80’s (Nick Martin used to work at the Uni I went to).
Spent most of my formative years on Protel - which has now morphed into Altium. Have never used Altium - can’t justify the license cost for the sort of stuff I do. When Protel 99 became too old and unsupported (probably in mid 2000’s), I moved over to Eagle.
Eagle was relatively easy to pick up from a Protel background. Did pretty much all the things I needed. But, like everything, had it quirks. The way Eagle did measurements on a PCB v how I was used to doing it in Protel was one that comes to mind (clearly it annoyed me that much!)
Eagle is now all bundled up in Fusion360, and I shudder with fear everytime I need to fire up Fusion360. So, decided to look for something to replace Eagle as I wanted a more up to date package.
Tried DipTrace (because a customer had a design that used that they wanted me to update) and didn’t really like it. Admittedly I didn’t give it much time.
Tried KiCad, and find it pretty much ticks all my boxes. Does it have some quirks? Yes, of course. But you learn to work with them.
I found it relatively easy to pick up, there is lots of info on the forums to help when you get stuck and there are plenty of features I’m yet to find the time to learn how to use.
Having access to lots of ready-made libraries and 3D models has saved me heaps of time.
I’m happy with my choice to jump on board with KiCad. I’ve no regrets, and have been encouraging others to do the same.
Mike
My 2cents. Stop wasting your time comparing and just use KiCAD. The program you know how to use the best will be the best. Design a few boards and have them made. Note that any errors are yours and KiCad warned you. With KiCad there are no down sides, no company mergers, divestitures, new CEO wants to make things more efficient, a subscription plan is what our users want or uninformed investors decided some crap BS.
The only problem with KiCad is that it is so good and so cool that it’s no longer free for me. I feel I have to donate an amount commensurate with my enjoyment and productivity each year.
There are things that KiCad does not do, eg field and thermal solvers, but commercial packages that do these features are really expensive
An important caveat to running google search queries like “kicad vs eagle” is that the returned search results are a mere snapshot of the situation as it was at that point in time. The nature of software is that it evolves, and software that is still being maintained and/or actively developed naturally sees changes over time. Usually, these changes are improvements. We all appreciate the rapid development cycle of KiCAD, and that somewhat reduces the value of google searches for comparisons because the current KiCAD state no longer reflects the state when the comparison was made several months or even years ago.
While I agree Google often cites some very obsolete article, it is for the reader to determine if the information is recent enough for their purposes. However when you have no information, often older info can be useful.
As someone who did PCB design 30 years ago, but then became a software guy, and have recently come back to PCB design, what I have to offer is that I was able to pick up and become productive with KiCad in about a week of full-time effort (so let’s just say 40 hours). I wasn’t productive out of the gate, but I’d say that for a task as complex as PCB design, that’s really not too bad. Who learns AutoCAD in a week?
I don’t remember what package we were using back in the day, but for me it was on done a Mac SE30 with a Radius Pivot monitor, IIRC. In other words, many people have better hardware on their wrists these days, and almost everyone in the first world has FAR better hardware in their pockets. When I was there, the people who were really good at it got Quadra 950s (which cost as much as a premium economy car at the time).
The one thing I’ll say is that whatever we were using back then totally wasn’t broadly as good as KiCad, but it did have auto-routing that beats the pants off the ‘no routing’ built into KiCad, and beat the pants off of Freerouting for quality of output… but, when you started a routing job it was usually at 5pm and would run all night, or all weekend for more complex boards, so… you be the judge.
I’m extremely grateful that KiCad exists. I paid nothing and I got something that got me 95% of the way there with far more “intelligence” that what I’d used previously (i.e. no schematic designer, so obviously no “Update PCB from Schematic”). We drew schematics by hand, and when they were mostly nailed down we’d print them on a 6 foot tall plotter that took up most of a room, and took most of an hour to plot it. Then we’d create the PCB layouts completely by hand. (Using whatever software, but there was no link between the PCB layout and the schematic.)
In other words, I’m counting my blessings for KiCad.
Off topic estimation of CPU progress in the last 30 years…
Quadra 950 apparently used and 68040, and apparently that is somewhat on par with 80486.
When I look at PassMark CPU Benchmarks - Low End CPUs Then the slowest / oldest processor the mention is an Intel Pentium 4 1300MHz, and it has a passmark rating of 77. 80486 had a max speed of around 120MHz (For the DX4) so a rough estimate would be a passmark rating of 7.
Currently at the top of the list is the AMD Ryzen Threadripper PRO 7995WX with a passmark rating of 153173. and 153173 / 7 = 21881
A weekend from Friday 17:00u to Monday 08:00u = 7 + 24 + 24 + 8 = 63 hours, and that is 226800 seconds.
So a rough estimate is that your weekend job from back then would now take 226800 / 21881 = 10.3651 seconds.
Moore’s Law at work!
Everything is relative. In order for you to find the best design program, you need to try and evaluate and compare with others.
I compare according to the following criteria:
I don’t like commercial products that are too expensive and not everything is convenient. Moreover, if Altium were now free, but not open source software, I would still give more preference to KiCAD, according to the criteria that I wrote above. For me Altium is too complicated and not convenient. In my opinion, the best products are created when they are created by a large community and when there is feedback from the consumer.
and yes - taste and color, no friends. everyone chooses what he likes best.
That is very helpful and an excellent account of your position.
I suppose the definition of “best” is crucial, here, and there will be many different interpretations of it. For myself, there are two really important elements: the look, feel and operation of the UI; and a speed-bump-free workflow. Obviously there are other important elements, not least of which is the capability and functionality.
Open source is not important to me, per se, although I obviously appreciate not having to pay a licence fee. But there is no “matter of principle” thing for me, as there obviously is for many people.
Cross-platform is also not important to me, as I always use Windows (I run Linux Mint in a virtual machine, but prefer Windows). In fact, I think cross-platform can be a problem because such programs hardly ever adopt the look, feel and design language of the OS they are running on. Most cross-platform programs look like orphans that don’t belong anywhere, using generic visual elements that look slightly “foreign” to whatever OS you are using.
The other disadvantage of open source is that the applications rarely get as much developer time as the commercial products, so they lag behind when it comes to functionality and keeping up with the operating system design languages. Libre Office is an example - to my eyes it looks ugly and clunky (like it was written for Windows 95 or earlier - oh look, it was), and misses out on some really basic functionality (text selection, for example, is a bit rubbish). Compare its look with Softmaker Office (a proprietary offering in the same space), which is so much cleaner and more modern. Another example: take a look at the Affinity suite from Serif and compare it with any open source graphics editor (or vector or DTP application). It started as a clean sheet design, launched in 2015 after six years of development effort. It’s outstanding, and makes GIMP look amateurish.
So, it turns out we disagree on cross-platform and open source! Having said that, I think KiCAD does well on the UI front. Yes, the menus, toolbars and dialogs look like Windows 98 or thereabouts, but it is learnable, internally consistent and the consistency between different platform versions is excellent.
While OpenSource is especially valuable for software people that want to tinker with code, the code’s “open-ness” is also realted to open documentation of file formats.
A very nasty caveat may be a software that locks your designs in a version/license prison: you design stuff only to find your OS upgrade broke compatibility with the version of proprietary software you did paid for, and now your current option is subscription model with license allowing to only store your project in the “cloud” (and being able to store projects locally is only available for the Pro Ultra license at 20k/yr)… Although this example is excaberated, it’s not THAT far what companies behind closed-source, proprietary code are doing now.
I have had several instances during my career where I lost access to proprietary software for various reasons (graduation, job changes, software supplier was acquired, etc.). This resulted in a lot of pain and lost hours. Open source software can have its own problems and pain points, but I have never lost access to years of my own work because of it, unlike expensive proprietary software.
John
Many commercial products have turned into monsters trying to combine the functions of other programs. As a result, work productivity drops, errors appear and this is clearly not what the user wanted… Quite simple functions turn into a quest and search for solutions with the opening of 10 menus and 50 settings sheets… This trend is observed not only in engineering programs but also in operating ones systems.