EasyEDA Two Years Later

Some people want everything on the cloud, while others refuse to put even the smallest scrap of data on the Internet. Most of us fall somewhere in between. A few years ago, we talked about a few cloud-based PCB layout programs including one called EasyEDA. We were impressed because it was a full package: schematic capture, simulation, and PCB layout. It was free to use, although they would give you a quote for producing your boards, though you were under no obligation to buy them. Of course things change in two years, so if you are curious how EasyEDA is doing, [Yahya Tawil] posted an in-depth review.

Some of the new features include an autorouter and the ability to order parts from a BOM directly, not just PCBs. The cloud aspect is handy, not only because you don’t have to install and update software to use it anywhere, but because it is very natural to collaborate with others on projects. We did notice, though, that the autorouter can run in the cloud, or you can download and run it local because it apparently loads the server significantly.

Unlike some online PCB layout programs, EasyCAD produces industry-standard Gerber files, so although they will offer to produce your boards, you can take your files anywhere. Two years ago we did have a short run of boards made just to see how they would come out and they were fine and the price was very competitive.

If you missed our original EasyEDA post, we looked at it and — no kidding — MeowCAD. We also know [Hari] used them for his giant LED cube boards. If you prefer your reviews in short videos, [MickMake] has part one of his review, below.

45 thoughts on “EasyEDA Two Years Later

  1. ” A few years ago, we talked about a few cloud-based PCB layout programs including one called EasyEDA. We were impressed because it was a full package: schematic capture, simulation, and PCB layout. ”

    I can see the later two benefiting from the cloud because there’s more horsepower on that end.

    1. “More horsepower on that end.” Maybe if you’re on a phone, tablet or low-end laptop, but anything better is probably more powerful. While the hardware the “cloud” application runs on is powerful, that power is shared across many users. Especially for a free product, where there won’t be as much hardware backing it. So if you only look at the relevant part, namely your tiny share, it doesn’t actually work out that great. For instance, the server might have dual 8-core CPUs (@ ~100 W each) = 16 cores (@ ~200 W) – share that over 10 users (1.6 cores @ ~20 W), and a dual-core laptop looks about even (2 cores @ ~15 W – ~45 W).

          1. Nothing teasing about a relatively cheap dynamic pool of computing resources to draw upon which is usually greater in a data-center. Laptop isn’t as easily to adjust if dual-cores isn’t enough.

          2. Sorry, vaguely saying “scale” just isn’t enough. PCB design doesn’t need scale. It is a single-user problem that is not currently limited by personal computer specifications. What do you need “scale” for? Why do you need more cores? How will drawing a trace from point A to point B utilize 75 CPU cores? The only relevant application could be renting a bunch of cores to crunch on a tough autorouting job, but the major problems there don’t exist simply because we don’t have enough CPU power. So I ask again: how to you intend to convince me that “scale” is relevant to the discussion? Why is the vague promise of “scale” enough for me to not only take my files off my own computer, but farm out the CPU work offsite so that I couldn’t work on the files even if I had them?

        1. Define scaling. The only relevant meaning in context is increasing hardware performance available. So you mean, it is harder and/or more expensive for me to upgrade my laptop, or switch to and/or upgrade a desktop, than it is for me to convince the provider of some cloud service to upgrade their hardware? Sure, if the “cloud” instance is something I’m using myself for my own things / services it is easy to pay a bit more to get better hardware. But not when using a third-party service like here.

          1. The provider of that cloud service will probably focus on building their application and run it on a fleet of rented virtual machines so they don’t have to bother maintaining hardware/networking/etc while having more time to work on their actual application. And if you are renting capacity on one of the larger clouds you can get machines that might get terminated prematurely but are very cheap. This way you can automatically get more computing power for shorter burst loads like enough users burning the autorouter and delete those machines when the load goes down.

            For instance on google’s cloud these ephemeral instances cost on the order of 40% of what you’d be paying for the same compute capacity normally.

  2. My only concern with an online EDA is when the rug is invariably yanked out from underneath users. If Eagle goes belly up, you can still use whatever the last version was and eventually switch to a new EDA.

    If EasyEDA goes belly up or the company behind it figures out it’s no longer profitable, users who’ve integrated their workflow into easyEDA are left out in the cold with no guarantee of an alternative.

    Why would I spend hours working with a tool that could disappear on someone else’s whim? Local control will always be preferable to the cloud if there’s no guarantee of availability and uptime.

    1. Except that now that Eagle is subscription, you won’t be able to keep using that (unless it is a cracked version, of course), you’d have to be using an older version – and probably only be able to open the older files too.

      I think EasyEDA has promised to open-source everything if they do go out of business, but I wouldn’t trust that – depending on why that happened it might not even be possible.

      But I totally agree with your point. I don’t use “cloud” apps for any kind of work, and I don’t want anything else that depends on someone else’s server to work either (particularly hardware).

        1. That falls under “using an older version”, as far as I can see – and that being the case, it is probably not available to anyone that doesn’t already have it. My guess is even the free version of 8 will still phone home to check if it is allowed to run.

        2. Oh really, that’s actually all I need, huh? What about the times I’ve needed to make four-layer boards, or boards bigger than 80 cm²? Or when I want to organize my schematic into a bunch of sheets?

          I thought I needed those things, but sure enough, I don’t! Thanks for the tip, jouni! /s

      1. Promising to open-source something is no promise at all. You have to ask: why would they go out of business? The most likely answer to the question is that their software/service isn’t competitive, is making a loss, and few people are using it, and if you find yourself being one of the few you’re up the creek without a paddle.

        Because nobody else cares.

        That’s the main falsehood of Open Source Software: anyone can in theory work on it, but in reality the ratio of plain users to able developers must be exceeding 10,000 : 1 and so when the pool of people who are still interested in some particular piece of software starts to shrink, it is very likely that none of the remaining know how to keep it running, and so the software dies.

        1. Very niche or poorly maintained OSS projects die all the time due to the reasons you mention. But there are some standouts that have enough momentum to keep going no matter who gets hit by a bus. I don’t see KiCAD ever going away unless another project does everything better someday, or if the traditional PCB CAD approach becomes useless for future technology.

    2. Well said.

      The main author’s assertion that the only downside of using cloud apps is that’s it’s “on the Internet” is utterly simplistic and completely inaccurate and quite frankly, if intentional, somewhat offensive as well. The downsides are considerably larger than “just being on the Internet”. As Ethan states, one of the bigger issues has to do with what happens when the software is invariably no longer being supported. You go from an old but still working version with normal software to something that literally stops working, taking all of your data and work with it. It’s as if suddenly every CAD file you spent years making suddenly not only stopped working but you couldn’t even open the files at all anymore or even use the software at all.

      That is not anywhere close to just “being on the Internet” and that’s only a small part of the issues with cloud based software.

      Characterizing the majority of people as perfectly ok with that seems unrealistic and misguided. Even more so if one relies on said software for their income and well-being as a professional.

      1. I’d say the “majority of people” are indeed “perfectly OK” with the so-called “cloud”. But only because they don’t realise that, or they (foolishly) trust the providers to keep supporting what they’re using.

        It’s also an issue with hardware, especially as it is totally unnecessary in almost every instance and deliberate to make people re-purchase something they don’t need.

        1. I have a Hello Sleep device. Makes a nice paper weight now that they switched the servers off.
          The cloud part of the system was not advertised and had I known at the time I would probably not have bought it.

          My wife has a fitbit which at some point, after newer models are released, will magically stop working because it will not be profitable to support it.

          1. We are very very smart people. I think we are smarter then the average person by far.
            I really think that we should try with all are might to stay off cloud software.
            When And if you were to really think about it Cloud base software is really bad for you.
            Like MartynC and PreferLinux sayed. And there are so many other reasons.
            I do use 1 cloud based program and 1 day I do expect all the files I did in it to be locked up tighter then a can of cat food in a box at the bottom of the ocean. I am using this one program to see what really happen when a server goes down. I have never gotten anything or used anything before so I think that this would be a good test.

    3. Sounds like an argument for format standards, as in one can play an MP3 in pretty much any player they want. PITA if one’s favorite encoder or player goes belly-up, but there’s still some hope. There’s of course going to be one loss, but could be said about any software change as well in the learning invested in a particular tool (Wordperfect to MS word).

      1. Absolutely, but how are you going to force these “cloud” providers to use it? In fact, how are you going to force them to even let you download the files in the first place, irrespective of format? There’s no way to do either – otherwise the whole discussion would not exist.

        1. Do you need to force “them”. The cloud services aren’t this homogeneous entity who offer you a single combination of characteristics. You are free to pick a service that allows that if it’s what you want. Just like you’re free to pick your local EDA package based on whatever is your criteria. And just like in a cloud service you can’t count on something being available online ad æternum, you can’t expect Eagle to keep that sweet sweet version <7 online. Have a copy of the installer? Hooray. You don't? Tough luck, now you have to use whatever CadSoft is offering you now.

          1. Sure. But you’re missing the point. Why would any of them let you? You can pick the one that does what you want, but only if they exist!
            And I use KiCad, so couldn’t care less about Eagle. I can count on that being available, because, being free software, there’s many mirrors and no one person or company can take them all down.

      2. Format standards is a moot point when software is deliberately introducing idiosyncracies and differing features to differentiate themselves on the market.

        The standard becomes ill-fit for the software and its users’ purposes, and translating between two pieces of software through an intermediate standard which neither of them implements perfectly becomes an act of futility – or as the video points out: importing is never perfect.

    4. Hello,
      I tried Easy EDA some weeks ago. I’ve convinced myself to put the cloud disadvantage apart because of 2 things : the desktop version and the Altium export.
      The desktop version no longer exists and I just have an Altium viewer which seems to be fine but I’d like to have a real Altium user to confirm.
      All in all, it’s nice for quick (small) boards with the community footprints “ready to use”. But for bigger projects, I would use a non-cloud based solution.

    5. You still have issues with local software which becomes unsupported or when provider is closing.
      It could be your license file that becomes rotten due to hardware change (failed motherboard, hardware update, graphic driver issues), or new OS not supported (and license forbid VM).
      I had this with CAD software (Alibre), I give up trying to resuscitate my old license as it’s not supporting win10 and my new hardware.
      The only way to overcome this is open data standard.

  3. Two or three months ago, I tried EasyEDA. The real killer feature for me was the BOM Export/Component Selector.
    I only chose Components from their big library of available components.
    The result of this was, that I was able to order produced PCBs along with the needed components – in the same package, no extra shipping fee. That was really great, as I was able to start soldering right away – No need to use three other onlinestores to source the components with the additional waiting time and shipping fee.
    Also, I would never have gotten the Components for that price here in mainland europe…

    Of course, I exported the design files, so I could reuse them even if they went offline – Although I doubt I will find the cheap-as-hell, chinese-datasheet 7 segment display drivers here in local stores…

  4. I use EasyEDA because my design files are consistent across multiple platforms. With the design in the cloud, I can work on it from ANY platform or system that has a browser. At home, I can use my desktop. At a coffee shop, I can use my laptop. In the office, I can use any computer. In every case, I am always working on the most current design files.

    Further, EasyEDA is easy, fast and inexpensive. For a new design, creating a schematic, placing components on the PCB and autorouting can be done in as little as 2 hours. If the resulting PCB is smaller than 10cm x 10cm (4″ x 4″), it can cost as little as $0.20 to $0.50 per board. Boards arrive in about a week. If the circuit doesn’t work right, just spin the board and order a new batch. The first PCB can be the breadboard.

  5. I’ve just used it last weekend for the first time to design an ESP8266 based sensor node.
    I’ve used Eagle and Kicad for small maker projects before. It was really easy to get into EasyEDA knowing the general ideas behind these types of programs. The really nice thing that stuck out was the library of parts and footprints already built by them and other people. In KiCAD I had to pretty much build footrprints and symbols for all the ICs I used because the library is quite outdated, or just not ment for stuff that makers use, here I just had to verify 2 footprints that were user-submitted and not from their huge library.
    The schematic editor was ok, the routing was pretty nice (not sure how accurate their DRC is but for my basic requirements it did the job fine). I didn’t use the autorouter cause they usually suck. Changing stuff in the schematic and updating the PCB worked flawlessly every single time even with major design changes, something that always seemed to go wrong in KiCAD.
    I liked the fact you can order stuff from their shop straight from the BOM and the fact it’s really cheap mostly: an ESP module was only around 20 cents more expensive than on Aliexpress, and I could buy tiny things in exact quantity instead of the usual 50-100 pieces that you can get on Aliexpress. My BOM asides from the module came to a grand total of 1.24$, I could not get those pieces for under 5$ on aliexpress/banggood/taobao due to small parts being bundled to save shipping. I also loved that you can combine your orders if you already placed a PCB order and remembered you needed parts too afterwards (save shipping).
    All in all I think it’s absolutely great for makers and people tinkering around with basic stuff, it’s simple enough to be easy to learn, does not have any stupidly complicated menus either.
    And if it goes away tomorrow you can just add another hour of fiddling and switch to KiCAD, they’re quite similar, the principles anyway.

    To people glorifying the fact that it’s on the cloud so it MUST be faster: 90% of the stuff you do in it is handled by your own computer – it’s just javascript. Only some advanced functions like the save functions, updating the board after a schematic change, the library of parts and maybe the autorouter (didn’t check) are cloud based, those are things that would run super fast locally anyway. So there’s no super-performance benefit if you ask me, it just runs fast because it’s simple and not a jumble of 100000 rewrites by 100 different people like KiCAD and Eagle.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s