Autodesk buys Eagle

[uncle_bob]

I did this super fast demo that shows an example of how SolidWorks goes about 2D geometry. Not only do you tell it dimensions, you can also tell it about relationships - vertical, horizontal, parallel, coincident, perpendicular, etc, etc.

That's definitely a powerful tool, no question about it, but I wouldn't call it super fast.  You still had to do an absurd amount of clicking and dragging to enter an utterly trivial amount of information: X,Y centers for five pads with two unique W,H dimensions.  If the drawing had simply stated that information directly, entering it into EAGLE or any other package would take 10 seconds.

Hi

Every time I see a story about "universally adaptable all problem solving autonomous robots take over XXX next year" ... I think about how long we have been struggling to get a seemingly stupid problem like this solved. There is a page in the data sheet that gives the information. It is in PDF format. Click .. PDF open ... click ... page selected ... click ... import ... click ... select page with pin numbers ... click .. import ... click .... select page with "schematic" ... click ... import.  That's a total fantasy. It's been a fantasy for at least 4 decades, despite it being a "millions of dollars a year" sort of problem.

Bob

[rx8pilot]

That's definitely a powerful tool, no question about it, but I wouldn't call it super fast.  You still had to do an absurd amount of clicking and dragging to enter an utterly trivial amount of information: X,Y centers for five pads with two unique W,H dimensions.  If the drawing had simply stated that information directly, entering it into EAGLE or any other package would take 10 seconds.

It is not at all optimized for component pad layout, it is a general purpose tool for creating vastly more complicated geometry that will be used in the generation of 3D. There are some tools that could have sped that up, but really the point was simply to show that you can pick a center,point, or line and tell it how far from another center, point, or line. In Eagle, you can only directly know where the center of the pad is and have to do the math manually to figure out where the edges are. When trying to build it around a center point to have a proper centroid for P&P, it is a lot of manual calculator activity. I have been drafting on computers for decades now, and Eagle is never 10 seconds for anything since the datasheets never call out the dimensions in the way Eagle needs to see them. It should not require the use of a calculator to make a footprint.

Every time I see a story about "universally adaptable all problem solving autonomous robots take over XXX next year" ... I think about how long we have been struggling to get a seemingly stupid problem like this solved. There is a page in the data sheet that gives the information. It is in PDF format. Click .. PDF open ... click ... page selected ... click ... import ... click ... select page with pin numbers ... click .. import ... click .... select page with "schematic" ... click ... import.  That's a total fantasy. It's been a fantasy for at least 4 decades, despite it being a "millions of dollars a year" sort of problem.

Bob

It is bizarre how crude it is. Some very simple format could allow all manufacturers to make available the data needed to import directly into any system out there with no translation needed at all. Strangely, the way it is done by millions is to download a PDF and manually figure it out. In 3D, it is very common to have IGES, STEP, and SolidWorks files for every nut, bolt, fastener, spring, etc that I ever use direct from the manufacturer - saving huge amounts of effort that would be required to model them.

[nctnico]

Every time I see a story about "universally adaptable all problem solving autonomous robots take over XXX next year" ... I think about how long we have been struggling to get a seemingly stupid problem like this solved. There is a page in the data sheet that gives the information. It is in PDF format. Click .. PDF open ... click ... page selected ... click ... import ... click ... select page with pin numbers ... click .. import ... click .... select page with "schematic" ... click ... import.  That's a total fantasy. It's been a fantasy for at least 4 decades, despite it being a "millions of dollars a year" sort of problem.

This is getting offtopic: One the biggest problem I see is that many drawings (especially for connectors and other mechanical-ish parts) are drawn by mechanical engineers and not electronic engineers. Sometimes it takes a lot of work to derive the pad locations from a drawing especially when the drawing uses 2 points of origin and no clear relation between those. But even the more sane drawings never use the centre of a component as a point of origin but that goes into the debate whether pin 1, the centre of the part or a mechanical mounting point should be the centre of origin. IMHO for some components I like the centre of the part but for components with a mechanical mounting point which needs to line up with a casing I rather have the mounting point.

[PCB.Wiz]

this looks quite neat:

Wow, yes, quite nifty, not seen that before.

I find a thread about it too
https://forum.kicad.info/t/this-is-what-the-footprint-editor-should-look-like/1499/18

Just had a quick play in Chrome, of the standalone download (Chrome only) a little rough, but the base idea is very clever and sound.
You can re-size pad(s) by ctrl-select to turn red, then the Size entry applies.
(but you cannot edge-select and change PAD Size that way)

Looks like you can delete any oops on dimension-lock by select-delete.

Nice little tool, for where everything is edge-defined.

[EEVblog]

Eagle to the defacto standard PCB package in the OSHW industry because it had a usable free version, and was popularised by many early OSHW advocates.
Autodesk would be well advised to release a statement affirming their commitment to the free version.

[djacobow]

This is getting offtopic: One the biggest problem I see is that many drawings (especially for connectors and other mechanical-ish parts) are drawn by mechanical engineers and not electronic engineers. Sometimes it takes a lot of work to derive the pad locations from a drawing especially when the drawing uses 2 points of origin and no clear relation between those. But even the more sane drawings never use the centre of a component as a point of origin but that goes into the debate whether pin 1, the centre of the part or a mechanical mounting point should be the centre of origin. IMHO for some components I like the centre of the part but for components with a mechanical mounting point which needs to line up with a casing I rather have the mounting point.

Yes, OT, but this, 1000x. I think it would be cool if manufacturers could put right into the datasheet a QR code or obviously machine-intended table of numbers that any cad tool could ingest to generate a pad layout. Yes, I understand that there are good reasons for wanting slightly different layouts for the same part, depending on the PCB and/or manufacturing process, but this would be a start and would represent the "recommended" layout we always see in the datasheets. It doesn't need to be put right into the data sheet of course, but that's not a bad place for it at all.

I mean, who designs anything without the data sheets for all the important parts in the design? How cool would it be to create a folder with all those .pdfs in it and your CAD tool would automagically have all the associated footprints ready for you?

[langwadt]

This is getting offtopic: One the biggest problem I see is that many drawings (especially for connectors and other mechanical-ish parts) are drawn by mechanical engineers and not electronic engineers. Sometimes it takes a lot of work to derive the pad locations from a drawing especially when the drawing uses 2 points of origin and no clear relation between those. But even the more sane drawings never use the centre of a component as a point of origin but that goes into the debate whether pin 1, the centre of the part or a mechanical mounting point should be the centre of origin. IMHO for some components I like the centre of the part but for components with a mechanical mounting point which needs to line up with a casing I rather have the mounting point.

Yes, OT, but this, 1000x. I think it would be cool if manufacturers could put right into the datasheet a QR code or obviously machine-intended table of numbers that any cad tool could ingest to generate a pad layout.

I've seen several manufacturer now pushing Ultra Librarian, it can generate symbol and footprint for a number of CAD tools from a single file per part

http://www.analog.com/en/design-center/packaging-quality-symbols-footprints/symbols-and-footprints.html
https://webench.ti.com/cad/
http://www.microchip.com/development-tools/resources/cad-cae-symbols

I had a quick try generating for Eagle and it was a bit sketchy, some were ok other were down right wrong with swapped pins

[EEVblog]

Interview with Autodesk:
https://blog.adafruit.com/2016/07/05/exclusive-interview-with-autodesk-about-the-cadsoft-eagle-purchase-autodesk-cadsofttech-technolomaniac

LOL at Matt's non-answer to this question:

Autodesk and a lot of other companies tend to focus on ‘subscription based’ and ‘cloud based’ software. Cadsoft EAGLE is firmly neither. Will you transition Cadsoft EAGLE to subscription or cloud based? What benefits to engineers would you see to either if you will?

At this time, we’re still planning to continue to develop the EAGLE technology, while making it more tightly integrated with other Autodesk products to enable a better workflow from design to manufacturing. In the short term, I’m particularly interested in taking a closer look at EAGLE’s core, such as hierarchy, modularity / reuse, routing, revision management and libraries. And of course, better mechanical integration!

Matt should have just been honest and said it's going cloud and subscription based.

[rx8pilot]

I have been thinking about what @mikeselectricstuff was writing. Maybe subscription is not as good for my business as having a perpetual license. I have access to all my old SolidWorks and MasterCAM data that is totally in my control and I don't need to pay or update unless I decide to. For now SolidWorks 2013 is just fine since all am doing is maintaining existing mechanical designs. I don't have to pay anything more or worry about an unwanted change in the system.

Adobe offers stand-alone and subscriptions and that allows me to have access to the full Adobe suite that I otherwise could not afford right now. It's an interesting concept to think about. I am mainly curious if Autodesk will deliver a major spanking to the software and how long that may take.

[PCB.Wiz]

Interview with Autodesk:
https://blog.adafruit.com/2016/07/05/exclusive-interview-with-autodesk-about-the-cadsoft-eagle-purchase-autodesk-cadsofttech-technolomaniac

Hmm, there seem to already be slight creepage on the limitations of the new 'free' version.

So yes, we will continue to make the freeware version of EAGLE available. We’ll also be zeroing out the educational license and making the 6-layer version available free to students and faculty for non-commercial work. We’ve done this in part to bring this into line with all Autodesk products (Fusion and Inventor for example are all free to students and startups making less that $100K / year) but also because the 6-layer license finally makes it possible for students to begin designing wireless (impedance controlled feed lines for example tend to be pegged against a plane layer) and likewise use more sophisticated memory busses (like DDR) which also require impedance controls.

 Seems to first say "students and faculty for non-commercial work" then says "startups making less that $100K / year", but that's for Fusion and Inventor.
 Students have always had quite open CAD tool access.

  If you are not a student, faculty or employed by a startup, are you now in the license wilderness ?

 Earlier comments on here, suggested the free version at 6 layers, was simply that.

 Publishing designs using Eagle, just got a lot more complicated, as you have no way of knowing in advance if your user base falls into those pigeonholes ?

[LabSpokane]

It's the end of the world.   

[nctnico]

Eagle to the defacto standard PCB package in the OSHW industry because it had a usable free version, and was popularised by many early OSHW advocates.

How many days will Eagle stay the defacto standard? The clunky user interface is not something Autodesk is going to fix. From all angles Kicad looks like a much better deal if you want to design truly open hardware (especially with CERN pushing it). For professional use you can get good deals from Altium and Cadence.

[EEVblog]

How many days will Eagle stay the defacto standard?

As long as major players like Adafruit, Sparkfun, Arduino et.al keep using it.

The clunky user interface is not something Autodesk is going to fix. From all angles Kicad looks like a much better deal if you want to design truly open hardware (especially with CERN pushing it).

It doesn't matter. People will use what they see other people using, and/or what they start out with. e.g. if you start out with an Arduino and design a shield etc, probably a 90% chance you'll use Eagle because everyone else is.
So it will be a long slow switch to Kicad.
Altium could have killed Eagle in fairly short order if they hadn't goofed up the positioning of Circuit Maker and played their cards right.

[kaz911]

I dropped AutoDesk completely after they sent their lawyers after me for BUYING a student license (not using it - not installing it - but just BUYING a student license on eBay - which never actually got installed)

They tried to force me to BUY a FULL license. But my lawyer put them in their place - but at a cost. eBay actually handed over contact details of the PURCHASERS to AutoDesk's laywers. I don't know what they did to the seller.  Apart from that I OWN(ed) a Fusion 360 full license - which I have now terminated.

So FU AutoDesk

[vzoole]

CERN pushing KiCad but they are using Altium and Eagle mostly .
Farnell pushing CircuitStudio but all the CS stand was empty in Embedded World while Altium stand was fully crowded. (And now I understand why Farnell didn't promoted Eagle.)

So I think Eagle will lives long because the 75% of designes made by Eagle and Altium (in Europe).

[technolomaniac]

Altium has been after me for $10k license - maybe I should be patient. I don't design a big volume of boards, but when I do - I am in a huge hurry and they get more complex every time.

Of course I'm a bit biased (being the guy building out the new feature set on EAGLE within Autodesk) but I would suggest waiting just a bit until you see the forthcoming new feature list.  It will be worth the wait.  Promise.   (Then make whatever decision you like...I'd just hate to see you overcharged for something you need only 10% of, at best.)

[technolomaniac]

Altium has been after me for $10k license - maybe I should be patient. I don't design a big volume of boards, but when I do - I am in a huge hurry and they get more complex every time.

IMHO a higher end package also saves a lot of time when it comes to the logistics part. This is often overlooked by the lower end offerings. The Orcad package I'm using produces a ready-to-go bill of materials (including manufacturers parts numbers, order codes, etc, etc) with one click. AFAIK this is impossible with Eagle and Kicad.

This is in EAGLE's future for sure and was something that having been tied to a distributor made difficult.  Relatively quickly we will be adding much broader reaching BOM capabilities with a reach into downstream processes as well. 

[technolomaniac]

Eagle supports attributes for devices. This way you can store all kind of information.
In the *.sch or *.brd files it wil look like this:

Code: [Select]

<attribute name="C-TYPE" value="E"/>
<attribute name="DIGIKEY" value="NC7SZ125M5XDKR-ND"/>
<attribute name="FARNELL" value="2453005RL "/>
<attribute name="MANUFACTURER" value="Fairchild Semiconductor"/>
<attribute name="MAN_PN" value="NC7SZ125M5X"/>
<attribute name="MOUNTING" value="SMD"/>
<attribute name="MOUSER" value="512-NC7SZ125M5X"/>
<attribute name="PACKAGE" value="SOT23-5"/>
<attribute name="PRICE" value="0.284"/>
<attribute name="RS-COMPONENTS" value="670-9807P"/>

When you export the BOM in Eagle,  there will be a colon for every attribute, e.g. "Farnell" with in that colon the ordercodes of
all parts available at Farnell.
It can store it as txt, html or csv. This way you can easily import it in Calc (or excell).

Yep, but that's not the best approach in my situation.  A BOM generator is useless without a parts database.  An unstructured key-value store isn't a real database, and in any event, a parts database doesn't belong in a particular CAD program's schematic, board, or library file.   EAGLE isn't a database manager (not that my text editor is  .) 

The various BOM generation options for EAGLE also aren't very flexible.  That includes third-party contributions, some of which are fairly decent -- Xess's comes to mind -- but still hardwired to someone else's way of managing inventory.

All of which isn't to criticize EAGLE, or to argue that Autodesk should add a database module to it, but to point out how it's actually a pretty great EDA tool for people with specific needs but limited budgets.

Duly noted.    (I'm a part of the dev group for Autodesk / EAGLE)  Having some means to store Key:Value data in a more substantial structure is something we've got plans to do.  What I would avoid at all costs is anything that wraps up user data in an unknown or unfriendly data model.  Something like a "Vault" just sounds ominous.  We'd prefer to favor open source, shareable content and by definition this will mean that we have to support a database model / data structure in which the user has unrestricted access to all of their content.  No use attempting to swim upstream.

[technolomaniac]

Interesting times.     Circuit Studio just got a price cut to $995.  CircuitMaker is now running under Linux via wine so CircuitStudio should work as well.

http://circuitmaker.com/blog/Blogs/3-steps-for-installing-circuitmaker-on-linux

Autodesk is really going to have to step up or EAGLE will die on the vine at its current price.

Wine is a poor substitute for native support, in my experience.  And unless the wine method is officially supported, it isn't viable for any kind of professional use.

As bad as many people think Eagle is, they still have multi- native OS going for them.

EAGLE is completely developed under Linux with native development for Windows and Mac as well.  This means we are not emulating anything to get it working in these environments.  As we move forward, we'll only strengthen the multi-platform support because as you point out, this is a unique ability of EAGLE.  Sprinkle in some new features like better routing, better DRC, polygon handling, etc. and you have the makings of something super special!  At least that's what we're betting on.  Stay tuned!

[technolomaniac]

You should be able to take the numbers directly off a data sheet an plug them in in less than a minute. Very simple, yet Eagle fails big time in this area.

Frankly I blame the draughtsmen who create those data sheet drawings for this.  Every time I need to create a new package, I always seem to have to reach for a calculator to derive the most fundamental dimensions like pad sizes and centerlines.  The dimensions I need to plug into EAGLE often seem to be the only ones that aren't given explicitly.

Duly noted.  Seems it would make more sense to just treat this with a calculator of some sorts.  Like what the IPC are doing.  We'll add this to the feature list and have a crack at this down the road.  (Speaking as an EAGLE developer)

[technolomaniac]

I did this super fast demo that shows an example of how SolidWorks goes about 2D geometry. Not only do you tell it dimensions, you can also tell it about relationships - vertical, horizontal, parallel, coincident, perpendicular, etc, etc.

That's definitely a powerful tool, no question about it, but I wouldn't call it super fast.  You still had to do an absurd amount of clicking and dragging to enter an utterly trivial amount of information: X,Y centers for five pads with two unique W,H dimensions.  If the drawing had simply stated that information directly, entering it into EAGLE or any other package would take 10 seconds.

I agree and as the guy from Autodesk / EAGLE on the board, the thing I'd say is we should look to approach this from multiple directions.  1)  We need to handle input data better.  Ok, fair enough.  The standards for this including IPC, JEDEC, etc - along with what the mfg's have been producing - however, mean there's just SO little consistency in how this data shared.  This I think is the elephant in the room.  This wreaks havoc on anyone building parts.  Grids and reference points and such are all good, but let's call the input data what it is...messy!  (some mfg's being MUCH better than others of course)

Bottom line, we need to flex a bit from the tools side to meet the incoming data in the middle or we are attempting to swim upstream against 40 years of information that's all over the map. 

2)  I think for data that can be mapped into a standard, we need to look at how we ensure users can create good, professionally- / consistently- solderable parts.  This includes looking closely at what IPC have been doing and starting to view this more as a calculator-type problem, like the guys at PCB Libraries have been.  They have awesome tools for building parts and I guess there's just heaps we can learn from the approach IPC and PCBL has taken.  We of course also have some wicked tools for mechanical.  So the intersection of this data is really one area where we as a company can hit some home runs (or kick some goals or score some tries or whatever other metaphor you prefer

[technolomaniac]

Eagle to the defacto standard PCB package in the OSHW industry because it had a usable free version, and was popularised by many early OSHW advocates.

How many days will Eagle stay the defacto standard? The clunky user interface is not something Autodesk is going to fix. From all angles Kicad looks like a much better deal if you want to design truly open hardware (especially with CERN pushing it). For professional use you can get good deals from Altium and Cadence.

As the guy at Autodesk / EAGLE, do us a favor and let us decide what we will / won't do.  "The clunky user interface is not something Autodesk is going to fix." is just fiction.  So stop spreading an agenda please and let's stay constructive.  Best to tell us what you'd like to see and we'll slate it into the development pipeline. 

[technolomaniac]

Being able to grab edges and points and define a distance would be a massive improvement. I feel like I am in 1987 all over again. Everything in Eagle is defined from the center of the object (pad) and referenced only to the grid zero.

Awesome suggestion, though I'd love to riff with you on the specific features that would make this legendary.  Before taking on EAGLE development at Autodesk I had used the EAGLE software for years and I have a few of these also that I think would be big productivity gains.  Feel free to hit me up here or on hackaday.io as @technolomaniac or twitter at the same. 

[technolomaniac]

Interview with Autodesk:
https://blog.adafruit.com/2016/07/05/exclusive-interview-with-autodesk-about-the-cadsoft-eagle-purchase-autodesk-cadsofttech-technolomaniac

LOL at Matt's non-answer to this question:

Autodesk and a lot of other companies tend to focus on ‘subscription based’ and ‘cloud based’ software. Cadsoft EAGLE is firmly neither. Will you transition Cadsoft EAGLE to subscription or cloud based? What benefits to engineers would you see to either if you will?

At this time, we’re still planning to continue to develop the EAGLE technology, while making it more tightly integrated with other Autodesk products to enable a better workflow from design to manufacturing. In the short term, I’m particularly interested in taking a closer look at EAGLE’s core, such as hierarchy, modularity / reuse, routing, revision management and libraries. And of course, better mechanical integration!

Matt should have just been honest and said it's going cloud and subscription based.

@Dave, it's not going subscription.  So there.    At this stage, that isn't anywhere on my roadmap.  Thought about it.  Decided against it.  Can I say that we will never in the life of any product do that?  No, of course not.  That would be at best unfair, at worst dishonest.  But I have so many things that are more pressing.  The point of my response - which I agree was unclear was - routing, real-time DRC, some improvements to polygon handling, better revision management and versioning, better BOM tools, better interface to manufacturing, some library improvements, interface to 3D, etc are all good things to worry about today as they drive value for the users.  Those are the priority.  We'll shelve the other stuff until get to a place where that makes sense.  That was the point of that comment.  I've got other stuff on my radar.  And I think that the shortlist today is pretty much a who's-who of what folks have been asking for for some time.  Only now we have a combined development team that can really drive some of this home.  Thanks for calling me out...I sounded like a politician and it was totally fair. 

[jpanhalt]

@technolomaniac

I have been an Eagle user since 3.x, got an academic/educational license at 4.03.  I certainly appreciate and agree with you that Autodesk will do what it sees fit to do, and doesn't need to be told.

FWIW, I disagree that the current interface is "clunky."  In fact, I rather enjoy having control rather than having some piece of software telling me what it wants to do, which gets back to my second sentence above.  I also enjoy standard shift sports cars.  Please don't fix it to be like SolidWorks.

The real point of this post is to ask that another class of customer be considered -- maybe something like a loyalty discount for long-time users or retired users.  The last time I looked, Eagle had a package for about $169 that offered 6 layers, a reasonable size board, and was only for non-commercial use.   Since the education package is gone, I hope you can keep a similarly priced (i.e, <$200), very functional package available.

Regards, John