Autodesk buys Eagle

[Karel]

Personally I prefer IDF but probably other formats will do as well. What is important to us, is that we don't need to deal with 3D-models in Eagle.
What we do is drawing a contour in layer 57 (tCAD) in the footprint. The linewidth is used to indicate the height of the contour above the board (linewidth is height / 1000).

That's a very clever idea - is that a widespread convention ? Google suggests that is somewhat Eagle only ?

Read this: ftp://ftp.cadsoft.de/eagle/userfiles/ulp/generate_3d_data_eng.pdf

and this: https://www.element14.com/community/thread/43949/l/idf-export-broken?displayFullThread=true

[mikeselectricstuff]

Definitely agree that an industry-standard format for part footprints is long overdue - isn't this the sort of thing that JEDEC should be doing?

Ease of part creation is one of the most important features of a PCB editor as there will ALWAYS be new parts that you need to create yourself.

I'm not sure DXF is a sensible answer, as it is a pretty poor standard, with many version and compatibility issues.
 

[PCB.Wiz]

Definitely agree that an industry-standard format for part footprints is long overdue - isn't this the sort of thing that JEDEC should be doing?

Ease of part creation is one of the most important features of a PCB editor as there will ALWAYS be new parts that you need to create yourself.

I'm not sure DXF is a sensible answer, as it is a pretty poor standard, with many version and compatibility issues.

DXF is already established, and it is not so much a poor standard, as no rules around using it for more intelligent PCB footprint extraction.
It works fine already, for outlines, and PAD centres and diameters.

Which is why I suggested tiers of DXF - the simplest DXF, are the most portable, and have similar IQ as a HPGL or Gerber file.

Next step is to have polylines, Arcs, width and layers meaning something.

Again, simple rules with examples help here - and a lite-2D package with Footprint helpers could help vendors create rule-meeting DXFs.

Then the top level uses blocks, and that can insert full drill/mask stacks at pin locations, with enough inbuilt rules a compatible PCB tool can load directly as footprint.

To get this from Vendors, I think Autodesk needs to supply a footprint-2D package, where they can view the stackup in quasi 3D.

I have seen complete PCB's exported as DXF, so you can embed enough information to manage complex structures.

[LabSpokane]

Definitely agree that an industry-standard format for part footprints is long overdue - isn't this the sort of thing that JEDEC should be doing?

The fact that there isn't a standard for footprints by now amazes me. Clearly, part footprints must occasionally vary for process reasons. It would be nice to have a library where the standard footprint would be the base to which footprint variants could linked. The user could view the variants versus the base using the visual comparison that rx8 suggested earlier.

[EEVblog]

Definitely agree that an industry-standard format for part footprints is long overdue - isn't this the sort of thing that JEDEC should be doing?

[uncle_bob]

.....  The notion of locking up people's libraries and design data into their format and their systems is just really outdated and it's a shame really.  I recall when Altium was brave enough to add OrCAD export and folks feared everyone would flee to another tool.  ......

Hi

The gotcha is that library lock *is* the way it works for every PCB program I have ever seen. Go over to the layout guy(s) and simply imply you might do *anything* with their pet libraries and watch the reaction. I have worked multiple places where each person doing layout had *their* libraries and refused to touch the ones done by anybody else. You might say I simply have a bunch of paranoid co-workers over the years. If you sat down with them ... not so much. The issues were very PCB specific.

Bob

Hi Bob -- I totally agree that people are (and should be) careful about managing changes to libraries...Especially when something's gone to production!!!  So I'm with your PCB team on this one!  But what boggles the mind about Circuitmaker (and this was really the point of that comment) is that the libraries are published to a public library system in which your components are then a part of the larger parts 'ecosystem' and in a format that isn't easy to get at.  I am all for sharing libraries but:

Yes.
Forcing publicly libraries is fundamentally and demonstrably a stupid idea.
Public libraries are great for seraching and getting parts, essential in fact.
But then once you have the parts and used them, you want to lock them away locally so no one can ever touch them except under expert adult supervision.

Hi

I would not *force* a public library in any situation. That's not how it's done in any other area (except Windows Update). In any rational setting, you have the ability to use standard libraries *or* overwrite them. You also have the ability to archive the whole project and/or it's libraries. There are working models for all this stuff. They do indeed have their issues. They are decades ahead of what we have now in terms of productivity *and* quality. Rewriting library entries over again 8,000 (or 16,000) times at 8,000 locations has its risks....I have never worked anyplace that got the make an entry stuff right *every* time.

Bob

[nctnico]

Yes, if you have a text based file format then integrating version control in the software makes little sense. It's just trying to feature-match Altium.
With Matt being a long time ex-Altium guy I think he needs to be careful here to not take the same do-everything mindset into Eagle. It doesn't need that, it needs to simply be a better and more usable basic PCB tool.

IMHO by the time Eagle can do the same as Altium it will cost the same as Altium, just be as clunky to use (I have used Altium for a few projects and I found the UI horrible) and Eagle probably got rewritten from scratch.

[uncle_bob]

Yes, if you have a text based file format then integrating version control in the software makes little sense. It's just trying to feature-match Altium.
With Matt being a long time ex-Altium guy I think he needs to be careful here to not take the same do-everything mindset into Eagle. It doesn't need that, it needs to simply be a better and more usable basic PCB tool.

IMHO by the time Eagle can do the same as Altium it will cost the same as Altium, just be as clunky to use (I have used Altium for a few projects and I found the UI horrible) and Eagle probably got rewritten from scratch.

Hi

I have used a number of programs over the years. At work these days we are part of the Mentor Graphics empire (for a lot of bad reasons). None of these programs are perfect. They *all* could use a good through scrub of their UI and "legacy features". Doing that takes a very single minded approach to the task. A startup can have that kind of focus. A big company often looses the ability to do so. An acquisition is one of those *rare* occasions that a large company might have the ability to come up with the needed focus. We can only hope ....

Bob

[station240]

@technolomaniac

Can you do something with the size limitations on the various licences, they are a bit silly when compared to what the PCB houses offer.
eg the free version is limited to 10x8cm, PCB house is 5x5 multiples, eg 10x10cm

Can we please have the ability to copy/clone entire parts inside/between libraries ?
The current system is annoying when the only difference between the part in the Library you have, and the part you need is the part number, or the labels on the pins (or a better pad outline).
I had to revert to manually editing the existing library down to that one part, and messing around.

[technolomaniac]

We use git for revision control (and backup) and it works perfect with Eagle's XML based files.
We have absolutely no need for any integrated revision control.

Yes, if you have a text based file format then integrating version control in the software makes little sense. It's just trying to feature-match Altium.
With Matt being a long time ex-Altium guy I think he needs to be careful here to not take the same do-everything mindset into Eagle. It doesn't need that, it needs to simply be a better and more usable basic PCB tool.

I agree.  Having lived it, the risk is always getting caught in the trap of ticking boxes because someone brings a spreadsheet of nice-to-have's to a sales call, or instead building an 'enterprise tool' with every bell and whistle you can imagine, all in hopes of landing a few whale-sized customers.  Though it sounds alluring (certainly if you ask your sales team) this ignores what's of more value to the wider audience of *engineers*...simple, easy to use, and not compromising where it counts.  Elegant simplicity.    I wont underestimate the undertaking but I think it's achievable or I wouldn't do it.  Thankfully I actually *use* this stuff.  And we're not afraid to ask the community when it's a choice of option A or option B and both are good options.

[LabSpokane]

We use git for revision control (and backup) and it works perfect with Eagle's XML based files.
We have absolutely no need for any integrated revision control.

Yes, if you have a text based file format then integrating version control in the software makes little sense. It's just trying to feature-match Altium.
With Matt being a long time ex-Altium guy I think he needs to be careful here to not take the same do-everything mindset into Eagle. It doesn't need that, it needs to simply be a better and more usable basic PCB tool.

I agree.  Having lived it, the risk is always getting caught in the trap of ticking boxes because someone brings a spreadsheet of nice-to-have's to a sales call, or instead building an 'enterprise tool' with every bell and whistle you can imagine, all in hopes of landing a few whale-sized customers.  Though it sounds alluring (certainly if you ask your sales team) this ignores what's of more value to the wider audience of *engineers*...simple, easy to use, and not compromising where it counts.  Elegant simplicity.    I wont underestimate the undertaking but I think it's achievable or I wouldn't do it.  Thankfully I actually *use* this stuff.  And we're not afraid to ask the community when it's a choice of option A or option B and both are good options.

I'll 2nd Dave's request to not try and be all things to all people.  The Yugo and Mercedes ends of the market are covered well.  We need a Ford. 

And if you brought some AutoCad-y ability to specify coordinates, offsets, etc, that would be incredibly good.  The snap-grid-only drawing of eCad is horribly inefficient IMO.  Altuim - at least CircuitStudio - seems to have gotten that horribly wrong. 

[technolomaniac]

What made EAGLE so attractive [to Autodesk] was the spirit of open source HW and the community around it.  I just happen to be one part of that community and when we were looking to make a move in electronics - having come from Supplyframe / Hackaday - I was super excited to see EAGLE as a prospective target.  It's built on a great foundation and it already has a completely open file format (something we intend to keep moving forward).

For sure, we're a big company and could make a run at some other ecad tools but the goal from the outset was to tap into what's different these days than say, 10 years ago or more.  So expect we'll push to make things even more open and not try to lock folks in in a way that means they lose their right to their content.

Sounds great.

One thing that Eagle could do, to both prove those credentials, and get early access to a nice Shove router, is to add a Import from KiCad button, (and test the KiCad Eagle importer that is already in place, under File.Import).
Maybe add a command line launch & button that says 'Open in KiCad'

To the few boards I've tried, the KiCad Eagle importer already looks pretty good, (as does their Altium importer), but if you want to round-trip this into a Router, that needs another level of  detail checking.

Definitely import and wider file conversion in general is in the cards (routing is also very high on 'the list' but will depend on real-time DRC so there is a sequencing thing that has to happen...it's happening now).  Our feeling is it's essential for any good tool to leverage content from multiple formats.  After all, if TI publishes a switcher design in an OrCAD format, I want to be able to use it.  If nothing else, let me import it and scrape the libraries for use in my own designs.  This is the shortest path to victory with any design (ie build on what's already there).  So making the file import for KiCAD only helps us. 

Likewise other formats.  For example, I'd love to chase down libraries in multiple formats as well.  We'll get there for sure, just time.  Even if you have to edit the parts, it's much easier to start with a component that converts 90% of the way there than it is to enter pin names for a Spartan 6 or an iMX or an M4 manually, one at a time.  This is what truly sucks about closed-source file formats.  The guys using the MFC stream writer to create binary files because they're too skittish about an open and transparent format. 

Thankfully in the mechanical world we have formats like STEP or dare I say, IGES (nothing wrong in principle, just huge .  Shame really we don't have the equivalent for ecad parts.  My hat's off to Cadsoft for making it easy to read the files.  That was a big move and a big reason we (Autodesk) really considered them.  It's on the surface so benign to change your file format but to make it open has huge significance all around. 

[technolomaniac]

@technolomaniac

Can you do something with the size limitations on the various licences, they are a bit silly when compared to what the PCB houses offer.
eg the free version is limited to 10x8cm, PCB house is 5x5 multiples, eg 10x10cm

Can we please have the ability to copy/clone entire parts inside/between libraries ?
The current system is annoying when the only difference between the part in the Library you have, and the part you need is the part number, or the labels on the pins (or a better pad outline).
I had to revert to manually editing the existing library down to that one part, and messing around.

Noted.    Related to copying parts, just open the library, then from the Control Panel, select the same library from the list, right click the part you want and select "Copy to Library".  it copies to selected component from the Control Panel to the open library. If its a duplicate name, it will  prompt you to rename it.  That'll give you the same part with a different Device Name and you can tweak the MPN attribute or whatever part number attribute you're using from there.  Checkout http://electronics.stackexchange.com/questions/16030/new-eagle-library-reuse-standard-package-symbol for more detail.  Hope it helps! 

[ehughes]

Yes.
Forcing publicly libraries is fundamentally and demonstrably a stupid idea.
Public libraries are great for seraching and getting parts, essential in fact.
But then once you have the parts and used them, you want to lock them away locally so no one can ever touch them except under expert adult supervision.

In CircuitMaker,  all parts are revision controlled.  You never lose yours even if some does a revision.   Your revision has your name on it and just because there is a different revision, it does not affect your design (unless you want it to).   

Master libraries are very nice when you have to manage many designs.    Being able to see where a part is used, etc is very nice.

Also,    having your design instantly ready to get prices for all the common distributors is very nice. 

My experience is that the PCB design is about 10% of a project.    There is a ton of work that happens post design and tools like EAGLE totally ignores the management aspects of the design.

I used Eagle for several years (4.12).    It took them ***YEARS** to simple add the ability to have user definable parameters.     Then it took another 5 to get basic dimensioning tools.       The excuse was always that you could write your own ULP....   That sounds nice except that my time is valuable. 

I hope that Autodesk can bring EAGLE up to date.   I do like the Fusion 360 tool.  Most of the idioms are simply ridiculous and it has minimal design rule checking.      The problem I see is that the amount of resources they will have to dump into it to get it usable will drive the price north.    It already costs $1145 for the full version.    Where will the pricing go once all the features are added to recoup the cost? CircuitStudio does more out of the box for less money....      The best thing I could see happening is a pricing war.

ECAD is a tiny market compared to MCAD.     The business side of things is going to be tough.     If they want to pull in people in the pro market for another tool,   It needs to improve an order of magnitude.       I will be curious to see the pricing model.

[Monkeh]

Yes.
Forcing publicly libraries is fundamentally and demonstrably a stupid idea.
Public libraries are great for seraching and getting parts, essential in fact.
But then once you have the parts and used them, you want to lock them away locally so no one can ever touch them except under expert adult supervision.

In CircuitMaker,  all parts are revision controlled.  You never lose yours even if some does a revision.   Your revision has your name on it and just because there is a different revision, it does not affect your design (unless you want it to).

But there's no point in it, because you can never trust the parts. In my experiments with CM, I tried a couple dozen parts people had drawn up, from simple resistors to rather more complex ICs. None were worth using, many would never have worked.

[uncle_bob]

Yes.
Forcing publicly libraries is fundamentally and demonstrably a stupid idea.
Public libraries are great for seraching and getting parts, essential in fact.
But then once you have the parts and used them, you want to lock them away locally so no one can ever touch them except under expert adult supervision.

In CircuitMaker,  all parts are revision controlled.  You never lose yours even if some does a revision.   Your revision has your name on it and just because there is a different revision, it does not affect your design (unless you want it to).   

Master libraries are very nice when you have to manage many designs.    Being able to see where a part is used, etc is very nice.

Also,    having your design instantly ready to get prices for all the common distributors is very nice. 

My experience is that the PCB design is about 10% of a project.    There is a ton of work that happens post design and tools like EAGLE totally ignores the management aspects of the design.

I used Eagle for several years (4.12).    It took them ***YEARS** to simple add the ability to have user definable parameters.     Then it took another 5 to get basic dimensioning tools.       The excuse was always that you could write your own ULP....   That sounds nice except that my time is valuable. 

I hope that Autodesk can bring EAGLE up to date.   I do like the Fusion 360 tool.  Most of the idioms are simply ridiculous and it has minimal design rule checking.      The problem I see is that the amount of resources they will have to dump into it to get it usable will drive the price north.    It already costs $1145 for the full version.    Where will the pricing go once all the features are added to recoup the cost? CircuitStudio does more out of the box for less money....      The best thing I could see happening is a pricing war.

ECAD is a tiny market compared to MCAD.     The business side of things is going to be tough.     If they want to pull in people in the pro market for another tool,   It needs to improve an order of magnitude.       I will be curious to see the pricing model.

Hi

ECAD up to this point has mostly been a "big whale" business. You go out after a few hundred outfits per country and get your "share" of the business. Each of the outfits buys a few dozen licenses for the big stuff (on average). There are 10 or less who buy a few hundred licenses for the big stuff (layout) and a few thousand licenses for the little stuff (schematic etc). Multiply it all up and you get some number X. Since these guys don't mover around a lot, the real number is the support on that share.

If you *don't* already have a player in that contest, wiping out that business (protecting the base !!) is no big deal to you. Sell to a few 10's of thousand nerds named Bob in the same country. Your share of that business is likely 80% if you do it right. The pile of cash you take home at the end of the day could easily be a multiple of the X above. For every person doing it with a pro license, I'd bet there are >> 10 people hanging around wishing they could do it as well.

Bob

[PCB.Wiz]

  For every person doing it with a pro license, I'd bet there are >> 10 people hanging around wishing they could do it as well.

  Yes, the old market is changing quite rapidly, and some will become EDA Nokia's.
When there was little choice, users had to cough up, and suck it up.

 Now more choice is out there, what sane user is going to select closed/captive/hostage situations ?

 Even some of the 'high end' features are evolving.
I still see images selling meander lines to match delays, but the newest MPU/DRAMs have inbuilt delay tuning ability, so that tech is not needed, and much smaller PCBs can result.

More module-level products are coming out, in RF and RaspPi areas, which drops the complexity per seat of the end users.

 As for possible seat numbers, look at info like this - these are annual numbers !
United States graduates roughly 70,000 undergraduate engineers annually, whereas China graduates 600,000 and India 350,000.

Now very, very, few of those will be actually employed as full time PCB designers, but a shipload more will be interested in designing PCBs

[Warhawk]

Yes.
Forcing publicly libraries is fundamentally and demonstrably a stupid idea.
Public libraries are great for seraching and getting parts, essential in fact.
But then once you have the parts and used them, you want to lock them away locally so no one can ever touch them except under expert adult supervision.

In CircuitMaker,  all parts are revision controlled.  You never lose yours even if some does a revision.   Your revision has your name on it and just because there is a different revision, it does not affect your design (unless you want it to).

But there's no point in it, because you can never trust the parts. In my experiments with CM, I tried a couple dozen parts people had drawn up, from simple resistors to rather more complex ICs. None were worth using, many would never have worked.

I have the same experience. I am not sure how it is now, but when I tried CM last time, there was not any revision changelog. You had no idea, if the newer revision fixes a bug or why it was created. For instance I took FT232 in SSOP package where somebody created a new revision with QFN package instead of creating a new component.
That's why I gave up on CM.

[ehughes]

But there's no point in it, because you can never trust the parts. In my experiments with CM, I tried a couple dozen parts people had drawn up, from simple resistors to rather more complex ICs. None were worth using, many would never have worked.

Why not make your own if you are unhappy.    The "Build your own" link in the library window creates a part that is only visible to you.

[rx8pilot]

Today, I was trying to panelize some designs in Eagle. No way.

I had to design the panel outside of Eagle and use various third party tools to get a proper panel done with v-score, partial routing, tabs, mouse bites, tooling holes etc. Like everything else in Eagle, it can be done but only after spending far too long with a manual calculator and copy/paste coordinates and whatever ULP's are laying around.

Maybe the future of Eagle will have geometry tools that are appropriate for designing a panel layout. 

[twistedresistor]

@technolomaniac

What I'd like to weigh in:

Give Richard Hammerl and Jorge Garcia a promotion or at least the ability to further on spend time on the eagle newsgroups.
Those two guys are saints in the support newsgroup, I can't even imagine the patience they have. Can't beat the support there. (compared to the Altium Forum or their support)

[KE5FX]

@technolomaniac

What I'd like to weigh in:

Give Richard Hammerl and Jorge Garcia a promotion or at least the ability to further on spend time on the eagle newsgroups.
Those two guys are saints in the support newsgroup, I can't even imagine the patience they have. Can't beat the support there. (compared to the Altium Forum or their support)

+1.  Those guys have been very helpful for years, especially Richard (Jorge is a newer hire.)

[Karel]

@technolomaniac

What I'd like to weigh in:

Give Richard Hammerl and Jorge Garcia a promotion or at least the ability to further on spend time on the eagle newsgroups.
Those two guys are saints in the support newsgroup, I can't even imagine the patience they have. Can't beat the support there. (compared to the Altium Forum or their support)

+1.  Those guys have been very helpful for years, especially Richard (Jorge is a newer hire.)

+1.

[nctnico]

Today, I was trying to panelize some designs in Eagle. No way.

I had to design the panel outside of Eagle and use various third party tools to get a proper panel done with v-score, partial routing, tabs, mouse bites, tooling holes etc. Like everything else in Eagle, it can be done but only after spending far too long with a manual calculator and copy/paste coordinates and whatever ULP's are laying around.

Maybe the future of Eagle will have geometry tools that are appropriate for designing a panel layout. 

I'm not sure whether other CAD tools have panelisation tools. When I need a panel I let the PCB manufacturer sort that out. The one I use (Eurocircuits) has a very elaborate web based tool to do that.

[uncle_bob]

Today, I was trying to panelize some designs in Eagle. No way.

I had to design the panel outside of Eagle and use various third party tools to get a proper panel done with v-score, partial routing, tabs, mouse bites, tooling holes etc. Like everything else in Eagle, it can be done but only after spending far too long with a manual calculator and copy/paste coordinates and whatever ULP's are laying around.

Maybe the future of Eagle will have geometry tools that are appropriate for designing a panel layout. 

I'm not sure whether other CAD tools have panelisation tools. When I need a panel I let the PCB manufacturer sort that out. The one I use (Eurocircuits) has a very elaborate web based tool to do that.

Hi

If you let them sort it out ... it works fast and easy on their gear. If you try to do it, it takes two to four times as long for them to hog out. Guess which one costs you more in the end?

Bob