Schematic/PCB Design for Linux?

VanessaDannenberg asks: "Occasionally, I have been known to design the occasional circuit board. I've been using Eagle, but with the board size limit of 3x4 inches in the free version, and a $400 price tag to exceed this limit, it is time to consider a Free Open Source Software alternative. Not being a Linux programmer myself, I have checked into and ruled out gEDA, KiCAD, Electric, XCircuit, and a host of others as being too incomplete to replace Eagle. My requirements are pretty basic: Draw a schematic, make a board out of it, edit and autoroute it, export to Gerber, and do it all natively within Linux. So, with this in mind, what suggestions do you folks have?"

Comments

Good luck!

[harrkev]

I do not know of anything like what you ask. AFAIK, gEAD is the best thing out there.

Unfortunately, this is a bit of a niche application.

KiCAD looks good

[Anonymous Coward]

KiCAD Check it out.

http://www.lis.inpg.fr/realise_au_lis/kicad/

Re:KiCAD looks good

[Kirmeo]

no damn it, the poster is a linux guy mod it down because its windows shite.

From the install.txt:

Kicad binaries exists for Linux and Windows (XP, 2000) Files (schematic, boards, libraries) created on both systems are totally compatibles.

Re:$400? Get real

[Intron]

Note that he's also asking for a free autorouter! While you're at it, why not ask for a free 21" color monitor? For a library, router, gerber out and everything on your list, $400 is peanuts. Its paid off on one board design.

The free way to do this is get a resist pen and blank copper-clad and just draw your circuit right on it.

This is what pisses me off...

[RingDev]

about the whole open source movement.

"$400 price tag to exceed this limit, it is time to consider a Free Open Source Software alternative."

So in other words, you're not willing to pay the programmers who support their families for this product, and you are not willing to donate anything to an Open Source project.

You sir, are a leech. You want a product for free, not because of a moral issue, a desire for community support, accessible developers, or any other OS reason. No, you want an OS product because your greedy little heart wants something for free.

-Rick

A little over the top? maybe, but I've had a crappy week. I'm going to go home, get drunk, and forget the last 4 Mondays.

Re:This is what pisses me off...

[Hosiah]

A little over the top? maybe, but I've had a crappy week. I'm going to go home, get drunk, and forget the last 4 Mondays.

God, I hope you took double shots! Uh, lissen, we can't keep it free(freedom) without keeping it free(price). And we can't give it away and then cuss people out for accepting it.

We couldn't possibly fit all the users of every product into it's developer base, anyway. You'd spend 90% of the release cycle answering emails.

Re:This is what pisses me off...

[RingDev]

"God, I hope you took double shots!"

Workin on it.

"We couldn't possibly fit all the users of every product into it's developer base, anyway. You'd spend 90% of the release cycle answering emails."

Imagine having 90% of your users donate: Code, Money, Hosting, and/or Bug Reports. Wouldn't that be a dream!

-Rick

Software isn't the only open entity

[Anonymous Coward]

Don't be so pessimistic. Some people give back to the community in forms other than cash and software. Maybe if he designs something useful, he'll share it with the world. Commercial tools present a high barrier for entry to the hobbyist, which discourages open source hardware.

Re:Software isn't the only open entity

[VanessaDannenberg]

Don't be so pessimistic. Some people give back to the community in forms other than cash and software. Maybe if [she] designs something useful, [she'll] share it with the world. Commercial tools present a high barrier for entry to the hobbyist, which discourages open source hardware.

And this is precisely why I asked about an open source replacement. It's one thing to pay for a product if you're going to use it to make money. It's a far different thing to expect to pay the same money for a product only to use it for hobby work.

I design gadgets for the Commodore 64/128, and a quick estimate shows that at the prices I've seen around the web in the last few days, I would spend more on the software alone than I would spend on making one production-ready unit of every board I've designed since I started fiddling with this stuff (that's only about 10 unique designs), and if I tried to sell, there's no way I'd ever break even. Several years ago there was a slim chance, but today, forget about it. Today, all of the stuff I write or design is free and open source, and stuff I have written in the past I have since declared free also (where the source code still exists). I think that fits anyone's definition of "giving back." I must stress - I do not program for Linux, just Commodore.

Oh, and to the other gentlemen who have mentioned auto-routing and other high-end features as being too much to ask of FOSS, let me see..

• PCB, the very PCB editor I started with years ago, is a nice board editor with autorouter (which I have yet to use) and some other nice features, but that's only half of the needed setup.
• KiCAD has a decent schematic editor, 3D viewer, and some other stuff, but it just has problems on my box (apparently poor integration, very slow board editor, crashy).
• gEDA attempts to integrate PCB with schematic capture and other tools, but it's buggy on my setup (missing config files in Ubuntu, schematic doesn't get translated over to the board properly, no component-onto-board auto-place function to go with the schematic capture)
• gschem2xpcb looks like it would fill in well to convert those gEDA/gSchem schematics over to PCB in a way similar to Eagle's autoplace feature, but this is just a stand-alone command-line program with only the one function, and the author seems have a major aversion to the GPL. *shrug*
• The GIMP of course has tons of features and a really nice UI, and in particular it has vector graphics capabilities and multiple layers, but of course it's not adapted for PCB/schematic work.
• Eagle, for this particular list, has wonderful parts libraries (for which utilities exist to convert these to other formats) and good integration between schematic and board, but it has some serious screen refresh bugs, plus the aforementioned 4x3 inch board size limit.

Along with these, every other open source program I've looked into has at least one of the features I need. I was just hoping for a program that combines all of these already-existing, already written features into one Eagle-killing FOSS program.

Hmmm... sounds like your priorities are a bit off

[rco3]

I'd try to feel sorry for you, but having just spent $775 to get the full-on Professional version of Eagle (with Linux and OS X licenses), I just can't muster any pity. I earn money by using Eagle. You can get a 100mm x 160mm, 4 layer version of Eagle for $125, as long as you aren't making a profit using it. If you are making a profit with it, either you can afford to upgrade to a for-profit version, or else you aren't charging enough for your work. You can also upgrade at any time with full credit for previous versions.

Now, I'm sure that all the gEDA people will tell you that you can help make their project do all you need it to... but I'm not really a programmer, nor do I have the time to become one - I'm busy earning money to feed my baby. I've contributed (a very small amount of) code to the kernel, I've contributed financially to open-source projects... but there isn't always a viable open-source solution to your software needs. That's when you need to pay someone for software that already works.

You want to design boards using Linux, you probably need to be using Eagle. Sorry. Consider either a) using the non-profit version or b) getting the for-profit version but not the autorouter - Eagle's is very good indeed (FAR better than Protel's, IMHO), but you'll almost always get better results hand-routing anyway.

Frankly, even at $400 Eagle is a bargain.

Time to evaluate what your time is worth

[stevesliva]

When I find myself using some crappy free version of software that I know I can just buy, I often just wish I'd spent the money and saved the time.

If you're designing PCBs, $400 should be chump change-- right? I'm used to EDA packages that cost well over six figures per seat.

It's probably not for a hobby

[technoextreme]

Reading Eagle's website I realized that they actualy offer a cheaper (Read $125.00) version for people who are using Eaglecad for non-profit purposes.

Re:$400? Get real

[alienw]

Bleh. The Eagle autorouter is next to worthless. It can't do autoplacing, and it's not very smart. Unless all your boards consist of less than 10 through-hole parts, it generally can't route them worth a damn. Real autorouters cost a lot of money for a good reason. The one built into Eagle is a toy.

I like the PCB program. It integrates well with gschem (much better than how Eagle does it). It's a lot less buggy than Eagle (no annoying boogers and redraw bugs). It resembles a professional PCB layout tool a lot more than Eagle does. With a bit more work, it could easily compete with Protel.

PCSB/Schematic CAD for Linux

[goodie3shoes]

I think the simple answer is that there is no package available that meets the poster's needs. Those that are available at no cost (FOSS) don't have the ease of use and power that's requested. Eagle seems to be the only thing close to the price range. I haven't used it; I have used "pcb" with success. Its interface is a little funky, but not counter-intuitive. And it is under active development, with a user group list that sends me 20+ emails a day. In particular the requirement for autorouting presents a high barrier. It's a high-end feature, but not really necessary for simple boards. A human being will almost always come up with a better layout than a machine, and for analog boards, autorouting is useless, because the design rules cannot be taught to the machine. If, as some have surmised, the person who originated the inquiry is a woman, the sometimes harsh and dismissive tone of some posts on this topic leaves me no doubt as to why there seem to be so few women involved with Linux. My thanks to those who have posted helpful suggestions.

Re:$400? Get real

[Austerity Empowers]

I've never met a good autorouter, although I've only seen the Orcad one and whatever it is that's built into Allegro. In every place I've worked they've been forbidden to use with good reason.

There's some magical force at play where the more complex the board, where routing by hand gets more painful, the more likely an autorouter will screw you up and cause a respin. At home, if I'm going to pay hundreds out of pocket just to get the board fabbed I'd rather not risk it. (Cost aside, there's the agony of hand solder, the blatant bribery and/or questionable ethics of convincing our contract mfg that it's "part of our high volume project, do it for free, please?", and the difficulty convincing my wife I've got to spin again rather than buy that couch she wanted...)

Routing by hand is almost always better, and once you get good at it, doesn't take all that long. The thing that NEEDS automation is creating schematic symbols & footprints. For real now, there's just no excuse for that.

Re:$400? Get real

[rco3]

I've used Protel and Eagle on the same basic design, and found that Eagle's autorouter is both faster and less crappy than Protel's - especially the first few updates of Protel DXP, which had an awful autorouter. Protel kept leaving loops, routing all the way around the world to get next door, and took forever to do it. I wouldn't use autoplacing for anything important, so I'm not really concerned about that "lack".

Basically, I have to cordially disagree completely with your assessment of Eagle. It isn't the best interface in the world, but the number of bugs and the performance are far better in my experience than you've described. I intentionally and preferably use it over Protel all the time.

I guess that's why they call them opinions.