Best Value for Schematic Capture & Layout Software ???

gEDA is free for any size project, not just trivial starter boards. A number of companies have used it for big projects, and not only avoided license fees, but also avoided vendor lock-in. If you're on a budget it's worth a look-see.

The PADS LE suite from Mentor Graphics is pretty good, but it starts at about $4000.

Mark Borgerson

What you consider "pretty cryptic" is no problem in real use. Eagle is easy to use if you are willing to read the docs. The number of needed workarounds got much smaller in the recent versions.

Well, that's indeed limited. Pulsonix costs more but is worth a look IMO.

Oliver

--
Oliver Betz, Munich
despammed.com might be broken, use Reply-To:

You don't say what kind of boards you have in mind but for simple (up to 6 layers, 5thou track/gap, 0.5mm pitch QFP, occasional 1mm pitch BGA) then Easy-PC will do you and its well within budget.

If you want more, look at Altium Designer which used to be way outside your budget but after a recent price drop is just about in. You get all sorts of stuff thrown in (like FPGA and HDL simulation) and it will cope with quite fancy boards.

Michael Kellett

You'll find all of this software is "pretty cryptic" in the final analysis; the processes for doing things take a learning curve, and the processes for doing things _right_ have an even steeper curve. For example you can just sit down and start drawing in OrCAD, but if you want to do it properly you need to spend a lot of time defining everything in your custom library parts so you can do a DRC and ERC and get meaningful results. And if you want to be able to export that PCB to the mechanical guys for an interference check, there's even more work to be done with the library parts.

This analysis is more complex than just the initial purchase price. Luckily you don't have a legacy of existing projects, so you don't have to think of how many hours will be required to migrate library parts.

One of the reasons for buying a "big name" commercial package is so that other people - parts vendors - will do the hard work of generating library parts for you. Unfortunately they draw them only for the latest versions of the CAD packages. So unless you're willing to fork over cash in truckloads on a regular basis, you'll be behind the curve. Cadsoft's policy is no "maintenance fees", upgrades of minor version number are free. In practice this means upgrades only cost money about once every five or six years.

I started using EAGLE (the full version) eight or nine years ago and have had no good reason to change; it's not onerously copy-protected, it has all the features I need, and the price was right when I bought it. I have looked at gEDA recently (since I don't want to upgrade to EAGLE 5) but it seems to have no autorouter - an autorouter, even a fairly primitive one, is very handy for simple boards, or for finishing off the noncritical bits of more complex boards. So though I'd like to go open-source, this is kind of a non-starter at the moment.

I have performed layouts on *many* boards using EAGLE 5.3.0. I started with 4 and upgraded to 5 which has many good fixes. Multilayer, Single layer, double sided, whatever, the autorouter is cool for boards that you are banging out, the manual routing is totally reasonable. The learning curve was no different than ORCAD was and way-y-y nicer than my "powerlogic" experience was.

-Chris

I've used gEDA for projects before and it could autoroute at least

6 months ago. I think your knowledge is old.

-pete

Hmm, I'll go look again. Unfortunately I can't test any of the binary distros where I am right now - fink doesn't work properly on my Macbook, and I've given up trying to fix it, and I don't have a Linux distribution suitable for running it here (only non-X Ubuntu Server editions). I'll try it out at home.

I used gEDA from the point of view of a hobbiest and I must say the pipeline from schematics to a printable PCB board is quite complex--but it worked.

The documentation for gEDA is vast and a bit disconnected, but mostly what you want to do at any stage in gEDA is available in Google somewhere.

Good luck.

-pete

That's what discouraged me from using gEDA for a small "hobby" board: the workflow wasn't at all obvious, and after puttering around with it for a while, I was stumped.

I opted for Eagle where the workflow was dead-simple (and since I very rarely do board layout, it needs to be). The board-house I picked accepted Eagle project files, so I didn't even have to go through the exercise of generating Gerber files.

I can't comment on how Eagle is for larger "real" project use.

--
Grant Edwards                   grante             Yow! Of course, you
                                  at               UNDERSTAND about the PLAIDS
                               visi.com            in the SPIN CYCLE --

I won't lie. gEDA is *complex*.

It took me maybe 2 weeks to figure out how to use it. But keeping my monetary costs low and having the ability to do Gerber files so I can be free with who I chose to do the board production was more important to me than the time I spent learning it. Not everyone will make that tradeoff though.

Thank you.

-pete

Another option that may be less complex than gEDA but better integrated would be Kicad I've used it both for homebrew (toner transfer) boards and for commercial boards through Alberta, Advanced, and Sparkfun's batch house.

It's free (libre and gratis), pretty easy to use, and does hierarchical schematics.

--
Rich Webb     Norfolk, VA

I didn't mean that Eagle can't generate Gerber files. Eagle can generate Gerber files just fine -- it's just that I didn't have to (which means one less step I could screw up).

--
Grant Edwards                   grante             Yow! Is it clean in other
                                  at               dimensions?
                               visi.com

Oh, generating Gerbers from EAGLE is *more* than one step. Many more, in fact, especially for multilayer boards with mask and silkscreen :)

Er, gEDA/PCB has had an autorouter s> I used gEDA from the point of view of a hobbiest and I must say the

Complexity is a side-effect of flexibility. Most of us have either scripted or Makefiled their process, so "make pcb" or "make web" does

100% of what's needed. I'm also working on simplifying the common gschem->pcb flow.

Or on the mailing list, where we're glad to help people use the tools.

Yup, but it can be used in many different ways because of that. We do have a new tutorial that helps you get started with the gschem->pcb flow, though:

Hi Alfred,

Over the past 20-25 years, I've played with several Schematic+PCB toolkits. My impressions of each, unfortunately, are rooted in the timeframes when I was actively using each and, as such, don't reflect how they may have evolved, since. Often, the tool was something I *had* to use for a particular client and my exposure to it was limited *to* that client's needs. Most of these I would only go back to "kicking and screaming" (more later) :>

E.g., I enjoyed DASH/STRIDES (mid 1980's?) as an incredibly quick -- an intuitive -- way of knocking out a schematic in short order. I'd probably still use it (even an ancient version!) today if it was a bit more integrated with other vendor's PCB and CAD tools. But, DASH-PCB was a real *dog*! (I wonder what it would run like on modern x86 hardware today? Might be hard to find an ISA machine for the hardware key! :-/ ) I think their (DataI/O) paranoia about copy protection probably sank that boat!

[Perhaps the *worst* tool was an early version of AutoCAD (like 2.9?) which was little more than an electronic etch-a-sketch :-/ I.e., it didn't embody any knowledge of the issues that pertain to electronic drawings]

Wintek made a naive little "drawing tool" for PCB layout but it was to PCB's what the aforementioned AutoCAD was to schematics (i.e., *you* did all the work).

[Again, each of these were very early experiences in "desktop tools" -- I'm avoiding talking about dedicated systems like ComputerVision, etc.]

Tango (?) and EasyTrax (?) were a (noticeable) step above Wintek's product. I'd place them a notch below Eagle (but it depends on what features you are specifically looking for).

I've had most success with Protel and OrCAD (in no particular order). Both have bugs and/or annoyances. And, tend to get expensive pretty quick! But, you can go back and forth between Schematic and PCB in both tools with *relative* ease (the process isn't always foolproof, though :< ). OrCAD has (I should technically say "has had" as I am unaware of the status of the current release) some annoying bugs but most can be worked around *if* you are careful. I've been able to layout some truly intense designs with it that other tools would have forced me into additional layers, etc. (e.g., I did a board that rightfully should have been 6 layers on just 4). And, much of OrCAD's user interface is non-intuitive at times. But, I think that is probably true of most tools (?)

[I remember OrCAD's Layout had an annoying habbit of not letting me place a track where *I* wanted it to go; instead, it would keep ripping it up and trying a different routing -- which would *fail*... tools should never think they are smarter than their user's!]

When I looked at gEDA it was in its infancy. And, I want a tool that I can use *now*, not something that *might* eventually do what I want (sorry, I've had lots of experience with open source software over the years that has taught me what I need to know before I embrace a particular FOSS "product"). If you are a hobbyist and can afford to bleed a little (i.e. where you have more time than money), this may be a *great* option! But, for me, "the layout needs to be in fab in 6 weeks" tends to limit how adventurous I want to be! :> I liken it to having a hammer where the handle isn't firmly attached to the head; every once in a while it flies off and you lose time retrieving it, reattaching it and repairing whatever porcelain vase it happened to fly *into* :-/

The trouble with most of these tools is moving between them (i.e., between vendors) is essentially impossible. There are just too many quirks that prevent you from using one vendor's tool with another vendor's "existing design" (even trying to use PCB tool B on a netlist prepared from Schematic tool A).

But (and this is an important one!) the same is also often true about moving within versions of a particular vendor's product! I.e., layout a board using version X of tool A and you might not be able to modify -- or even *view* -- that layout using version Y of that same tool!

The lesson to learn (at least the lesson *I* learned) from this is to pick a product that does most of what you think you will need to do and embrace it. *That* version of "it". Otherwise, you may find the schematic symbols that you had to hand create aren't recognized by the next version (building symbols is an annoyingly time consuming chore unless you just draw a box with N signals attached without regard for appearance, functionality, standards, etc.).

I got bit after upgrad> The high end is of course Orcad, for $10,000.

Ouch! I hadn't realized the price had crept up so... :<

I think EasyTrax may also be "free", now?

Do yourself a favor and think of what your real CURRENT needs are and what your FUTURE ones are likely to be. (i.e., you really don't want to be changing tools midstream as your business grows -- you'll have other more important things to deal with!)

In no particular order:

Do you do analog, digital or mixed signals designs?

Are you likely to be involved in any "power electronics" or high voltage designs? (i.e., how much flexibility do you need in the DRC's?)

Do you do mostly thru-hole or SMT designs? Or, mixed?

Do you need to interface to other tools (e.g., to design bed-of-nails clamshells, test vector generation, etc.)?

Do you need to interface to other vendors? E.g., does your prefered board house have certain requirements beyond photoplots and drill drwaings?

Any more *exotic* layout problems that you might typically encounter? (e.g., components on both sides of the board, "reversed" pinouts, components *under* other components, etc.)

Do you need forward and back annotation between schematic and PCB? Or, are your designs small enough or infrequent enough that you could do this (reliably) manually?

Do you need to integrate physical packaging with the layout process (e.g., ensuring tall components are not placed in certain areas of the PCB)?

Will you need many special (schematic) symbols (most decent PCB tools support most of the standard "packages" -- but the same is not true of schematic symbols especially as you may have to deal with custom and semicustom chips)?

How will you handle (inevitable) bugs? Can you discipline yourself to rigorously implement workarounds (which you may have to identify on your own!) or do you expect others (i.e., vendor) to do this for you? Note that this is particularly important if others (besides yourself) will now (or eventually) end up using these tools! I.e. *you* may be aware of a particular bug/workaround but will everyone who works with/for you be equally cognizant?

What are you going to do when your current tool stops meeting your needs? How much (*time*) will you be willing to reinvest at that point?

I'm sure I can think of a dozen other issues that you probably should consider. But, if you actively think about your environment and what the design+layout process entails, I am sure you can stumble on them as well!

Good luck! Remember, there is no "right solution" for this question; just what's "most right" for *you*!

I wouldn't like anyone to get the impression I'm maligning gEDA. I really do think it is a great tool. The complexity is there for a reason and I can see over time the complexity is being hidden during the common use cases but is always available when needed.

And that is a very nice tutorial!

The thing I ran into maybe a year ago was mostly a disorgaznization of information, such as I type gEDA into google, get to the wiki, and the first tutorial I come across is:

and not the new one you just posted.

The documentation seems to be spread across a few revision of gEDA's tools and there is no really good way to figure out what matches what until you run into a command use or menu item which doesn't exist and then you move to another piece of related documentation where you can continue your work.

The information really needs to be streamlined from when someone types in gEDA in Google to what do they see and how is it navigated.

Thank you.

-pete

Depends entirely on what your technology/production timeline will look like. If you are going to move into multiple layers, and high speed design rules fast, then it may pay to do more research.

Also look around at the various Spice packages - if you expect to simulate, that's another selection question...

Download versions a shortlist of candidates, and also get a 'good' package as well - eg Mentor's PADS eval has no time limits, and only a database limit, so you can use that for many startup projects.

Then, try them all, and see how much you can tolerate the shortcomings of the cheaper ones. Use a free one, on the first project - real use is a great way to find out what you can tolerate :)

-jg

I didn't get that impression. I agree that it's complex, but there are some (perhaps obscure) reasons behind that complexity, and some cool things you can do *because* of that complexity. It's a tough balancing act.

We're working on integrating PCB's web presence with gEDA's (PCB was on sourceforge, now everything is on gpleda.org) so hopefully this will improve over time. Also, we're phasing out gsch2pcb for the normal use cases as part of our Linux Fund project:

The first bullet item replaces gsch2pcb with an "update from schematics" button in pcb :-)

Yup.