Best PCB design tool?

[Jeffw5555]

I checked into Eagle years ago, saw no reason to spend more money for something I don't need. Certainly it is a cool feature to click on nets and see the link. Cost and usability is a big advantage. I upgraded to the latest version of the program I currently use (Sprint Layout) 8 years ago, cost me about $ 60 USD. What is Eagle these days since they got bought by Autocad, who is running it into the ground? $ 700 a year? $700 times 8 is $5,600. It would have cost me almost 100X more to use a program that is horribly unintuitive, has a huge learning curve, gives me no benefit. That is the advantage.

I never said you were inexperienced. The feature you talk about benefits the inexperienced, not so much the experienced, especially for an old guy like me that can leverage old procedures while gaining benefit from the computerized layout and accelerated manufacturing process. Please note that my "formative years" were done back when PCB layout errors were hellaciously expensive. Tricks learned to result in zero errors are exactly the same if done by layout on computer versus on vellum.

And I am sure you are aware that Autocad is deprecating Eagle in a couple years? I can keep using the software I have today, probably another 10 years easily. And not spend another dime.

Remember, this group is focused on DIY. A program like the one I use is so easy to learn, almost no learning curve. The average DIYer could lay out his first board in the time he could rack his brain on how to even get started on the Eagle ecosystem. And it cost 49 Euros. Within the budget of many DIYers.

 

[Cqwet Dbdfte]

Yes Eagle is dropping into oblivion 2026, no reason to go there. It is old, quirky, with deep DOS roots. Autodesk improved it a lot, then gave up and started the Fusion360 mongrel, with PCB layout hidden behind somewhere.
A change nobody asked for.
Today I'd use KiCAD, which has improved miles.
I would have to transfer gobs of libraries, set up new DRC , layer stackups, CAM profiles, and BOM generation, and test that they don't mess anything up. It will be a pain but I'll do it before 2026.
The "free" version is available but very limited. Can be used as a viewer and possibly a library export tool.
I hate sending money to Autodesk.

 

[Newmarket]

Sure, schematic to layout checking is obviously important for inexperienced designers, or if you are dividing the work among others. I do it all, so I don't need it. I wouldn't trust even a high-end layout program to create the most optimal design anyway. An example is a GPS module I designed that my company makes. It needed special ground plane considerations, along with proper heat relieving in the underside of the daughter module to get it to flow properly during manufacturing. There is so much to know about the balance of circuit performance versus manufacturability, as well as repairability, that cannot be captured by having a high end layout program with schematic integration.

You are now confusing schematic/layout integration with auto routing. Integration does not affect the ability to define your own physical layout taking all relevant considerations into account. And any PCB designer / engineer / technician would be a fool to not have eg back annotation if given the choice.

 

[SwedishWings]

Sure, schematic to layout checking is obviously important for inexperienced designers, or if you are dividing the work among others. I do it all, so I don't need it. I wouldn't trust even a high-end layout program to create the most optimal design anyway. An example is a GPS module I designed that my company makes. It needed special ground plane considerations, along with proper heat relieving in the underside of the daughter module to get it to flow properly during manufacturing. There is so much to know about the balance of circuit performance versus manufacturability, as well as repairability, that cannot be captured by having a high end layout program with schematic integration.

Few, if any, experienced engineers agree with this.

 

[PermO]

If you are a hobbyist, like me, and you just want to be able to do fairly simple dual layer boards for audio projects and not overcomplicate things, I can highly recommend EasyEda.

It's free, you can order your boards from the software and it's fairly easy to use.

Learning software is not my hobby, if I'm still staring at a blank screen after one hour it get's deleted instantly, I have binned many, and EasyEda just clicked with me within minutes.

 

[Jeffw5555]

You are now confusing schematic/layout integration with auto routing. Integration does not affect the ability to define your own physical layout taking all relevant considerations into account. And any PCB designer / engineer / technician would be a fool to not have eg back annotation if given the choice.

The specific topic point I was responding to was schematic net verification. Not auto routing. The point is that I see a ton of bad boards these days, and too many people rely on these automated tools. i.e. "the schematic says to connect point A to point B, and I did that, and the program verified it". Again, if I could find a cheap schematic and layout tool that achieved the integration, especially renaming components on schematic due to the layout evolution, of course I would like it. It is just that I haven't found that tool, don't want to invest an inordinate about of time on the learning curve, and don't want to spend a ton of money on it.

 

[Cqwet Dbdfte]

I never used autorouting, my typical boards are very compact, extensive use of polygon pours, stripline and microstrip, controlled radii of RF traces, thermal vias, selective "stop" layer on thermal areas, phase matching, etc I do not feel comfortable leaving up to the included Autorouter. I also try to make schematic like the board, avoiding putting vias in RF traces, and keeping them short.
I'm not saying autorouting is useless.

 

[Jeffw5555]

Few, if any, experienced engineers agree with this.

Is this the IEEE forum? No, it is the DIY forum. I am giving counterpoint to those that insist you need schematic integration. I am an "experienced engineer" that uses a cheap, but super accurate and easy, layout program, to show the DIY people here that they too can use a cheap program to achieve excellent results without a lot of learning curve or expense.

 

[Newmarket]

The specific topic point I was responding to was schematic net verification. Not auto routing. The point is that I see a ton of bad boards these days, and too many people rely on these automated tools. i.e. "the schematic says to connect point A to point B, and I did that, and the program verified it". Again, if I could find a cheap schematic and layout tool that achieved the integration, especially renaming components on schematic due to the layout evolution, of course I would like it. It is just that I haven't found that tool, don't want to invest an inordinate about of time on the learning curve, and don't want to spend a ton of money on it.

But you veered off into specifics of the layout eg ground plane and thermal design. Not related to connectivity.
afaik most (all ?) ECAD software now offers PCB/schematic integration to some degree. Even the budget conscious ones.
OTOH auto routing is often not included. At least not in the base version. And it's often not great unless you do pay big money for the "Advanced" tier.

 

[Jeffw5555]

But you veered off into specifics of the layout eg ground plane and thermal design. Not related to connectivity.
afaik most (all ?) ECAD software now offers PCB/schematic integration to some degree. Even the budget conscious ones.
OTOH auto routing is often not included. At least not in the base version. And it's often not great unless you do pay big money for the "Advanced" tier.

Yes, my bad, I strayed from the key point myself, sorry about that.

I agree that most "proper" ECAD software should (and do) offer integration. I would love it that my current program supported it. I might give KiCAD another look, I like the idea of an open source program. (I looked at it a number of years ago, and the UI wasn't very good, it didn't have good resolution and easy to switch grid, I like to change grid on the fly during layout)

 

[SwedishWings]

Is this the IEEE forum? No, it is the DIY forum. I am giving counterpoint to those that insist you need schematic integration. I am an "experienced engineer" that uses a cheap, but super accurate and easy, layout program, to show the DIY people here that they too can use a cheap program to achieve excellent results without a lot of learning curve or expense.

You referred to "inexperienced designers" in your post that i replied to, hence my comment. But perhaps you meant something else?

I really don't understand what you are trying to convey here: "cheap, but super accurate and easy, layout program, to show the DIY people here that they too can use a cheap program to achieve excellent results without a lot of learning curve or expense."

"cheap program/without ... expense": There are free CAD software packages that have excellent integration with schema and layout. What cost are you talking about?

"super accurate": I have never had any issues with resolution in any CAD software. How do other CAD software lack accuracy?

"easy": a strange term for a software that forces you to transfer a schema to PCB manually.

"without a lot of learning curve": The schematic capture is usually the easiest part to learn, and most modern CAD programs make transfer of netlist from schema to PCB no more complicated than pressing a button.

I started doing PCBs before there were personal computers (we made the designs on transparent film using light tables), and I have welcomed every single step of progress of CAD tools, and encourage every aspiring engineer to learn the tools and benefit from the awesome help they offer. These tools are designed the way they are to make the engineers work easier, and they certainly do, even for the beginner once they learn how to use them. Sure, some tools have a more steep learning curve than others, but given the huge amount of online resources that provides great tutorials there is really no reason not to learn the proper tools.

 

[MidnightArrakis]

OTOH auto routing is often not included. And it's often not great unless you do pay big money for the "Advanced" tier.

[auto routing is often not included. And it's often not great unless you pay big money for the "Advanced" tier] -- I once used a $25,000 per seat PCB-design program when I was designing 10-layer PCB's at the NASA/Goddard Space Flight Center and let me tell you.....-- BIG BUCKS -- doesn't equate to "overall wonderfulness"!!!

Our typical board-design process was: We would spend Monday up to Friday at noon just working on component placement. When you are working on boards with 3,500 to 5,000 nets, you wanna make certain that you have "all of your ducks in a row", if you know what I mean. Then, we would go to lunch and when we got back, we would spend all afternoon getting all of the auto-router parameters setup. There were a "bazillion" different parameters available that could be tweaked a "gazillion" different ways!!! And.....THAT took some time to work through.

Then, you would hopefully have the auto-router all ready to go as you were ready to walk out-the-door on that Friday evening. After a quick review of all of the auto-router parameters, you would get up from your terminal and press the "Auto-Route" key and leave. As you were closing the door to the lab, the NASA mainframe computers were beginning to analyze your board-layout by looking at your components placement and going through all of your nets. After the computers had digested all of the data about your board-layout, the auto-router would kick-in and begin its job. And, WHAT A JOB IT WAS!!!

Since these were all Surface-Mount boards mixed-in with some high pin-count ASIC's and BGA's, there were thousands of component pads that first needed to be "fanned-out" for vias. The auto-router may have gone through a dozen different routines of "fanning-out" vias before it came up with what it considered an "optimum" fan-out. Then, the actual -- auto-routing -- would begin. The auto-router would literally spend all weekend, 24-hours around the clock, with auto-routing these 10-layer boards. What I thought was pretty cool was that we could access the NASA computers from home and check-in on the process with our home PC's. The auto-router would also be creating a "Progress Report" as it was routing a board and if things weren't going as well as we had hoped from our original auto-router setup, we could further tweak the setup from home and then resume the auto-routing process.

When we got back into the lab the following Monday, we would load-in the finished board and check it out. And, here is the part that relates to the comment made by Newmarket. After reviewing the results of a $25,000 PCB design program that has just spent 60 continuous hours of high-powered computer time in auto-routing a 10-layer circuit board.....we would then -- HAVE -- to spend the next following week "cleaning-up" the auto-routed board because of all of the basic mess and other shenanigans that the auto-router introduced into the overall routing!!! I suppose you could just call this extra work "job security", huh??? All of the "clean-up" was necessary in order so the board could be fabricated. Otherwise, I am more than certain that the PCB-fabrication shops would have rejected the GERBER files as being "non-manufacturable".

Here's a couple of examples:

>> TOPSIDE COMPONENTS (PCB-1)

>> BOTTOMSIDE COMPONENTS (PCB-1)


>> TOPSIDE COMPONENTS (PCB-2)

>> BOTTOMSIDE COMPONENTS (PCB-2)


>> ONLY 3 INTERNAL ROUTING LAYERS SHOWN (PCB-3)




So, Kids!!!.....when you get really, Really, REALLY good at using DipTrace, Eagle, EasyEDA or even KiCAD.....then maybe YOU, TOO can design PCB's like these examples!!!

/

 

[SwedishWings]

Since these were all Surface-Mount boards mixed-in with some high pin-count ASIC's and BGA's, there were thousands of component pads that first needed to be "fanned-out" for vias.

Things got a lot more complex with FPGA's where flexible pin-out allow routing signals to different pins, BUT, you have to make sure you pass FPGA-tool DRC, use banks with the right voltage, high speed logic get placed in the right bank (and/or SLR) with respect to clocks and clock regions, avoid congestion, be able to physically fan out all high speed signals and that the FPGA P&R tool eventually can solve the design you throw at it - and all this BEFORE the RTL design is finished. FPGA flexibility adds a new dimension of complexity for board level design and takes a lot of time to get right. I know of no CAD tool that helps in this process. I'm still trying to find ways to be more productive in this area.

 

[MidnightArrakis]

FPGA flexibility adds a new dimension of complexity for board level design and takes a lot of time to get right. I know of no CAD tool that helps in this process. I'm still trying to find ways to be more productive in this area.

[I'm still trying to find ways to be more productive in this area] -- So.....this then opens up an opportunity for you to create a new program that will solve this issue for you!!! You could end up being a "Software Billionaire" someday!!!

By the way.....isn't it "Tuesday" right now where you are and maybe around midnight? It's Monday at noon here right now.

/

 

[C H]

...Much like all of the "Climate Change" foolery that we are having to deal with these days. A cow farts somewhere and the "Green New Deal" fanatics go into hysterics about how the methane gas is going to kill us all because that fart raised the global temperature by some imperceptible amount!!! GET A LIFE!!! (not directed at you, gswan).

/

It's a bizarre theory, isn't it, that farting cows somehow melt the Polar icecaps and cause wildfires!

Other bizarre theories that the overwhelming majority of scientists who have tested the evidence and accept as scientific fact:
-you get heavier as you move faster
-light is both a particle and a wave
-we are the product of evolution

 

[SwedishWings]

[I'm still trying to find ways to be more productive in this area] -- So.....this then opens up an opportunity for you to create a new program that will solve this issue for you!!! You could end up being a "Software Billionaire" someday!!!

By the way.....isn't it "Tuesday" right now where you are and maybe around midnight? It's Monday at noon here right now.

/

If I tell you it's Tuesday, around 1.32am, you can probably figure out roughly where I am given latitude of 14.558

 

[Cqwet Dbdfte]

[auto routing is often not included. And it's often not great unless you pay big money for the "Advanced" tier] -- I once used a $25,000 per seat PCB-design program when I was designing 10-layer PCB's at the NASA/Goddard Space Flight Center and let me tell you.....-- BIG BUCKS -- doesn't equate to "overall wonderfulness"!!!

Our typical board-design process was: We would spend Monday up to Friday at noon just working on component placement. When you are working on boards with 3,500 to 5,000 nets, you wanna make certain that you have "all of your ducks in a row", if you know what I mean. Then, we would go to lunch and when we got back, we would spend all afternoon getting all of the auto-router parameters setup. There were a "bazillion" different parameters available that could be tweaked a "gazillion" different ways!!! And.....THAT took some time to work through.

Then, you would hopefully have the auto-router all ready to go as you were ready to walk out-the-door on that Friday evening. After a quick review of all of the auto-router parameters, you would get up from your terminal and press the "Auto-Route" key and leave. As you were closing the door to the lab, the NASA mainframe computers were beginning to analyze your board-layout by looking at your components placement and going through all of your nets. After the computers had digested all of the data about your board-layout, the auto-router would kick-in and begin its job. And, WHAT A JOB IT WAS!!!

Since these were all Surface-Mount boards mixed-in with some high pin-count ASIC's and BGA's, there were thousands of component pads that first needed to be "fanned-out" for vias. The auto-router may have gone through a dozen different routines of "fanning-out" vias before it came up with what it considered an "optimum" fan-out. Then, the actual -- auto-routing -- would begin. The auto-router would literally spend all weekend, 24-hours around the clock, with auto-routing these 10-layer boards. What I thought was pretty cool was that we could access the NASA computers from home and check-in on the process with our home PC's. The auto-router would also be creating a "Progress Report" as it was routing a board and if things weren't going as well as we had hoped from our original auto-router setup, we could further tweak the setup from home and then resume the auto-routing process.

When we got back into the lab the following Monday, we would load-in the finished board and check it out. And, here is the part that relates to the comment made by Newmarket. After reviewing the results of a $25,000 PCB design program that has just spent 60 continuous hours of high-powered computer time in auto-routing a 10-layer circuit board.....we would then -- HAVE -- to spend the next following week "cleaning-up" the auto-routed board because of all of the basic mess and other shenanigans that the auto-router introduced into the overall routing!!! I suppose you could just call this extra work "job security", huh??? All of the "clean-up" was necessary in order so the board could be fabricated. Otherwise, I am more than certain that the PCB-fabrication shops would have rejected the GERBER files as being "non-manufacturable".

Here's a couple of examples:

>> TOPSIDE COMPONENTS (PCB-1)
View attachment 122742
>> BOTTOMSIDE COMPONENTS (PCB-1)
View attachment 122743

>> TOPSIDE COMPONENTS (PCB-2)
View attachment 122744
>> BOTTOMSIDE COMPONENTS (PCB-2)
View attachment 122745

>> ONLY 3 INTERNAL ROUTING LAYERS SHOWN (PCB-3)
View attachment 122746
View attachment 122747
View attachment 122748

So, Kids!!!.....when you get really, Really, REALLY good at using DipTrace, Eagle, EasyEDA or even KiCAD.....then maybe YOU, TOO can design PCB's like these examples!!!

/

Well I hope I never have to.

How many people were involved in designing and verifying this board?
How long between project start to delivery of first revision?

 

[SwedishWings]

It's a bizarre theory, isn't it, that farting cows somehow melt the Polar icecaps and cause wildfires!

Other bizarre theories that the overwhelming majority of scientists who have tested the evidence and accept as scientific fact:
-you get heavier as you move faster
-light is both a particle and a wave
-we are the product of evolution

I see reasons for doubt, but not because the scientists working in the field get things wrong. The problem is rather that research of views challenging the current conclusions is non-existent as funding of such research is virtually impossible. In a field where there is limited concrete evidence, theories are difficult to prove or falsify, where simulations with huge number of variables is a fundament for the conclusions - very hard scrutiny is absolutely necessary, but that is not happening. We probably suffer from similar problems in theoretical physics where we have been standing still for a few decades now. Perhaps the evil polarisation we see spreading like a cancer in our society has spread to science more than we are willing to admit. Don't want to paint the town black, but it's hard not to see parallels to John B. Calhoun's behavioral sink.

 

[C H]

I see reasons for doubt, but not because the scientists working in the field get things wrong. The problem is rather that research of views challenging the current conclusions is non-existent as funding of such research is virtually impossible. In a field where there is limited concrete evidence, theories are difficult to prove or falsify, where simulations with huge number of variables is a fundament for the conclusions - very hard scrutiny is absolutely necessary, but that is not happening. We probably suffer from similar problems in theoretical physics where we have been standing still for a few decades now. Perhaps the evil polarisation we see spreading like a cancer in our society has spread to science more than we are willing to admit. Don't want to paint the town black, but it's hard not to see parallels to John B. Calhoun's behavioral sink.

Fair point, but is this the place for this conversation?

 

[SwedishWings]

Fair point, but is this the place for this conversation?

Agreed.