EEVblog #527 - Altium Entry Level PCB Tool Rant

[mikeselectricstuff]

Autorouter is a must

Totally disagree on that one - a good autorouter takes a lot of effort to develop ( or buy in), and will only be useful to a small proportion of the low-end market as it's of limited use until you can afford enough laers to make it work well. Resources should be concentrated on making manual routing as quick and easy as possible (e.g. interactive routing).
Probably best to be purchasable as an extra (as opposed to being part of a big step up with other stuff).

[TerminalJack505]

Eagle is SHIT and - sorry for being rude - anyone who pays any money for Eagle is an idiot. Whatever you can do with paid Eagle you can also do with free KiCad (or Inkscape...). If you have to - go for DipTrace - it's currently closest to being decent pcb tool at an affordable price.

I'm not going to defend Eagle but an even bigger idiot would be someone that pays $8000 to be a software beta tester. 

I've never had Eagle crash and I've never encountered any show-stopping bugs.  It's just clunky to use.  From what I've heard of Altium there are complete versions that are "to be avoided" because they are so full of bugs.  All I can say about that is "wow, just wow."

[EEVblog]

Actually there is something in between - its called the EAGLE Hobbyist version. Its $169 for a 160x100mm board size. it requires filling out  (and signing) a form for them, though.

And as the name suggests, you can't use it for commercial designs. There is even argument over that I believe if you use it to produce an OSHW project and then someone else goes and sells it.

[EEVblog]

From what I've heard of Altium there are complete versions that are "to be avoided" because they are so full of bugs.

That is true.
Many people in the industry (including me) stuck with 99SE for a long time because several subsequent versions didn't cut the mustard.
I think Altium at one point even had a dedicated marketing campaign to get these luddites to upgrade 

[krivx]

Now, i have another question for the general audience here.
Here is something i have a hard time understanding...

One thing I feel might have been missed here - for a hobbyist/start-up/"budget" market Altium also needs to compete with a wide range of products outside of the PCB tool market. We're now living in an era when a cheap DSO can be bought for a few hundred dollars, a Hakko for $100 (in the USA) or clone for even less, a decent handheld DMM for $50 etc and this does include the second-hand market flooded with gear from liquidated businesses. If I'm outfitting a lab on a budget buying the best tool is completely irrational if it means I lose out on the opportunity to buy something else which is just essential. What good is a fantastic PCB package if you don't have the tools to build, power and debug the boards?

[free_electron]

From what I've heard of Altium there are complete versions that are "to be avoided" because they are so full of bugs.

That is true.
Many people in the industry (including me) stuck with 99SE for a long time because several subsequent versions didn't cut the mustard.
I think Altium at one point even had a dedicated marketing campaign to get these luddites to upgrade 

I did , for a looong time ! Main reason : the loss of global edit. The new way required typing search expressions. Only in a later version did they implement 'find similar'

[metalphreak]

I tried out Altium after always using the free version of Eagle. It is so much better even for basic designs it's not funny. If Altium do a cheaper version for <$200, I'm definitely in. And none of this yearly subscription stuff. For something I would use maybe once every month or two, it has to be a price for a version I can keep using for as long as I want.

A free non-commercial version would be amazing. I'm sure plenty of people will end up making real products, even if its small production run, low profit stuff. It's not hard to justify a few hundred dollars for software when you are actually making a bit of coin from what you are doing.

[PeteInTexas]

For the hobbyist market, they should develop something web based that PCB houses can deploy for themselves and be a one stop shop those that in reality only do something once or twice a year.

That would be a winner: for the hobbyist access to all features of the best without paying a dime for it; for the PCB house, pcb orders.

NO ! absolutely not!

web based is crap. you need to be online to use it. that sucks and the user interface is slow as hell.

as for pcb houses : i want choice. china is cheap ! ( hehe. now i'm the guy asking for cheap ) and you are tied into a closed format.
give us at least gerber data so we can go where we want.

html 5 is promising.  certainly capable for hobbyist needs.

the idea is, where ever you want to go for a pcb house, they will have this web based Altium tool.  Altium makes money by selling it to the pcb house; the hobbyist gets to pick which house they want and because its all Altium (for free), the files are portable.  Its a win-win.

[PeteInTexas]

For the hobbyist market, they should develop something web based that PCB houses can deploy for themselves and be a one stop shop those that in reality only do something once or twice a year.

Arghh!!!! Noooooo....  Most web based applications suck.  I don't want to be tied to the availability of the Internet (in general) or some web server that may or may not be up.

Most web based apps don't use HTML 5.  Web servers have very excellent up times, even ones with high traffic/bandwidth.

Complaining about availability of Internet connection, in 2014?  If you don't have wired internet in the garage, consider you geek card revoked.

[PeteInTexas]

For the hobbyist market, they should develop something web based that PCB houses can deploy for themselves and be a one stop shop those that in reality only do something once or twice a year.

That would be a winner: for the hobbyist access to all features of the best without paying a dime for it; for the PCB house, pcb orders.

oh god no. most web-based things tend to be slow as crap and forget about 3D manipulation.

check out HTML 5 demos

edit: I was really also thinking of WebGL

[elgonzo]

That is exactly the reason that quite some time back Altium moved to the "All you can eat" deal and (as mentioned in the video) actually made PCB "optional extra".
The basic version only came with schematic and embedded/FPGA, in order to push their FPGA hardware dream.
To get PCB to you had to pay top dollar. Of course, practically 100% of people needed the PCB, so they were forced to pay top dollar. I also meant that management could no longer see who was using just PCB/schematic and who was using the embedded stuff. So internally, and externally for marketing, they could say that everyone was buying into the FPGA vision. It's was actually a very clever move from that aspect.
Of course they have now admitted that (well) over 90% of the business is PCB.

They did? (Sorry, i don't work in this industry and didn't know that.)
That is quite an embarrassingly desperate move...

[arvidj]

I suspect that's just available to actual schools and students, not people teaching themselves at home?

At last a term that fits me ... "people teaching themselves at home".

I wonder if a letter from my wife would be considered "adequate documentation" to qualify for the student discount? 

[hikariuk]

For the hobbyist market, they should develop something web based that PCB houses can deploy for themselves and be a one stop shop those that in reality only do something once or twice a year.

Arghh!!!! Noooooo....  Most web based applications suck.  I don't want to be tied to the availability of the Internet (in general) or some web server that may or may not be up.

Most web based apps don't use HTML 5.  Web servers have very excellent up times, even ones with high traffic/bandwidth.

Complaining about availability of Internet connection, in 2014?  If you don't have wired internet in the garage, consider you geek card revoked.

Or you're living in parts of the UK...

[andtfoot]

For the hobbyist market, they should develop something web based that PCB houses can deploy for themselves and be a one stop shop those that in reality only do something once or twice a year.

Arghh!!!! Noooooo....  Most web based applications suck.  I don't want to be tied to the availability of the Internet (in general) or some web server that may or may not be up.

Most web based apps don't use HTML 5.  Web servers have very excellent up times, even ones with high traffic/bandwidth.

Complaining about availability of Internet connection, in 2014?  If you don't have wired internet in the garage, consider you geek card revoked.

Or you're living in parts of the UK...

Or Australia... 

[walshms]

For the hobbyist market, they should develop something web based that PCB houses can deploy for themselves and be a one stop shop those that in reality only do something once or twice a year.

That would be a winner: for the hobbyist access to all features of the best without paying a dime for it; for the PCB house, pcb orders.

oh god no. most web-based things tend to be slow as crap and forget about 3D manipulation.

check out HTML 5 demos

edit: I was really also thinking of WebGL

I think this is pretty unlikely for a lot of reasons... but here's just two off the top of my head:

First, the PCB houses need to be sure that the cost of doing something like this would make sense.  From Altium's point of view, this would effectively be a huge number of seats, which would make it extremely expensive for the PCB house, which would then have to pass on the cost... not a good thing.

Second, HTML5, WebGL and what -- Javascript?  Java?  Either way, coding that would be a huge nightmare.  Not to mention that it would run very slowly.  Maybe I'm missing something here, but I can't see that even being an option for most coders, and it would certainly take some significant investment in time and money to make that work.  In the end, I can't see a web-based PCB design tool having any traction.

What does that leave us with?  Remote desktop?  That might work, but then you not only have the seat licenses for Altium, but you have the seat licenses for RD services... even more expensive.

I don't know.  From my point of view, I can't see this as being something that runs anywhere but on a local PC if you want to get any kind of performance out of it.

[hikariuk]

I think if Altium had a lower priced hobbyist version - in the £100 to £150 arena - I'd probably buy it after a little consideration.  Especially if they also provided a sane upgrade path for people who might need it.

[walshms]

want fpga,ip cores and compilers? add 2K

Yeah but then nobody would buy that option.
...

... Of course, practically 100% of people needed the PCB, so they were forced to pay top dollar. I also meant that management could no longer see who was using just PCB/schematic and who was using the embedded stuff. So internally, and externally for marketing, they could say that everyone was buying into the FPGA vision. It's was actually a very clever move from that aspect.
Of course they have now admitted that (well) over 90% of the business is PCB.

BTW: Dave, this thread finally convinced me to go ahead and start participating in the conversation.

Maybe not so clever... market penetration is everything, and Altium hasn't learned that lesson yet.  I can quote a classic example of this in the form of the Microsoft/Novell history. 

Novell always had the model of selling their licensing and support through a partner network.  Great, and at first, it worked really well.  Novell owned the network market.  Once Microsoft jumped in, and their marketing machine made it mainstream, Novell found itself entirely unable to compete on name recognition alone.  Their software is second to none -- I've been supporting it for over 30 years -- but it's a hard sell, especially where the people who make the choices aren't the ones exposed to the technical aspects.

If Altium were to offer a free tool that would enable the masses to get familiar with their software and focus on the PCB market, they'd get the penetration they need.  But as important as that is making sure that the "free" version isn't crippled to the point of being useless... the right way to go would be to make it possible for people to actually make use of it to design reasonable boards, and get them produced.  I don't mind manual routing -- and I suspect that most of us would rather avoid autorouting anyway, so that's a perfect thing to leave out.  FPGA compiler?  No, thanks.  But let people export Gerber files, do four layers, and create boards that are in the neighborhood of about 100mm² and I think they'd have a winner.  Make it reasonable to do that much commercially, say in the neighborhood of US$600, and I think that forms the basis for a winning formula.

They'd win by selling in volume -- the lesson Microsoft taught everyone very effectively, if they were paying attention.

[grego]

I don't see a free altium doing 4 layers.  It's just too much flexibility.

The reason Eagle has been successful is because it's free - and it's crippled.  However, anyone who says it's crippled to the point of being unusable I just point towards the Arduino.

I'm kind of with free_electron on this one - everyone wants this to be the perfect free tool.  It won't happen (or I'd be surprised if it does).  I full expect it to be 2-layer with either board size restrictions or pin count restrictions.

Altium isn't going to give away MORE than Eagle.  Why should they?  AD is clearly better than Eagle the problem is the price.  If I had to do it I'd match what Eagle can do but do it via AD.  Then tier the pricing up for larger boards/more layers and compete in there .. for example the Eagle tiers pretty much suck - there's a lot of room for someone to come in with a good tool for free to get people to load it and then pick their tier.  You minimize the support options so it doesn't cost you more headcount and then just let people upgrade through the tiers.

Fortunately Altium canned their CEO.  Unfortunately they have a history of being colossal screw-ups when it comes to product launches.  So we'll see.  I think Dave pretty much nailed it though on where/how Altium can really grab the lower end of the market by the short and curlies and shake it up.

[resistor]

I think the important thing to remember, which it's easy to lose track of as a professional, is that for most hobbyists the convenience of the tool is not the highest priority.  If a hobbyist only makes a handful of PCBs per year, it doesn't really matter whether it takes them 10 hours or 20 hours to design it.  A professional PCB designer has a business interest in being about to turn out design revisions as quickly and conveniently as possible, but the hobbyist doesn't have that pressure.  The hobbyist only cares about being able to do it at all.  Convenience is an added bonus.

[walshms]

I don't see a free altium doing 4 layers.  It's just too much flexibility.

...

Altium isn't going to give away MORE than Eagle.  Why should they?  AD is clearly better than Eagle the problem is the price.  If I had to do it I'd match what Eagle can do but do it via AD.  ...

Four layers ought to be a standard starting point... for the kinds of things people would want to do today, that's almost a necessity, not a luxury.

It really all depends on what they're trying to accomplish.  If they just want to get some seats, fine -- the approach you're advocating is the one most companies would take.  Sure, much better product, and yes, they could certainly do no more than Eagle and likely win in the longer term... but they'd still have a relatively small user base.

If they want to own the market, that's just not enough.  If they want to do volume -- which is where the real money is -- I think they need to get aggressive.   I don't know what their installed base actually is, but I think it's perfectly reasonable to think about what it could be -- and that's probably at least an order of magnitude increase at a more accessible price.  Maybe two orders of magnitude if they really do it right. 

Now -- if you reduce the price by an order of magnitude and get two orders of magnitude increase in sales... I know what I'd do. 

[0xdeadbeef]

I think the important thing to remember, which it's easy to lose track of as a professional, is that for most hobbyists the convenience of the tool is not the highest priority.  If a hobbyist only makes a handful of PCBs per year, it doesn't really matter whether it takes them 10 hours or 20 hours to design it.  A professional PCB designer has a business interest in being about to turn out design revisions as quickly and conveniently as possible, but the hobbyist doesn't have that pressure.  The hobbyist only cares about being able to do it at all.  Convenience is an added bonus.

I disagree. The opposite is the case. If you work every day with a tool, you can learn your way to achieve things, even it the complete GUI is a little weird. If you design just a few PCBs a year, you need an intuitive GUI, else it is much too frustrating to learn weird non-standard ways to reach simple things again and again since you forget all of it in some weeks. That's why I like DipTrace so much more than e.g. Eagle. In DipTrace, most things works as you would expect from any other windows program, in Eagle, everything still works a bit the way that some electrical engineer thought 20 years ago it would be easy to implement in a completely self-designed GUI.

[JoannaK]

Autorouter is a must

Totally disagree on that one - a good autorouter takes a lot of effort to develop ( or buy in), and will only be useful to a small proportion of the low-end market as it's of limited use until you can afford enough laers to make it work well. Resources should be concentrated on making manual routing as quick and easy as possible (e.g. interactive routing).
Probably best to be purchasable as an extra (as opposed to being part of a big step up with other stuff).

This I agree with Mike. Autoroute can be good for simple boards (as a ballpark: like Arduino-level, with plenty empty space), but for  more complex layouts it's usually more-or less useless alone. When the board complexity increases, designer will end up spending more and more time o patching up the results and re-doing everying by hand anyhow.

That's said, some high end autorouters (specttre, fireblaze) can archive quite some amazing jobs, but usually the setup-time (all parameters needed on trace/netlist basis) take a *lot* of time and expetise.

As for real world pics.. These are top and bottom pics  of over 10 years old board.. made with Freescale (orig Motorola) MPC555 BGA -microcontroller, to 4 layer board. Entire board size is 10cm*10cm and it's quite well packed.

Routing is (by layer) horizontal, gnd, power, vertical. The wire/cleareance widths are both 5 mils if I remember correctly. All passves and some smaller support IC:s at the bottom side, top side dedicated to bigger chips and connectors.

[JoannaK]

I think the important thing to remember, which it's easy to lose track of as a professional, is that for most hobbyists the convenience of the tool is not the highest priority.  If a hobbyist only makes a handful of PCBs per year, it doesn't really matter whether it takes them 10 hours or 20 hours to design it.  A professional PCB designer has a business interest in being about to turn out design revisions as quickly and conveniently as possible, but the hobbyist doesn't have that pressure.  The hobbyist only cares about being able to do it at all.  Convenience is an added bonus.

IMHO for hobbyist ... most important things are learning and archievement. What's there is to learn on Autoroute, besides pushing a button? With good hand-route tool even the hobbyist could learn about layout, board design, signal integrity, astethics etc..

[arvidj]

A really dumb question that may get me kicked off the thread. Just trying to think "outside the box" ... no pun intended.

One of the constraints often mentioned as based on board size as defined by length and width.

Would an area based constraint be a feature rather than a fixed length and width constraint?. As in "Eagle has a 80 sq. cm" free version, shape it however you would like [SIHYWL]". For $169 it has a 160 sq. cm version, SIHYWL.

The thought being that if Altium is looking for some way to constrain the application yet not look like they just copied the Eagle marketing material, would an area based constraint be an opportunity to achieve that goal and also be of benefit to the end user?

[walshms]

A really dumb question that may get me kicked off the thread.

I certainly hope it doesn't get you kicked off the thread.  Me either, for that matter, since I'm replying. 

Area is all that should matter.  This is how it should be done by everyone.  Who cares if the board is exactly whatever size on a side?  That's an artificial and arbitrary nonsense IMO.  Not everything you might want to do fits neatly into a rectangle.

Let the PCB houses worry about that -- if you design something that requires a lot of routing out when drill time comes, you pay for it.

Of course, it's a lot easier to calculate area when you're not dealing with anything but parallel lines... but that's down to lazy programmers, I think.  It's just not that hard to do... and even if it were, I think it wouldn't be such a big deal if they just took parallel lines on an outline and calculated that way.  If you have a board that requires a lot of cut-outs, then maybe you should rethink it anyway.