Free Altium is Coming

[Simon]

My other thought that if you are small business who needs to get things private,  you need to come up with a workable business plan that can buy a real tool.    Expecting a free tool for your business without having any restrictions is not reasonable.

Now,  it looks to me that they are planning on options for closed design but right now it is looking like a community product to start with.

No one is arguing that a business wants a free software. I have purchased the 500 pin version of diptrace, I'd not have wanted to buy the full version as I don't need it.

It sounds like altium maybe leaving the door open for small business and new startups, we shall see.

[Wilksey]

Maybe not even startups, hobbyist at a push I would say by all of the stories.
I get the feeling if what has been discussed is true, that this "free" version is merely an introduction to the Altium software, I could imagine myself getting very pee'd off with it, maybe thats because I use the "full" version.

The interface from screenshots look like it was made by the same guy that did AutoTrax 2020, some kind of beastly toolbar which takes up 1/4 of the screen space!

I will of course give it a whirl when it does become available, but I can't see the benefit over using this vs Eagle or Diptrace, or even KiCAD, apart from the fact it introduces you to Altium, but is a "free" Altium necessity or a nice to have? I mean to require the power of AD you need to be serious about what you are doing, not just hobbyist, which is why I use Eagle as well as AD as the licenses are so expensive for AD to become even conceivable for hobbyist use. I understand the need for an Altoum "viewer" perhaps, so we can look at designs and print them (not even plot them to Gerber), but thats about all I can see beneficial.

BTW, Dave, do you use the latest and greatest AD or an older version? I have found 14.x.x to be a bit bloaty.

[nixfu]

>Expecting a free tool for your business without having any restrictions is not reasonable.

Yeah, except KiCad has no restrictions and is free.  If the community grows a little and a few improvements are made,it could become pretty dominant.

[mikeselectricstuff]

What exactly does that mean ?

2.)   When you create a part,  you are creating an entity that is keyed to an specific manufacturer part number.    (The correct way iMHO).    So,  you do not a create a 555 timer.  You create a TLC555CDR with all the metadeta.

There _HAS_ to be a way to make generic parts, otherwise you'd need to have seperate part numbers for every value of resistor, capacitor etc. which would be utterly dumb

[Precipice]

There _HAS_ to be a way to make generic parts, otherwise you'd need to have seperate part numbers for every value of resistor, capacitor etc. which would be utterly dumb

No, this actually works fine. Especially when you've got a library of R_0603 you can pick from, which I'd imagine the Octopart tie-up (and Altium libraries) will provide.
If it's a different part, it's got a different design ID, with all the attributes bound to that part, which go into the BOM.
If you want to edit it in place to pick a different part, that's easy enough too. 
I see people kicking off about this, but, to me, it's the right way to do it. One reel, one part type. It's not a resistor that happens to be 12K, it's a very specific 12K 0603.
(If I don't know the value yet, it's an R_Undecided, which sticks out like a sore thumb at BOM time, and very rarely gets placed by mistake

[zapta]

My other thought that if you are small business who needs to get things private,

Do you know when things become public? continuously as you work? Only at certain events (e.g. gerber generation)? Manually when you declare a version as stable?

My hobby work is open against github but I do have the ability when to publish. Same when authoring a post in these forums. Otherwise you get partial and broken files.

[ehughes]

There _HAS_ to be a way to make generic parts, otherwise you'd need to have seperate part numbers for every value of resistor, capacitor etc. which would be utterly dumb

Making a unique part number is the proper way to implement a library workflow.
Why would put a generic part?  You have to know a value and part number to actually build something.    You just do at the front end to make your BOM generation sane.   If not,  you are repeating a lot of work!       

Think about this.... Is there really a generic part?     When you create a BOM to get a board stuffed,     You can't list a "10k"

If you actually get this made,  you have a line item for an a Panasonic "ERJ-3EKF1002V"   and possible substitutions.      It really takes little extra work once you get a good workflow

I always thought the generic part only kicks the can down the road.  At the end of the day there needs to be a source-able part.

With the full version of Altium,  I made a database library that has all resistors and caps in all values and all packages from 0201 to 2512 in all values.    This is actually quite easy with the DBlibs.   Once you figure out the part number scheme everything can be generated via a script.  Since different parts are just lines in a database or spreadsheet,   it is easy to enumerate all possibilities

We even made a part scraper that went to Digikey grabbed all the SMT diodes, fuses, etc for certain manufacturers.    We got all the SOT-23s,  SMAs, SMBs, etc.

So,  in the end it makes for much cleaner design artifacts.

That all being said I believe that Altium is including there generic passive libraries from the vault.

[Bassman59]

Eagle is the hobbyist standard today because it was free for small simple stuff a decade ago.

This is the whole truth. The rest is commentary. (Now go read it.)

[mikeselectricstuff]

Generic parts make complete sense, they have a value attribute to specify the actual value. That way you can treat them as the same or different depending on who is using the data.
The problem with making parts too specific is it makes it harder to select between equivalent parts when ordering when items are out of stock, discontinued or more expensive than equivalents.
Another problem with being too specific is making sure everything is defined in a consistent way. Maybe some sort of heirarchical definition system would be useful.
Another issue is you don't want to have to tell your pick/place machine vision system about all the different parts - all it needs to know is "0603 resistor".

Specifying generic resistors down to a specific manufaturer part number is just insane. More so when you are trying to target a wide audience, who will be buying parts from all sorts of places.

Making a unique part number is the proper way to implement a library workflow.

there is no such thing as "Proper" or "Correct". Only more or less appropriate to a particular set of requirements
 

Why would put a generic part?  You have to know a value and part number to actually build something.   

In many cases you don't.
If I don't care which 10K 0603 1% resistor is used. I do care if purchasing or a subcontractor delays produciton because a specific make of a generic part happens to be out of stock.

Think about this.... Is there really a generic part?

Yes, many.

    When you create a BOM to get a board stuffed,     You can't list a "10k"

I can list a "10K 1% 0603"
At the very least the "0603 1%" and the "10K" should be merged at a late stage into a specific part number, not at an early one at library creation.

What might be nice is to have an automated way to generate manufacturer part numbers for generic parts, so given "0603 1%" and "10K" it would generate the correct part numbers for a few manufacturers, then filter by stock and price to select which one to order.

Of course it is important to know which parts are generic and which ones  are more specific.
 In practice we're mostly talking about R's and C's, but these make up a rather large proportion of parts (as opposed to part types) used in most designs.
 


 

[Simon]

Eagle is the hobbyist standard today because it was free for small simple stuff a decade ago.

This is the whole truth. The rest is commentary. (Now go read it.)

Altium want it both ways, they want to be seen to be giving something away but are scared shitless of loosing a couple of cents profit. They want the big name but they won't pay for it. It's just PR wankery. I'm not buying it!

[mikeselectricstuff]

Eagle is the hobbyist standard today because it was free for small simple stuff a decade ago.

This is the whole truth. The rest is commentary. (Now go read it.)

Altium want it both ways, they want to be seen to be giving something away but are scared shitless of loosing a couple of cents profit. They want the big name but they won't pay for it. It's just PR wankery. I'm not buying it!

The whole thing will live or die on what lies on the upgrade slope between the free version and full AD.

[zapta]

2.)   When you create a part,  you are creating an entity that is keyed to an specific manufacturer part number.    (The correct way IMHO).    So,  you do not a create a 555 timer.  You create a TLC555CDR with all the metadeta.

Can you define your own manufacturers and part numbers? For example 'Generic'.

I find it useful in Eagle to place a resistor/capacitor in the schema and assign values in the schema without dealing with libraries.  Often I deal with the various footprints only when the schematic is stable.

The least and the latest I need to deal with libraries and exact specification the better it is for me (I use the schematic capture as a brainstorming napkin, iteratively refining the design) so a loose flow will be a better fit. I also liked Free Electron's idea of being able to modify the library in the schematic capture, for example by repositioning pins around the block.

(all the above is for hobby work).

[T3sl4co1l]

There _HAS_ to be a way to make generic parts, otherwise you'd need to have seperate part numbers for every value of resistor, capacitor etc. which would be utterly dumb

Making a unique part number is the proper way to implement a library workflow.
Why would put a generic part?  You have to know a value and part number to actually build something.    You just do at the front end to make your BOM generation sane.   If not,  you are repeating a lot of work!       

Think about this.... Is there really a generic part?     When you create a BOM to get a board stuffed,     You can't list a "10k"

If you actually get this made,  you have a line item for an a Panasonic "ERJ-3EKF1002V"   and possible substitutions.      It really takes little extra work once you get a good workflow

Another reason came up recently, and puts yet another big heavy nail in the notion of prepared libraries...

Second source.  You've designed the board, put together the BOM, and manufacturing says they can't get X.  Or, they can only get so many of X, and the lead time is impossible.

The library-centric approach would require you to add those fields to the library part.  But this is against the whole concept of a second source: what if that temporary source dries up?  What if it's project specific, and you must call out substitutes on a need-to-know basis?  What about when the substitutes aren't exact, so you're making a direct judgement about the suitability of a part for a project that may not be suitable in other projects?  Putting that in the library could be disastrous!

Having a part in the library to begin with, you encounter the same pitfall, because the part originally selected for that library item will eventually disappear.  You don't want to keep using the same part for five or ten or twenty years just because it's in your corporate libraries.  That leads to all sorts of suboptimal decisions.

Tim

[Bassman59]

Another reason came up recently, and puts yet another big heavy nail in the notion of prepared libraries...

Second source.  You've designed the board, put together the BOM, and manufacturing says they can't get X.  Or, they can only get so many of X, and the lead time is impossible.

This isn't a schematic/PCB-design problem, it's a purchasing problem. Hopefully before you design the part in, you've got your purchasing person making sure you can buy the parts? (And yes, for the one-person shop, the buyer is the PCB designer.)

The library-centric approach would require you to add those fields to the library part.  But this is against the whole concept of a second source: what if that temporary source dries up?  What if it's project specific, and you must call out substitutes on a need-to-know basis?  What about when the substitutes aren't exact, so you're making a direct judgement about the suitability of a part for a project that may not be suitable in other projects?  Putting that in the library could be disastrous!

What my various employers have done is to split the difference. Schematic symbols don't have vendor part numbers embedded, they have company part numbers. A BOM is generated from the schematic, and that's run against the list of parts which includes vendor part number(s). If purchasing notifies engineering of a shortage, someone will look for a substitute which can get added to the allowed parts for that company part number.

Having a part in the library to begin with, you encounter the same pitfall, because the part originally selected for that library item will eventually disappear.  You don't want to keep using the same part for five or ten or twenty years just because it's in your corporate libraries.  That leads to all sorts of suboptimal decisions.[/quotes]

That's true, and it's why it's incumbent on the designer to work with purchasing during the design phase. And it helps if engineering is kept aware of which parts are going into NRND or last-time-buy. The purchasing staff should tell all of engineering, "hey, someone put an XXXX on a board, that's NRND, what do you want to do?" and then engineering works it out.

Yes, I know that purchasing departments can sometimes be opaque. 

But yes, there's always the question about using something new because it's new, or using the old part because it's already in the library and we have a few reels of parts in the warehouse.

[Precipice]

Second source.  You've designed the board, put together the BOM, and manufacturing says they can't get X.  Or, they can only get so many of X, and the lead time is impossible.

The library-centric approach would require you to add those fields to the library part.  But this is against the whole concept of a second source: what if that temporary source dries up?  What if it's project specific, and you must call out substitutes on a need-to-know basis?  What about when the substitutes aren't exact, so you're making a direct judgement about the suitability of a part for a project that may not be suitable in other projects?  Putting that in the library could be disastrous!

Once a project is off for manufacture, I regard it as frozen - unless there's an ECO, what's designed should be built. Any second source and obsolescence issues are to be managed on that project, by me and the manufacturer. They'll be rolled back into the library if it makes sense.
There's a unique in-house code for every part, which is what is used as an index for external assembly shops / purchasing teams. There's also an explicit manufacturer and manufacturer's part code.  There are also courtesy catalogue codes, to let me easily order for prototypes. It's possible this should lie in an external database, but I'm happy enough to have it in the library.

No doubt it's imperfect, but it does work OK, and manufacturers cope well.
So- in my opinion and experience, it's not insane to have a library part per BOM line.

Mike - the footprint is just another field in the library part. All my R-0603s are called R_0603. This works perfectly for PnP.

[IanJ]

I find it useful in Eagle to place a resistor/capacitor in the schema and assign values in the schema without dealing with libraries.

Same here..........I coudn't image having to go back to the library to place each and every different value of resistor. With Eagle all you do is go to the library once, place the part on the Schematic then duplicate to your hearts content, and just change the VALUE attribute......and from there everything is reflected in the BOM.

Ian.

[Rufus]

Altium put a lot of effort and money into 'Altium Live' their Facebook for EEs  - it's horrid. They put a lot of effort and money into 'Vaults' their cloud data storage and management stuff - hardly anyone uses it. They put effort and money into their 'Content store' iTunes/App store for EEs which has never had an item for sale.

It seems the only way they can get anyone to use this stuff is to embed it in a cut down version of their main product (which people do use a lot) and give it away to people who don't have any money.

The last figures I found for sales of Arduinos (darling of the maker scene and feature of their screen shots) was 700k with another estimated 700k of clones and derivatives. 1.4 million Arduinos sold (bit out of date, about 3 million Raspberry Pis sold but that is more computer than electronics). Microchip alone sell 3.2 million microprocessors a day. Makers are but a pimple on the arse of the global electronics market, a very vocal and publicised in 'new media' pimple but pimple all the same.

I just can't see a good business plan for Altium in this.

[etzz]

I just caught up with this long topic, and was about to go away, just an old curmudgeon in disgust, but Rufus' post made it all worthwhile.  HA
Yes, they are a pimple on the arse, and one devoid on decent funding.

The future source of great ideas?  I know that large companies regularly troll the "maker community" boards looking to commercialize the good ideas.
Eric

[mikeselectricstuff]

No doubt it's imperfect, but it does work OK, and manufacturers cope well.
So- in my opinion and experience, it's not insane to have a library part per BOM line

Not insane but far from ideal.
Which makes more sense, let's say  for resistors 0402 to 1206, E96 over 6 decades, 1% :
2,304 individual library items, or 4 items with an attribute that gets merged at the latest possible time?

 

[Precipice]

Same here..........I coudn't image having to go back to the library to place each and every different value of resistor.

This really isn't onerous in Altium. I don't know how much hassle it is in other packages, but flow would be:
Shift-drag to copy the part
Double-click it to open the part dialog
change the design ID if it's obvious (like just changing the value), and you're done
or pick another part from the library chooser, if you want something more elaborate.
Honestly, 10 seconds, and you've got a part from a controlled library on your sheet. There are annoyances with Altium, but this ain't one...

You're used to something else, obviously, but this works, and works well.

Edit: Massive libraries just aren't any kind of a hassle. Database controlled, nice and quick, easy to search. Who cares if there are loads of parts? It's not as if you have to create them or tend to them.

[mikeselectricstuff]

Honestly, 10 seconds, and you've got a part from a controlled library on your sheet. There are annoyances with Altium, but this ain't one...

10 seconds just to add one part...? That's just ages....
One big advantage of a part+value model is that instead of going to a library, you can  quickly make copies of parts that are already onscreen and just change the value. Maybe 3 seconds.

[Precipice]

Maybe 7 seconds, having just tried it.
Shift-drag to get a copy of the component
double-click to get the parameters
edit field
done, get back in the schematic.

It sounds like the same level of difficulty - ie, none.
If you want the same component, then it's just the shift-drag.
I'm OK with this, and it really, really isn't a slowdown.

(What I'd like from Altium is, when going through the place->part dialog, is to be presented with a list of parts already in this design, rather than most recently placed, possibly in another design, since I often run many jobs at the same time. Being hinted to reuse parts already on the BOM would be handy, if I'm placing something I don't particularly care about.)
But all this 'ooh, picking from the library, that must take ages', Nah.

[ehughes]

For me,  this flow ensures that I can generate a BOM from my outjob that has

1.)   A fully populated Assembly Sheet that references a real manufacturer part number.   An out of stock part number is a *supply chain* problem, not a *design problem*.   
2.)  A list of Alternates
3.)  An order list for suppliers.

That can be then sent to purchasing to manage the acquisition.  In many cases, "I" am purchasing for the prototypes and I really like having a ready to go BOM to automate my orders.   (From Outjob to Order is very quick...)

If in the case you have multiple parts that can work on the same land pattern,  you need a internal part number for a layer of abstraction (which cannot be handled by Circuit Maker).       

For me,  it is more than value & package.     Value, Package, tolerance, voltage rating, temperature coefficient, etc....  At that level you need a real part number.

It is just easier to spec a manufacturer part number as the primary key.    If that doesn't fit then the next level of abstraction is a company part number (which then you become the "manufacturer" as you are adding a value add stage of sorting a input product).

Having a schematic with a generic resistor with a value just means  someone has to do a bunch of manual labor to fill in a BOM later (i.e. kicking the can down the road).   For something simple,   not a big deal.   When you have many  line items,  I know I don't want to do it.  I like having my tool do it for me and prefer to do the work in the front end to save me in the backend so everything is automated in the design tool.  I am unware of any way (without a custom script, etc) in Altium to Specify something generic and have it fill in the BOM with a *orderable* part number.

Now, I understand the desire as it just shifts the problem to another point in the process.    Everywhere I have been that has a controlled,  lifecycle managed design flow does it this way.     Everything has a part number...  Doesn't matter if it is a resistor, a piece of wire,  a dildo.    BOMs need real data for a proper design to print package.


I think part of the problem is that some are only thinking of how to handle it in tools like Eagle or Diptrace.    In that case, yes, making a library of resistors is difficult as their model can't handle this level of abstraction very well. (i.e. one symbol  being linked to a table in a database to generate a large amount of parts).

For Altium, It is simple to make parts database to handle this.

The guys working on Circuit Maker just need to do some more work to figure out how to cover the 6-sigma of use cases so libraries are sane.    If you can make the 6-sigma of users happy,  then your job if done :-)

 

[os40la]

I think some people seem to be missing the bigger picture here..... Like CircuitMaker new on the scene causing DipTrace, Eagle, Etc. to compete with offering better freebees for us Hobbyists.. 

It seems to me that some people on here seem to want to use CircuitMaker as a cheep tool for their business and can't or don't like the way it may work for business processes. "Hello McFly"... Did they not say that they are Marketing for the MAKER community? 

Does anyone wonder why Microchip offers a PIC C compiler for free now? Do you think it has something to do with other free tools from their competition?

[mikeselectricstuff]

For me,  it is more than value & package.     Value, Package, tolerance, voltage rating, temperature coefficient, etc....

Except a large proportion of passives will not need that level of detail, and where they do, that's when you specify a particular part, otherwise it's a jellybean.