A presentation given at the 2012 Electronex show in Sydney by the SMCBA.
IPC-2581 Open, neutral global standard for efficient PCB design data transfer.
Presented by:
Gary Ferrari - Executive director and co-found of the IPC designers council.
Andy Kowalewski - IPC Designers council Master Instructor
EEVblog Main Web Site:
http://www.eevblog.com
EEVblog Amazon Store:
http://astore.amazon.com/eevblogstore-20
Donations:
http://www.eevblog.com/donations/
Projects:
http://www.eevblog.com/projects/
Electronics Info Wiki:
http://www.eevblog.com/wiki/
IPC-2581 Open, neutral global standard for efficient PCB design data transfer.
Presented by:
Gary Ferrari - Executive director and co-found of the IPC designers council.
Andy Kowalewski - IPC Designers council Master Instructor
EEVblog Main Web Site:
http://www.eevblog.com
EEVblog Amazon Store:
http://astore.amazon.com/eevblogstore-20
Donations:
http://www.eevblog.com/donations/
Projects:
http://www.eevblog.com/projects/
Electronics Info Wiki:
http://www.eevblog.com/wiki/
Okay, uh, I'm actually filling in for someone on this Consortium in the 2581 but I have been somewhat involved uh with it in the in the sidelines. Uh so I have a short presentation that will introduce you to what's going on. uh and basically uh. the 2581 is a data format.
It's not software, it's a format. uh and it's used to communicate a lot of information touch base with some of that information within the slide set. I Do want to give you a little bit of a history of data transfer and why are we looking for another format? Why are we doing what we are doing as a group, who the players are and some of the Uh history of data transfer The Good The Bad and The ugli and Clint Eastwood is not part of that right? First thing I Want to talk about as little is the design process ba. Basically the Uh design process.
You design your board, you do your then you pass it out to a prototype manufacturing and then from that you might be doing some in circuit test and assembly. It goes right on through the whole product development cycle. And what be nice is if we had one data Highway that covered everybody's needs and it's a two-way street you you place something down some information. Someone needs that information to take it off.
They put new information on uh as we stand right now, every one of those Evolutions have their own formats, their own data uh format, some It's Paper some of it's electronic Etc So we really would like to have something that's more economical and let's see what happens. The process involves. Uh, today? Uh, multi-format uh delivery? Well, that means very simply is that I have a formats for everything I have the RS 274x for for with the Gera files coming out, uh I have maybe some excellent drill files in their format I have the D350 uh Six format for the electrical test and maybe I have the drawing information for the 2511 and asky PDF Hpgl, Jpex. all these different formats and and every time you have a different format, you have to have a translator to create the format and then you need something to to create a back into the format that you need in a particular piece of equipment.
So there's plenty of room for error. and there are errors. Errors happen every day. and uh.
I Remember, at one point in time uh, we did analysis, it was over over $250 million a year goes down in the waste can because of poor data transfer and that's basically poor communication. Downstream So we've been chasing this thing for quite quite a number of years. Okay, and today you know we're trying to do more with less and and and data transfer and errors and things like that are not helping us at all. Uh, the Cad companies have been very reluctant in the past to uh apply any resources to post processors.
They make money selling you the basic tool all right. and uh the basic uh, post processes are done by the entry level uh programmer just to get used to using the the tool and sometimes that data is just not not correct. Scale is off. Uh elements are in the wrong place missing Etc So while companies undertaking do more or less they want to reduce Reliance upon paper. Okay, uh. we're looking to go more electronic data transfer we see in our shop. we see a lot of electronic d uh transfer just about all of our inputs coming electronically. Uh, we have many solutions uh, but they all require custom extensions and and they're costly to maintain.
So at the OEM level the fabrication level at you know the contract manufacturers, all of those have their own specific requirements and and we need to streamline that process somewhat and and it gets a little bit worse as well. Let me give you a little bit of a data transfer standard history: Uh way. Back in 1970 I know he you weren't born then Okay, but Back in 1970 Uh IPC Uh. put together this IPC D350 format.
Uh, it was an intelligent version of Gerber format. Now Gerber format is very simple. It's the least common denominator in data formats. It's used to drive a plotting machine.
It's proprietary format says turn the light on, drag it over here. shut the light off. I Just created a line on a piece of film. Okay, uh.
we needed more information and the D350 series of formats added in the aperture list. They added in a bunch of other items. That was okay, but Back in 1970. Why? it didn't really take off? It's very simple.
The Cad companies did not want to support any kind of a format that would allow you to take your your work from my to from this tool and put it into someone else's tool they wanted lock you on in so they didn't support any kind of a a data format that would allow interchangeability between Care Systems That just that wasn't in their CS in 1970. In Uh 1980, the US military adopted the Mil standard Uh 275 designed for for for the boards and data transfer uh in an attempt to also standardize the information that's being placed out there. the IC I in 1988. Uh, they adopted the 350 and anything that goes internationally with the IEC usually gets their own number, their own numbering system, and a lot of times it's the same document that we have I Know at IPC we we we we submit W of our documents I for adoption and and and they had the equivalent Uh which is the 61882 D1 in 1988 for that 350 document uh in 19, 92 and 94 in that time frame, the Jpca, the Japanese printer circuit Association uh.
They created the K Toad transfer and they had different concepts for codes uh and and that that was Uh. released in the I as well as an IEC 611 18210 and then we come up into Nowaday 1998. Uh IPC had the 25x 1X series the Gen Cam. Now Gen Cam was a format.
Genad was a format that was owned by a private organization. They donated it to the interest into the industry and the reason we selected that Gen CAD format is because most of the electronic side of the business through the the the Uh Uh engineering side was all all in in this particular type of format and it would take very little for us to adopt the basic format and then add the things that were unique to the printed wiring board side of side of the business. So back in 1992 we started Uh I'm sorry back in 1998 Uh, we built that Gen Cam. We called it Gen Cam so people had a relationship to the the other one the Gen Kad product that was sitting out there Uh and then we formed some revisions on that and then came the problem. Okay, it's not on the slides and I'll give you the problem. Uh, we had a company out there called Valor and Valor created Uh has the Uh the cam station software for most of the board shops. about 85% market share worldwide. Uh and they have a format called ODB Plus+ All right and it's a robust format Uh and it's an intelligent format.
It contains a lot of the database. The only thing is it it it's it's not really structured very well internally. Okay, Number one: Number two. It wasn't as robust as Jen Cam and number three, it was proprietary to Valor.
Okay, whereas the IPC format is industry owned, it's developed by the industry and distributed by the industry. and it's all free of charge. Okay, uh, uh. they they went to to uh, the Nei organization.
They had us put a hold on the Gen Camp product for 5 years. Okay, and uh. basically the Marching Ules are very simple. make a new format that makes ++ look like Gen Cam and then we can adopt it.
So after all of that and we enhanced it, we enhanced the format. Uh, and we came with the 2581 series. Uh, in in 2003 Immi recommended that 2011 we we we formed this consortia and so now you have a a 2581 series Uh formats that gives you the best world of Gen C and ++ We'll talk a little bit more about as you go along the line. Uh, it eliminated the the Uh the dedicated company's ownership.
However, Uh, ODB Plus+ still exists. Valor still exists under different names. We'll talk about that in a little bit. Um, they support the 2581 however.
Uh, they will modify their ODB output to kind of lock you into their their types of of thinking and we'll go through the merges and all that other stuff later on. But basically then we have the Consorti 2011. We'll talk more about that. Uh, this was what we were talking about a little earlier.
We froze the uh uh, the revision for 2 years to let the C companies or let the software writers write up the format. So let's face it, I have a format. You know what a format is. All it is is a structure.
Okay, each one of these little squares on the ceiling could be a format. So maybe I'll take all of the whole information. put it in this square like a mailbox. I'll put all the line information over there.
I'll put the test information over here and on and on. all structured format. So now if if if your company needs the whole information, you'll go to that particular box. You pull out the whole information and if you need to do something, you put in another thing and the guy in the back he needs that information, he pulls it out and put something else in so everybody knew where it is. It requires the Cad system to be able to Output all of the information within the file to the appropriate box in the mailbox. and then it involves uh, your the the manufacturer assembler, the tester to have software that will also extract it from the prop box and then move on forward. That's basically a data data uh transfer, uh type of uh road map. Okay, in the meantime, uh, the Committees looked at, looked at what was out there and it looked like the XML type of description data transfer description would be taking hold.
Okay, and then the Worldwide Web committee. Finally, this locked in XML schema as a way that they were going to transfer all kinds of data. So so we figured listen for for an open- Source type of of format and then we're going to make the new version all XML schema based so that we're compatible with the rest of the world and everything else that's going on data transfer. So that's what we did and what that scheme permits us to do.
It allows us to look at the numerical, the Ang, angular and other descriptions and we can check for for correctness through software compliance. So we're We're using the tools that everyone else is using in order to create data. Uh, data transfer software. Uh, the need for data intelligent data transfer as we've moved on over the years is because we're trying to get more efficient.
Trying to get more efficient efficient. We have much more complex uh electronics that we're trying to uh uh, document and transfer the data. So again, we want to have that intelligent data transfer. Um, I I'll give you a couple of Quicky items on what I feel this means to me.
Okay, let's compare it to Gerber how many you familiar with Gerber most of you are familiar with Gerber Okay, Gerber being the least common denominator. Will Gerber tell you what the whole sizes are I'm going to tell you what the whole sizes are, the tolerances or the locations. Will Gerber tell you how many boards to order? How about the stackup? Would it tell you what the layer stackup is, what the copper is on each of the layers? probably not. Will tell you which conductors have controlled impedance? Will it tell you about the the The Beboy test fixture, what kind of fixtion, and all everything about building that or or or the uh uh ICT test fixures.
We give you all the test vectors and on and on and on. All of that information is information that we require someplace. A long development cycle. All right.
The TW The Intelligent Data Transfer will contain all of that information. It allow you to have all the information. you can segment it and only send the fabricator what he needs and assemble what he needs. Etc Or you can send a hold on file. What's nice about it. It's good for archiving. Uh, your your product for long term. Everything that's NE necessary to build that product would be in this archived file so you can bring it back 20 years from now and duplicate what you had to do as long as the materials and components are still available.
obviously. So what's the Uh solution to the problem? The 2581 is the single source file containing the entire PCB structure. everything for parts of Erp systems C data cam, data models, you name it. Uh.
The improvements over the existing formats you have very rich schema defined which defines all of the the intricate parts of everything. Uh, it could support a part polarity, part orientation. Uh gives you consistent definitions Uh improves the support for drilled and Mil content so blind and buried vs. back drilling VG Groving slots all that stuff is contained.
You won't find that in a Gerber file at all. Okay, and a lot of that you won't find in an ODB plus file either. Okay, uh, However, it's supposed to be totally Rob robust. Okay, improve embedded components, stacked component support.
It'll give you improved support for your functional in J tag and ICT uh and it gives you automatic viewing with the with the with Uh viewing software and Andy will show that afterwards to show you some demonstrations of that. Uh. Summary Benefits: I've taken this from several presentations that I didn't create so you know I'm guessing at some of it, but we're we're doing fine. Uh, the benefit summary is it's open.
It's vendor neutral and standards based. When I say it's open open, it's owned by the industry IPC is an industry based trade Association that's Mcba is part of IPC Uh, they have association with it so it's it. You own it. If you see something missing, you want something, you just let them know and it gets.
it gets placed in there whereas something that's private. If you want something you need something, it has to make a business case to get it. so it wants to be an O Open Source Uh improves your efficiency in design fabrication manufacturing by P by passing accurate data you don't have to assume when you're not losing. uh a a uh, an aperture list Something like that.
everything's coming out right directly from the data file. There's no misinterpretation. okay, uh and results in lower cost basically. Uh, your information transfer wants to go from what your cash system is is looking for all the way through to the electronic data transfer for side.
I'll just kind of skip this slide. There's a couple of slides in here that are necessary or not anyway. I Have a C specific design, I have a C independent golden file, and Manufacturing specific processes. The ideal ultimate goal is to create this data transfer and directly drive your end machine. Give you an example in in Bbard test. Okay, we say you have to have a 350 6 file which defines the test the test parameters for the for the Bebo tester. Uh, but the software that we sent the 356 doesn't really drive directly. the machines.
The test machines maybe the new ones accept it and drive it. But basically what they'll do is they'll have a 356 file that you gave them and then they'll take the Gerbers that you gave them. They'll extract a net list to compare that net list against a 356 and they'll make that net List look look like the 356. Then they drive their machine.
So we're trying to eliminate all of this conversion and driving and stuff like that want to drive it. Ultimately, just drive it directly makes a lot a lot of sense. I I would think okay, uh, this is a very busy slide and I'm certainly not going to point to the little letters cuz I can't see them I'm sure you can't see them from in the back, but basically it's a it's a road mapap of all the vendor merges and I just want to focus on one set of merges for you as an example and and why. we're at this stage with the Consortium Okay, uh, you had uh Valor with the ++ okay and then uh Valor was purchased by Menag Graphics Okay, now Menag Graphics now owns owns Valor now uh Valor was also supported by Cadence and several other CAD manufacturers.
They all had Valor output Valor helped them write it, do all that stuff like that. but when me took Graphics took it over. they're Cade manufacturer. their competition is Cadence and the rest of those guys so they immediately went to work to make it so that those other guys didn't have access to the right format cuz they wanted everyone to use the Obb that was the mentor format.
Well, you know, Cadence and Zukin and everyone else got together and said, listen, you know we're very nervous we have a private company going to own the only data format that everyone's using That's not really a healthy situation. We need to do something different. So they formed a Consortium and the Consortium consisted of a lot of supplies. You'll see them all coming up on the next several slides, so go up one more slide so at least you could be looking a little bit.
Uh, they formed the Consortium to adopt the industry standard of 2581 from IPC because it's owned by the industry and and in the Consortium they said okay, what do we need to do to make this thing successful Since we're adopting it Number one, we have to make sure that the format itself coming out of a particular piece of equipment is accurate to the standard. It doesn't hiccup if there's something missing or something that's additional. Okay, we have to make sure that the the software that does that is validated. We have to make sure that when it goes into the next piece of equipment or into assembly wherever it's going, that the input software can accurately record and take from that database all the information that it needs. It doesn't hiccup if it doesn't see something or does something in a special way. They have to make sure that all happens. So when They adopted it, they adopted and they said we're going to create the software. We're going to create the models we get the industry to support and effort to iron out all of the data transfer stuff.
So if you go, if you went to the last IPC conference down up in Shamberg Illinois a few weeks ago, if you're going to PCB West on the west coast of of the United States the week after next you'll see they'll have a booth and they're going to be exchanging a design file that they create. In Cadence They're going to move it over to Zukan. they're going to move it out to to Valor They're going to move up and down the whole product development system. Uh, system.
They're moving back and forth and any hiccups or enhancements. they're taking control and they're making their recommendations to IPC committee so that we can roll the dawn things in. So here we have an industry Consortium that's sitting out there and and basically it's a free Consortium There'll be a slide in in here how you can join and what the requirements are for joining because it's it's a freeb thing. so you have different work working groups within the Consortium.
All right, and the whole purpose of that evolution is to develop proposals and work with IPC to extend the standard to new technologies methodologies. Okay, and the leader for that is Gary Carter from Fij Jitsu He's the lead in that particular working group and Harris Nvidia and Erikson are all part of that particular working group. From the technical side, Uh, you had uh, you need to validate the 2581 files that were consumed by The Tool Uh, you need to identify the areas that needed to be extended and uh uh Cadence Hon, he's the lead Downstream Ad Diva wise artwork conversion Uh, Zukin and Easy Logic are all part of that that effort or that particular technical team. Then the awareness and promotion uh will come from Uh published articles in all the Publications at the at the different conferences PCB West Uh, Cadence uh uh and and and Hermon Sha is the lead Zuk and Down interet Up media.
They'll all be down at Down at the At the conference in BO showing how this thing and demonstrating how everything is working right? Uh, who can join the Consortium is open to any PC design supply chain company who's prepared to adopt the Uh consortia's goals and objectives. and that's just a slide talking about the goals coming up. Uh I Committed to the road map doesn't cost you anything. All you have to do is make a public station a public state says listen, we support the 2581 and we're going to try to help Pro promote it any way that we can.
either uh, ask for our suppliers to accept it in or or or put it, place it out whatever the case may be. Uh, and you get your logo thrown on there that type of thing I Mean it's just simple as enough you go a website and and you just uh uh, you go ahead and join it all right? Uh, actively support that 2581 one you you talk it up in your Dfms and your Fab you ask for it uh your your C company if it's not one of those your board suppliers, anybody else down the stream. If you ask for it and you ask for it long enough and and and you're threatened to not give them the job or something like that, they'll put pressure on their suppliers to give them the support. uh, the the software support that's necessary. All right. The mission statement for the Tech Techical Working group: Simple. The goal of validation process team is to Define and document the validation of the 2581 export against current exported data formats like Gerber ++ NC Drill Etc So we want to validate against those particular individual programs, make sure that it truly replaces those programs to ensure that 2581 data is identical and complete for the fabrication assembly and the test of printed circuit boards to promote the ensured industry adoption of that particular standard. We you know, with the 350 format and all the previous efforts, we really never had an industry Consortium promoting and pushing it and trying to do a lot of the work.
This is the first time in many, many years that I've ever seen where the industry actually getting behind and pushing it and they had selfish means because they didn't want one company to control everything, they wanted the industry to stay with it. And that team validation flow. uh from the creation tools whether it be Cadence Lego Zuken or Intercept or Alum whatever it may be, they will output the files in Gerber or the 2581 ++ and see they'll output all the files and then the validation tools from Ad Diva or about D wise or Downstream Those guys. they will actually read the data in and see if it works, see if it's doing the right thing.
Okay, uh. the technical team uh, will. The comparison process is going through the 2581 out through the Cad tool, outp, put individual files back through the cam tool and keep doing that type of of setup. So the road map.
The road map is is very simple. create and test a suite going into the push and test avilable all the way out to board fabrication. You can start from the beginning out to the end. It's as simple as that.
Okay, test cases and viewers. You'll notice there's a link that you know you notice there's a link sitting up here. Okay, and that that link will get you up to the test cases of viewers. You can see the results, what they had, and you could download the viewers I think Andy's going to be showing us one of the viewers.
Is that correct? Andy A viewers how how it's working and you can see what was done with the Cadence Nvidia and Fujitsu They've been transferring files back and forth between all of these groups. We really makes it pretty good. Okay Current Consorti: Uh uh uh. Members: uh. sitting up there, you'll see media companies? You? You? you'll see Nvidia Cisco Intercept Technologies Harris Corporation And and growing much growing since since then I mean my company's not even up there yet. We haven't done that yet. All right, it's just a matter of just saying hey yeah, I support it, They know that I support it. And then next several slides try to give you some kind of a brief description uh of of what's going on in there.
You have a A A A A standardization strategy uh, for your generic stuff, your administrative data, your drawings, your bedboard tests all way down through Library symbols or Alternatives and how complex the data is for those very simple dat is just real basic. Everything that we have today. All right complexity B Will now go with the 2581 Series where we're adding all the extra information here and then it's a a 2511 series series that's supporting that with the documentation. Etc Okay, and what they did with the segmentation: Concepts You have all your build materials hierarchical, uh, conductors, your drilling routing LS All that information and what they did they said okay, you have design, fabrication, assembly and test who who contains that information or or provides it or uses it and get the yeses and those so to try to give them a little bit of a road map of what's going on.
Uh, they all adopt one standard viewing convention that we've had for many years. You view it from the from the face you're looking at and go on down one, two, 3, four, five, six as you're moving on down. That is the viewing convention that's been around for many, many years. Uh, the standard and used dictionaries have been defined.
We had to Define things like butterflies and circles and Contours and rectangle rounds rectangle Corners All of that standard User Dictionary has all been defined as something standard because you know when you're dealing with electronic format, dealing with computers, you have to have a common name when you're transferring different data and everyone has to have the same thing. If you're going to have a slightly different name and you're going to have a slightly different name, we can't really get software to to correlate slightly different names. it has to be the same thing. So everything moves on down the line.
You know how it is. You type the wrong letter, the computer shows it back to you, says just spell it right. Okay, so it's the same type of thing and they cover all the different types of Primitives that you might see out there and we can keep on adding it all right. We can keep on adding it.
Uh, the new tool schema was released in the first quarter of 2012. All right, and they're still working on enhancing it. And then the last slide that we have is the standards hierarchy. Here are your data formats your 2581 series, The your generic requirements for the board descriptions and the administrative and Manufacturing data in in the two 82 implementation of Design Characteristics in 83 the 84 is the board data description and it keeps on going on down the line. all right. and your documentation. You 26 guys down that way so there's an awful lot that's in there. I'm give just a brief snapshot of what's going on I because I know it's late I'm kind of done with mine.
we're going to transfer the over to to Andy Andy will now show show you the actual viewer of the software viewing a data file from 2581 and I don't know what else you're going to do so you do the talking. Did anybody notice what was seriously missing on the Consortium members there? The uh, yes, the 600b gorilla of the Cad CAD bedos met Graphics they're not in the Consortium is El good sorry I I I need to correct that. Uh, me. the graphics is missing from the slide, but they were forced into joining the Consortium to protect their interests.
I Don't believe that they're providing any resources to further that they just want to see what it's going so they could conter somewhere. but they are part of the Consortium ah I stand corrected. Fantastic. What about Alum Are they part of it? Uh, I was I thought they were, but I didn't see them up there.
so I really don't know? Okay, so that's a good thing, right? Because traditionally Gary said traditionally the vendors didn't want to do this right because I could take an ALG of Mile and put it into Expedition and vice versa. The vice versa will no, no, no right, it costs 120 Grand or whatever it is costs an Al is 4 5 grand or whatever it is. they don't want that tool to go down there cuz that will do 90% of what this to do. Okay, so there's a whole bunch of issues about how this is not a good thing is from a centage point of view.
Anyway, we uh, don't have liftoff. Yes, we did. Of course we did. I Got to have a quick look at the standard.
This is free. Uh, you can download it from the from the Consortium. Somewhere in there was a um, um, somewhere in there was a um a link to the Consortium and you can download this stuff from there. The standard is open free, available to anybody and if you look through it typical standard, right? right? you got that.
Ask questions later you getting this. Okay, it's this is really great. I've had a quick look at this I Don't know it in great detail, but it's fantastic. Let me quickly describe how this works.
We have a file. It's a text file. It's basically a text file. XML Very well defined I have a pad and I Define a pad and it's got to search certain diameter and down here I have a pad that's a rectangle and I Define that as a circle, one, rectangle, one, something else, one, something else but further down somewhere down there I defined an integrated circuit and it says give me 4 of pad one and put them there and there and there and there and there and there and there and there, and so so on. So now I've got a little group down here. that's that calls up something up there, calls it up, inst, instantiates it 14 times and says here's an IC 14 pins. Okay, the further down in the file, it says put one of these down there at this location Okay, not only do you put it there, but rotate at 90 and flip it. Okay, so what you're doing is you're describing something once and calling it up multiple times.
Okay, if you think about that system, that's almost infinitely extensible right whenever I get a different pad shape and there's a whole bunch of them as you saw. If I get a different pad shape I can just slotted in there P it by a unique name and then just call it up every time I need it. Okay, so the people that thought through this thing I think I think did a really brilliant job. So it's really great and the standard is free.
This business rotation is a big issue. Okay, if I pick something up out of a tape this way and I rotate at 90 Did I rotate at - 90 or plus 270. What happens if I put it on the bottom side of the board and rotate at 90. Did I go this way or did I go that way? This is not a trivial exercise.
Okay, this is defining right and make and cast it in stone if you like so that there are no possibility of Errors right? Okay, so that's that's the. that's the standard. You can call it up. It's free.
Have a look at it. Okay I have a let me let me show you the file itself. Oh I digress I need to digress I I got so R in this thing I Have to show it. Okay, O21 Um.
O201 Okay, 402 02115 right? Just last week, a new capacitor from Marata Point4 by .125 mm. Okay, they have trouble with this Okay, and a lot of people are reluctant with this. What are we going to do here? Okay, this is about the size of a full stop on a sheet of paper. Okay, what are you going to do here I Had to show you that.
Sorry, it's going to be a long time before I design something like that. Okay, first thing to do we're going to this is this is the file Right. If you look at the file, well, it's a test case. Something rather XML Okay, this file is a one of the test cases that you can get off the Consortium website.
They have six test cases generated by different CAD systems, right? So the the file is presented like this, right? and it's very hierarchical. Okay, it's 44 MB long. Okay, so it's a reasonably serious file. If you want to look at it in text, you'll need a reasonably serious text editor.
Okay, I have a thing called notepad++ Search. Do a Google search on it that handles this really well. Not only that, it's you can uh, shrink things right? So what I can do is I can view uh, where are we Collapse Collapse level collapse level one right? So what I'm doing is this hierarchical file that has a multiple steps I can collapse it so I can look at the main headings okay and I have a slow laptop and it wish I wish it was quicker but it isn't. Uh, feel that or I did it wrong I did. No, it's still working. Okay, so this is serious text editing I guess. But if you look at the the structure of this thing very very hierarchical. Okay describes a circle, then it describes a rectangle, numbers against them names Etc so forth.
So it's sort of very very step stepwise. While that thing's working, it's probably using up every CPU cycle it can lay its hands on. This is a Viewer Free Viewer from Downstream Technologies Okay, this sucks right to use this thing is like something nobody's ever seen in their lives before. It's a shame at home.
I've got the wise viewer W that's from the Consortium as well. You can down they all that for free but you go through a little rigoll to get a license. Okay I've got a license at home? What? I didn't realize was you need a separate license for your laptop and so I of course sat off the file and I have had reply bu So I can't show you the wise one. Let me tell you, the wise one is infinitely better than this one.
This one was almost certainly going to improve. Downstream Technologies are the people that have uh, PC Gerber Cam 350 I think it is. You know they do those sort of processing tools. Okay, wise do the same sort of stuff and they did a much better job of this.
Um, so if you got to use a tool, start with Wise and I'm sure these people would catch up. But you get this sort of typical sit things. the highlighting sucks. I should be able to highlight it tool and if you really look really closely and I'm looking and I can't see it that C4 in is highlighted somewhere in there.
Okay, I can zoom in I tell you I tell you if if this was C this is almost like Al possible to use I'm joking. Okay, so you can just keep zooming in until finally it looks like something you understand. Okay, you can turn holes on it off. uh what? we do drills different sizes I can turn them off.
somewhere in there is a come on, just turn it off. In fact, turn some, turn some of them off right. So you get a look at the sizes you can look at nets. there's a list of all of the Nets in there.
Technically if I click on there probably highlighted, why doesn't Zoom to that not I don't quite know. So I'm not sure what they what they were thinking when this happened. Um, you can look at layers I change the colors of layers Etc and so forth. I should right should be able to right click there and say turn them all off and then turn on only the ones I want to.
No, they don't do that right. So I've got to click them all off one by one. Etc the Wise One infinitely better. Okay I just wish I had a license parts.
You've got components. What we got here. these are footprints. So I guess you can look at different types of footprints. They should be highlighted wherever they appear And the components typical. Okay, so that's Works a bit like your CAD system. I much worse in this case, right? So I would imagine this, this type of viewer is going to be fairly common. It'll be like your PC Gerber or like your Gerber editor.
You know you do a you do a Cad design. You've been looking at this for 3 weeks, right? Everything's perfect. The Dics are clean. You call up the G is just for one final.
Checker The very first thing you see is an error. Boom right there. Okay, you see things that are G reditor that you have missed in your your in your CAD system right? because of the way that you're looking at things you're looking at Fresh eyes when you're looking at a Gerar because you're just looking at one layer. Generally looking at two layers is is is fairly complex and difficult to to grasp.
So suddenly you're looking at one layer with fresh eyes because you haven't seen it this way before. And that's where you start to pick up little strange things that the Drcs did not pick up and we will get viewers like this that will actually do that for us. Also like we'll be able to turn things on and off that looks at assembly information looks at at C at um, fabrication information test information should be able to isolate all of those things independent of our CAD system, right? Because now we're going to look at it Fresh eyes yet again. And these viewers no doubt will also check the file.
that 44 megab file. That's this, this this one 44 megabytes. It will go through all of that to make sure the syntax is correct. Okay, and if the syntax is correct and it's going to go Downstream without any problems, that's the plan.
Okay, okay, um, there's not much more to say with this because it's it's basically an extension of what we do with CAD only much better. Now we got the information one big file and that's all there is. and I don't know I'm not sure about the segmentation I Don't know that these these viewers do that, but what I should be able to do is say take extract out of here only the information for the board fabricator right? Because if you think about it, this entire file can go to a competitor and you make your product right in in there. Yes Ma'am I I I Think the segmentation and Export that you're looking for really comes from the Cad side.
This is just allowing you to view the file that you output. So if you if you output only the Fab information then you'd be able to view that. If you output only the assembly information, you should be able to view the assembly information. Uh, so the Cad vendor is going to say C You can click on a button that says Fab and it'll generate it 2581 file for just the fabrication.
Correct at this end, Not at this end. Okay, that makes sense. Okay, so if you y um, just on that is it. Is it? Uh, Is it possible or is? is? Is it intended that this file format be the native description of the PCB that your CAD program is actually using not explain It Can Be it can be used that way. Yeah, at the moment. right? If you if you have, let's say, matter, Graphics or Alum Alum has its own structure, right? So the the way that Alum describes its database appears in schematic files PCB files. And then there's some background files too that that sort of extract some sort of information. So it has a way of describing the way it works, right? and Metagraphics has the same sort of thing but in a totally different way.
and Zuk and and and Cadence Etc and so forth so none of them are compatible. Those native file formats are not compatible. What this is going to do is it's going to extract ract the information in a format that everybody can use in a neutral format which is text based XML based which people can use worldwide because it's well defined so in ter in terms. What? What you're trying to say is it converts a Cad file to a neutral file or it can do right.
P Had years and years ago had a what they call the neutral file format and it was a completely text based file. it was. It was fairly hierarchical also and you could take that and you could manipulate it any way you wanted because it was B done in text. Okay, but there were no translators to translate that neutral file format into something different into a different CAD system.
but this goes a long way to that. So if U if you can generate a file that carries the whole thing, then that becomes a file that somebody can manufacture your board from. I Guess the point I'm um, thinking about is could you take one of these if you had the complete Um description? Could you take it into a C system, manipulate it, change components, whatever, and then send it off to the manufacturer, right? Yes, hang on there, There's a provisor. There's a caveat.
Okay, let's say I'm sitting at Altium Okay, Altium is very happy to supply an output of 2581, right? But it may say no. I won't let you bring in a 2581. Okay, suddenly the the whole business of translating for one cat system to one other falls apart. Alum just might do an input also because they'd be quite happy to snatch work away from Expedition for example, right? But I I Can almost bit my house on the fact that ex.
that Metagraphics might have an importer, right? Happy to give an exporter? That's fine, right? Because that's all. Downstream Stuff an importer? Maybe not. Yeah, let me to that. Okay, at the conference, you'll see where they created a file in Cadence and they moved it over and read it into Zuk to do with the changes that you're asking about.
Put the file right back into cables. so they've got utters. Yeah, they're importing we are. We are talking purely at, um, a manufactural level rather than schematic capture level. I Mean this file does not actually incorporate what we regard as a human readable schematic. Actually, that's a good point because it's only the production files, the assembly files, the test files. So it's actually an output format. It's not a true CAD format and it probably I assume loses a whole lot of intelligence that are built into the parts and the design.
That design rules. all the design rules, all that sort of stuff and keep outs and all that sort of stuff. Building their parts well, all the key bouts. all that stuff.
That's all part of the 2581. All that information is in 2581. The only one I'm not certain about is the schematic information itself. just a straight schematic.
There's nothing in here about schematics. Yeah, I I Didn't see anything in that. That may be that another 25 series thing 2583 or 13 or something. I'm I'm not sure that, but I know from the board side all you could archive the total construction including all of the test fixtures, everything involved with manufacturing assembly, test all your test specs coming out of your CAE system all captured within the format to be able to drive your your ICT test else.
So it does as as if between look like capture things like 3D mechanical models because that's part of the archive because the mechanical engineers are going to need to sit there and deal with that. So is the format. Extent is it does it currently cover that or is it can't that? No, not yeah. I'm not sure about almost certainly have a step file reference somewhere in there.
I'm sure I could do a search, but it take too long. Um, you're right about schematics. It's something I hadn't thought of. um maybe it's not the all things all people that we thought it might be I'm not positive that it doesn't I just don't know I didn't see anything like it.
Yeah, Anyway, it's a big file. Um, let me just see. when that thing's finished, its ruminations. There it goes, right.
I've collapsed it down to the top level. Okay, so there's two lines in that top level if I if I unfold it for the next one I think I have to go to there right? I Spent another half an hour waiting. This laptop is slow, right? What I've done there is I've actually uncollapsed it one cut up. Okay, here we go: Things like Content roll function equals I'm not sure what that means Function Mode full Level one step name PCB Layer Reference Top Layer Reference Ground two Layer reference L10 obviously not in order Etc and so forth.
Coming down here, we've got a dictionary standard equals inch entry standard Circle One Now we start to Des describe Primitives Okay, ID Circle1 it's a circle of diameter That right We defined inch up here. this is inch okay, Circle two is that many Ines right? So once again, we've missed an opportunity to actually get the world on the single standard. Are you going to tell me that We here are going to tell the United States to adopt metric I I think I Think the simple Logic For this is if you find everything in units of angstroms or nanometers or whatever you want to, you can at least get exact inch measurements. If you just find things in inch measurements, you can't bring it back to exactly Gary Could I just point out the joint is a stir I I Do want to I Do want to say one thing about okay one thing I do want to clarify the the documentation committee that's doing all the documentation and it's not part of these slides they are going through hammering out all all the the the name conventions and everything else for the from the schematic drawings down to the B drawings assembly Jo all that documentation side and in preparation for the 2581 committee to be able to now take that on in. So so I I I I'm I'm positive that the schematic is not there yet, but it's all that information is being okay formatted in such a way so that the consorti can then take it and start start working stuff in so it's working Pro in projects more I'm thinking about this: I I know what's going on with that other with that other group and uh, I'm sure that you get the eventually in the metric and and the conversion day. I'm sure that's an issue that's trying to deal with somewh, right? Okay, but but this is the this is typical of the structure of the way this file works. so you got circles and if you keep going down further you get you come to polygons right Here We have a shape and this is what we're calling that shape. Okay for for reasons and I'm not quite sure why that happens.
It's a contour, which means I don't know what polygon that mean. Start there, keep going with those segments and end at the same point I think that's the same thing and that finishes that section and then I go to another Circle 134 Somewhere down here are rectangles. Uh, there's a rectangle by Center Okay, this the rectangle Center is there and the height is. Or the width is that.
and the height is that, right? So they're describing all these basic real PR We saw a slide of some of those. If you keep going down here further and further and further and further you come to the bill of materials. Uh, you come to things like a net, right? There's a net name that has a net name of that, it has a net Point X and Y as a layer references bottom right. that's referring to a layer description that was up in the top of the file, etc.
etc. and so forth. So it describes Nets and and it picks up shapes right. It picks up that shape which we described up there somewhere.
Okay, so it's it's ping all those shapes up by that net name and somewhere down here is a bill of materials. It'll start from The Primitives and work its way up Primitives the P Stacks P Stacks Yeah, I somewhere I've missed it. It's up there somewhere. So this is what do we got here right? 1, 34,9 76 lines, right? So this is not a trivial file. Okay, if you look at a typical it's it's closing off all brackets of the hierarchical structure right at the end here. So if you look at this is Level 1, 2, 3, 4, 5, 6, 7 for all I Know there's other levels further up. Okay, so that's basic sort sort of it. But you know if I open a Gerber file uh We've looked at Gerber files before four whole string of X and Y coordinates some of them with with leading zero suppression, trailing zero suppression, X number of decimal points, digits out to the decimal point, etc etc.
There's not much different, right? It's all text based and it's just in a different format and describing a whole lot more stuff. Far smarter than any Ger file I ever saw which started 30 years ago. Turn a light on, draw it and and then turn it off again. it's pretty D there Any questions for my limited knowledge: all yes, the Um Bill of Materials.
Is it is it a um call a smart Bom that includes for example it's a an Oem, my approved vendor list or is it simply um, more of a a Cad liary generic? This is a one in 914 Di and the for that component are in some other file. yeah see I told you resist. Oh here we go sorry I don't want to make light what you're saying cuz you you are quite right right there. This the Bugg is bringing up questions that I Haven even thought of.
There's a let me see if I can find another one of those. The simple answer is yes, yes Yes, it does all of that All you know what else will let you do that you can't do today. You can take two different boards M them together to the fabricator to put on in one pan. Right now, if you send the fabricator two different files say I want to both on the same panel.
He won't do that. They oh, it's two separate jobs, but you can Nest it together and send it there and it's treated as one unit, build together at the same time and allow you. That's good. I Came up to the bomb.
There's there's the bomb reference, right? This bomb name is all bomb. There's a header and a revision and a step reference on quite sure what that means and then further down here it says it. It looks at specific components. Okay reference designated name: Le1.
Do you want to populate it? Yes, we do. Okay, somewhere down there is may be false as well. Um, and you can keep coming down to C1 C3 C1 C5 Etc So it's calling up. It gives you a quantity.
It gives you a number of iOS Why that's important in a bomb? I'm not totally sure. Category: Electrical That's probably really important for some reason or other. This is the part number, right? That looks like a a manufacturers part number. Go back up and there's a package reference and that's the footprint that we're going to use and that is described up there somewhere.
Sorry. go back up a little bit. You scroll down. Yeah, here you see it says a Legro definition source coming from Legro. Okay, so they even identifying it came from. Okay, I guess that's important. Well, it's probably important, but if it's a neutral file format, it doesn't really matter where it came from, doesn't It doesn't matter there there there you are. That's the short story.
I' watch this space I Wouldn't be surprised if in a year's time we'll actually be doing this using it. Whatever customes they need. oh would think so. Yeah, you look at those.
Primitives we looked at some of Gary had a slide that you had some Primitives you know, Op Oblongs rectangles, ovals, a whole bunch of stuff. Sure, Um, there's no reason why I couldn't invent widget which is something else right? there were there's capacity H here for a polygon so you can make any shape that you like Call It Whatever You Like provided it's a unique name and then call it up anytime you want. Very, very extensible. This This is this hierarchical stuff you can slot stuff in anywhere you like.
What are you stirring about now? John The beauty of standard. There are so many to choose from. Precisely, that's exactly right. But are we going to get purchasing integrated into this? since they like their own database? It's in there already.
No, they're not using this. They're using their own database. Oh, they're noted in designators and crap like that I me everybody else needs that. Not them.
Okay, traditionally right, you have the Cat system and the development team. and over there is operations. and operations. Want a whole different way of doing things.
I tell you have some translators of some sort. Oh, imagine I give them this file and somewhere somebody's writing software to convert it to that to their format right until they accept this as as a definitive article also. but you know there's there's a lot of management. Uh, what? Itm ECM Can't remember a whole bunch of different types of software that will a company can use to control its purchasing control.
It's a whole bunch of stuff that runs the company whereas over here is the R&D group that does the development and over there operations make the things in in volume right. And there's always been a Div. Well, always been a divide and sometimes a really big divide between the two. It's the all over the- wall stuff you remember.
we used to do that with assembly and Fabrication we don't do that anymore. This is not going to solve that problem because this doesn't have doesn't contain the pro. But you know there are attributes right? You're can attach an attribute to your parts that he's your your company part number and as soon as you've got that that key that can key into the the database that's over there right? So it's extensible in that sense that you can put in attributes your own attributes if you want right that will link into your system. Is there any way we can convince the Um Prime component vendors to use this common package numbering sche. We for one moment we about years ago I thought we were heading that direction and all of a sudden they ml. Well, you know it's interesting you asked that because we're going through a big deal. Now with this with the uh uh, where the pin one is okay and you're getting all the different standard organizations three of them work, the other two don't and uh I I I I Think we would talk about maybe trying to incorporate one way to be consistent and then if you're using one of the oddb ones and you know what you have to do to make to break from the consistency because it's obvious at this point in time we can't satisfy all those organizations because they're not consistent themselves, do Jed Every just not. Well, you know, you see.
You get this feeling that if you sort of think about the ramifications of what's Happening Here we're down at this level and we need to be at that level, right? if we go to sort of start incorporating everybody that's in the loop of the electronics Industry as such. Okay, and and this business of control within a company is part of that issue, right? So when you get Digi key part numbers and then you get Element 14 part numbers and then you get A manufacturers part number and then you got equivalence and you got Alternatives right. And and the the meaning isly subtly different. Suddenly there's a can of worms that's getting bigger and bigger and bigger right? Sooner or later, and possibly long after, I'm dead, this might all come together.
But let me tell you this. This first step is a really good worthwhile step. But you know you know all your concerns and things if you you went up on the website there I Think put concerns in there the committee looks at. We miss that we can put that in and put it in the que.
No, it's it's as simple as that because it's industry own. It's just not putting it into uh and it has nothing to do with any kind of business case or anything. Because there is no business case, you need it, You need it. Good idea.
Yeah, so what it needs is support, right? So if your CAD vendor doesn't do this, so I I need this I want to give it to me? give it to me yesterday. Okay, and I'll say probably no. I'm kidding. I'm kidding.
There's vested interest involved. so if it's if it's really easy to use one cat system to get a board made and this guy doesn't do that, you have to go through all these hoops again. this guy's going to be out left out in the cold soon or later. Okay, so I think that's there.
There's a strong ground spell. Much, much more than we've had before. Yeah, because of the consorti ex the first time you ever seen themall getting behind something like that I Think that's very encouraging, right? it was. It was.
in fact. Gen Cam fell apart. didn't fall apart, but Gen Cam sort of came to a grinding Hulk because ODB Plus Plus had the market here and Jenam was trying to do the right thing and getting the two together was fantastic and everybody said what a great idea B said no, we're going to keep the perform proprietary. We'll give you what we've got, but we want to be able to change it. You can't do that with an industry standard. Okay, so they buted heads and eventually they went to Nist and Nist said right, this is the way we're going to do it. So the two come together and now we've got a good system and a system that people are getting behind. Wonderful! Bravo Oh I should turn this off? No, no, no, why don't come me, turn it off.
me: "Oh man…. This is going to be so dry." 2 minuets later "Oh hey this is really interesting! :3"
Make sure IT people think completely, otherwise it becomes useless.
xkcd 927….
Most of the Engineers today are no more than "Technical Clerks"
Interesting insight into a typical engineering industry discussion and good taste for budding young engineering students who likely don't yet understand the application of all that theory yet. Thumbs up for the free insight !
People who find this video boring:
Engineering is at times (I'd say most times) this way. If you find that boring, there are plenty of other cool videos around, you don't have to watch this.
But just think about how the world would look like if there weren't people who cared a lot for such "boring" things ;).
0:28 Touchbase: Woohoo crossed off the first word on my bullshit bingo sheet. I bet i get bingo before the 30 minute mark.
edit: 1:20 information highway, let's make it bingo before 15 min are over
I'm using TinyCAD and I love the XML data format – it's so easy to script something that automates tasks in your schematic.
Using FireFox, no problem with videos!
he says "ahh" way too much.
I watched it through, and this is a great idea.. Dave we need a vid about what do you think bout it…
If everyone thinks that this kills competition… Look at HTML(it's standard, and any major browser vendor MUST SUPPORT latest version). We have many MANY great browsers out there Like Firefox(stop using this crap), Chrome, Internet explorer(That one is not great, you should stop using it), Opera, Safari…..
The unusual projector set up is the most interesting part.
Fine on Chrome.
Competition is good for all aspects of business, sounds like they started a consortium to kill off competition. One company โowningโ all the CADs not going to happen, someone will always build a new and better CAD or mouse trap.
YAY for open, free, accessible and portable file formats!
34:30 the highlight of this video for me
Buggering license issues foil us even today. Hyper fast customer service for even those looking to promote their software at forums and conventions, genius. Wise might be better but it's the same head-smashing cretin machine we're all used too.
If you want to watch it now use IE
On Opera same black screen.
This man doesn't have a grey beard to twiddle? This can't be!!!
It's a bit on the boring side of things, but it's important!!
Interesting Lecture! I like it!
Geees I'm gonna save this vid for when I can't sleep. Turn it on……wait 2 mins…….then ….ZZZZZZZZZZZzzzzzzzzzzzzzzzzzzz
Some professional gentelmen ๐
I just bought some 'chewy middle' chocolate chip cookies.. their great, If I did not have them on my desk, I wouldn't have watched 10 minutes of this video. Love the EEV blog.. this vid.. ehh not so much.. but they all cant be winners.
This might be a good candidate for the AltZone channel.
please do another episode of the power supply series !!!! ๐