Stop Getting Stuck - AltiumLive 2022

Carl Schattke
|  Created: February 3, 2022  |  Updated: February 23, 2022

Order from chaos. Train your engineer. Scope the ECO. Breaking down barriers and predicting and assessing and solving the puzzle.

Using design and collaboration tools effectively Design PCBs with a Free Trial of Altium Designer Here: https://www.altium.com/altium-trial-flow

Highlights:

  • Importance of a sequenced scene of the changes
  • Areas of communication that present challenges with changes to boards
  • Risks that come with changes
  • Looking at density to approach difficult design changes
  • Tools for design changes

Additional Resources:

Transcript:

Carl Schattke:
Hello. Welcome. I am Carl Schattke and we are going to have a presentation today about how to stop getting stuck. We're talking about getting stuck on PCB design changes. One of the nasty areas of PCB design is when you have changes and bad things tend to happen if these changes get extensive. Or if there's poor communication or a whole host of problems. We're going to look at those.

We're going to examine what kind of changes there are and what it takes to prepare for those changes in a winning way. How to communicate without breakdowns and then how to break out of different design problems. Break through them and then how to create a good plan to escape out of your changes into a completed product. We'll look over some finishing thoughts after that.

So welcome. Let's get started. I'd like to start by telling you a little story from my youth. When I was about five years old, I went to the top of this building. It's in Chicago. It's called the Prudential Building and that's not the reason I'm telling this story. My father was an engineer and his father was an engineer. He worked on this building and he was the chief electrical engineer. This building was a new skyscraper in Chicago. It was the first one built after World War II and it was the tallest building in Chicago at that time.

It needed lights to show how beautiful this new building was and the chief electrical engineer was my father but the outdoor lighting was a different engineer. He had specced some lights that went on the outside of the building and they weren't quite bright enough when they finally got the building built. They were all worried and they were like, "Oh, we're going to have to tear up the grounds and put new heavier cables in, to support that much current." They talked to my grandfather about it and he said, "It's okay. I already put the heavier cables in. I kind of figured you might be needing them."

So then this building got this beautiful lighting to it and in the time since this building was built, now it's no longer anywhere near the tallest building in Chicago but when it was built, it was. But the thing for us to remember is that if we expect the unexpected when we're doing our original design, then we're much better prepared for the problems that might come to us. We're a lot less likely to get stuck, so that's the key takeaway from that.

The other thing that's hard to design is when things aren't going well, so here's a picture of a ship. It sunk. It made international news because cruise ships don't usually get stuck and beached like this. But there was a captain that did something that was not safe and it was a danger to everybody on board. He took this large ship into an area he shouldn't have and he didn't have plans for getting it out. It was an environmental mess and it was bad news. It ruined a really expensive ship and so we don't want to be doing that with our designs. We want to make good, clean, clear choices that are safe.

Here's another project that didn't go well at all. This is a dam in Oroville, California and a couple summers ago, this thing, we had some record rains. They came down and overflowed. They had not decided to put the right foundation on the spillway and it started to erode. Then it eroded the area that was the earth for the dam and this was a very, very large concern because they have a city right below this. They had to get rid of ... They had to evacuate the city for a long time. It was a long time for this to get shored up and very expensive.

When we have problems on our boards they can be just like this where massive emergency and things. So we want to avoid these kind of problems and how do you do that? Well, problems are bad. We want to avoid them. So I was involved in martial arts for a long time during my life and one of the things that we would practice would be locked positions. We'd practice on putting our body into manipulations where we would be able to manipulate our opponent and get what would be like a locked position. If they move one way there's pain. If they move the other way there's pain. Really their best option at that point is just to tap out because there's no way to move.

We want to avoid getting stuck like that so when we do our designs we want to way out and we don't want to design ourselves into a corner. One of my favorite things that I like to do is ride a motorcycle now. When I took the safety training class, they taught us this process about how to avoid problems when changes are coming. They call it the SIPDE process. The idea is that you scan and look at all of the areas around you that could potentially be a problem. So you keep your eyes farther ahead and you look left to right. You keep high awareness of what's around you.

Then you interpret that and you see, "Well, this could be a problem or that could be a problem." Then you predict what's going to happen in your mind. Then once you've made that prediction, then you decide what would be the best course of action and then you execute. So all of this happens very fast when you're riding a motorcycle and the faster you go, the faster it has to happen. So there's a lot of decisions that are made very quickly but this is a plan ahead of time for making smart decisions. Best way to get out of trouble is just don't get into it in the first place.

So this is some background about why it's really important to plan when we make changes. Changes are like the weather. The more we prepare for it, the better it is for us. So what do we need to be prepared for with circuit boards? Well, let's take a look at that. First a little quote from Mary Shelley, "Nothing is so painful to the human mind as great and sudden change." So let's hope we don't have too much pain from that sudden change. So what are the reasons people make changes on circuit boards? Let's take a look at that.

So scope changes. There's added or we add or we take away functionality. The size might change. The environment might change. In this case, we got a funny looking bicycle that they said, "Well, keep the bicycle secure." Well, that's a scope change but they didn't really look at the whole scope, right? We want to be able to move the bicycle later. A block of cement's not going to really work that well but it will keep the bike in one place. Nobody's going to take it. We need to be aware of unintended consequences.

In the world of change, we look at our scope but then we also look at schedule and cost. We're going to encounter problems with scheduling because things have to happen in the right time, in the right place in the right way. We might have a sourcing problem. We might have a safety-driven challenge. We might have legal considerations, why we can't sell something. There might be commercial reasons why we have some problems. So we might have real rapid schedule changes where there's either a market opportunity if we move quickly or a challenge.

We might have manufacturing lying down in which case we need to respond really quickly to that kind of a situation. There might be critical path situations in the whole timeline of a project where what you're working on has to be done right away in order for the whole project to come in on time. So there's a lot of reasons for accelerating schedules and we'll look at some ways to help do that later. Then we also might have changes because of cost or availability. So big problems right nowadays with part shortages and things like that. Or there might be a massive increase in cost or massive lead times.

I'm hearing lead times sometimes in a year-long, in excess of a year right now for some particularly hard to get parts. So real constraint on business and you have to redesign your way out of those, so a big reason why we see changes. Then there might be a safety-driven change where something's just not safe anymore and we found that out. There might be legal or commercial implications with why that would be. We also might have quality reasons for change. So if we're building something and we find out we need to increase the amount of current it handles, well, that's going to be a redesign.

Or we need to go to higher voltage. That might be a redesign. Or might be finding that there's arcing or there's CAF problems with a particular circuit and we're getting shorts there. We also might have thermal issues that we need to resolve. We also have reliability issues that could be coming up and these could be due to vibration or solder joint failure, a part cracking, via and trace failure. There's a lot of different reasons why there's reliability problems on a circuit board. We might have to fix something like that or there might be manufacturing yield problems and come to find out it's not making money from particular things.

We might want to improve the ease of assembly or improve the automation at ICT test points. Maybe we were building in low volume and now we need to build in high volume. Or we might find ourselves needing to add in circuit test points or functional tests. So all these are never an accident. It's always a result of an intelligent effort. That's a quote by John Ruskin. So we also have quality changes in issues that come up from our schematic. Might be logic errors. Might be compiling or netlist errors. There might be a need for increase in a memory storage or a decrease in it.

We might find that better parts are available and we want to put those in. Or there's we built it and found out that the issue was incorrect. Or there might be obsoleted parts, so a number of different schematic quality issues, not necessarily the engineer's problem but it becomes everybody's problem once those kind of issue arrive. There's PCB design quality issues. So maybe there's a DFM, a DFA or DFT type of problem. We might find that there's signal integrity problems. Maybe the clocks aren't working right or the diff pairs aren't with the right ... They don't have the right impedance.

Or there might be impedance discontinuities. Or there might be problems with the way the crystals were done. Or the high speed bus. Or a lot of different signal integrity problems might drive a PCB change. Or power integrity problems where we have voltage drops or ground bounce or something like that where it gives us a problem maintaining our logic levels on our circuit. Or we might have termination and it's at the wrong end. Or getting reflections that are causing failures in the circuit. Or opamps or power supplies that weren't right, weren't routed the right way.    

They're noisier and we're not passing specifications that we should be for that. Or may have just incorrect impedance on things. Lot of different PCB quality issues can come up and cause changes. Then we get into PCB manufacturing issues. Those can also be a reason for a lot of changes. We might need to add extra layers to support something. We might have extra steps in the process. We might want to reduce the number of layers. Drill sizes might be wrong for the parts that are on the board. We might have plating that's wrong. Maybe it's too thick and it's a more expensive board than it needs to be.    

Or it might be too thin, not supporting the current carrying capabilities. It might be plating. Maybe we used a certain kind of plating and we find out that we have problems with that. Maybe we chose the wrong finish or something like that. Or the copper balancing isn't effective. So a lot of reasons for manufacturing, driving our ECO process. Then we also have assembly quality issues. So if you're trying to build something and the board's got a problem, then we have issues with assembly. It's not really working for us and we might have poor drawings.    

Or documentation might be wrong. Or there might be missing ICT or they're too close or they're not usable. Or we have a clamshell fixture and it's expensive to maintain. We might want to go to all the ICT pads on one side so they can just build a less expensive fixture. Or pad sizes might be too small. There's a lot of reasons that we might want to update a board because of assembly quality issues. Then labeling and tracking might not be done right or in the right place. Or maybe the nomenclature's not right on our board, so a lot of changes can be driven from quality issues.   

Then whenever we make changes we need to look at the risk. So we don't necessarily want to do things the way they were done because that's the way they always did it. No. I think that's kind of a cringe worthy statement. We want to really mitigate our risks and look at changes to the process, control. If we do make a change, we want to know if we made that change the right way. So we can cover that by either checking the process really carefully like, what was done, keep track of it. Or we can load in the old board and the new board and use the Altium Tools to compare the old to the new.    

Then we can also validate those changes with a third-party, maybe with some automatic tools for validating of the new board. Then when we do have a change on our horizon, we want to make a quick decision so that we can still go with a go or no-go decision on doing the project. That's calculating the possibility of success. If we tried to pack 20 pounds in a five-pound box, pretty much know it's not going to work. So we have to come up with a different solution but a lot of changes are in the marginal kind of an area. We may or may not succeed and we need to make that decision quickly.    

So we're going to look at some tools to help with that later too. So now, the best change is the one we don't have to make. The best fight is the one we don't have to engage in. We want to look for alternatives, right? So we want to design for change. When we do a design, we don't want to block the ability to modify that design with our future efforts. We want to predict growth in functionality and predict the change likelihood when we do the initial design. That's going to help us make a more logical choice in the changes and make it more possible to do that.    

So when we look at changes, one of the ways that, that's done is the iterative process. So we see the constant changes be ... Different organizations are going to vary in their risk profile so there's going to be different resistance to change depending on the type of company or organization you work in and what the total equation is as far as the reason for making change. Some organizations, you're constantly changing. It's just known that's what you're going to do. It's going to always happen. Then a certain amount of scope creep is natural but that's I see often very different based on different organizations.     

Then scope creep gets pretty predictable if you're with an organization for a while or working with a team for while. You kind of know how much they're going to be changing things so you can kind of figure that into your process. Then we always as designers need to be really flexible as far as what kind of changes that we come. We want to adapt to those as quickly and flexibly as possible. Not present any additional obstacles based on our beliefs but based on like, "What's the best thing for the product?" Let's make those changes right away and be flexible.    

Then with continuous improvement we want to go through a plan, do and review. So we're going to plan our change. Then we want to do it. Then we want to review it and make sure it was done right so that we can continue and carry on. You can do this with a good cooperation between the electrical and mechanical engineers and the PCB engineer to make sure that you're always doing continuous improvement. Then go through a change and check and repeat kind of process. Then like we eat a meal and it's great when we have it but the next day we're still hungry.    

Change is kind of like that. We do a change and then a few months down the road we're still going to be hungry for a better product. We're going to make a change. So it's not uncommon to see boards change a lot of times based on any of the reasons we covered earlier. Edwards Deming said that if you can't describe what you are doing as a process, you don't know what you're doing. So let's learn what this process is.     

All right, so we're going to have some winning preparation to get there. All right. So I like to put changes into four different buckets as a designer. So first it's small, easy, simple, drop in, connect, done. Same day kind of stuff. These are typically real small changes. Then there's the moderate changes. A little bit more than simple but still relatively minor, maybe 10 to 30 components. Limited active part kind of changes, small. It might be one or two days get a redesign done on something like that.    

Then there's major changes. Now, we're talking active parts. We might have limited room. We're going to have to move a bunch of things out of the way. This might involve maybe a 10 to 30% change to the circuit board. It might be half a week to a week kind of time frame. Then there's like next level changes, which are complicated and involved and complex and hard and challenging, way more than the other three. Then this might because a processor changed or parts in different circuits need to relocate. We need to clear out new room for the things that we're adding in.    

These would be typically your multiple day or multiple week, multiple month kind of a change where you have pretty massive changes. So all these different kind of changes go into different buckets. We kind of think about how we work on these based on the bucket that we find ourselves in. So the first couple ones are relatively minor. We can just kind of go right with whatever it is. But the major and next level changes are what we're going to be talking about for the rest of our presentation today.     

So the General Sun Tzu said that if you know the enemy and you know yourself, you need not fear the result of a hundred battles. If you know yourself but not the enemy, for every victory gained you also suffer defeat. If you know neither the enemy nor yourself, you will succumb to every battle. So be smart and know yourself when it comes to what you're trying to do in the area of PCB changes.    

What are the levers? What can we control? Well, we can control placement, the area that we place it in, the rotation, the proximity of one group of placement or parts to another. We can control the order. We can control the arrangement. We can also control the priorities. What are the critical timing delays? Any noisy or sensitive areas? Same layer routing, the via bound areas, the stack tolerance areas. How much room do we physically need for the circuit, these changes? Are there magnetic components in the sensing and monitoring power components, the sense line?    

So all these are the ways that we can control our circuits and oftentimes these are leading to changes in the design. We can control that and then the sequence of the changes. So we can control whether we work on the easy stuff or the hard stuff, the big stuff, the small stuff, the most likely, the least likely, the lowest risk to solve, the highest risk to solve. We can work on one schematic page at a time. We can choose to work from left to right, top to bottom, high voltage to low voltage, like whatever makes sense for the overall design is generally how we want to arrange the sequence.    

A sequenced scene of the changes is extremely important to having success because if you do things out of sequence then you're going to have to redo things as you get down in the process. So we can also change the rules. So sometimes if we have a change, we can change the rules in order to adopt a new technology or change something. We can sometimes reallocate pins and relays or series circuits or pull ups or pull downs. Maybe they can go to a ... We can actually request changes in the schematic to make changes easier. Oftentimes that helps us and we need to know that, that's one of the tools and levers we can pull as designers to request changes in that area.     

Then board technology is another change that we could make if we find that we've greatly increased the density of a particular design, we could switch it to start using microvias or pad cap vias. Or we can increase the layer count. Or we can increase the ground count or copper weight if it needs to carry more current. Or we can change line widths. Or there's a lot of things that we can do in order to make this change a little bit easier and not get so stuck on it. So there's usually a trade-off between the speed of the change and the cost to make the product.     

More design up-front generally is going to result in a lower cost product but it might result in a change that takes longer. There's going to be a calculus involved around that over do I want this changed super fast and get to market right away or fix a problem right away? Or then come back and drop cost out of the product later? Or do I want to lower my cost initially in the design and shoot for the lowest cost that the most efficient design from a cost perspective right away?     

That's going to be a business decision based on what you ... But like Archimedes said, "If you give me a lever long enough and a fulcrum on which to place it, I shall move the world." You can move into your changes if you know the things you can control. You can make a big difference on how well that process is going to go and find yourself not getting stuck. If Archimedes can move the world, we can move the stuff on our circuit board.    

All right. Now, we're going to look at communication breakdown. We're going to break down some of the different areas of communication that present challenges and opportunities within the communication aspect of doing changes. This is a big part of the success of a project is the communication that takes place between the team members. So when we think about working on something, we want to grab the right tool and like Maslow said, "To the man who only has a hammer, everything he encounters begins to look like a nail." We need to be really flexible with our communication options.    

As designers, we see multitude of different ways people like to communicate. Email. I like email because you can look it up. You can look things up quickly and get things quickly. I don't like email because there's so much noise and garbage to get through just to get to what you need. So search is real important to use email effectively. Word documents are nice to keep track of changes in one place kind of useful for that. Then spreadsheets are useful. You can track them and keep track of what was done, what wasn't done. So I like spreadsheets for that reason.    

You can also add images to spreadsheets and Word documents but spreadsheets tend to be pretty nice from a checklist kind of a way. PowerPoint slides pretty easy to create with just cut and paste. Then keeping like one topic per change on one slide is a pretty handy way to do things. Then there's a lot of collaborative working spaces where whenever it gets updated, then whoever's working on it is going to see that update. You can add or subtract things to the project based on what they're status is as opened or closed. Very useful working tools. Then Concord Pro or Altium 365 very handy to let your coworkers know what you're working on and keep track of things.    

Then there's notes that you can put right on the PCB or schematic. You can put notes into the message. Altium has a message window where you can put things in there for people as well. You can also import things like 3D shapes. Sometimes it's just physically easier to see the constraints around something by importing a 3D shape and then you don't have to ask questions. It's just you know how big the case is. You know whether things are going to fit or not. Not so much on the questions but just having the right data at hand.    

Then sometimes we can get physical parts in hand and we can look at those parts by what ... We can just see what it is if we can hold it. Sometimes that's better for complex parts or things like that. Then we can see like, "Well, if we did this or did that differently or like this might be usable space. The image shows it's not usable space but it probably is." So we can design denser products if we sometimes have those things in hand. We can check relays and things like that to make sure that they work in the right priorities things like that.    

We can also use extra mechanical layers to put things down and one of the things I like to do is have a spacing layer. If I have a minimum distance between ICT pads, I can put a circle the same diameter as that ICT pad and the same distance that it has to ... I can put the arc the distance that the ICT pad has to be away of it. I can put two or three of those circles next to each other. It becomes immediately clear whether or not I've got the right space for all the test points that I need. So I might put those circuits on like a solder, on a clearance layer. You can use arcs and circles for soldering clearances.     

So if I have a selective soldering probe coming to do the soldering on through hole parts, I need to keep my surface metal parts like three millimeters away or five millimeters away from those through hole pads and not put any surface metal parts next to them. So that's a way for me to easily keep track of that visually in the design by putting that data on just a extra mechanical layer. High voltage clearances, I might put down a trace for patient isolation so that if I know I need eight millimeters or 10 millimeters between these two different areas on my design, I can just use it mechanically or put that down and it becomes real easy to see what no traces or no components can pass fast.    

Then I can also import shapes into 3D things if I have like complex cooling plates or something like that where I have complex geometry in mechanical situations I can also import that. So different ways to communicate, all good ways to make it easier to get that job done. Now, I might have said something controversial earlier. I want to train the E.E. not to wet or go on the schematic. What do I mean by that? As designers, one of the hardest things that we get when we do changes is to get a design that's not been planned for change. Then the engineer basically rips out and puts in whatever he wants and doesn't keep track of that for us.    

Then reuses all the reference designators and what happens is if I've got a change that's extensive, two, three, 400 parts and then we update that board and the designators were reused. Well, then all the new parts go where the old part was whether they belong there or not. You have a massive hodgepodge of junk and you have to get all the parts off the board that were new. All the new parts have to go onto the board in a different circuit. Well, now, you've created a mess that can be several days of delay on a large product. This I've seen over and over and over again.     

I try to train the engineers so they know not to do this and one of the ways to not do this is to just slide those parts that you're taking off to a corner of the schematic. Throw a discard page into your schematic for compiling. Get your new reference designators by keeping the old reference designators on that discard page. Now, all your new reference designators have a unique reference ID and they will not go where the old part was. They will be coming in on the side of the board instead of right in the middle of the board where they don't belong. So big deal for people doing really large changes.    

If you're doing those simple and easy moderate boards you don't need to worry about this at all but if you're doing hard complex kind of boards, this is a really big deal. All right, so I put this quote down here, "How many times shall I forgive?" said Peter. "I tell you not seven times but 77 times." That's just a little reminder for me to be nice when engineers do this.     

All right, so communicating around the world and remote. So a lot of problems happen when we don't understand what other people are trying to tell us. Language barriers could be a reason for that. When you have language barriers that is going to mean that there's probably some translators. The translators might not understand the words that are written or they might not understand engineering. Speaking engineering in language is sometimes a different language. So using text with supported images, I find really, really helps when there's language barrier differences.     

Then there's cultural differences. So different cultures respond to problems and situations differently. So what I like to do is employ a tell-ask-tell method. I tell them what the change is. Then I ask them to tell me what it was that they understood from me. If they can tell me what I told them in a way that makes sense then I have a much better idea that they really understand what I have asked to be done. Or I might be the one following the tell-ask-tell and I'll repeat what I heard in the way I understand it just to make sure I understand what that engineer or that manufacturing partner might be telling me.    

Then the other thing that I see out there is the automatic yes. So in some cultures they might just say, "Yes." What yes means to them may not mean the same thing to you as what it means to the person who was saying it. So be careful of the automatic yes. Then local customs can be quite different to what we're used to, so in different parts of the world don't think of things or relate to things the same way that we do. Our culture that we grew up in and around might not be the culture that somebody else grew up in and around. We need to be sensitive to the differences in that and sensitive to how that might change the way that engineering would be done based on those cultural differences.     

Then there's conformity variances around the world. Some cultures are very conform. They have a lot of conformity to different things and then other cultures are ... They're very low conformity. There's also a difference in the directness. Like how direct I am might be very different from how direct somebody from a different culture would be or a different background. This can also be based on experience levels. So if I say something that's wrong, I should be challenged on it. I should be open to challenge.    

But if I say something and it's wrong and somebody's rather new even though they know there might be a problem, they might not be willing to challenge that based on my status or their status. Sometimes we're a little bit in fear of making a thing but really we need to have the best interest in mind of everybody involved and make sure that we're not wasting money by making mistakes in our engineering. Then there's time zone constraints so we want to present multiple options in our feedback.     

So if we make a request and there's a time zone involved. We know there might be two or three different options. Sometimes it's going to save us time if we can present the two or three options in our request and say, "If you do this, I'd like to know this or that." Then that sometimes can save a day in communications and then that's going to finish the project faster. It helps if we can preset our hand off times when both parties are awake if we're dealing with time zone constraints.    

Then also sometimes with 24/7 projects where you go around the clock and you're trying to get things designed as quickly as possible. So employing two or three time zones or people working in two or three different time slots during the day is going to be another communication challenge. We need to be aware of what goes on with that. Frank Zappa said that without deviation from the norm, progress is not possible. So we want to be open to having that deviation from the norm and looking at the unexpected in a way that can bring us progress.    

So now, we're going to go into breaking out and breaking through. We're going to start talking some circuit board stuff here that get into the meat of this. So what is it about changes that's going to make you successful making them? Well, we want to let the math tell us what's possible. So when we look at component density, we can do density maps, we can do density studies. How much area is there? What kind of area can be expect ... What kind of area can we expect to route based on the density?     

We can look at past projects and look at that density compared to what density we see on this new project or on this request. Obviously, if there's more components in the area ... If there's more area for components than component area you have on your board you have a density problem. We have the same thing with routing density. So how many traces are going to fit through this section? Can we fit those? Is that going to actually work? We have to do the math and find out. Via density. Do we have room for the vias that are going to be required? In this particular area, am I going to be able to make a transition from my components to the inner layers to route to other components?    

Fine via density is probably the limiting factor on changes for extremely high dense boards. It's like you become via bound and that's one of the tougher problems to solve than PCB design. Then grids. We want to make things easier for ourselves and we can use grids in a lot of ways to make them easy. If we keep BGAs on a major whole number grid then all of the traces are going to be exactly centered if they're also on that whole number grid. So it's easier to set up BGAs and other large parts on grids so that it's easy to tell if things are routed in the middle or not between them.    

Then when we work on BGAs, there's a ring around that BGA. We can put components around that BGA and I like to use a ring of ... Usually there's a lot of termination to BGAs so Rs and Cs are routed. Then there's usually some power supply support for that. So by using rings around that, that's going to create routing channels that are going to be perpendicular to the BGA. That's going to be where we can do our north, south and east, west routing. Then we put our vias into channels between the BGA and the first row of termination. Then between the first row of termination and the second row of termination, we need to leave an area for via transitions there.  

Then if it's a really big BGA maybe a third ring or something like that but that would be for like really, really big complex kind of BGAs. Then the IC routing the power into that, usually we want to bring that from one side or two sides and then let that power come in from one side and try to eliminate our other circuitry from blocking the power from effectively supplying that BGA. BGA change is probably one of the harder PCB changes that there are because the signals from that go all directions. They're not as easily put into rows and the escapes are harder because there's multiple rows and multiple layers involved.    

We also want to be aware of room for ICT pads so it's always important to ask up front, "Do we have an ICT type design or is this going to be low volume or we're just going to do functional tests?" So there's different strategies based on that and we can use grids for that. Then routing channels are open under rows so we already talked about that. So sometimes the questions are complicated and the answers are simple. That's what Dr. Seuss told us. We know how simple Dr. Suess likes things.     

All right, so the best way to escape from your problem is to solve it, Robert Anthony. So routing channels top and bottom pin out escape. So our most challenging problems are usually escaping from top or bottom parts with a lot of pins in a high density area because we have to fan out. We don't have room so now you've got vias that you have to drop in but those take room. So and even if you have pad cap vias, if you have a fine pitch part, you have to escape from that because the pad cap via pitch is smaller than your pad. So these are the type of problems that are generally the harder routing problems is escaping from very high density connectors or parts.    

We find the top and bottom layers are usually where that's the most challenging. Then we also have high voltage boards that have very large clearances. Those can give us real hard challenges with routing escapes. Then routing density, we need to pay real close attention to how we bias the layer directions. Then also seek out to even that density on our board, especially if we're doing more changes later or we expect that. If we keep to the X and Y locations and drop vias as needed to maintain routing channels north and south and east and west on a board, that's going to make a big difference for our ability to complete that board.    

Then we also need to look at the via density in different areas. Do we need to slide things over so we have enough room to drop the vias into a particular area? These are the main tools that you'll use to solve those really, really difficult design changes. Then we need to be aware when we do drop vias in to make sure that we're not building a fence of vias where there's an opening in the ground plane and now the return path has to take a long windy path back instead of just directly underneath the trace. That's going to increase the impedance and then you're going to have fields overlapping.     

When fields overlap then that's going to create more crosstalk and potential circuit failure. So I find that vias are like the tourniquet on a dense design. Like they really are where we get stuck and things don't flow. So real, real important to look at that when we're doing our changes.     

Now, let's look at the basics of routing. So to me, there's like only a few different types of traces and we use all these in conjunction with each other. But each has its own unique advantages and disadvantages. So the I shape, the straight connection from one place to another, is going to be your highest most density route that there is. So a direct connection from one connection to another on one layer is always the best. It's going to have the best signal integrity. It's going to have the tightest coupling or return path that there is.    

Next is the J shape. So this is the next highest density after the I shape and it's basically going past a pad and coming back. We can use this to keep all the nets on a particular circuit on one side and they can be nested. These are going to reverse the order of the nets by coming from one place to another. Doing the J shape is going to reverse that order and then obviously, we can reverse the order of nets if we pair this with an I shape type net.     

Then C shape type nets are the next highest density after I and J shape. This is also going to reverse the order. Great for nesting. It makes sense on the outer edges so if you have a layer of bias on the outer edges like a C shape, kind of shape of routing is going to be very effective for those type of areas. It's going to also give us nice tight coupling to the return path. It makes sense around the edges. Then the S or Z shape, this uses up more area but it gives us more flexibility in the routing.     

But this is also going to keep the same shape for the design. Leonardo DaVinci said that simplicity is the ultimate sophistication. I believe that a simple routing shape is going to be the best for us as well. Always looks easiest when it's the hard ... The hardest things look easiest when you're done.     

Okay. So principles of routing. Let's seek the design flow and look for the most overall efficient use of the space. So this is going to always result in fewest number of vias, the fewest crossovers and it's going to result in the fastest completion rates because you won't have a bunch of extra junk in the way. Fewer vias is going to be fewer blockages, especially on a through hole design because once you drop a via it 100% blocks all layers on that board for that area, whereas, if you have a sequential lamination board, you can drop a via and still get routing underneath it.    

So it's less of a block depending on how many layers that via is going to be involved with. So higher density boards we're almost always using blind and buried vias. That's what I'm talking about here but in all cases, fewer vias is going to be fewer blockages on other layers. So when we route, there's so many different ways to route that board. We can wrap things. When I think of wrapping traces that's overlapping. Strapping where we go on one layer, come back. Nesting where things are building up on top of one another. Folding, like if we want to change the order, we can fold those nests, sweep and then come back on another layer, so that's going to reverse orders.   

Twisting, basically turning over, under, around, in, out, up, down, like we route all directions always to the best advantage. But keep that bias on your layers. That's going to be helpful. Then another tool in our belt is to use pushing and shoving. So this is basically a way to clear out a particular area that the Altium Tools have a pretty powerful router for this but I find that I like to do it manually because I get a lot more control over the results and I like what ... I tend to do this manually more than automation but I'll turn the automation on for a particular area if it makes sense.    

Then we can also use circular boards that have arcs or circular shapes in them. We can use arcs to get the greatest routing density on those. To route around those we can drop arcs and space the arcs at the right distance to get that done. So whatever routing we do, we want to follow Einstein's principle of, "Everything must be made as simple as possible but no simpler." After we break out our boards and figure out how to route them now, what's the way to get the great escape? Well, first we want to have a flight plan.    

So a few years ago there was a crash in a river in New York and Captain Sulley brought an airline safely down in the river and saved everybody's life. He said something about that, that stuck with me. He said, "Having a plan enabled us to keep our hope alive. Perhaps in a similar fashion, people who are in their own personal crisis, pink slip, foreclosure, or in circuit board design change can be reminded to, no matter how dire the circumstances or how little time you have to deal with it, further action is always possible. There's always a way out of even the tightest spot."   

He landed a plane on the river and saved a lot of people in ice-cold conditions. Kudos to Captain Sulley but I really like his idea of having a plan ahead of time for what we have to deal with. So when we think of a flight plan for our printed circuit board, this can be written or not written but you always want to have a plan in mind even if you're not writing it down. You always want to know what your starting place is and what your ending place is. Then sequencing your actions so they build upon one another that's going to create that start to end finish.    

I like to oftentimes get the easy things out of the way first because while I'm working on the easy things I can be thinking about how to solve the real hard problems. So sometimes I like to just do the real easy stuff first and then come and encounter the hard problems. Now, when I get complex changes I like to locate and group the components that are unplaced by schematic pages and circuit functionality and use that to co-locate those components. Also arrange them in a way that's going to be the most dense for that particular circuit and what I would consider to be like the ideal placement for those group of components.    

Then I have a block of components that I can move around and I know how much area I need to clear for it. I know how much room I'm going to need to drop that into a particular area. Then once I've got those groups of components localized then I can take and put them on the board closest to the area that I think they're going to go around the perimeter. Then I'll work on each of those one at a time to get that into the circuit. You always want to work on your sequence and your flight plan. You want to do first actions in a way that's going to make the later actions easier.    

So if you can work on something and get it out of the way, now when the next thing is working on it, you don't have to avoid that or plan for room for it or overplan or under-plan for it. So do the things that are going to be blocked first and then or do things that could be potentially be a block first so that you clear up the room for later. Now, the next thing you want to do is don't reuse your reference designators. Reannotation can take place after the new parts are placed. It should never be done before.    

Do use room because that's going to help you sort your components out. Group like functions on the same page of your schematic and then group different types of circuits on different pages. Then keep an ECO list because that's going to help you track why things change or why they were added. So tracking progress and sharing your updates. So hard changes can be tracked with a chart that logs over time. You want to have your parts placement completion, your connection completion and then your total DRC.    

If you put a chart like this together, you can see and predict how long it's going to take to finish that design up. You can see the completion percentage is going to get better and better as you route more of the project. Your parts placement is going to get better and better as those parts are on the board. Then the total number of DRCs is going to come down as that board gets more and more finalized and it's ready to produce. But that's going to be rather predictable and you can save time communicating with management if you drop out.     

If you have a standard place that this gets posted then you won't be following it. It's going to save you time in communicating with management what the status of the project is instead of having meetings all the time. So this can be a real time saver. One thing, when I design a board and I'm going to get paid for it or the company is going to have a benefit from that is you can take the total value of that project being done, divide it by the remaining work to be done. And each task that needs to be done is worth that much money. So if you have a board that's like a $25,000 dollar design and you've got 10 traces left, they're all worth 2,500.     

If you get down to two, they're worth 12,500 and that last one, you'll get it in because it's worth 25 grand. Nothing's going to stop you from getting one trace in for that much. All right, so kind of a motivating thing and Michael Dell said that anything that can be measured can be improved. I agree with him on that.    

Some finishing thoughts. Where there's a will there's a way. If you believe it to be possible you will achieve it. Begin with the end in mind. Always have an idea of what you want to do. All things are possible to a willing mind. Most important tool you have is sitting in between your ears and on your fingertips, how you deal with things. Then do the next thing. Whenever you're working on changes, do the next thing. Do the next thing. Do the next thing. Don't stop until it's done. You have to have that next thing mentality.    

Then focus on one thing at a time. We want to conceptually think of all things at once but when we do the work, we want to just be totally focused on one thing at a time. Multitasking's a myth. You don't jump all over the place. A man who chases two rabbits catches none. We want to have an indomitable spirit. We never want to quit and never give up. Really focus, focus, focus with complete attention to what we're trying to do.    

Then always realize motivationally who you're going to help when you get finished. This is a tremendous thing to keep you going when the days get long and the nights get short. If it's more about you ... If it's about more than you then there's more of you. I believe that if what you're doing is going to help others and you realize it's going to help them, then you have more resources at your disposal to help them out. So thank you very much. I hope you learned a lot about doing changes on printed circuit board design and being open to change and making change and learning how to manage change. I hope you never get stuck now because you took this course. Thank you.

About Author

About Author

Carl Schattke, CID+ likes sharing the PCB Design skills he’s worked on for the last 44 years.He’s been fortunate to work on several thousand circuit boards supporting all types of industries. He ran TLC design a PCB design service bureau for ten years supporting over a 100 clients. At ASIC Designs he designed Modems, Notebook computers, PDA’s, and many embedded systems.

Carl designed reconfigurable memory cards for startup Nuron that was acquired by Intel. At Intel Carl moved from the Network Equipment division to Sort Test Engineering where he designed bare wafer sort test solutions. In his current role he designs computers, controllers, sensors, power electronics, and high speed communication systems. He’s also been active as a PCB Design consultant working on a wide variety of other projects. Outside interests include martial arts, cycling, music and family activities with his wife and four children.

Recent Articles

Back to Home