Description
Key Learnings
- Understand and perform the benefits of simulation in the electronics space
- Learn about the perv iteration process between EAGLE and Fusion 360 Simulation environment
- Learn how to make design choices based on the simulation results
- Understand and learn how to use the collaborative nature of A360
Speakers
- James YoumatzBeen a Technical Support Specialist at Autodesk for 4 years supporting Fusion 360 with a background in Mechanical Engineering and FEA. Been experimenting with EAGLE as well as Generative Design workflows. My current role is expanding and curating our knowledge database to better help customers self service any issue that may come up. Now our support team has grown internationally and I am helping to train new Fusion support folks as the Knowledge Domain Expert for Fusion. Always feel free to reach out and say hi!
- Edwin RobledoEmbarking on my career in the communications industry, specializing in implementing fiber optics communications and data management, I transitioned into a more intricate sphere that allowed me to engage with circuit board designs through my association with the EAGLE PCB Design team. I extended my professional journey seven years ago by joining the Autodesk Fusion 360 team. Since then, I have created marketing content encapsulating various media, including blogs, videos, articles, landing pages, and webinars. This multi-faceted approach has been targeting promoting the electronic design capabilities of Autodesk Fusion, reflecting on my experience of electronic design and implementing marketing strategies that align with the industry's contemporary demands.
EDWIN ROBLEDO: Thank you everybody for joining us, greatly appreciate it. Sorry for the late start. Of course, that's not going to work on the first hand. So we'll do our best with what we have.
Now, my name is Edwin Robledo I'm part of the EAGLE support team. EAGLE was a company that Autodesk bought last year. We basically just do electronic design circuits. So that's all we're all about.
But we also now have a full integration. For the past several months, Autodesk developers from the Fusion Team, as well as the EAGLE team, have now been able to integrate EAGLE and Fusion, giving us certain capabilities that we could actually interact with Fusion and EAGLE, including certain analysis.
Now, two weeks ago, EAGLE was actually added a new feature. This is just two weeks ago. So we're working with betas and alphas here. So now we also have some Spice simulation. And we'll be going into that a little bit more. So I'm Edwin Robledo. And James is with us today as well.
JAMES YOUMATZ: Hi, thanks you all for coming. And I'm James Youmatz. I'm part of the Fusion 360 Team. I work with technical support.
If any of you guys have submitted a support request, I'm sure I've chatted with you in the past. And so today, we're going to be talking about like what Edwin just said, and with EAGLE in the Fusion 360 simulation environment.
So EAGLE was pretty new in the Fusion. We have that PCB environment. And with it, there's a lot of new things that you can do. And we're going to be exploring one specific workflow today, which is just for us it's going to be analyzing a board with a single heat source, and kind of trying to determine if the load on it is going to cause the board to fail, and what we can do to kind of fix it.
So, really, the key takeaway that I want everyone to focus on is not really as much the, how I do things, because with engineering and simulation, there's always many ways to do a how. You can set up your board in a lot of different ways. And, realistically, the question you're trying to answer is whatever you define that to be.
So we're going to be defining our own questions today, and seeing, what are our unknowns? What do we want to solve for? So, really, the takeaway today that I want you guys to focus on is the why I did something, and the what, as opposed to really the how, and kind of the powerfulness of the collaboration between the two softwares.
I'm going to show you that we can kind of speedily change things and update our model. So I quickly polled the class in the beginning to see who had a background in things. And traditionally with simulation, it can be a little tough.
So without this integration, what you would do is send it to your ECAD guy. He'd do your simulations in his program. And he'd set up the board in such a way.
He'd send me a file somehow. I'd upload it into my simulation software. You make all your CAD-ready geometry. And then you have to do your setups. And it takes a while. And it's an iterative process.
EDWIN ROBLEDO: Actually, there's a real good possibility I could maybe send him the information. You might have to set up the whole thing, because there's nothing really compatible before this, of course.
JAMES YOUMATZ: And it can take awhile. So what I'm going to demonstrate to little later on is that with this kind of interoperability between the two softwares, and the plug-in in Fusion, you don't need to keep setting up your board. It should update, for the most part. Obviously, any changes you make, will have to make those updates. But you don't have to go back and make those changes again.
EDWIN ROBLEDO: So before we actually start, since the majority of you may not have a background in electronics, I wanted to describe to you, what is EAGLE? I want to tell you, what is EAGLE? And how does the workflow of EAGLE does? That way, we set the foundation. That way you could have somewhat of an idea what we're doing.
So EAGLE basically consists-- it's a really simple to use circuit board design software. We've been around since 1988. So we've been able to build a really nice, strong community. And that community has followed us over to Autodesk as well.
In addition to that, there are many companies that actually have based on EAGLE files. In other words, if you've heard of development boards such as Arduino, or if you've heard of maker spaces, such as Adafruit, or SparkFund, you'll notice that all the files that they currently have available are in EAGLE format, because we were the standard.
We've been around for a very long time. And there's actually a version of the software which is actually totally free. So you could actually download it today and start designing today. You only have to have an Autodesk account.
Now, on my next slide, I want to give you an idea of what the type of platform does EAGLE actually consist? So EAGLE is a platform of multiple editors, very different from your workflow and workspaces that you use in Fusion. Fusion is basically a drafting application with certain areas which you actually use external.
In EAGLE, you actually use something called libraries. Those libraries are categorized by functionality, such as transistors, capacitors, inductors, linear. Or we actually have them categorized by manufacturers.
And that is one of our editors, which is called a library editor. Now, when it comes to actually designing, we're going to be using something called the schematic editor. The schematic editor is usually the engineer will start there, because you use symbols and shapes that represent behaviors of the component.
Now, that will transfer over to the circuit board. Now, the circuit board is actually the physical aspects of the schematic. Now, EAGLE has the capability of that.
When you're working with a schematic, it actually works with the board layout simultaneously. We're one of the only programs out there that has what's called true, real-time annotation. So as you edit your schematic, your board layout.
Now, I'm going to go through a simple workflow that I did a short video that I did here to give you an idea. Here, basically, I'm using my libraries. I'm just grabbing I think it's a diode. I just needed an LED.
I'm placing the diode. I'm going to define some connections on that diode. From there, I actually create my board. I actually create a shape for my board that I actually needed.
Now, all the components, as you see, came out outside of the board. It does not come out laid out. Now, I'm going to go ahead and move those components inside my board outline. As you could see, I'm always working in a 2D format.
All the connections that you see here, all those yellow lines, are connections that were defined in the schematic. EAGLE actually transferred. It's call a netlist. That got transferred from the schematic over to the board layout. The next step I'm going to do is I'm going to actually define copper connections for those traces. And I'll do that manually.
Now, we also have a tool in EAGLE called the outer router. It's an algorithm by which the signals will be connected through an algorithm called the outer router. And it will make its own decisions.
It does approximately 100 runs for every line, looking for the most optimal. Now, the nice thing about outer router is it understands how many cores you have. So it will run a thread of the outer router, of the algorithm, on each core. You could have up to four of them, each running on each core.
So you could really determine what are the results that you want. Now, let the outer router run really quick. That way, they understand. So here, I'm setting up the outer router.
I'm doing some manual routing here initially. And it became really intelligent recently. We added whole new features to this. And this is the outer router algorithm really quick.
And I push it to Fusion. And that's my final results in Fusion, once I had it done. Now, let's go to the next slide.
Now, let's go to the meat of the matter here now. Around, I would dare to say two or three weeks ago, we added something into EAGLE called Spice simulation. This is extremely different type of simulation that maybe you're used to seeing in Fusion.
Spice simulation, it actually stands for Simulation Program with Integrated Circuit and Emphasis. There's a couple of simulation tools in the market. Another one, very popular, is called PSpice. We decided to go with Spice this time.
It's very common. And getting models for your components in Spice format is actually pretty simple. Now, what we're going to do here is we have a-- so on this design, what we're doing is you're able to plug in your power supply here.
You plug in your power supply here. And then here are my variable resistors, which allows me to modify the current that I'm looking for. I'm looking for current.
The LED display will give me the reading that I'm looking for. If I don't get the reading I'm looking for, that means the power supply I'm using will not work for what I need. That's basically that design that we're doing here.
Now, we're going to go ahead and do some simulation. So I'm going to go live with EAGLE. That way you could see a little bit how this works.
Now, here where we're concentrating mainly is on the behavior of this MOSFET, which is actually triggered by voltage. Now, this MOSFET, the voltage being triggered here is actually being controlled by these two ops. I wanted to use that 3D render, because visually it was better to use that.
Now, I'm going to go to EAGLE, because it's not going to look like this. We're going to go 2D now. So I wanted to do it this way first.
So just to get your bearings, remember I showed you the MOSFET on the 3D render? Here it is. It's right here. This is the MOSFET and the 3D render.
And here's the two screws I showed you, where I put my load through here. Now, let's take a look at what it actually looks like in the schematic now. On the schematic side, we will always actually divide it by sections.
Let me explain these sections, just momentarily that we understand what's going on. It's slanted. So my mouse keeps on moving on me.
Now, this is my on and off switch. That's all this. Is my on and off switch on the circuit.
And this is actually my transistors, and some digital stuff, just to move my LED. This is to move my LED. And I'm using this microcontroller, which I have a program or for it as well. And this is on sheet number two.
Now, I want to concentrate. So here's my power supply. Here's my LED, and the microcontroller driving my LED.
Now, here is the key what we're going to be analyzing. This is where we're going to be mainly concentrating. Right here, remember I told you the variable resistors? They're right here.
I have my operational amplifier. And I have my MOSFET right there. MOSFET is triggered by a gate. And it has current flow through a source to a drain.
That's where we're at. So what I'm going to do is I'm not going to simulate all this. That's not realistic.
So when you're working on a schematic, you're not going to really be simulating your entire. That makes no sense. You're going to be actually simulating portions of your schematic, because you need to actually subset it.
That way, you could understand, what are your lines of operation? So I'm actually going to extract this, and actually turn it into a subcircuit now. So I'm going to extract this information, and actually simulate that.
I will find it today. Of course, I had it all set up. And now, I don't.
OK, now this is our MOSFET. Here's my gate. Here's my drain, and my source. And this is going to be the main concentration of this class, because that, actually when it begins operating, actually creates an amount of heat, which can be outside of its range of operation.
Now, I did not simulate the variable resistors, what they also call [INAUDIBLE]. I did not simulate those. So what I'm doing is-- because it's not necessary. I'm actually substituting them by a voltage source.
Now, I'm going to be using a DC suite. I'm going to be going from 0 volts to 0.05 volts. I'm just going to do a sweep on it. And I'm going to measure my input, versus my current that is on my load. You guys remember where you screw in your load?
This is the load I'm using for now. I'm using four volts. I want to see, if I have on my load four volts, how does this actually react? So let me go ahead and invoke the simulation in EAGLE. Now, I got the models from this operational amplifier.
And I got this one from Afinion. So I downloaded those models from the actual manufacturers. And it allows you to map them out.
And it's actually really easy to map them out with this new version of EAGLE. So I do a right click here, and I go to add model, I just go here where it says, map. And if you downloaded the map from the manufacturer, you could actually use select now which model it is, assign it the pin out for that model.
And now, you have a component that you could simulate. Now, it takes into consideration all of the connections that you actually have on the schematic. Sorry about that. I'm taking a lot longer than I expected to do here.
So let's go simulation. Now, we added this brand new bar that you see right here. On top, you see right here. This is all about our simulation here. I put my probes in, my voltage probes.
And for those that have an understanding when you're going to measure current, you've got to open up the circuit. So we open up the circuit. And we have a library called NG Spice where we have an app meter. So I actually open up the circuit right here. And I put my app meter there, because that's where I'm going to take my reading from.
Go ahead and my simulation. As I told you earlier, I'm not using the pots. So I'm going to go ahead and do a DC sweep from 0 to 0.55 volts.
All my models have been mapped out. Let me just go ahead and simulate. It automatically will give me readings for all of my currents, as well as my voltages.
But what I'm only interested right now is in my V in, versus my V out, my current right here. So I want to take a measurement on this current that I have right here. So with this new simulation tool, I'm able to actually hide my signals. That way, it's not so cluttered, the actual information I need to see here.
So I could apply my changes. Let me see if I get this right. And there it is. That's what I meant.
And they actually rise at the same time. So I got my voltage in, which rises. And my current actually rises as well.
So I'm getting a current reading. At the end, I get a 5.5 volts. So as you could see there-- I zoomed in. Let me refresh.
So it's the purple one, the one that we are observing right here. And if you notice, on the purple one, I get a reading of 5.5 volts. Now, that means I have approximately 5 amps running through from my drain to my source.
Now, I go ahead. And I took a look at the Afinion specification sheets for this. And I put the information right here to make it that much easier.
I'm getting a calculation from the load that I have, which is five volts. Remember, I actually put a load of five volts. I'm getting a read of five amps. That means that MOSFET has a total power of 20 watts.
That's how much power that is. That's actually really, really big. Now, according to the specification sheets from the manufacturer, for every watt, it will raise 75 degrees Celsius, letting me know that if I do the calculation, I get 75, times 20.
I have a temperature of 1,500 Celsius rise, because this is not even counting ambient. So my true reading is 1,520, because my ambient is 20. So it's a rise from the ambient.
Now, we're going to go ahead, and I'm going to pass this information to my partner, which is in a totally different state. And I say, look, these other readings I'm getting. Let me know how that reacts on my circuit board. And that's where he takes over on Fusion.
JAMES YOUMATZ: Thanks, Edwin.
EDWIN ROBLEDO: You bet. Any questions so far? Cool, man.
AUDIENCE: [INAUDIBLE]
EDWIN ROBLEDO: I'm going to get there. I'm going to get there. I'm getting there. I'll be getting there real fast.
JAMES YOUMATZ: Spoiled the ending to the movie.
EDWIN ROBLEDO: Spoiler, spoiler, yeah, man.
AUDIENCE: [INAUDIBLE]
EDWIN ROBLEDO: Well, what we do is that we have a plot, where you could actually see it. And then the data can also be seen, so that you could actually export it, and take that into a spreadsheet.
So the data can be exported. So I can't really text it. But you could extract it. I could show that to you later, if you want.
I don't know my schedule today. But I'll be at the answer bar. So if you have any questions about this, you could go visit me there as well. But it can be extracted, though.
JAMES YOUMATZ: Perfect, any other questions before we move into Fusion? So real quick, I just wanted to show this slide. Just keep this image in the back of your mind.
This is what the actual board is going to look like in Fusion when it comes in without a heat sink. So right here, we have the transistor. And it's our singular heat source. As you can see, there's a lot of geometry on this page right now.
And just keep that in the back of your mind. We have all these. We have the power source, the LED, the serial numbers. Everything comes with it.
So real quick, just before we get started the, one thing I just did want to know, especially since some people have not seen this before is that EAGLE and Fusion are now integrated. There is a plug-in in both programs now that allows you to move on a two way street in between the two.
And there's obvious limitations. So in our specific example, we started in EAGLE. And we're pushing the board to Fusion to use to do a thermal simulation. You can start in Fusion and push the board to EAGLE, and vise versa.
But in this specific example, we're just pushing from EAGLE to Fusion. So to do that, within EAGLE there's a Create a New Fusion Design, which we can go into. But trust me on this one.
EDWIN ROBLEDO: We did that yesterday. We kind of did that one yesterday morning.
JAMES YOUMATZ: And what you'll do is you end up creating the design, as Edwin showed. And then there's just a single button you press. And it will push it to Fusion.
On the Fusion inside, I wanted to put this slide, especially since I work in support. I've seen these questions a lot. These are the throw your computer at the wall moments. And I just wanted to touch on these real quick.
So the first thing you need to do in Fusion is since the PCB environment is a beta feature right now, you actually have to go in and enable it in your preferences. It won't automatically be there. So you guys go to use Fusion later and you're not seeing it, that's why. You got to enable at first.
The other thing is since we're doing a simulation on it, at this time there's a current limitation where you can't use the board directly. You actually have to use it as a reference component in an assembly. So what I suggest is just save the file out, create a new blank design, and just pull the board in.
And then that way, you'll be able to update the board, and push it back. Otherwise, you can't send it back to EAGLE currently. It's something that's being developed still.
So now, let's get into the meat of the presentation. So you guys remember, the first thing I said was, you need to define your own question, your own unknowns. And you have to figure out what you're trying to solve for. That's what simulation is all about.
And our question is, will the transistor overheat? And as he mentioned, it's probably going to overheat. Let's be real.
But what we can take away from this is analyzing, how it's going to do it, and why it's going to do it, and start to figure out what we need, and kind of do that iterative process back and forth till we can find it. And that's what simulation is perfect for.
You're not testing this in real life. You're not blowing components. You're testing it. You're making accurate assumptions, just to figure out the information you need, which in our case is, is it going to overheat?
We don't really need the specific value. We're just trying to prove our theory. So in order to even start, as I mentioned, here's a picture of the board with the heat sink. But as you can see, this is really detailed.
We have those serial numbers in there. There's pins and connections everywhere. That's a lot of data to simulate.
And as he was mentioning in EAGLE, it's unnecessary. You really don't need all this data. To be honest with you, half of these components aren't even going to be affected with temperature.
And not only that, the biggest reason you need to kind of simplify is when you go to solve a simulation, you have to mesh it first. It creates a surface mesh around all the surfaces.
And it creates little facets. Once you start having serial numbers, that count goes up. And it really drives up the simulation time. So anyone that's done simulation knows you got to simplify that CAD geometry.
And it can be a pain. So Fusion introduced a nice tool. And I'd love to thank the guy. I know he's sitting in there somewhere, Mike Snell, for putting this in here with a simplify environment.
So within Fusion, there's now an environment that allows you to do it all within the CAD program, as opposed to having to have CAD ready geometry, or sending it back to the CAD program. So in here, you have some really nice features, such as remove features, which as you can see here in my little screenshot-- and I'll demo this one live-- you click a body. You specify the features you want, the size.
And it'll quickly identify them. And you can remove them really quick. So in this specific example, I can get rid of all these connection holes in one click of a button. It's nice. It's quick.
It allows me to kind of move through it. The other one are remove faces, and replace with primitives is a big one. As you can see in the example on the right, all these are now blocked.
There's no holes anymore. And now they're just very primitive shapes. And this will allow me to get a nice coarser mesh. It won't be as much information. This will drive my solve time down, and give me, hopefully, accurate results.
So I will move into a live demo in a little bit here. But I just wanted to show just a quick video of that remove features working, because, personally, it's one of my favorite tools. As you can see here, they're just grabbing the body.
And they're kind of changing the feature size. And they hit Delete. And everything updates.
So now that we have a simplified model, it's now simulation ready. So we need to set up our thermal situation. So we need to define our knowns and our unknowns.
And in this, are knowns are, we have a transistor. That is our singular heat source. We need to define, what are these loads that are acting on it?
In my specific example, I sat there for a while, trying to think of what I really needed, and what I wanted, because in real life, you're going to have the board on a mount. It's not just going to be sitting in free air. So there's, obviously, convection acting from the open air to the board.
And, realistically, there would be conduction. But in this example, I assume no conduction to the board. And what this is going to do is it's actually going to give me a result that is kind of over-analyzed a little bit, and over-constrained.
And by doing that, I have a simpler simulation. But it gives me that result I need, which is knowing what that final value might be. And I also have an internal heat, which is what I forgot to mention.
The transistor is that internal heat value. That's what's driving it. And I chose 20 watts, which I believe is from the spec sheet data on the transistor. So that's the max wattage that this transistor can put out.
So that's my heat source. As far as convection goes, just the one note I want to make-- because I spent a lot of time kind of debating this myself. Five to 25 watts per meter squared kelvin. Is an acceptable range.
And without a really higher end package or some sort of flow analysis, you're never going to be able to accurately identify kind of what the heat transfer coefficient is actually acting on the surfaces that are exposed to air. So in this specific example, I personally chose 12. It's kind of right smack dab in the middle.
It's a nice number. And it's still on the conservative side of it a bit. So I'm not under engineering. I'm a little over engineering this.
And the final point I want to make always is make good assumptions. This is what's driving your simulation, are your assumptions. So then, are we ready to hit solve?
In Fusion, we have a nice little thing called the prechecker. And I love it, because it saves you that time of, you think you have everything properly defined and constrained. You hit Solve.
You wait a few hours. And then you find out that you forgot to set something up. So we have this prechecker. And it's just a little green check mark.
And sometimes it will pop up warnings. But you analyze those at your own. But for the most part, in mine it was very straightforward. I got the green light, and I go.
In Fusion, you have two ways of solving it. We introduced a cloud option, so you don't have to use your local resources. And there's also a local option.
So, personally, I like to use the cloud solver. It does cost Cloud Credits. But it allows me to free up my compute resources. So I can send the job to the cloud, let it solve in the background. And I can go on using my computer.
So the one note I just did want to make, especially when using EAGLE boards, is you do have pinned connections. So what you're going to get is, picture your board here. You have a pin connection coming through.
So there's actually now an interference on the board, because these are two separate bodies. So when it goes to mesh it, it's not identifying it as a surface that gradually goes into it. It has overlapping phases.
So that's something you're going to want to clean up in the geometry. And it's something that the prechecker will pick up on. So, again, just pay attention to the little green check mark.
It's not necessary to get it. You can ignore the warnings. But just make sure your results make sense.
So we hit Solve. Then it's time to analyze the results. And this is really the crucial part here, is, what are you interpreting? What are you seeing?
So in my specific example, like I said, we had 20 watts, which was the max heat going into it. We know from the spec sheet data that I cannot go above the critical temperature of the transistor, which is 150 degrees Celsius.
Edwin in his Spice simulation pointed out that it should be around 1500 degrees, 15, 20. The thing is going to melt way before it even hits that temperature. And that's really what we want to know.
We don't want to know the final temperature in this situation. We just want to know, is it getting anywhere within that threshold criteria, which is 150 degrees Celsius? And then the final thing that I just like to check is kind of, do your results make sense?
You could get a value. And it doesn't necessarily mean that it's correct. So one thing that I like to use is the temperature distribution.
So you can see on the right here, we have this slider. And I'll go into this a little bit. And we can see my min and my max temperature. So my max temperature is on the transistor, which is kind of expected.
It might be a little anomaly on exactly where it is. But for the most part, my simulation told me that the max is on the transistor, which is good. And it makes sense.
And we can see that the temperature is gradually cooling as it goes here. And then on the other end is where it's the coolest. And that's where I wanted to point out about the results and the values. You can see on the nobs for-- those are potentiometers, right?
EDWIN ROBLEDO: Yeah.
JAMES YOUMATZ: You can see on those nobs that it's about 100 degrees Celsius, which doesn't really make sense. And I sat there for a while pondering if this was the kind of correct results that we're looking for, and what we're trying to achieve.
And that's when I came to the conclusion that they are right, because, realistically, before that ever hits 400 degrees, it's going to fail in real life. So that potentiometer is not going to hit 90 degrees Celsius. And the board is still kind of transferring the heat in a correct way, a way that makes sense. So pretty accurately, I could say--
EDWIN ROBLEDO: And just to point out, when we were pushing and pulling this, those nobs traveled this whole board. At the beginning, they were over here. And we said, no, that's too close. They went back here. The LED was somewhere too close as well. So we went back and forth, like I would say like 90 versions until we actually got here to this point.
JAMES YOUMATZ: Right. Now, I threw a lot of numbers at you. So why don't we just actually dig into Fusion, and kind of take a peek at what I'm saying? So let me go ahead and bring Fusion up.
Unfortunately, it can't bring the results up. That was the issue I was having in the beginning. For some reason, it did not want to load in Fusion.
That's OK. So go ahead and bring it up. And let's go into the first one, which is this simplify environment.
So you can see here, here's my board, the pins, the LED numbers, the serial numbers. This is just too much. This switch, very unnecessary.
So we go here. And in Fusion, a nice tip that I always like to tell customers is work left to right. They always make their menus in such a way where you just can work left to right.
So the first one is, we defined our study. Now, it's time to simplify everything. So we go in here.
And you'll notice that I now have a unique timeline. And I've done a little bit of stuff already just to get it ready. But this timeline is separate from anything in the model environment. This is unique to its own.
So the ones that I want to demonstrate, and my favorite, is this remove features tool. So as you can see, I have a body. And I've gone ahead and grouped things together to make this just a little quicker.
But I can click this. And as you see, when I change the feature slider, more things are getting selected, less things are getting selected. And you can see that I can actually kind of toggle on and off what I need.
So in this example, I'll go ahead and just hit the Delete button. And it's gone. And you can keep going on an iterative basis until you finally get what you're looking for.
And you can also right here kind of choose the manual features. And you can see now, there's really nothing left to remove except for the whole board, which is what that color is highlighting. The next one is this replace with primitives.
And, really, this is a powerful tool. And I didn't really realize how powerful it was until I started digging into it. So as you can see, there's tons of little resistors, tons of very similar instance components. And what this is, is it's actually smart enough, you'll see-- let's replace this with a box.
And that's roughly the same shape. So that's OK. We'll go ahead and hit OK.
You'll notice other ones are turning with it. So I don't need to apply this to the whole board. So I can just quickly start going ahead. And I'll say, OK, let's grab those.
They all change. And we can keep on going. The only the other thing to keep in mind is just, what do you actually need as far as geometry goes?
So take the switch, for instance. Do we really need the switch in here? All it is, is adding extra geometry.
So what you can do with Fusion is if you do a quick double click, it'll highlight the whole thing. You can right click, and say Remove. And it's gone.
It's not on your board anymore. It's not something we need to analyze. You can very quickly kind of go through it.
The last one that I really kind of like is Remove Faces. Where this is kind of powerful is let's take the plugs here, for instance. For anyone that's used Fusion before, and kind of removing faces, you'll know that Fusion has a heal command.
So when you delete a face, it will try to automatically heal it. And what that does is, so you can see here, let's say I hit Delete on that face. That's not right. Look, put a hole in it.
That's not what I want. So what this feature is really smart about is it tries to grab all the faces that are relative to it. And it tries to delete what you need.
So as you can see, I don't see the holes. I really just need a block here. So what I can quickly do-- and you can see it starting to highlight it. And it's showing you what you're going to delete.
Just start grabbing faces. And now you can see, it's selected 92 with four clicks. I only chose four clicks.
I hit Delete. It's gone. And that's now ready. That's a good looking piece.
So now, we have our simplified geometry. It's time to set everything up. Now, the first thing that you need to do when you set it up is go over your materials.
And this is something that, personally, you work with it. When you're trying to analyze your results, maybe it's not best to use a custom material. So for this specific example, we used FR4, which is a material that's not in the material library.
And I don't want to show it too much. But in the material library, what you can do is you can go in here. And you can start to assign materials.
And as I said, I've kind of group things together to make this quicker. So I didn't have to go through and kind of grab all my resistors, and grab all the components. I can just very quickly grab pieces at a time.
So I've grouped it together. And you can see right now, the board is an FR4 material, which I've gone ahead of time and kind of created. If you're more interested in information about that, if you download the PDF handout of this, I kind of go into how to create your own materials a little bit.
But so for this, the board is FR4. If I can remember my alphabet, there it is. Oh, this is Edwin's computers. So FR4 isn't in here. But it's on my computer.
And on this one, we can see I'm going to go ahead and assign everything to be aluminum 6061. Let me just hit OK. Now, my materials are defined.
The one thing I just want to mention is check the values. Go into the table values, and look, and see if those thermal properties are correct. The next thing that we need to do is define-- and again, we're working left to right. I've done my materials.
Now, it's time to define the loads. These are what's driving this simulation. And this is a very interesting menu.
And I think it's really neat. You have these selections. But you also have these object types that can really speed up the process. You can either apply a load to a face or the entire body.
So in this specific example, I really just want to apply a load with an internal heat of 20 watts. That's setup. It's quick. It's done. My loads are displayed here. We're good.
The next is we need to define the convection. And like I said earlier, we're defining convection to the entire board. We're saying that this is exposed to air.
And the reason we're doing that is, this is going to show me the highest temperature profile. What I mean by that is air is not a good medium for transferring temperature. And we know that.
It doesn't dissipate to air nearly as efficiency as if you were to put it on a metal plate. It would conduct to the metal plate. And the metal plate would take some sort of that heat away from the board. It would kind of draw the heat away.
So for this, for simplification purposes, I said, let's just expose it to convection, worst case scenario. So this is a nice little tool. It says, Select All Faces. So what I can do right here, I turn on Select All Faces.
I click this, real quick, it has every single face that's exposed to the air right now. It's just floating. And this will change it to convection, which my value, which I defined myself, was 12 watts per meter.
Now, on my board, since I've gone ahead and done this manually, you would need a define contacts. What contacts is saying is, kind of my example with interferences, is you have your transistor with the pins. And it's trying to go through the board.
You need to find that contact to let it know that, hey, these two touch within some sort of range. And Fusion is good enough, if you were to hit Solve, it will throw up the automatic thing and automatically detect it, which is for the most part, pretty accurate.
But if you wanted to manually define it you can actually go in the context here and start to define those. And I'll go into this a little bit more when we get a heat sink. But this can come in handy when in real life anytime you're going to put something against a heat sink you're going to use a thermal face.
And you do that because when you machine these parts, there's always little imperfections and dents. And it doesn't sit perfectly flush. And there's not the best heat transfer because of that. So what they do is you can actually go in.
And you can edit your contact set. And you can start to define those resistances through it. And that's if you wanted a more accurate result. If you don't define it, which is what I ended up doing, you're assuming that the heat transfer is perfect. The temperature on the back face of the transistor is going to be the same temperature on the front face of the heat sink.
So now that we have that solved, I just like to do my self-checks. And this is really important is this degree of freedom here. And this is true for any simulation you're going to do in Fusion. It's especially helpful in FVA analysis when you have assemblies.
But this degree of freedom view kind of shows you what the board is doing. And this is a good way to check your contacts. For instance, if the transistor was green, and everything else was red, I know I set something up wrong, because that's saying that the board is fully insulated. There's no heat.
And like I said, with the stress analysis, that's even more important when you have moving parts, because what you'll start to see is that something might be fixed with relation to the other. But the other has a free range of motion. So you know your setup is wrong.
And we're doing this so we don't waste the time and the money and the solves. So check my prechecker right here. And we can see, she's ready to go. I got the green light, let's floor it.
So I'd go ahead and hit solve here. And I have my options. I would choose Cloud. It's simulation ready. And I would hit Solve.
So I wish I could bring up the results on this. Unfortunately, for some reason Fusion was acting really funky this morning and wouldn't let me pull them up. But we can see from the example in the PowerPoint that it's going to overheat.
It's too hot, way too hot. So at this point, I need to send this information back to Edwin. And since Fusion is a cloud product, it's really nice. When I save the file, I can create comments in it. And he would see these comments reflected in his A360 browser.
And in this example, Edwin is out of Florida. I'm out of Boston myself.
I said, hey, Edwin, just like you thought, it's going to fail. Let's add a heat sink. Let's add a big one.
EDWIN ROBLEDO: I would get that notification in EAGLE as well.
JAMES YOUMATZ: So let me bring up the PowerPoint. OK.
EDWIN ROBLEDO: So now he bounces it back to me. So return to sender, I'm the sender. Now, he bounces it back to me.
So he verified what we knew, too hot. So now, I have to look into the specifications of the manufacturer, and see is, what do I need to modify on my board? That way, I could get within a range of operation for this component. Could you go to the next slide?
Now, what I do is I actually look at the specification sheets of the component. And I find that there's a form factor of three if I use that heat sink. So before, we were going up 75 Celsius for every watt. Now, what we're actually going to do is that we're going to go up three degrees for every watt. So that gives me-- the three times the 20, gives me the 60 Celsius temperature.
Now, in between the MOSFET, as well as the heat sink and the MOSFET, there's this glue. It's a type of thermal grease that they use to glue them together. And that gives me a factor, a junction case. This is called a junction case of 0.084. If I take into consideration the glue that is going to be used between both of them, I could actually assume this is going to be one.
So between the body and the heat sink, I get 1, times 20, gives me 20 degrees. With the thermal, I get three kilowatts. Multiply that by 20. So I get my 80 degrees.
Now, I have to add my 20 from ambient, because, remember, this is a rise. So I get a grand total of 100 degrees Celsius, fully within the range of operation, fully within the range of operation. So for then, through this method, I've been able to prove that if I add a heat sink, I could go ahead and actually get it operating.
Now, he's going to go ahead and continue doing more simulation through Fusion. And I'm using that heat sink, specific because of the high rise. But you'll see the differences. If we use different heat sink sizes, what is the thermal reaction on that? So, basically, I'm assuming that with this heat sink, with these values, I'm fully operational. And now I send it back.
JAMES YOUMATZ: And just the one that I want to make is that we can see from the simulation, the 80 degrees, that's actually a delta-t. As you mentioned earlier, that's without ambient. So, realistically, his simulation is telling us around 100 degrees Celsius is what this transistor is going to operate at with the heat sink, which is within that allowable range.
Keep in mind, 150 is where that thing maxes out. So we don't really want to be too close to that. We kind of want to include a factor of safety.
So now we need to update the board in Fusion. And this is the part that I cannot stress enough. This is my bread and butter. This is what I love about having this feature in Fusion.
i don't need to go back to Edwin, and say, hey, can you export a STEP file and send it to me? Give me a Dropbox link. This will now update my linked assembly in Fusion.
And what you'll notice, we can do this pretty quick. And that's really the why of this presentation, is everything can be done quickly. My specific example, it's obvious.
You need a heat sink. You do. But the question is, maybe that's not the right size. Maybe we can go cheaper. Maybe we can make it a little smaller.
These are all questions that engineers, designers, ECAD guys, MCAD guys, they ask. And it's this very iterative, collaborative process between the two that makes it really nice and really fast. As you can see here, this is what Fusion looks like when he makes his board change, because it will automatically apply to my board.
And keep in mind, it's its own file. So as you can see by the chain link, I'm using this board in an assembly that has a base feature that can be machined, et cetera. So all you have to do is right click it.
You hit Get Latest. It'll update your board. That's in the model environment.
So now, we need to move into the simulation environment. And when you go in there, I'll demo this in a sec. But when you go into the simulation environment, it's going to recompute everything because it has a specific timeline again for Simplify.
And this is where it's powerful. It completely blew me away on how accurate it was. And what it did was I didn't have to set up any of my geometry again.
It knew it. And the loads were still there. My constraints were still there. I didn't have to spend that 20 minutes, 30 minutes, an hour, two hours, setting everything up, simplifying my geometry again. It just updated.
The only things I really had to do was clean up any reference areas that were there. Since we added a heat sink, we extended the board length a little bit, which caused a few things to be out of place. But it's as quick and easy as just redefining what the reference was.
And then we're good to go. And then it will recompute it. We add our load to the heat sink now, because this is a new element. So we obviously have to add convection to the heat sink, and show that this is now exposed to air, and then just set up all our new value conditions.
And then we can hit Solve again. And this is what I really wanted to show. And keep in mind, this is not flow.
Fusion does a steady state analysis. So this is just capturing one moment in time. And what I'm doing on the right there with the slider is I'm actually just starting to filter temperature values out.
And you can see, when this restarts real quick, that, really, the hottest point is on my transistor. Remember, we're engineers. You got to keep that in the back of your mind. Am I seeing what doesn't make sense?
And that's what I'm trying to show is, as the temperature distribution is kind of laid out as the hottest source is our heat source. That's accurate. Then the heat sink takes a bunch of that load.
And it's kind of evenly distributed, which is good. The area on the board due to our contact has all of the temperature here. And then the board, being the material that it is, is now dissipating heat very efficiently to the point where when we get to the potentiometers in the back we're hitting an ambient temperature.
So that's good. Realistically, there might be a little bit of heat applied to those. But it's so far away. And it's its own component. Depending on the materials, and things like that, we can expect when you touch that potentiometer, you're not going to get burned. That's good.
So real quick, just before I hope into Fusion, let's go into a little bit of math. So in this specific example, since we are pretty sure now that this heat sink is going to work. We saw the results.
Actually, let me go back. i don't think I pointed that out real quick. I was able to do a probe real quick.
I didn't want to take this value, because it's kind of an anomaly. It's pointing to the hottest point being that contact right in between the pin and the board. That's realistically not going to happen.
And it's only off by a few degrees. So the average temperature on the transistor, which is really what we want to look at, is about 102, 103, degrees Celsius. That's good.
That's under the 150. We've hit our design criteria. But now we need to check that.
We need to verify that with a few hand calculations just to make sure that we're kind of in place. So as he mentioned, he did a Spice simulation of this. And you can kind of see the outline here. And I'll let Edwin talk to it a little bit more.
V means temperature, not voltage. And what he did was, it does the math. And what it was saying was--
EDWIN ROBLEDO: So one is the heat sink. One is the grease between the heat sink, as well as the MOSFET. So you get the 80 degrees, which is the 80 volts, plus the 20 from the grease as well, giving you a grand total of 100 degrees.
JAMES YOUMATZ: Correct. So hit was pretty close. His simulation was about 100 degree Celsius. Mine said about 102, 103, in Fusion. And then so I just wanted to verify mine real quick with a very simplistic approach.
This is just getting you kind of an average convection value. And that's what that equation is. It's convection.
it's q, equals your heat transfer coefficient, where q is your heat source. And then that delta t value is the difference between ambient and your actual temperature. So I assumed h was 12, just standard value that I chose.
The area, which is what makes this calculation a little bit tougher to approximate by hand is, you're not really sure what the area that's being done by convection. So I just kind of assumed the heat sink area.
That would give me my lowest amount of area. And it's saying that this is the only amount of area that it can dissipate with. And, obviously, my ambient, I chose 20 degrees Celsius. And q, which was the heat that we defined from the spec sheet, was 20 watts.
My hand calc said around 87 degrees, which is a little bit on the low side. But, again, there's some values in there that are kind of specific to making good assumptions. So a little bit on the lower side, but it's still within that range of being accurate.
It's within 10%. So it's pretty good. So let me just hop into Fusion real quick. And I'll show kind of the final portion.
OK, so you can see in here, here's the board. And I had just updated this. And that's what I wanted to show with the simplify again, is that I had just brought the board in. It's now been updated in the model environment.
I go into simplify. And this is what it looks like. And this is what I was trying to demonstrate, is that you don't have to do everything over again. It remembered everything.
The only thing that I need to add now is just all my values to this, which I can easily go ahead and do, just by doing the convection again. And we'll quickly go into our results here. So as we can see, if I just real quick turn on the min and max values, that we have our temperature here.
This is ambient. And we've answered that question now, that underlying fundamental question, which was, do we need a transistor?
EDWIN ROBLEDO: Do we need a heat sink?
JAMES YOUMATZ: And I think we can say accurately that we do. And at this point, now is when you can start to make those design changes. Maybe this is a little over-engineered.
We're about 50 degrees off of what it can handle. So maybe we use a smaller heat sink. Maybe we go back, and we make these changes. And we push it back and forth.
EDWIN ROBLEDO: And modify the enclosure.
JAMES YOUMATZ: And that's really the why that I wanted to point out in this class. This is meant to be a quick process. And it can be. It really, truly can be.
And we can even go back into Simplify, let's say. And we wanted a shorter heat sink. You can very quickly start to drag these, and make your modeling changes in Fusion. And you can do it in EAGLE as well, and bring in an actual component from a website, and get them spec out. And you can run through these process multiple, multiple times, until you finally get that question and the answer you're looking for. That's our presentation.
EDWIN ROBLEDO: Yeah, we're way over our time. So, anyway, so as you could see, the work flow of going between EAGLE and Fusion, as we did yesterday, it's pretty nice. And I didn't know much about the simulation portion.
So I really relied on James. And we just added simulation to EAGLE. So now, you can see what the workflow, and how this affects your product.
So if you're working on a product, all this analysis now actually work together. And now you can determine what your product is going to look like, or what modifications. So, basically, that's what, in a sense, what we want to do. So if there's any questions for us.
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: I'm sorry, you're saying the holes?
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: No, they are misaligned. That was something that I--
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: No, no, no, no, not at all. And that's, like he was saying, those are the little design adjustments.
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: No, no, not at all. And he's right. And that was something that we had done. And, again, part of the difficulties were these pins.
These are very small components. They're a little hard to mesh. And what I was finding was that the location-- because I had done a few simulations on my own to kind of prove that. It really wasn't changing that anomaly too much.
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: Because it's not filtering too much. And I actually cheated a little bit. I went into the convection services, and started filtering them out.
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: Perfect. And I know simulation stuff in the dead of morning is a little tough. So are there any other questions to open it up?
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: One more time?
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: No, I didn't add radiation to it, because I wasn't really too interested in how that would radiate heat, or at least personally. And this is what I was kind of getting at is I didn't want to get too specific of an example for this class.
I really wanted to keep it open, and kind of show that the question that you're going to pose to yourself as an engineer or designer is that's up to you. That's on your own. What do you want to figure out?
And in this case, radiation, I played with a little bit. It was really increasing my solve times. It was making things a little bit more challenging.
And like I said, this is kind of exposed to air. We're doing a very hypothetical situation. In real life, the original model, which I can show in a bit, actually has a big machine case around it, which now starts to affect your convection values and things like that.
And it makes it a little tougher to kind of figure these values out. You might need a program like fluid dynamics or something like that to kind of understand what the flow is acting on the source, and same with radiation too. So I just want with a very simple kind of over-constrained example of that. But you're right. In real life, it would radiate heat a little bit to the outside sources.
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: Yeah, it's a little funky, to be honest. We ran into that a little bit. And I think he can speak better to that. But he was saying if something's incorrectly mapped. How does it come into Fusion?
EDWIN ROBLEDO: You mean mapped in what sense? You mean the component itself if it's mapped? No, so we're talking about in the simulation aspects of Fusion being mapped. Is that correct?
JAMES YOUMATZ: No, he's just saying--
AUDIENCE: [INAUDIBLE]
EDWIN ROBLEDO: No, so that information is not transferred. That only survives in EAGLE by itself. So that information does not get transferred over, no. No, so those models are-- so, in other words, Fusion has no electrical capability, or nothing like that. All that stays in EAGLE. So by the commentaries, I let him know my values, because that's what we're interested, is on the results, basically, correct.
JAMES YOUMATZ: Any the other questions?
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: Yeah, I posted them. I was having a little issue uploading them. I was able to do it this morning, luckily.
And the one that I did want to make is if you're looking for more specific information, I kind of went all out in the PDF handout for the class. So definitely take a peek on that, because I tired, for the presentation at least, kind of tailor to my specific example, and kind of stay within the rails a little bit.
But I could talk all day about this. You can define your own simulation. That's what it talks through. It goes a little bit more into the theory, and kind of, what are you looking for? What are you trying to define? What are you trying to do?
And it gives you the step by step on how to set up your board in EAGLE. And it gives it very kind of in-depth. And I tried to make this not a four-hour class at 8:00 in the morning.
EDWIN ROBLEDO: Yeah, we could have easily done a boot camp out of this, probably.
JAMES YOUMATZ: Perfect.
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: Yeah, it is.
AUDIENCE: [INAUDIBLE]
JAMES YOUMATZ: Right, let me hop into Fusion. I can explain it a little bit easier. So I'm just going to toggle over the model environment. And this will spell it out a little bit easier.
So I'm assuming based on the badge, you have a little bit of Fusion background. So as you can see here, this is linked. This is my board that came in from EAGLE.
And here's the original file right here. The reason you have to do that is there's two criterion that we found out when we were demoing this. And since this is a preview feature, it's something they're still working on.
What we found was that if you were to add something to the board that changed-- I'm going to butcher this response. But it's pretty much along the lines of the way that Fusion handles components, and the information that's going through it.
So, for instance, you can't have a linked assembly. So that board can't be on its own in a linked assembly. So if you have any x reps in the file-- so kind of the vise versa or this situation where the board has its own file, and you start to throw components into the board file. What we found was when you start to pull, references are getting lost and component misinformation.
So what we did is we disabled the feature. And that's what I was saying I found out the hard way, was I had set everything up on my original file, and got that big warning that popped up. And it will tell you, it won't let you do it.
So my suggestion is, you create the board as your own file right there. And then you make all your changes to the file. So when that updates, what you'll get in your assembly file-- because now I've pulled that file into my machine part assembly. And I've lined it up and everything.
And what you'll find is you'll get a little out of date warning when he updates the board. So now, all I have to do is hit that Get Latest button. And it'll put the information in. Anybody else? Perfect.
EDWIN ROBLEDO: And also, if you look up for the handout of the class we did yesterday all about the pushing and pulling between EAGLE, it's pretty well defined there. Even though this handout, we added some of that information as well. So there's details there as well. OK.
JAMES YOUMATZ: Thank you all for attending. I hope you enjoyed the class.
EDWIN ROBLEDO: Really appreciate it.
Downloads
Tags
Product | |
Industries | |
Topics |