Internet Explorer is not a supported browser for TI.com. For the best experience, please use a different browser.
Video Player is loading.
Current Time 0:00
Duration 42:38
Loaded: 0.00%
Stream Type LIVE
Remaining Time 42:38
 
1x
  • Chapters
  • descriptions off, selected
  • en (Main), selected

Hello, and welcome to industrial TI tech day. This is the buck switching regulator power modules get to market faster with reduced EMI and smaller power supply size session being presented by Alejandro, and myself, Stephen. I'll be the moderator for the session. All participants are muted for the session. Please use the chat function to ask a question and address it to everyone.

Questions will be addressed at the end of the presentation if we have time. Also chat if you're having any problems hearing or seeing the presentation. Thank you for joining.

And a little bit about ourselves, so my name's Stephen Ott. I'm a marketing engineer for the wide [INAUDIBLE] converter group with TI. I started my career back in national semiconductor in 2005 in power management, and I've had various roles across the power portfolio in National and TI. I'm based in California out of the Santa Clara.

Hi, everyone. I'm Alejandro Iraheta. I'm an applications engineer for the Wide Vin Converter and Modules Team at Texas Instruments. I'm a recent college grad. I started my career with TI in 2018, and I've been involved with power electronics since. I'm currently based in Warrenville, Illinois.

Thanks, Alejandro. Next slide. OK, great. So today, we're going to cover several topics. We're going to go into how modules speed up the power supply design cycle and save board space, various EMI sources and how to mitigate those using modules, the inductor selection and design trade offs when the inductor is fixed inside the module so looking at how the value compares to a converter. Then we're going to go into some new module package technology that we recently released and product highlights within those packaging technologies, and then we'll wrap it up with design tools and support.

And just so that we're all talking the same language here, stepdown bucks switching regulators, they're segmented into three categories. On the far left, we have a controller, which has just the control inside the silicon with external FETs and external inductor to make up the complete power supply. In the center is the DCDC converter or stepdown buck regulator.

That includes the controller FETs on a-- inside the package with the inductor external. And then on the far right is the module, which includes everything, the controller, FETs, inductor, and many of our modules also include internal caps and resistors for additional filtering and programming. So with that, I'm going to hand the presentation off to Alejandro to go into the technical topics.

Thanks, Stephen. Hi, everyone. So let's begin. So how exactly do modules shorten the design cycle? Well, when using power modules, you know that a power expert has done the heavy lifting up front, and by heavy lifting, I mean the engineering work involved with the power design. So to be more detailed, a power designer has picked up the optimal converter IC to cover across the design applications and specifications as well. Also optimal components were selected by the module design engineer to match up with the converter IC.

Afterwards, the solution was then characterized across all of the corner cases while implementing the latest packaging technology to achieve the best possible trade offs and also meet market requirements. In terms of the internal module component selection, TI has a very strict component level qualification due to the fact that not all inductors and capacitors are created equal. So when using module, the end user doesn't necessarily need to be involved with the following, like, control architectures schemes, the inductor selection, qualification, any layout challenges that we can all agree it comes with. Plus prototyping can be quick and efficient due to the simplicity of the solution. All of these factors can definitely help shorten the design cycle of DC to DC converter.

Here's one example of how simple the power design can be with the power module. In this example, we are using one of the newly released power modules from TI, the TPSM53604. This module is rated for 36 volts on the input and up to 4 amps of current on the output. In terms of design steps, the engineer can get a converter pretty much in three easy steps.

Step one, it's highlighted in blue, pick up a couple resistors and set the desired output voltage. Simple as the simple reference to table one in the data sheet. That can help with just setting the most common output voltages. Also I know it might be difficult to tell in this small table, but note how table one gives you a suggestion for the required output capacities needed for that particular output voltage. I'll talk a little more on that in a few seconds.

Step two in green, select an input capacitor. In this case, we can refer to table two of the data sheet since we know that the input cap is a very important component of any buck converter. We make sure to provide the information on several high quality capacitors we have tested in the lab.

And step three in red is to select an output cap. In this case, table three is listing several components from various vendors, and that's to help the engineer eagerly pick capacitors to meet the required value for the output voltage that we stet back in step one. And finally, as most of us know, the board layout is an important aspect of the converter, so in the data sheet, we provide a good layout example plus layout files are available for download on the product page on TI.com. These resources can definitely help speed up both the schematic and the layout of the design. Simple as a copy-paste from our already characterized evaluation or design.

So speaking of characterization, since you are not selecting an inductor, we are able to provide a few curves in the data sheet as shown on this slide, and that's to illustrate the performance of the module, what you're going to get when you purchase one. So here we have a few curves showing, for example, conversion efficiency, the safe operating area when using the EVM layout I mentioned about in the previous slide, the dropout characteristics of design, also some waveforms, startup behavior along with shut down behavior of the module and the low transient response. There are many more curves in a typical model datasheet. And, again, all of these closely represent the performance you're going to get since the conductor has been chosen for you, and it's already part of the integrated package.

The integration of the inductor brings additional benefits in terms of board savings. Here are a few examples to visualize the board area reduction. So, for example, on the top middle, we have the TPSM53603 module using the LMR33630 IC inside. On the top right, we have the TPSM265R1. As you can see in all these examples, the area of savings can be significant when using a module.

So one question that definitely came to my mind when I first joined the power module group was what about thermal design? And I'm sure a majority of you are having the same question. This actually is a topic that concerns many power designers and is often questioned by customers, and that's OK. It's valid, makes sense, especially when one is integrating all of the major power dissipating components into the same package.

So that is why we usually do spend a lot of time characterizing the thermal performance of the module package, and we end up providing the safe operating curves in the data sheet. These curves can provide the engineer an idea of the capabilities of the parts for different Vin, Vout, Iout, different ambient temperature conditions, and just basically show any potential derating for high ambient temperatures or output power conditions. Again, these curves are collected on a known good layout, such as the EDM board of the device, but-- and we also understand that not all designs have or need to board areas of EDM. We do some additional characterization to help with thermal design.

So that additional information we provide to help is the border area versus [INAUDIBLE] thermal performance curves. Essentially we characterize the thermal performance of the package on several board sizes and then provide this info to the end users, so that way, engineers can quickly get an estimate of the board area needed for their specific application. There's an application that goes into more detail on this, SNVA848. It's available for download at TI.com. I recommend taking a look at it. There's a lot of great info on there.

Basically, to give you a quick summary of the [INAUDIBLE] spoiler alert. The engineer will look up the power dissipation for their particular application. Then you would calculate the required thermal resistance from the passage with given the power dissipation ambient temperature conditions. And finally, plot the results, and you will get an estimate on the board space required to meet this thermal resistance requirement from the package. As you probably notice by now, thermal design is an important topic for us.

Another important topic for any switching converter EMI mitigation-- but before we look at mitigation, let's get a quick understanding of the noise sources in the design. Here is our favorite ideal buck regulator. At the input on the left side, we have some DC voltage and input cap. Also there are two power switches, the high side and the low side [INAUDIBLE]. Basically, through their switching action, they chop up the input voltage into a pulse voltage with a particular duty cycle in the output inductor, and the capacitor filters, and pulse voltage.

And depending on the duty cycle, we get a certain output voltage lower than the input. As you can see in this diagram, everything is clean and ideal in this case. But in reality, there are a few more components that show up in buck regulator in the form of parasitic elements. They're highlighted in red. They are free, but most of the time these are unwanted and cause some trouble.

There are parasitic conductors associated with the IT package, the input capacitor, and its placement with respect to the power switches in series with the high and low side switches. Also there are parasitic elements associated with the inductor in the output cap, and also there's a high di/dt loop on the input side of the converter. Most of the problems customers experience with switching power supply to noise are related to the parasitic component in the design.

As you can probably tell, having excessive inductance in the high di/dt loop can create unwanted ringing on the switch node. And since the inductor is not an ideal inductor, this noise can make it all the way to the outage voltage. I'm sure many of you have observed this with buck converters in the past.

Let's take a look at an example. As you saw in the previous slide, there is a high di/dt loop associated with the input cap and the power switches of the IC. In this example, we're placing the input cap a bit further away from the IC. The approximate loop is shown in red on the left side.

If we look at the switch node of this 12 volt input design system, we can observe rigging or spiking up to 18 volts. Also the frequency noise on the output is about 75 million volts peak to peak. And in terms of radiated EMI, we see that we are just violating the class B specs.

So what happens if we position this input cap differently? We can see that minimizing the di/dt loop area, which is reducing its inductors in the process, can have a positive effect on the switch node ringing and the radiated EMI. In this case, bringing the input cap closer to the [INAUDIBLE] and power [INAUDIBLE] into the IC. It reduces the switch node ringing, so 18 volts went down to 14.5. Also the output noise went down to 47 millivolts from 75 millvolts peak to peak, and also the [INAUDIBLE] EMI improved.

So I think this speaks very loud on its own, but I'll help it out a little bit. The input cap decision affects the output noise. Many people may not realize this immediately when looking at a buck, but it's a fact.

OK, so with this background in mind, how exactly do modules help with the EMI? Well, there are a couple advantages modules can bring, so by integrating some of the input caps extremely close to internal IC, it will reduce the high di/dt loop area. And by making the internal switch node as small as possible, it would also reduce the dv/dt node area and therefore reducing the capacity coupling of noise. So these two advantages can definitely help with mitigating EMI generation and passing the necessary EMI specifications.

On the bottom right corner of this slide, the visual example of what is going on under the hood of the two modules, TPSM53603 and LNGM33603. Notice how there's a high frequency filter cap placed in the ideal location along a shielded inductor. Here is a quick visual on how modules can shrink the size of the di/dt loop by integrating some input caps inside the package or just simply by carefully choosing to pin out of the module to facilitate good positioning of the input cap.

So here's one example of how the input cap placement can greatly affect the EMI performance. The blue curve shows the performance when the high frequency cap and the main input cap are as close as possible to the IC. The red curve shows what happens when the main input cap is further away while the high frequency small cap is close to the IC. The green curve shows what happens when both high frequency and the main caps move away from the IC. And the purple curve shows what happens when the high frequency cap is removed while the main cap is also far away.

Here's a demonstration of a power model using the same IP inside as using the previous side and integrating high frequency cap inside the module package. Clearly, you can see that the placement of the main cap is a lot more forgiving than the regulator. The blue curve shows the result of the good placement. The green and red curves demonstrate what happens when the main cap moves further away. In both cases, you can see that, even with suboptimal component placement, the integrated caps help with meeting the EMI limits.

And just to summarize, here's a table of the actual peak readings. So the module can get better EMI results, even when both solutions have an optimal layout. And the module can get better results when both solutions have equally poor placement. So there's a lot more details on this experiment in the white paper noted on the bottom here. This can be found on TI.com as well.

And here are a few more radiator EMI curves from several of our modules, typical data for 4 volts and 24 volt rails for four different products. All of these meet the class B limits. OK, as I said before, the power modules integrates the inductor inside, so let's take a look at some of the trade offs associated with this.

It's safe to say that the inductor is one of the most important components for a converter and also its performance. So there are several factors in which make it so important, so the inductor's value affects the overall size of the component, the ripple current, the transient response, the peak currents by the power set, and the overall efficiency. The winding resistance, or DCR, also affects the size, efficiency, and the dropout performance of the converter. Also the core material can affect the overall size and efficiency. The saturation currents spec, now that can affect the [INAUDIBLE] peak current and the overload production feature of the converter, and the shielding of the inductor can affect the EMI performance.

There are strict requirements when selecting an inductor or any other patented components that will be included inside of a module. We have approved vendors and use industrial grade components. Besides the vendor qualification, we also do our own [INAUDIBLE] component level testing and reliability. Some of the additional screening is listed on this slide, so we have lifetime tests, high temperature storage tests for inductors, breakdown voltage, and other tests that exercise and fully characterize the components we include in the BOM.

Another question that we often get is what are the trade offs with using a fixed inductor value? Well, a set inductor value can definitely put constraints on the minimum and maximum switching frequency of the module. Also the output voltage range can take a hit, and the overall conversion efficiency as well.

Here on this slide, I've listed two examples. The TPSM53604 is rated for 1 volts to 7 volt output, where the IC inside the module can support up to a 24 volt output. Same scenario on the LMZ36002. The output range is there and the IC inside. In general, modules are often optimized to cover the most common output rails with good performance.

So let's take a look at one example from our wide input voltage design. So let's say we were designing a 24 volt to 5 volts buck with a 4 amp output, and wanted about 30% ripple on the inductor, and a switching frequency of 1.5 megahertz. Well, if you look closely on the third table-- I know it might be pretty tiny. Apologies on that. We can find that a 2.2 [INAUDIBLE] inductor may be very suitable.

The value could also be suitable to cover the adjacent voltages like 3.3 volts, but we can see that at corner conditions, IVin, IVout, or low Vin and low Vout. That value may be really undersized or oversized. So what we can do is shift or tailor the sweet spot by turning the frequency knob and end up with a much better operating point.

For example, with the 6 2.2 [INAUDIBLE] inductor, it would be best to operate at 1 megahertz for 3.3 volts on the output and for 1.8 at 500 kilohertz. You'll save on some switching losses while still getting the same amount of ripple. And if in your application, you are not able to change the frequency from 1.5 megahertz, then the trade off for using this inductance at lower output voltages would result with lower efficiency.

And let's say if we were trying to get to higher output voltages, let's say 12 volts, ideally, we would have to switch to a higher switching frequency like 2 megahertz to keep the ripple under control. So these spots here include inductor ripple current as a consideration, and there are definitely more factors involved in the design process. But as you can see, having a fixed inductance can put constraints on the operating conditions and the outward range of the module. Typically, the module data sheet will include a table to help guide you on achievable switching frequencies for your desired output voltage while taking into account your input voltage rail.

Here's a comparison between a module and a regulator in terms of efficiency. This particular module, TPSM53603, it happens to use the same inductance value as the regulator for this condition. And as you can, see the module can achieve a smaller solution size and similar efficiency at the regulator.

Here is another comparison. In this case, the module is using a much smaller inductor case size inside. I've noted the inductor sides with the blue square as a reference point. You can see the delta, and the solution size, and efficiency for the same operating conditions. So the regulator can achieve better efficiency, 2% to 3% higher, but the trade is obviously solution size.

So thanks, Alejandro. That was a lot of really good detail. So to summarize what Alejandro just reviewed with you, the trade offs between the module converter and controller vary in terms of availability, the amount of products we have. In terms of breadth of portfolio, of course, converters and controllers have been around since the beginning, so there is a ton of options for you to select from.

Modules are relatively new on the market. There is fewer options, but it's emerging using our newest technology. Design difficulty, there really is no argument that a module is easiest to use. Depending on the product that you select, we do have some modules that have everything except a single resistor integrated so-- sorry-- a single resistor integrated for just setting [INAUDIBLE] voltage. That makes a really easy converter.

Of course, you have to select the conductor as well as depending on the specific converter the way it was designed. It could require external compensation. And of course, the controller is the most flexible, but you're going to have to do some pretty substantial design work to make a controller design.

In terms of solution size, usually the control is the smaller option, especially with the 3D construction techniques that puts the silicone under the inductor. The converter can be made pretty small by increasing the switching frequency or doing dual-sided layout techniques. In terms of EMI, the module is good out of the box. Converter controller, you can also optimize those to have very good EMI.

It's going to depend on the selection of components and layout. And cost, modules integrate more of the bill of materials, so the cost is usually a bit higher than the converter controller. But as technology progresses, the modules are getting more competitive on costs.

So which is the best fit for your application? Really depends. I'd say consider a module when you have limited time and resources, your design is limited on board space, or you have limited engineering resources. Consider a converter or controller when the actual component cost is your number one priority and you have a lot of ample engineering resources and time to spend on the design.

The application has some very specific requirements. Sometimes a converter controller is necessary due to the limited ranges of the module like what Alejandro showed. And as well as just to close that, the newer emerging [INAUDIBLE] technologies were released, and the cost delta is diminishing.

So I'm going to take a quick pause here and open up a pole to help us understand your requirements better for our roadmap development, and then we'll continue on with the presentation. Should be pretty quick. Just two questions. Thanks a lot.

OK, thank you very much for taking the poll. Alejandro, can you move forward here? So now we're going to go into our new package technology and a few of the products associated with those packages.

So the newest module package technology that we've recently released here is Enhanced HotRod QFN package. It integrates a HotRod silicon on a lead frame along with bypass capacitors and passes around the IC. In this case, the inductor can be stilted over the rest of the circuit like in this example here, which provides two main benefits, smaller footprint, and we can also use a higher performance inductor with lower DCR.

The magnetics are allowed to occupy a larger area of the package without sacrificing the actual footprint area. The package is overmolded with mold compound, and the bottom manualization allows for a very friendly customer footprint with a large thermal pad to facilitate the heat transfer. This footprint, it looks like a [INAUDIBLE] with a large brown pad in the center. All the signal pins are on the perimeter, so as you can see in this layout example, really easy to get an optimal layout in this design for good thermals as well as EMI.

And comparing to our competition here, the excellent thermal performance with the silicon that we're using translates to a smaller footprint and higher output power and higher ambient temperatures. In this example, the TPSM53604 has more thermal contact area and lower power dissipation to deliver more current at higher ambient temperatures. And the product itself, TPSM563 family, there's three current options, 2 amp, 3 amp, and 4 amp. Covers input voltages between 4 to 36 volts and output voltages between 1 and 7.

Really small package. It's only 5 by 5.5 by 4. That makes it the smallest option you have for-- between any option in the market, converter or module. And this model uses the LMR33630 if you're familiar with that silicone.

And, Alejandro, so evaluating our modules, this an example of the EVM board for the TPSM53604. All of our modules of excellent design tools to help you quickly prototype and do your evaluation. Pro points to access the critical nodes of the circuit. You can quickly set the voltage to common values with jumpers, and we have additional footprints around the module for customizing the capacitors in your design. Next slide.

And the other package I want to write here is the system and package with MicroSiP. We have two different footprints, MicroSiP, MicroSiL. MicroSiP uses a BGA. MicroSiL is LGA. Looks like a QFM footprint.

This is our highest density package option mainly for lower currents. It embeds the silicon inside a substrate, and then we mount the small passives on top of the substrate. This package would have a 36 volt module family, the LMZM2360100 up to 1 amp and [INAUDIBLE] volt. In that case, it only requires an input and output cap, so you can see in this photo it's an incredibly small footprint and easy to use.

For 4 voltages mainly for point and load, say, from 12 volt input supply to power processor [INAUDIBLE] rails, the TPS82130, 40, 50 are really popular on the market. [INAUDIBLE] compatible 3, 2, 1 amp family, high efficiency, full load and light load, and [INAUDIBLE] for doing sequencing. [INAUDIBLE] things soon here, this section might already be released. The TPSM8282122 is our newest 5 volt input family for up to 2 amp, and believe it or not, this is actually smaller than in the previous generation in a 2 by 2.5 millimeter footprint.

Then a little bit higher current [INAUDIBLE] the 4 amps [INAUDIBLE] three, TPSM828130. This also has a couple of additional features like frequency sync and spread spectrum option. And this package is a little bit larger, but it's still insanely small for the rated current. [INAUDIBLE] package, this product is going to be [INAUDIBLE] end of-- by the end of December. The TSPM265R1, this is our first 65 volt input embedded module with up to 100 milliamps of capability, and this has very low quiescent current at no load conditions, which translates to high efficiency in the milliamp range for very light load applications.

And to wrap it up, that was just a few of our hero products and our newest products. This is a overview of our portfolio segment by voltage, low Vin, mid Vin, wide voltage. And as you can see here, we have a lot of parts in our portfolio, very broad option, [INAUDIBLE] multiple families with [INAUDIBLE] current options, different package types for different requirements from the embedded. We have a leaded package, QFM and new enhanced HotRod QFM as well as the open frame packaging. [INAUDIBLE] for low voltage, low current all the way up to 55, 60 volts, and we have some products for powering really high current consumption FPGAs that can deliver 160 amps like the TPSM831B31.

And just to finish things up here, we have a lot of desirables and supporting documentation available up on TI.com. Not going to go through these, but we have a lot of app notes on EMI and noise mitigation as well as videos and articles. In terms of inverting output, that's a really common application for a lot of-- see this pretty often. Customers like using modules because inverting buck boost is a little bit tricky, and it's relatively easy to set up a module for inverting output.

And then lastly, we have thermal app notes and soldering app notes and how to use modules. And to finish things up here, what I'd like you to walk away from this presentation is that you do not need to be a power expert to design a really good power supply. Modules almost always save board space, even for the most experienced power designer. Modules enable you to have a good board layout and EMI as well as modules optimized for typical-- they're easily optimized for typical voltages that are common in many applications, and a converter is more flexible. Since the inductor is not fixed, you can usually do more optimization with a discrete circuit.

So I want to thank you for listening in, and we have a few links here going to the TI.com module portfolio landing page as well as some of the highlighted products. And thanks a lot for listening in, and if you have any further questions, please just send us an email. We're happy to answer those questions, and have a great, great day.