Upal Sengupta, Staff Applications Engineer
Texas Instruments, Battery Management Solutions
For most of my career developing portable equipment, I’ve been taught that “less is more. The fewer integrated circuit (IC) devices on a PC board, the better. We’ve been on a quest for smaller, lighter and cheaper since the first portable computers and mobile phones showed up 20 to 30 years ago. Much of the integration trend, of course, is due to rapid advancements in silicon process technology. I almost hate to admit that I vaguely remember when “submicron technology” was a big deal. Now I can’t honestly remember the last time I used an IC that had a feature size anywhere near 1 micron.
I’ve worked on teams that developed highly integrated PMIC devices for mobile phones over the years, and observed a fairly consistent trend starting with the first digital (2G) products in the 1990s up to the 3G products that became common around 2007 – 2008 or so. The more we could put into the system’s PMIC the better – our goal was reduced chip count which equated to lower cost. Even back in 1996 we knew that “someday” the mobile phone would just become a commodity item, so it really needed to be cheap. Of course our ultimate goal back then was to develop the “single-chip mobile phone.” I will admit that many of us thought that really meant at least two chips, one analog and one digital, but the story isn’t over yet.
As phones added more features, and correspondingly more internal power rails, our PMIC devices kept getting bigger and bigger in terms of functionality. The actual chip / package sizes did not, however, because while we added functions, the advances in IC process and packaging technology let us put those features into (roughly) the same amount of real estate. In the mid-1990s, our “state of the art” PMIC had (maybe) five DC output power rails, a couple of audio amplifiers, and a linear battery charger that drove an external power FET. I think it was around 8 mm by 8 mm with 48 pins. Ten years later, we had 100- to 200-ball BGA packages that were around 8 mm by 8 mm or smaller, with no external FETs, 15 to 20 power rails and switch-mode charging. So yes, we certainly did more with less, from one generation of PMIC to the next.
But in the last few years, a different trend continues to emerge. While we make things smaller year after year, decade after decade, we may at some point realize that not everything in the world really follows Moore’s Law. After all, it isn’t literally a “law” but more of an observation. It also turns out that at some point, smaller and smaller silicon geometries aren’t always the way to go for power ICs.
So how does power consumption break down in a mobile product? Roughly speaking, the big loads that consume power are the processor(s), the radio, the display, and the battery (when it is being charged). As mobile systems have gotten more complex, moving from 3G to 4G, we see the trend toward higher-integration PMIC’s starting to reverse just ever so slightly. While the system PMIC remains a relatively large, complex device with numerous functions, some of these high power functions are being offloaded into external circuits. In particular, the RFPA power supply, the display backlight driver, and the battery charger IC are implemented externally. When you have a 3 Amp-Hour battery, and want to charge it in an hour, there’s just no other way to do it if you don’t want smoke coming out of your PC board.
As engineers, we sometimes forget that everything we do is driven by what our customers (whether individual consumers or corporate customers) are willing to pay for. With the overwhelming boom in mobile technology recently, we know people want high speed data, big bright screens, and the ability to use their devices anywhere and everywhere. So, to enable that, we need power. More power means bigger, higher capacity batteries. (Of course that leads to the never-ending-quest to get more energy from less space in your battery, but let’s not go there right now.)
Bigger batteries mean higher current chargers, and we’ve now gotten to the point in smartphones and tablets that it actually makes sense to disintegrate the power conversion solution into multiple components. (This would have practically been heresy back in the good old days.) To that end, some of us former PMIC-guys are working on stand-alone battery chargers again! Fortunately, with the high efficiency chargers, now we can still charge a battery at 2, 3 or even 4 Amp currents and maintain a 1 to 2 mm component height for the power components on the PC board.
So, particularly in the field of analog and power conversion circuits, we can’t always hold to the mindset that less is more. Sometimes, if you want more (of something), you just need to use more (of something else) to get it. So maybe less is not always more. And maybe Big is the new Small when it comes to power.
- http://www.ti.com/lsds/ti/power-management/battery-charger-solutions-products.page
- Silicon Technology: https://docs.google.com/viewer?url=http%3A%2F%2Fwww.eng.tau.ac.il%2F~yosish%2Fcourses%2Fvlsi1%2FII-1-Si-Tech-introduction.pdf
- Moore’s law: http://en.wikipedia.org/wiki/Moore’s_law
- Mobile phone thinness trend: http://www.gsmarena.com/mobile_phone_evolution-review-493.php
Upal Sengupta is a staff applications engineer with the TI Battery Management Solutions group. Since joining TI in 2003, Upal has worked as an applications engineer and technical marketing manager in support of TI’s portable power and battery management technology. Prior to TI, he worked as a system design engineer for OEMs developing mobile phones, portable computers, and consumer products. Upal received a BSEE from the University of Illinois, and an MSEE from Michigan State University. Upal can be reached at ti_upalsengupta@list.ti.com.
