ETech Preview: Why LCD is the Cool New Technology All Over Again

You may also download this file. Running time: 00:43:53

Subscribe to this podcast series via iTunes. Or, visit the O'Reilly Media area at iTunes to find other podcasts from O'Reilly.

In an early test of the OLPC XO in Nigeria, the student users dropped every laptop several times a day. Despite the laptops’ rugged construction, they occasionally needed fixing, and a group of six-year-old girls opened up a “hospital” to reseat cables and do other simpler repairs. Mary Lou Jepson, One Laptop Per Child project’s CTO, had this response: “I put extra screws underneath the battery cover so that if they lost one, they could have an extra one. And kids trade them almost like marbles, when they want to try to get something fixed in their laptop.”

Mary Lou led the development of the OLPC’s breakthrough low-power transflective display, combining a traditional backlit color display with a black and white display that could be used outdoors. She left OLPC to form Pixel Qi, and bring the revolutionary engineering used in the XO to the broader consumer market. In this interview, she discusses lessons learned from OLPC and shares her vision of “cool screens that can ship in high volume, really quickly, at price points that are equivalent to what you pay for standard liquid crystal displays.”

At ETech, Mary Lou’s keynote presentation delves further into Low-Cost, Low-Power Computing.

JAMES TURNER: I’m speaking today with Mary Lou Jepsen, Founder and CEO of Pixel Qi. Dr. Jepsen previously served as chief technology officer for the One Laptop per Child program where she was an instrumental player in the development of the OLPC’s revolutionary hybrid screen. She also previously served as CTO of Intel’s display division. Dr. Jepsen was also named by Time Magazine recently as one of the 100 most influential people in the world for 2008. She’ll be speaking at the O’Reilly Emerging Technologies Conference in March, and we’re pleased she’s taken the time to talk to us. Good evening.

MARY LOU JEPSEN: Hi. Nice to speak with you tonight.

JT: So in some ways, you’re kind of uniquely qualified to comment on the current travails of the OLPC since you’ve been in highly influential positions both in the OLPC effort itself and at Intel, who some believe tried to sabotage the OLPC. Do you think that the OLPC would’ve had wider acceptance if the Intel Classmate wasn’t competing against it?

MLJ: It is interesting. I think the OLPC, and I haven’t seen the latest numbers, sold a lot more than the Classmate. I think head-to-head there’s no comparison which is the better machine, and I’m not saying that just because I’m the architect. But what’s really happened has been extraordinary. I think OLPC’s impact in sort of spearheading the movement to Netbooks is fairly undisputed, although OLPC is not the best selling Netbook; 17 million Netbooks shipped in 2008 and that’s through companies like Acer, Asus, MSI, HP, Dell. And that impact on the world is starting to be felt.

JT: What were the factors that led you to leave the OLPC program and start Pixel Qi?

MLJ: You know, I started OLPC with Nicholas in his office in the beginning, in January of 2005. And at that point, right after that Bill Gates, Steve Jobs, Michael Dell, all said it was impossible. So it became my job to sort of take that, create an architecture, invent a few things, convince the manufacturers to work with me to develop it, get a team together, and take it into high-volume mass production. And then it got to the point where my days were spent getting safety certifications for various countries.

And I just realized, it’s time for me to continue doing this; this is the best job I’ve ever done, but to keep going, why not make these components that are inside of the XO and let everybody buy them rather than just exclusively making and designing them for the OLPC laptop. If you make more of something, you can sell it for less. So rather than just serving the bottom of the pyramid, why not take the fantastic technology that we developed at OLPC and serve the whole pyramid? Everybody wants their batteries to last a lot longer. Everybody wants screens that are e-paper-like and high resolution and sunlight readable. So why not make these for the whole world?

JT: The new Netbooks that are now emerging don’t meet some of the OLPC design goals such as outdoor readable screens or low power usage, but they are being produced in huge quantities. Might it make sense to let the market drive the price of these down, and then use a Netbook as a starting point for a system for the developing world?

MLJ: Sure. But the problem in the developing world, if you’ve been to some of the development sites of say OLPC in the Middle East or in Africa or even any computer lab in the developing world, you fast realize that it’s not even the price per machine. And admittedly, while my initial goal at OLPC was to make a hundred dollar laptop, it fast became a goal to make a 1-watt laptop. It’s a lot harder and it’s a lot more important because the infrastructure of electricity and telecoms is thin, if existent at all. And so a Netbook at 10 watts–how do you power it? You’re okay for the reverse battery charge, but then how do you power not just one but a million of them? And that’s the problem we tried to solve at OLPC by developing a 1-watt laptop. Because then you can crank to recharge it, use a small solar panel. You can use all sorts of different energy solutions. And the XO actually takes anything from 10 to 20 volts as an input. It deals with spikes of weird power and the electronics don’t fry if there’s spikes and so forth. But that needs to be solved for the developing world solution. Netbooks are part of the solution. And certainly, the laptops are a lot cheaper now than before we started at OLPC.

The impact has been extraordinary, but it’s only the start. And the Netbooks that are being sold right now are essentially laptops miniaturized, regular laptops miniaturized. In the XO from OLPC, everything’s new. New screen. New architecture. New power management. New touchpad. New keyboard. You know, it’s like you can drop it from five feet. It’s rugged. It’s rain-resistant. New mesh networking. New software. It just goes on and on and on what is new in the XO. Everything’s new in it. But in the Netbooks, it’s a new size, but the motherboard and the basic design is pretty traditional. And so it occurred to me at OLPC with the demand, a lot of different Netbook companies tried very hard to buy, for example, the screen that I designed to use in their Netbook products. And we were just in the situation at OLPC of saying, “Well, our mission isn’t to sell components to the IT industry; our mission is the name of the organization: One Laptop per Child. And so we should concentrate on that and why don’t you just buy a screen from some other company?” But the truth is I think that we can make better Netbooks by making better components for them and helping to teach the Netbook companies some of the learnings we had at OLPC. And they are very anxious to get these improvements.

A lot of people want screens that have an e-paper look in reflection, that are sunlight readable, that dramatically improve their battery life. And everybody knows the screens are the power hogs in the laptops. And different companies say they have 8 to 10-hour battery life, but that’s with the screen at a brightness level that you can barely see. And so it’s not really 8 to 10 hours, but what we’re providing are screens that let the laptops have that 8 to 10-hour battery life.

JT: I actually had a question about the form factor of the OLPC because I had one, I got one from the Give One, Get One program. I had my wife try to use it. She was going back to school for a master’s and she tried to use it to take notes. And she’s a very small person. And even she had trouble with the keyboard. We eventually got her an Aspire One Linux version. And the question is it seems like there were a few minor tweaks like a 90 percent keyboard that would’ve made it something that could’ve still been used by kids, but would’ve given it a more general ability to be used?

MLJ: Right. In fact, the keyboard size and the dimensions of the XO are not that different than what the Netbook dimensions have come to be, a 9-inch screen or a 10-inch screen is actually our screen, just a little bit wider because its aspect ratio. And the keyboard really determines the screen size. But what we ended up doing is given that this was for small children and it had to work in multiple languages, is using basically more keys than are on some of the Netbooks, so that we could silk screen on different languages. For example, OLPC made the first Amharic keyboard. That’s the language they speak in Ethiopia. There’s about 100 million people in Ethiopia. They’ve never had a keyboard for their language before. So we had to sort of trail blaze that and several other languages as well. And just in order to do that, you need the extra keys.

And yes, it does admittedly make it uncomfortable for people that are used to a conventional QWERTY keyboard, for example, and used to typing on that. So there’s a switchover that’s required. Kids aren’t used to that. I mean they’re learning language for the first time. But admittedly, OLPC, in their next generation laptop, is omitting the keyboard, is using a multi-touch screen–two screens together, both multi-touch but you can use it like a book with two leaves of a book open, or put one of the leaves on the desk or in your lap and use that to type. And the advantage there is that you can make any kind of keyboard. Little kids need different big keys and probably less of them. They’re not typing a manuscript yet. They’re learning different things. And to learn chemistry maybe you need a different keyboard than to learn whatever–pick a language, Chinese, Arabic than you need to learn math or optics or what have you.

JT: That’s almost like that Russian keyboard design where every key had an LED or an LCD key cap on it so you could remap the whole keyboard.

MLJ: Yes. And so more generic than that, it’s a screen so you can remap it to absolutely anything that you want. And that’s what OLPC’s working on now. We’re helping them. We’re designing the screens for that. But right now, the advantage of this keyboard is, kids do spill stuff and drop stuff and you’ve got to make something–the XO is–it’s a tank. It’s hard to destroy it. It’s not a little delicate machine.

It’s got to last in the field in tough conditions with kids that will drop it sometimes. And I think in Nigeria we had a test pilot back in OLPC days, where the cables on the keyboard were a millimeter too short and there were these slanted desks with a concrete floor. And the desks were bolted to benches in front of them. And the laptops, every time the kid would jostle on the bench, the benches were supposed to hold two kids. So, of course, five kids were on the bench because more kids showed up to school when the laptop showed up. And every once in a while when the laptops would fall on the floor, a cable would come loose.

Nothing would break. Once after six months, I got a bug report that a screen broke. Once. Can you imagine? With laptops falling every 15 minutes for months during the school day. But these little girls opened up an XO hospital. The name of the laptop’s the XO. And they opened up a hospital and started fixing the laptops as if they were patients. And the boys really missed out. But you had six-year-old girls repairing laptops and reseating cables. And, you know, it’s really fixable. And it’s very, very different than a normal laptop that’s not designed to be fixed.

I saw that and I thought, you know, when I was a little girl fixing stuff, sometimes I’d lose screws on the floor and these girls might — or boys. And I put extra screws underneath the battery cover so that if they lost one, they could have an extra one. And kids trade them almost like marbles, when they want to try to get something fixed in their laptop.

JT: So these are the people I’m going to get the next time I call Gateway technical support, right?

MLJ: Maybe. I mean part of learning is also learning the hardware. It’s okay to open it up. You actually–you know, when you first get a car, maybe you change the oil yourself or do something else, but it’s a way of becoming sort of one with your machine, you know? It’s good.

JT: So you obviously are on top of current display technologies, to say the least. What is the current state of displays? Where is the market driving them? And what are the game-changing technologies we might see soon?

MLJ: It’s interesting. What’s happening in display–I mean post-economic meltdown is what’s happening everywhere. The factories are, relatively speaking, empty. Perhaps half-full. Perhaps half of them are shutdown. And there are some new really interesting display technologies. And you’re talking to somebody who spent 20 years developing new display technologies. And us display people think we have to invent our own new molecule, and invent the manufacturing process, and driving scheme and so forth. And I think that after 20 years of doing this, when I started at One Laptop per Child, I just said, “You know, the kids we’re trying to reach will be adults if we do sort of business as usual in the display industry where the original fundamental patents, in I think every case, have expired before the new display technology has shipped over the last 100 years, and patents last for 20 years.”

So I’m trying something different now, which is let’s do like what the silicon people do. This is what I learned at Intel [laughter]. Wow, the silicon people, you have these people designing to the manufacturing processes, and what we can ship in high-volume mass production, and they’re able to design pretty interesting silicon chips.

So I thought at OLPC, “Why don’t we do that with the existing manufacturing processes for liquid crystal display? Use liquid crystal display manufacturing processes, but in different new ways. But in ways that we can ship in six months to a year.” And so that’s sort of what I did. I was the first small entity that the large LCD manufacturers let into their factory. They said no many times first, but finally, I got one that said yes. And you know what? We went from spec to mass production ready in six months and shipped a million of them. And everybody, all the Netbook companies, wanted these screens that I made.

And so that was enough of a sort of hit where at Pixel Qi, my new company, we’ve been able to convince the largest manufacturers in the world to work with us to develop our designs into really cool screens, that can ship in high volume, really quickly, at price points that are equivalent to what you pay for standard liquid crystal displays. They use the same materials, the same processes. We just sort of look at how we put it together quite differently, just like a silicon designer uses the same silicon fabs and processes to create radically different functions in silicon chips.

JT: I thought my monitor had died tonight and I was actually pricing out 24, 25-inch monitors. I was shocked at how inexpensive they’ve gotten. You can get any number of the lower brands for under $400 now. Is that mainly because of excess capacity or has the production cost gone down?

MLJ: Well, the production cost has gone down. I mean the price of screens has halved in the last nine months or so. Yes, it’s been an extraordinarily difficult time. The manufacturers would love to sell them for more. One thing is amortization schedules. There was a lot of build out of LCD factories, five, six, seven years ago. And those are expensive propositions. They’re a couple billion dollars at the low end and at the sort of big — the way you sort of describe the sizes of these, they’re in the sizes of the mother glass, the glass from which screens are cut. And at the biggest end, the size of the mother glass is the size of a double bed or a queen-sized bed. The sort of smaller ones are 1.5 square meters. But they are fully paid off.

And the cost per screen goes down as a result of that. Plus, the LCD manufacturing industry, as DRAM and a lot of the hardware, are extraordinary at cost down. And that’s what they call it, cost down. And that’s the constant march. The constant focus is to deliver the same product for lower price. And so on average, the LCD industry has hit, I think, a 20 to 30 percent cost down year over year over year for the last decade. If you think about what, yeah, your 19 to 20-inch monitor cost what people started buying them in 1998-99, they were pretty pricey. And now you can buy–I think a standard 20-inch monitor you can get for $100-$150 right now, which is a nice thing to be able to leverage if you’re trying to design screens for high-volume mass production.

The price really does matter. And if you’re developing new manufacturing processes and plants, that development cost is extremely expensive, extremely. And in fact, to the point where in the last 50 years, there’s only been one new display technology that’s come in to really, really high-volume mass production and that’s been liquid crystal display. And as much as we love–we do really, I think, fall in love with these new prototypes, these new display prototypes, OLEDs, electrowetting, electrophoretics. I myself can tell you I developed a holographic video system as a master’s student at MIT in the late 80s, and it’s still running. And it’s still pretty far away from high-volume mass production, but it looks great.

And what really ends up delaying it is the cost of making these prototypes, that look so great, but commercializing them–it’s not like software where once you make it work, you can make millions of the things. And it’s not even like silicon, or a motherboard, or something, where you can tweak a couple of things. It’s really, historically at least, only worked a couple of times. And there’s a couple of notable exceptions. Texas Instruments did something called–well, they had many names, but it ended up calling it DLP, three initials. But it was a micro-mirror MEMS device that ships in about half of the boardroom projectors that are used in the world and that has a volume of about 3 million units a year. It’s an amazing technological achievement. It’s a marvel that they were able to do it, but the volume doesn’t come anywhere near what LCD does.

Another notable exception is plasma television. But the EU is trying to outlaw plasma because the power consumption is so high. It’s like, you know how a refrigerator is a really high power consumer in a typical home? Well, a plasma TV equals two refrigerators. And so if you just look at the environmental impact, in terms of the power it consumes, it’s a negative.

JT: One technology that has come along recently, and certainly Amazon is pushing it, is the new kind of electric ink displays that you see, for instance, in the Kindle. Is that going to mainstream more do you think?

MLJ: Oh, the electrophoretics–the founder of that company is a good friend of mine. We were freshman in college together. Joe Jacobson founded E Ink and it’s a phenomenal technology, but they’re working on color. They’re working on video. When you hit the page turn button, it takes–you have to stop reading until–I have a Kindle and I have to hit the page turn button when I’m three-quarters of the way down the page, and wait for it to refresh while I read the rest of the page. And then if I get stuck on something, then I have to go back. And we really need something with video, or a way to do fast page change and color. And I believe it’s actually a lot easier to do that with standard LCD. The screen that we have in prototype in March has the reflectivity about of the electrophoretic technologies, E Ink is one of them, There’s about 20 different companies working on that.

But in addition, it’s got color and video and a price point of LCD. And so I don’t know how–and here’s the thing; if the electrophoretics could get into high-volume mass production, at Pixel Qi, we’d be right there using the manufacturing processes. But right now, at this moment, the factories of the world are empty. And the deals that can be struck are extraordinary on price. And so if you can find a way to use the fabs that exist to make something, the price that you can hit is really hard to beat compared to developing a new manufacturing plant or technology.

And one of the advantages of electrophoretic is supposed to be the power consumption, in that you don’t need to refresh the screen every 30th or 60th of a second, with all of the pixels, and doing that takes power, because it holds its charge. But as a result, you have to unwrite the charge before you can write something in. And the voltage which you have rewrite it at is hard. And so if you actually look in at the details, the advantage, kind of when looked at from a systems perspective, according to everything that we know, it kind of disappears.

And so what we’re trying to provide are screens that look like the electrophoretics, when you have the backlight turned off. But when you have the backlight turned on, you get color and so forth. But all the time, you get video. And so, one of the interesting things, people want a backlight. So even if the electrophoretic technology beats us, you know, so be it. They still want a backlight, and they call it kind of the marriage saving device, so that you don’t have to turn on the light if you can’t sleep and you want to read in bed, you really don’t want to make a lot of light in the room. And so having the small backlight behind the thing really focuses the light towards just the person reading it, is a big advantage.

JT: So developing nations aren’t the only place that energy consumption is starting to become a factor as you mentioned. LCD and certainly plasma monitors are becoming a whipping boy for the idle power reduction fanatics. How much can be done in this area without sacrificing picture quality or startup time too greatly?

MLJ: I think enormous things can be done. For example, right now in front of me, I have my laptop on. Not a single pixel is changing on the screen and I run Linux, pretty nice. I run Ubuntu. And still the question, is what’s the motherboard doing on? What’s the CPU doing on? What is all of that doing on right now when nothing is changing? And I still want to see the screen maybe, but why is the rest of the motherboard on? And, of course, what we did at One Laptop per Child was say–I got to design the laptop. I had never designed a laptop before. I designed a lot of screens. And I said, “You know, let’s design it from the screen backwards. If the screen isn’t getting any new information and the motherboard’s not doing anything, the big secret to lowering power consumption is turn stuff off that you’re not using.”

So if you’re not using the motherboard or the CPU or the rest of the chips on the thing, turn them off to save power. So that was the big secret in lowering the power consumption of the XO. But I mean to do that, we had to turn everything on a dime, the way motherboards are made, the way the low level software works. Everything was turning on and off all the time. So even if you’re idle for two to five seconds, the motherboard turns off and then turns back on. And that is a lot bigger, than say, the efforts to make more efficient AC adapters or wall warts, which is laudable. But if you look, they’re already there at 85 percent efficiency, and so there’s some movements to make them 90 percent efficiency. That’s all great and important to do, but by turning off the motherboard and the CPU when you’re not using it, you can make the battery life go to 5X what is normal, and so the sort of bang for the buck in doing that is tremendous.

And what we’re trying to do at Pixel Qi, and I think other efforts are also taking shape. We’ve shipped a million units doing that with One Laptop per Child, my former project. And why not get other people to do this? Some of the green efforts–I work with Greener Electronics Council. They put together something called EPEAT, which is an award system for how green a laptop is. It sort of takes a spec from IEEE. It’s called IEEE–I’m going to forget the number–2006 and there’s four more digits. Anyway, it’s like this sushi card, where a laptop is green for the more checkmarks you have on this sort of sushi card of 20 to 30 different items.

The strange thing is nowhere in that IEEE spec, which was a thing that the industry got together to create, does it note the lifetime of the laptop. Does it last one year or five years? Or nor does it note the size and weight. And while Energy Star compliance is mandated by that spec, you don’t get extra credit for doing, say, ten times better than Energy Star.

And so if you really want–and I’ve mentioned this at several different meetings. I go to periodically some of the green electronics meetings to sort of speak and talk about different things. And a lot of the legislators say, “Well, what do we do? We can’t legislate that everybody has to do ten times better in a year.” And so the question is, how do you do something like Silicon Valley does, right? They fund ten startups; they know five of them won’t make it, but one or two will be off the charts, 10, 20, 100X success rates. And how do you encourage that kind of improvement, year-over-year, with some people washing out sometimes, and rather than this sort of incremental improvement? It’s a question really for the industry.

JT: Just out of curiosity, when you get, not so much the computational display market, but the consumer market, why are my LCD TVs drawing as much as they do? It strikes me there’s not a lot that’s going on when the thing’s off.

MLJ: Right. On the LCD TV, there’s really a motherboard. Your LCD TV is essentially a computer. That’s why there’s so many different things you can plug into it. Do you even know what half of them are?

JT: I was actually really amused that I got a big like a 50-inch last year and the last page of the manual has a GPL notice because it’s evidentially running a copy of Linux inside of it.

MLJ: Oh, cool. That’s pretty cool. But I mean there’s so many different connectors on it. And there’s a big motherboard that does something called scaling. So if your DVDs are a different resolution than HD, than standard TV, than the signal out from your computer for example. And so it has to translate all of those different video formats and display them on the screen. And so that takes some power.

But the other sort of big power draw is, you know, they say it’s the screen and updating the screen across that area has some power draw. But really, it’s the backlight, which is a big light and only about–well, typically five percent of the light from the backlight gets through to outside of the LCD, to your living room. Five percent. So let’s start there. We can improve that. Can we get to 10 percent? That would halve the power consumption, in very rough numbers. Shouldn’t be so hard to get to 10 percent. But it has improved a lot. It used to be one to two percent that got through. And so there has been improvement. But we’re really working on that here at Pixel Qi.

And if you look at where the losses are, there’s well, the color filters absorb, you know, they let red or green or blue light through. And so they absorb basically two-thirds or three-quarters of the light, the color filters, the way they’re set up. And so is there instead a way to spread out the color and give a different color to a different area of the pixel, so you can have the color pixeled out. But we’re working on that.

Another thing is something called polarizers. I don’t know if you remember back to physics days, they said light is a particle or light is a wave. Well, if you take the wave theory, half of the light is sort of vibrating up and down like an ocean wave. And the other half of the light, you can divide it into vibrating side-to-side, moving like a snake in the grass let’s say, side-to-side. And in an LCD, only half of the light is absorbed. Let’s say the snake light gets absorbed, and only the light that’s vibrating sort of like an ocean wave gets through. Well, you’ve thrown away another half of your light because, the way that we use liquid crystals now, the way that pretty much all liquid crystal displays throw away another 40 to 55 percent of light, because it is not totally perfectly efficient, because they have to modulate the polarization state of the light. And so we’re working on ways to stop doing that.

Another thing is it’s pretty hard to see your cell phone or camera or laptop when it’s outside because the light behind the screen is in competition with the light from the outdoor light or even a bright indoor light. Why don’t we use the light from the outside or the bright room light to enhance the brightness of the display? We’re doing that, too. It’s like there’s a lot of things you can do and you can use the standard manufacturing processes if you really understand what the layers are, you use these bottles of liquid crystals. Well, there’s a certain number of bottles of liquid crystals that are approved to use in the factory, that they’ve tested and tested for years, and they know they work. And there’s different layers of aluminum and glass and amorphous silicon. And there’s different rules about how you can pattern it, the size of the lines and the spaces between the lines, and what kind of blobs that you can put down.

But those are the basic rules and then after that it’s well, what can you think of? How can you instead rearrange this with existing–you know, there’s these different optical films that you can use that scatter the light in different ways, or bend the light in different ways, and make color in different ways. And with that, what can you invent that addresses the five percent throughput? We should be able to do better than that, right? And then actually, if you look at the LED efficacy, you know, the power in to the power out, that is not a 100 percent efficient system. And if you look at that, you lower the efficiency from power in to light into your light into your living room, dramatically even lower than five percent; it’s closer to two. Which we should be able to do better, shouldn’t we?

JT: What role do you see Pixel Qi having in the next few years of portable computing?

MLJ: Portable computing? I think it’s often said Apple and the iPhone and all of that’s really visionary. But I mean people do want kind of a big iPhone, a tablet. And when you look in an iPhone, all you see is the screen. There could be little green men inside of the thing, or the thing could be just steel with a screen. You don’t know. And with the advent of all of the services now that are doing cloud computing, and widely available different wireless telecommunications systems, what you still need is the screen, and a battery that doesn’t weigh a lot and that lasts a long time.

Where the actual electronics are, and how that works, whether the image is rendered in the device itself or exterior to it, it doesn’t matter. Touch is obviously coming, and there’s many, many different efforts in touch. An analyst I spoke with about six months ago was tracking 200 different nascent, and much more mature, multi-touch efforts. And we are working on a couple and have favorites of ourselves at Pixel Qi, but I mean I think that more and more it’s really about a screen that you can carry somehow, that’s not too heavy, that you can see that your batteries last a long time, and that’s more readable. What we do actually, eight hours or more a day as professionals, is we use our laptops. And we’re essentially staring into a flashlight eight hours a day, on a low resolution screen.

And the reason we prefer to read off of paper is A) it’s higher resolution by a lot, and B) we don’t have to stare into a flashlight to read the paper, so it’s easier to read. One of the things that we’re saying at Pixel Qi is that some people actually want to read off of their laptop screens. Right? It’s not–sort of 20 centimeters away from your eye, you want to read. You can see high resolution data at that text, at that distance. And we’re making a screen that is optimized for reading, and for multimedia. We can’t give away and we can’t go without supersaturated gorgeous color, but it’s hard to read a long legal document on the screen of your laptop and not miss some key words.

And we’re trying to make a screen–and the OLPC XO screen was the first step, but the next step is a nice improvement over the performance of the XO screen, both in color and in the sort of paper white, white state that we can get. And that will be on the market this year. And then for 2010, we’ve got a bunch of new things. But we think that the future of computing is really kind of all about the screen. The chip wars are over. The screen wars, I suppose, are starting.

JT: When you talk about not really caring what’s inside and cloud computing, that almost sounds like you’re heartening back to the thin client day where you would use WiMax or 4G and essentially, you’d be just moving all of your computing off somewhere else and just trying to do the display in I/O.

MLJ: Maybe. I mean thin computing never really got a bad name. And maybe cloud computing is the new name for thin client. I think you might be right. But you can do so much more. I mean isn’t the iPhone kind of like a thin client right now? I mean I don’t know. But even some of the Netbooks are kind of thin client, but with a 2 gigahertz processor, it’s not that thin. Ten years ago, the Sony Vaio I had ten years ago had a 500 megahertz CPU in it and it was the chichi expensive laptop of that whatever, pick your time period, of that six months or year or whatever. I’m sure there were other good laptops out, too, of course.

But yeah, maybe thin client. Not everybody is–well, maybe you and I are crunching numbers all of the time on our laptop. Half the time, a laptop is idle on average. Idle. And people need things to do, email and web browsing. Read books. People would like to read books and newspapers and magazines, blogs, and write, and watch videos. Yeah. Sure. Watch videos and so forth. But even for books, if you look at the education market, you need to have color for illustrations. It’s hard to learn geography without color for the maps, or chemistry without a way to sort of show molecules or whatever for diagrams. You do have to have the color in there in the books, especially for learning.

There’s a crossover coming right now between E-Readers and Netbooks. And I think that we’ll see a whole new class of–well, I guess the Netbook’s a new class. I guess maybe a maturing of that space, and form factors that allow more comfortable reading than your typical clamshell laptop, and a lot of more use, I think, as clearly pioneered by Apple and the iPhone, of multi-touch in that space.

JT: So can you give a brief overview of what you’re going to be talking about at ETech?

MLJ: Um, I would love to give a brief overview of what I’m going to be talking about at ETech, but honestly, I’m talking to the conference organizers this week to fine tune what the message is. And so it’s a little premature. But perhaps we’ll touch on these subjects and others of interest to the audience. I mean maybe the question could be back on you. What is the audience most interested in at ETech? This is my first time at ETech so I want to, I suppose, be provocative and get some conversations going, maybe talking about hardware, but maybe sort of generalizing that out to what’s happening more generally in IT, and the way hardware touches that.

JT: I’ll keep an eye on the comments for the podcast here and maybe somebody will suggest something.

MLJ: Okay. Great.

JT: I’ve been talking to Mary Lou Jepsen, who is the Founder and CEO of Pixel Qi. She will be talking at the Emerging Technology Conference which will be occurring in early March. Thank you so much for talking to us.

MLJ: Thank you.

tags: , , , , ,
  • http://highlycomposite2.blogspot.com Jaya Kumar

    Hi,

    I think there are a few statements in this article that require further explanation. In particular “And the voltage which you have rewrite it at is hard. And so if you actually look in at the details, the advantage, kind of when looked at from a systems perspective, according to everything that we know, it kind of disappears.” My understanding of current data completely contradicts this statement so I’d like to understand if I’m misunderstood something.

    The following “OLPC’s breakthrough low-power transflective display” may not be accurate given that independent analysis has not been able to show any special capabilities or properties. See this EETimes article.

    Thanks,
    Jaya Kumar

  • http://olpc.tv Charbax

    Thanks a lot for this interview! I’m really looking forward to the next generation screen by Mary Lou Jepsen, the one she talks about have a whiter reflective mode and better colors and that’s to be in commercial products..

    Sure the chip wars might be over soon, but I’d say not yet. Intel dominates that netbook market, AMD don’t seem interested in it (no successor to the Geode), so I think that we really need to be seing all of the ARM chip manufacturers come and introduce their ARM Cortex based alternatives runing all sortss of really thin Linux distributions such as Google Android and Ubuntu and have thosebe extremely optimized for cloud computing using Google Gears, Chrome V8, instant on and so that they perhaps dynamically use light to lower backlight intensity to always provid the same readabillity at the minimum power consumption.

    I’d really like to see the first Android ARM based laptops with Pixel Qi power consumption nearing the $100 already this year.

  • http://www.pixelqi.com Mary Lou Jepsen

    Jaya,

    James Turner (the author of the article) sent me your post and asked me to reply.

    Electro-phoretics don’t need periodic AC refresh of usually 30 to 60 times a second as most liquid crystal displays do – and thus the perceived power advantage that is widely cited. But, the voltages required for electro-phoretics are much higher and require more power to get there (power goes with square of the voltage – for example – P = C*V^2*f), and you need to unwrite the image before you can write a new one. It may sound like it’s not close – but let’s say the voltages are not 5V as in LCD but 10V (this is conservative – I believe the voltages that electrophoretics need in practice for real shipping product are higher). Let’s say that the voltage is 2X higher and thus the power consumption to switch the screen is 4x higher, plus you have to switch it off before you can re-write it – meaning you need 8X the power to switch a single frame. This gets us closer but at 30 frames per second at the standard LCD voltage – it would be about 25% of the power consumption if you had to update the LCD at 30 frames per second. But now – look at being able to update LCD more slowly – Indeed companies are shipping product right now with LCDs that don’t require any update – Even just slowing the update down makes the power consumption difference quite slight. And then – the advantage can gets lost in the overall systems design – in how well the display can work with the motherboard to reduce power consumption – and – cost while providing a delightful reading experience.

    - Mary Lou

  • http://highlycomposite2.blogspot.com Jaya Kumar

    MLJ, here’s my response:

    “Electro-phoretics don’t need periodic AC refresh of usually 30 to 60 times a second as most liquid crystal displays do”

    Yes. Agreed. But lets assert that electrophoretic displays (such as the one used in the Kindle-1, Kindle-2, Sony PRS-505, PRS-700 (basically all the ones that are on the market today)) do NOT require refreshes. Thus, there is no static power consumption by an electrophoretic display. In contrast, the standard twisted nematic LCD always requires static power consumption.

    “and thus the perceived power advantage that is widely cited.”
    Okay, I noticed you’ve said “perceived power advantage”. I believe that it is not just a perceived advantage. There is a measured and real power advantage with electrophoretic displays over LCD displays. I will continue and hope to convince you that the data backs this assessment.

    “But, the voltages required for electro-phoretics are much higher and require more power to get there (power goes with square of the voltage – for example – P = C*V^2*f), and you need to unwrite the image before you can write a new one.”

    There’s complexity here. I agree that an electrophoretic display would consume more energy than an LCD during an image update. The point to raise there is that the energy is only needed when an update is required. As opposed to LCD where energy is needed continuously. Further, electrophoretic power draw is image history dependent so its not entirely true that you have to flash an electrophoretic pixel to a start state before moving to final state.

    “voltages required for electrophoretic are much higher”. “this is conservative – I believe the voltages that electrophoretics need in practice for real shipping product are higher)”

    I have observed data so that we can talk in specifics. There’s the direct drive E-Ink display used in the Esquire advertisment and that used +/-15V DC. I have an AM300 E-Ink Vizplex + Broadsheet development kit (same controller and panel used in Kindle2 and Sony PRS-700), the panel voltage is still +/-15V DC.

    “power consumption to switch the screen is 4x higher, plus you have to switch it off before you can re-write it – meaning you need 8X the power to switch a single frame.”

    I would prefer to compare specifics but realized that I lack the data. I will dispute the statement that you have to switch it off before you can rewrite it. Each electrophoretic pixel has a current state, eg: white , gray or black. The transition does not always have to go to an opposite state. It can be gray-gray or gray-black or gray-white as well. There are multiple modes of transitions and so it is a difficult measurement to make objectively. I believe the worst case power consumption is around 50mWatts per update for a 6″ diagonal panel (about 20 sq inches). That’s 2.5 mW per sq inch. I think so. I’m not certain here. I should generate this data.

    “This gets us closer but at 30 frames per second at the standard LCD voltage – it would be about 25% of the power consumption if you had to update the LCD at 30 frames per second.”

    I got lost beyond this point. Are you comparing an electrophoretic display being updated at a steady rate of 30fps with an LCD? If so, why? That’s not an intended mode of use. I would never expect to use an electrophoretic display for content that is any more than 2fps. I think I’ll stop here. Thanks for your discussion.

    Jaya Kumar

  • H.K.

    Very interesting discussion (albeit lengthy). I am admittedly a novice in this field so I learned a lot. Particularly the objectives of the work in Africa and the energy consumption vs cost importance. Reminds me of the food donations-you can give the food (notebooks) but teaching one to be self sufficient with sustainability (adapting/building the electronic infrastructure) is really the challenge. I do hope that as they develop (I’m an optimist) that their internet security follows given they will surely have wide access to the net/computers by that point. They should get a site like http://www.justaskgemalto.com for those kids given the number of americans fooled by nigerians these days.

  • ST

    Jaya,

    I was also confused and surprised to hear MLJ state that “And so if you actually look in at the details, the advantage, kind of when looked at from a systems perspective, according to everything that we know, it kind of disappears.”. This statement seems to contradict what I read about electrophoretics device such as the Amazon Kindle, which apparently can be used for days with only one charge.

    But after reading her reply I understood what she is trying to say. She is saying that IF one wants to have a display that can play video THEN electrophoretics will not have much advantage over LCD in term of power consumption. So in the end electrophoretics will be confined to devices that only display static text and images.

    Mary-Lou,

    In your reply you stated that “Even just slowing the update down makes the power consumption difference quite slight”. Are you saying that slowing down refresh rates of the LCD does NOT cut down on the power consumption significantly because most of the power consumption is due to the back lighting?

    I can’t wait to see your new display. I own an OLPC and I really like the screen. The sort of tablet you described is exactly what I want: lightweight, reasonably large, easy to read, and hopefully cheap too.

    Regards
    ST

  • http://www.marybranscombe.com Mary Branscombe

    @Jaya
    ” I will dispute the statement that you have to switch it off before you can rewrite it”
    That’s the way e-ink works. There is no knowledge in the controller of how much the voltage in each cell has decayed from the level it was last set to, so in order to write to the cell for the next frame correctly, it must be flushed first. This is why the Kindle screen goes black when you ‘turn the page’.

    If you don’t want anything beyond 2fps, then you’re not considering the video applications that MLJ is looking at – and you can’t use it for a PC screen either!

    LCD screens without refresh; Intel demonstrated these on notebook PCs at the 2008 fall IDF (and possibly the year before). Roll on graphics controllers with the control circuits to do this.