Sony Crystal LED Display - new display tech - Page 3

IMO, "Quantum Dot" sounds better. After all many people have heard of Liquid Crystal Displays, so some may end up thinking that Crystal LED is some kind of LED LCD. I don't think the general public associates the word quantum with any particular existing consumer technology. (Quantum HDDs have been off the market for over a decade).

I'm fine with Sony calling it Crystal LED!

......Billy Crystal singing -> http://www.youtube.com/watch?v=r_pAY7X9e84

They've always had a silicon substrate.
Silicon carbide blue leds were made for a few months...about 15 years ago...they didn't work out well, short life, wrong colour.
White leds are blue leds with a yellow phosphor.

No I didn't miss your statements, just brain fart, sorry.

On your other question, I think that the issue is the same reasons that OLED has been in development for so long. For some time OLEDs have been possible to make, but, brightness, lifetime have been issues.

I think that the problem might be the same as OLED, Blue. Last I heard, blue was difficult to make on sapphire, now try to make it using CVD/PVD techniques with silicon and it gets tricky I bet. Also I think that brightness may be an issue as well.

Didn't mean to rub anyone the wrong way with this, but my point was that for every pixel on your LCD, there are three sub pixels that require three transistors on glass to control. Given that a diode is only the lower half of a standard NPN transistor, it would be simpler and cheaper to build just diodes, LEDs would be about the same or a little more complex. Gate metal on the glass surface could be used for address control lines instead of Gate control (Base). One upper layer of Moly/AL for power and there you are.

Everybody else here adds something to the conversation from their specialty so I thought I would add some from mine.

TNG, give what you suggest, I wonder what it is about OLED that has made that technology so popular as compared to non-organic LED displays? If the technical hurdles with developing a suitable LED TV are similar to those with developing an OLED TV, why has no one shown interest before? Not even in the small format market, where OLED has already proven itself viable, do we see LED screens.

There must be some reason why OLED was favored to begin with, no?

hopefully sony engineers a better product than their lcos/sxrd fiasco.

No wonder Sony gave up on OLED since they know how to do ILED. In theory this should cost less to manufacture because there's no more need for complicated, light-blocking LC layers, it's just a "backlight," although it's probably a super-expensive backlight right now. The fact they can add as many independent RGB LEDs to the panel as they want means there should be no significant size limitation and, even at huge sizes, this thing should be super energy-efficient. Sony said 70W for 55-inch display even for an early prototype.

^^ Heat. They have to address heat. Usually the carrier or packaging expire much earlier than the LED itself


So that's usual TFT right?

Yes everyone contributing is best, I'm also here to learn. What's not cool is this:

Anyway back to regular broadcast.

OK, so chew on this because, again, you know a ton more about this than I do.

Sony's own people are largely in the dark about what the heck they are demoing. But the few that knew anything know this:

It's not one giant substrate with 6 million LEDs at all. It's a bunch of little tiles wired together that are each some amount of inches across. How many? No one could say -- and it's possible no one there even knew.

In theory, this is nearly infinitely scaleable in either direction. You could make a huge one or a smaller one. You'd lose resolution going smaller, however. You could probably join some pixels if you went too big and didn't want millions upon millions of pixels (in other words, you could make every 4 adjacent pixels into one larger square).

The modules sit against the glass, but would be fabbed using some existing semiconductor-type processes combined with whatever new would be required to make bright even LED at the approximately 0.15 mm width of these (I did some crude math of 1220mm x 1920 pixels / 3 for rgb to get that number). So that translates to 150 micron LEDs? Clearly, the transistors and whatever electronics are required to join these together are trivial in terms of space vs. the LEDs.

I have no real clue how you'd make these "mini-displays" but ironically enough, the one display technology that Sony did really develop in the past decade -- LCOS -- gives them an interesting leg up here. Why? Not because it's actually very related as it isn't. But instead because it's sort of this weird hybrid semiconductor part and they at least have some knowhow in how to make that happen.

My overwhelming sense is (a) that this will cost at least $300-500 million to commercialize (b) as Natascha McElhone says to Jim Carrey in the library in The Truman Show "If we don't go now, it will never happen (c) this is Sony's absolute last best chance to get back into the TV market because it would give them something they could actually be proud of and call it the next great thing from the people who brought you Trinitron and revolutionized television (d) they are going to dither and pretend they are still interested in pursuing OLED and end up doing nothing.

Sony's people say that as long as the current execs are around, they will keep making TVs and keep "innovating". This is a real chance to do that and while it's entirely and fundamentally unproven, if they can fab the pieces, this has the chance to really do a lot of things OLED cannot. It could theoretically ramp faster because the yields would be at the building-block level not at the finished display level (those yields should approach 100% almost immediately as you could literally fix failures at the factory... fabulous) and it can scale in size. It's worth noting that larger-than-55-inch OLEDs are not only going to suffer the 8G fab problem but -- at least at Samsung -- the mask problem as well.

Soliciting Oh-so-many New Yearnings? Or Superior to OLED, Never Yours?

It's almost as important for them to commit to building these as to actually doing so. The fact that they are so tepid around it suggests that it's not so simple and / or that they are simply not capable of big steps any longer. I'd take the S-LCD check and bet it on the 3 new initiatives that will define Sony this decade. Starting with this one.

Your move, Sir Howard.

It seems to me that this technology has much better potential not only to grow in size, but also in resolution. Just add RGB LEDs. People at CES have been in awe of Sharp's 8K LCD. Imagine what would be the reaction if this was 8K ILED display. Also, sub-pixel LEDs finally open a door to a true HDR display. This tech just cannot arrive soon enough.

The thing I don't get here is that, at least people at Gizmodo, are completely blown away by Samsung's OLED while the reviews for the Crystal LED have been lukewarm. Why would that be? It's not like OLED should have any advantage over Crystal LED as far as color, response time, lowest black level, brightness gradation, contrast, viewing angles.

If price to produce these types of displays is comparable to price to produce LCDs, it will mean death for all transmissive display technologies. No more LCDs and also no point of developing MEMS displays. It's probably all about self-emissive display technologies from now on: ILED, OLED and maybe QLED.

That's what I mean. LED chips are usually made on 4"/6" wafers.

You come back to the same difference with WOLED and RGB OLED. You need to be top quality and precision to be RGB OLED or LED. Not so much if it is behind some filter. Piecing them together will not be a major issue if it is behind the "scene" Think about it vtms.


We are in agreement here:

Spec, here's the thing: You don't need exceptional precision with this Sony technology, not in the sense of Samsung's. Let me clarify.

In normal semiconductor manufacturing, you eventually improve the process to get yields into the 90s and beyond. Yes, if a TV takes say 200 of these LED "modules", there'd be losses, but they'll be "chip-size" losses. Those are tolerable. You aren't throwing away entire substrates, you're throwing away what amounts to a commodity chip. Beyond the chip stage, you are likely to have very close to 100% yield.

In LG's OLED -- which is looking more and more like unmitigated genius if they pull it off (admittedly a huge if) -- you combine a relatively simple multi-stage vacuum deposition of the OLED material layers, and very well understood TFT backplanes and color filter litho to make panels. At each of those, you are likely to have very high yields and get good results.

Samsung's OLED is going to result in entire failed panels and even entire failed substrates. Initially, tons and tons of them. (In fairness, even LG's simpler method will have yield problems at first; no one has done it before). The whole issue of whether they can make masks that hold up to volume production, whether they can make the scanning work effectively, etc. are legitimate issues.

Obviously, Sony is blustering here without even a promise to build one model, but their claim is "Look, we stopped making LCD because no one was making money doing it and there didn't look to be much more way of fixing that. This technology is making chips, something we know something about. We believe this will scale and be cheaper than LCD -- just like OLED." Now the bluster also claims it's better. I'm of the distinct opinion that it makes no difference. There is every reason to believe it will be comparable, with longer display life to boot.

The cost of the modules, the number required, and the amount of inter-module assembly are certain the devil's worth of details here, but it's plausible that this methodology could yield low-cost displays, where low-cost is a relative term.

I see your logic but correct me if I'm wrong: I don't think Sony is even a sizable LED maker. The quality of the LED has to be even higher for the LED lightbars used in backlighting. And I'm not even sure if gamut with D65 can be consistent. Longevity is also in doubt until someone tells me what is the status of the heat produced by these 6m LEDs.

Sony is going to depend on Toyoda Gosei or Epistar to meet its spec?

There is a big difference between putting say 9 32" motherglass to make 1 80". The edges will be major problems (talk about SDE ). It's not a major problem if it's hidden behind, but will be if it is front of the panel.

Technologically Sony CLED is extremely impressive and one point in this is that they apparently got uniformly emitting LEDs, normally they have variations in light output. BUT practically the chances this will be widely commercialized are zero. There is deep irony in the fact that the company which just moved out from the local dimming using a couple of hundred LEDs is showing display with 6 millions of them. If they were not able to make loc-dim commercially sound how on earth they could be able to do it with 6 mln LEDs???. They dared to put 6 millions LEDs but could not optimize how to put efficiently a thousand or so to have dense local dimming

This has broader aspect concerning LCD and OLED. LCD tech has still considerable headroom in applying dense local dimming, this is not used since there is no serious competing tech. If OLED would indeed started gaining some share based on PQ, there would be immediate response of LCD with dense, even RGB-based local dimming. Local dimming with say 1000 zones and RGB optimization would be next to impossible to beat by OLED based on PQ and it would be cheaper.

Very interesting, thanks for sharing. So it sounds like it is similar to a large format LED display (by "large", I guess I really mean "mega" - i.e. cowboys stadium, or any other stadium for that matter), where they build up the screen - almost Lego-style - from smaller individual LED arrays. The key with this TV, obviously, is they've got some very, very small LEDs. I still have to believe, though, that these smaller arrays (or tiles, as you called them) are not made up of tiny individual LEDs that must be painstakingly arranged and wired. Although I suppose that is still a possibility for a prototype such as this, it seems as though for a production application you would have to fabricate them TFT style.

It would be very interesting to know just how many individual pixels comprise each tile.

Thinking out loud:

If sony could do this (this being build LED displays up modularly, using a grid of smaller, easier to manufacture pixel arrays), then why couldn't they or anyone else for that matter do the same with LCD, or even plasma? I.e. if it is cheaper to do it that way, why hasn't it been done before?

I imagine part of the issue is getting consistent calibration (for lack of better word) across each tile, such that you cannot see the grid.

Regarding specifically your last sentence: They aren't. They are made using semiconductor type processes. This means fabricated on chip lines, like microprocessors, memory chips, or exotic parts like TI's DLP digital micromirror devices. I have no idea how many LEDs are on one "module" nor how many made up the sample display, but the analogy to the stadium-type displays is entirely valid. The difference is how the building blocks are made. That's the part that's different.


While I agree that they might well not commercialize it, I don't agree that 150 micron LEDs fabbed on a semiconductor line can't be made to output light uniformly. The current method of uniform light emission preferred for TVs involves beaming small numbers of LEDs through plastic light guides and relying on various quantum properties of light to get them to approximately spread evenly across the display. This works more than well enough. I see no reason why these LEDs -- assuming their variance in output was inside of +/-20%, would be a problem.

According to Sony, anything involving LCD is "more expensive". Using these modules would raise, not lower, the cost of LCD just as OLED starts to get less expensive.


The quality of the individual LEDs needs to be nowhere near as high as that used in lightbars. There is a "strength in numbers" phenomenon here as well as things like random distributions. Over 6 million, so long as your variance is low enough, a lot will disappear into the haze of the universe. I doubt very much that the OLED pixels will have perfectly uniform light output either -- or PDP cells for that matter -- but the variance shrinks beyond the point of anyone seeing it. And the odds of neighboring pixels varying in the same wrong direction become vanishingly small.

So yes, they will need good cool. The demo is good, but overall, I don't know how pure the red, green and blue are.

No idea.

The demo has no screendoor at all. The LED run to the edge of the module. The electronics clearly must all be behind. It appears seamless in the demo. And, yes, it will have to be in the release version. Is that easy? If the "packaging step" -- where the chips would normally be plasticized say in the case of microprocessors -- was done in a way that the inter-chip linkages could be reliably achieved by a robot / machine that slides them into one another before a final locking frame holds everything together, I don't see why not.

Perhaps inter-tile separation would be a problem over time or a chance for assembly errors. But assembly errors would not dictate throwing everything out; you could take the modules out and "go again".

Some more information here, including close-up shots of the display:

http://translate.google.com/translat...12_504057.html

I could be misinterpreting things, but it looks like the RGB "subpixels" are almost a single unit? (obviously it's three separate LEDs that make up a whole pixel though)

Link is not working, http://www.google.nl/search?q=http%3...w=1052&bih=262

seems to be this one -> http://translate.google.nl/translate...12_504057.html

That's strange, I tested it in a couple of browsers.

This is the original link: http://av.watch.impress.co.jp/docs/s...12_504057.html

I am just wondering if anyone who is at CES 2012 can tell me which panel impressed you the most.

The 55" OLED from Samsung / LG or the 55" Crystal LED from Sony?

Just looking for some subjective feed back on this!!!

Thanks

Paul

Well I will say that my former boss, now back in Japan said that this was made using LCD coaters that we sold a couple of years ago, not Gen 8 stuff, but Gen 5 or 6 (I would have to look up the model number). I remember when they were building this fab somewhere in Southern Japan, not because I had to go there, but because I had to stay in contact with one of my Japanese counterparts and the time shift was killing me.


No, 150 is a good number, good calcs on your part. I would imagine that they are looking at proof of concept, wiring together tiles of LED arrays is troublesome I would think and would also add thousands of connection points that could become an issue.


They also have the facilities to do it and now that they are probably scaling back on partnerships with Samsung, they have the personnel as well to do it.


Good thoughts. I don't want to see Sony drop by the wayside like Pioneer did, and you are right, they need something revolutionary, like the Sony of old would do.


LOL, yeah, still waiting on SED, FED..., but to be fair, OLED has a better chance than those at this point. Until I can replace my old 52" Sharp with a equivalent OLED though, I think it is all vaporware.

This one works but when i try translate i get blank page. I use IE9.

Of course the demo is seamless. We are discusing the logic of manufacturing in quantity ie whether it is plausible. OLED is plausible now, I am skeptical on CLED. SDE will be an issue when u combine panels, until the day comes when you can have technical ability to produce true borderless displays.

The biggest giveaway to me is the lack of chatter in the supply chain.

I've not seen it but I doubt it will be "packaged" in the traditional LED sense. And not ceramic encapsulated as in IC since it has to emit light.

Locking them is easy. But making sure they are uniform and driven properly is not. That's why Chip on Glass is fading.

And did I mention heat? Before anyone claim standing near it is not hot, that's not the point. It is the heat per square inch generated. That's why light bars are aluminum and not traditional PCB. And main cause of LED failure. So marketing material of 20 years LED lifespan is technically correct but not practical.