Light-Emitting Diodes (LEDs) have been around for decades, and they come in all colors. Now, you would think it obvious that video displays should be made from LEDs, to take on the various competitors. Certainly this has been done on a large scale, and I doubt I am the first to think of creating display panels full of LEDs for home viewing. However, achieving a good image will require smaller pixels (say, about the size of those phosphor dots on a good Cathod-Ray-Tube computer monitor). The "pixels" in those large-scale units are frequently off-the-shelf individually-packaged LEDs, each somewhere between a quarter to half a centimeter in diameter. No wonder the overall screen, with millions of such pixels, is only suitable for billboards and sports arenas!
So, how tough is it to make a video-display panel out of micro-sized and closely-packed LEDs? Rather tougher than I originally thought! There are two main problems. First, unlike a lot of electronic technology, LEDs are mostly not made from silicon. So, while silicon diodes are very common in integrated circuits, and large arrays of them are also made (for sensors), diodes that emit light have so far not been easy to incorporate into standard integrated-circuit manufacturing. Nevertheless, researchers have been pursuing and steadily making progress toward creating and integrating silicon-based LEDs. Even whole arrays, "suitable for large two-dimensional areas", of silicon LEDs have been made, as far back as three years ago (probably just one color, though). The apparent goal of silicon LEDs these days seems to be tied up with optical communications between circuits, and it seems they have forgotten the possibility of making high-resolution display panels using Light Emitting Diodes. HAVE THEY FORGOTTEN? So, perhaps this is a worthwhile thing to post here, to remind them....
The second main problem concerns the silicon onto which that large array of LEDs would be manufactured. If you want a High Definition TV panel that measures, for example, 18 inches diagonally (hmmmm, it occurs to me that I've never encountered monitor sizes specified in metric! -- but 18 inches is about 46cm), then you need a piece of silicon that is just about 16 inches by 9 inches (41 x 23 cm). (Pythagorean double check: the square root of [41x41 + 23x23] is 47-and-a-small-fraction.) Anyway, nobody makes such sheets of silicon yet! And, they will have to be single-crystal silicon, all of which is currently made by a rather slow and expensive and wasteful process. It could very well be that a whole plethora of alternate flatscreen video technologies will be saturating the market before LED panels ever get a chance to strut their stuff, especially in a price-competitive way. On the other hand, those alternate technologies (besides Liquid Crystal Displays and Plasma Panels) are just barely coming out of the labs, and so to whatever extent this Idea should be pursued, all is not yet lost. Small displays, using current-sized silicon, could be produced now, while ways are sought to make lots of larger-sized silicon.-- Vernon, Dec 01 2003 A Competing Technology http://www.ifire.comInorganic Electroluminescent [Vernon, Oct 04 2004] A competing technology http://www.iridigm.comReflective Interference [Vernon, Oct 04 2004] A competing technology http://www.eink.comCharged particles inside ink drops (still working on color) [Vernon, Oct 04 2004] A competing technology http://www.philips....icleId=2376&lNodeIdCharged colored inks\ [Vernon, Oct 04 2004, last modified Oct 21 2004] A competing technology http://www.gyriconmedia.comCharged rotating spheres (still working on color) [Vernon, Oct 04 2004] A competing technology http://www.universaldisplay.com/tech.phpOrganic light emitting devices (apparently NOT diodes) [Vernon, Oct 04 2004, last modified Oct 21 2004] About the Jumbotron http://www.svconlin...nstall_getting_led/NOT suited for home viewing! [Vernon, Oct 04 2004, last modified Oct 21 2004] A competing technology http://www.ece.uiuc...plasma/plasma2.htmlPlasma display panels may never become high-resolution enough for the desktop. [Vernon, Oct 04 2004, last modified Oct 21 2004] Making lots of flat silicon http://www.halfbake...con_20ManufacturingA couple of suggestions, to make this Idea practical. [Vernon, Oct 04 2004] Basic principles behind LEDs http://mrsec.wisc.e.../LEDs/LED_intro.pdf [Vernon, Oct 04 2004, last modified Oct 21 2004] Info on available LEDs http://wolfstone.ha...EmittingDiodes.html [Vernon, Oct 04 2004, last modified Oct 21 2004] Details about image quality http://www.sharpsma...e_image_quality.pdfI bet a lot of you didn't know how many details go into making a good image. [Vernon, Oct 04 2004, last modified Oct 21 2004] Pixels http://whatis.techt...9_gci212793,00.htmlJust in case you don't know the difference between these things and Tinkerbell clones. [Vernon, Oct 04 2004, last modified Oct 21 2004] An abstract describing arrays of LEDs http://content.aip....AB/v77/i1/10_1.htmlThis is where the quote above, "suitable for large two-dimensional arrays" came from -- with a Y2K date. [Vernon, Oct 04 2004, last modified Oct 21 2004] Progress has indeed been made http://www.eetimes....ry/OEG20021029S0027With respect to making and integrating LEDs into standard silicon circuit manufacturing. [Vernon, Oct 04 2004, last modified Oct 21 2004] The main current competitor http://magazine.fuj...vol34-1/paper13.pdfMaybe this is how LCDs got to be where they are today. [Vernon, Oct 04 2004, last modified Oct 21 2004] Another competitor for the wall http://www.samsung....ts/TV/ProjectionTV/Projection TVs need to be mentioned, along with all those other technologies. [Vernon, Oct 04 2004, last modified Oct 21 2004] One more competing technology http://www.candescent.com/showcase.htmThe flattened Cathode Ray Tube. [Vernon, Oct 04 2004, last modified Oct 21 2004] HP custom 6-digit red LED pixel display http://www.sphere.b...parts/hpparts5.htmlAbout 3/4 the way down. [phoenix, Oct 04 2004, last modified Oct 21 2004] Top marques go LED (.PDF) http://www.avintera...atures%202/merc.pdfSome information on MiPix displays. [phoenix, Oct 04 2004, last modified Oct 21 2004] Barco launches revolutionary intelligent LED pixel block at Frankfurt Auto Show http://www.barco.co.../show.asp?index=921More information about MiPix displays. [phoenix, Oct 04 2004, last modified Oct 21 2004] SuperSCREEN http://www.superscr..._panel.php?navid=15"...modular and lightweight LED panels..." [phoenix, Oct 04 2004] Signtek http://www.signtek.com.tw/lmms.htmLED moving message sign and display unit [phoenix, Oct 04 2004] I can't stop! http://www.shoutmusic.com/led/totally unrelated [RayfordSteele, Oct 04 2004, last modified Oct 21 2004] Back to the top http://www.halfbake..._20Display_20Panels [sufc, Oct 04 2004] OLEDs http://www.wave-rep.../tutorials/oled.htmOrganic LEDs offer displays made up of tiny light-emitting diodes made out of organic film rather than silicon. [kropotkin, Oct 04 2004, last modified Oct 21 2004] (??) Further report on OLED development http://www.wave-rep...-files/oled2002.htm'It was seriously suggested by Seiko Epson that the sweet spot for OLEDs will be in displays from 20" to 100". A sample display was outlined which uses an Active Matrix backplane, which would sell for $2,000 and be hung on the wall - as a flexible display.' Dare I say baked? [kropotkin, Oct 04 2004, last modified Oct 21 2004] OLED details http://www.kodak.co...display/SID2000.pdfYup; they are indeed diodes. [Vernon, Oct 04 2004, last modified Oct 21 2004] Another link Micro-Diode_20Display_20PanelsGotta jump in on the action! [RayfordSteele, Oct 21 2004] OLED http://www.oled-display.netA blog that talks about the latest news on OLED display, lighting and AMOLED technology. [Camslozano08, Jan 02 2010] DrCurry, I have searched and searched to find ANY reference to somebody actively pursuing LED display technology suitable for home use. The Jumbotron doesn't cut it! So, if YOU can find such development in progress, then and only then should this Idea be deleted. Good luck!-- Vernon, Dec 01 2003 Yeah, but now [Vernon] is just being silly. If we agree that LED displays exist, he's calling for small LEDs - nothing more. It's a "me too".-- phoenix, Dec 01 2003 I think we occasionally miss the point/appeal of an idea if it concerns the method of attaining something, rather than simply being a me-too. That's what I think this is about.-- yamahito, Dec 01 2003 [yamahito] I'd agree but LEDs are a well-established technology. LED displays exist. This leads me to believe there is some barrier to [Vernon]'s idea (probably cost).
Here's the way I read the idea: "We should invent ultra small LEDs so we can emulate the color phosphors on a traditional CRT. Furthermore, the LEDs should be manufactured as a single unit/sheet rather than a conglomeration of discrete components. I have no idea what the miniturization process would involve or cost, nor can I demonstrate a reason why this would be preferable to existing or emergent technology. Lastly, I know these exist on a macro scale, but let's ignore that."-- phoenix, Dec 01 2003 phoenix, this IS the HalfBakery, remember? And part of what I wrote was to indicate that we either have, or almost have, the technology to implement this idea -- but that nobody seems interested in actually doing it, as specified herein. Cost is likely the main factor, which is why I posted this only AFTER I posted that other idea regarding silicon manufacturing.
humanbean, you seem to be thinking that each one of those micro-LEDs was going to produce as much light as a regular stand-alone LED. WRONG. They only need to produce as much light as an ordinary phosphor dot --because that amount is KNOWN to be adequate. And, if the process by which they do this is more efficient than phophor excitation, then the total power will be less than a CRT, and not the kilowatts that you describe.-- Vernon, Dec 01 2003 I think the idea is to see how many links we can add to a page before the Halfbakery falls over.-- DrCurry, Dec 01 2003 [FF] I see this idea as "create a LED array on a single chip." This isn't just making smaller LEDs, it's a seperate idea.
I agree that this is an idea that seems simple yet hasn't been done. Does this mean it can't be done? I have no idea why - there's nothing about LEDs that would make them difficult to put on a chip.
I really hate the idea people have that anything that is at all difficult to understand should be MFD'd for the reason that if it were possible then someone else would have tried it. Please, if anyone has a problem with the technology or the cost involved with this idea, tell us what these problems are.-- Worldgineer, Dec 01 2003 freefall, those arrays of LEDs that already exist have HUGE amounts of spacing between them, relative to the sizes of those LEDs. That is the main reason why when you want to do images with current tech, you have to have huge displays. The ONLY place I've encountered any mention of essentially-adjacent LEDs on silicon is in that abstract. And those researchers were looking to use the light for inter-chip communications, not for video images.
Folks, the HalfBakery is full of oodles of ideas that are almost practical. I independently thought of this particular idea, was not surprised in my searching to encounter others with vaguely similar notions, but only I have put the effort here into trying to present it with as much detail as could show its practicality. If it seems "obvious in hindsight", tough. It does not currently exist in the form presented, and nobody out there seems to be working on making it exist in the form presented. THAT is what qualifies this for the HalfBakery.
So, unless Jutta steps in and denounces this, I shall be deleting those MFD references in the near future. Unless those who posted them decide to edit them out.-- Vernon, Dec 01 2003 [V] You've been here much longer than I, so correct me if I'm wrong. I always thought it was bad form to delete people's MFD's without a moderator's ruling.-- Worldgineer, Dec 01 2003 [Vernon] "phoenix, this IS the HalfBakery, remember?...the HalfBakery is full of oodles of ideas that are almost practical." Hey, all I said was that it sounded like a "Me Too". This isn't your best idea, but it's not your worst, either.
[Worldgineer] "I always thought it was bad form to delete people's MFD's without a moderator's ruling." Not at all. Vernon can delete the annotation. [Freefall] can notify an admin if that's a problem.-- phoenix, Dec 01 2003 humanbean, actually I think that LEDs can produce different amounts of light depending on how much power is fed to them. We WANT shades of each color (256 of them!), after all. About the efficiency thing, I did not know how LEDs compared to phosphors (I did put an IF in that prior annotation). I do know that LEDs are lots more efficient than certain other techniques, such as incandescent filaments, so it seemed reasonable to think they were comparable to phosphors.
Thank you, phoenix. At least I am offering fair notice, about impending deletions.-- Vernon, Dec 02 2003 Thanks [phoenix], I always wondered about that. It seems like a very civil way of handling things.
[V] - I've been thinking about the pixel density you'd be able to achieve with this process. It may turn out that pixels are placed so close together that you wouldn't need a large piece of silicon, you can just magnify the image. Who knows, maybe we'll end up going full circle and end up with a big lens on our TVs.-- Worldgineer, Dec 02 2003 Worldgineer, I do understand that we can make electronic components so small these days that a mere phosphor dot is enormous by comparison. On the one hand, what you wrote would reduce cost by not needing such large pieces of silicon, but on the other, one has to worry more about defective LEDs (and the transistors behind them). Larger parts, about the same size as phosphor dots, would likely be more reliable. Also, the smaller the LEDs, the more light they would have to emit, to compensate for being smaller. LEDs do not last forever, and the brighter you run them the more quickly they fade. So again larger LEDs would be better, because they could be operated at lower power levels, most of the time, and the whole display would last longer.-- Vernon, Dec 02 2003 I admit I'm far from an expert in the microelectronics field, but the concept of expected failure intrigues me. Do they design for failed transistors in processors?
It seems to me that unless the failure rate was quite high, you wouldn't need to make the LEDs significantly larger - in fact, keep them small, just stick a hundred of them together with the same function. If 15 of them fail, you still have the brightness of the remaining 85. In fact, if you had a way of verifying this (resistance across the circuit?) you could automatically adjust the brightness of this pixel.
The issue of brightness and lifespan is a tricky one. Letting the LED array be small will increase the brightness required. But I have to ask if this is even an issue. Isn't the brightness/area of a normal LED (considering only the glowing area of the LED chip, not the lens) brighter than a television screen by several orders of magnitude? Since normal LEDs can last 10's to 100's of thousands of hours, this shouldn't be a problem.-- Worldgineer, Dec 02 2003 Worldgineer, yes, they do design electronics these days with a certain amount of redundancy, to compensate for defects that happen during manufacturing. There are limits, though. Liquid Crystal Displays, for example, often have a bad pixel or six. This may not sound so bad when you think about a 1280x1024 LCD that has more than a million pixels, but the bad ones still look ugly. Now I don't know whether bad LCD pixels are more often due to failed transistors or to failed "cells", but certainly those transistors can be rather larger than the ones inside, say, a Pentium 4. Generally, see, the cleanliness at electronics manufacturing plants is so thorough because that is the ONLY way to make the multi-millions of tiny transistors that MUST function inside a processor chip. So, when those cleanliness techniques are used while making the much larger transistors for LCDs -- or for the LEDs in this Idea -- I would expect just about all of them to always work.
Regarding using lots of tiniest LEDs per pixel, well, you can't really have it both ways! Originally you suggested tiniest LEDs so that the overall display could be magnified. If instead you want to have lots of them to be a pixel, then the overall display panel is going to end up about the same size as if fewer larger LEDs had been used. I do see one possible advantage to vast numbers of tiniest LEDs, and this is that NO transistors may be needed. I expected that each LED would need a transistor to control the amount of power fed to the LED. With lots of LEDs per pixel, overall pixel brightness just means turning on some of the LEDs, and leaving others off.
More specifically, to get 256 brightness levels per color-pixel, then each green, each red, and each blue pixel should be an array of 16x16 LEDs -- probably mixed together as 48x48. For an overall display that is advertised as being 2048x1152 pixels (suitable for HDTV), then there would have to be (48x48) x (2048x1152) or 98304x55296 LEDs in the display. They would be directly and individually addressed from the edges of the display (the way pixels are addressed in a plasma or electroluminescent display). Now, if this overall 2048x1152-pixel display was physically the 16x9inches described in the Idea, then the actual physical size of the LEDs would be smaller than phosphor dots on a Cathode Ray Tube, but would still be quite a bit larger than transistors in a microprocessor. Making them smaller yet would provide for the redundancy that started this part of the discussion...Thanks!-- Vernon, Dec 03 2003 Rods Tiger, they do make specialized "display controllers" of various types. If the controllers are controlling subcontrollers, fine. I assume that for whatever type of controller this kind of display would need, something suitable can be devised.-- Vernon, Dec 03 2003 OLED (Organic Light-Emitting Diodes) are a relatively recent but still well-established technique for producing displays. They are constructed out of organic films rather than silicon, but like conventional LEDs and unlike LCDs can produce self-illuminating displays. They would seem to offer clear advantages over silicon-based LED displays, in flexibility, weight, thinness, and robustness, in addition to the fact that they already exist.
With all [Vernon]'s research, has he never heard of these, or (as seems to be the case with many of his other ideas) does he simply prefer to invent more awkward versions of pre-existing technology?-- kropotkin, Dec 03 2003 kropotkin, I do know of OLEDs. But see the sixth link on the list (universaldisplay.com), where they do not define the "D" in OLED as "diode", but instead as "device". It is my understanding that OLEDs are electroluminescent in nature, which means that they glow when a current flows through them EITHER direction. LEDs only glow when the current flows one way through them (and current flow in the other direction is blocked).
For some reason, the best info I've been able to find regarding the resolution of electroluminescent displays is that they are kind-of grainy. This would make them suitable for either small low-cost displays like on camcorders, or on wall-covering displays where people sit back far enough that the pixels are unnoticeable. Not much better than plasma display panels. I could be wrong about that. Still, this Idea is about achieving high resolution, and as far as I'm concerned, if it can't meet or beat the clarity of a good-quality CRT (cathode ray tube), then I'm not interested.
Yes, I am aware that there are probably better ways than LEDs to make a display. (So far my personal favorite is to be found at the second link, iridigm.com, where they are working on ambient-light devices.) However, as I've already stated, ideas such as this is what the HalfBakery is for -- and so here it be.-- Vernon, Dec 03 2003 OK, I agree that OLED displays are diode displays. I am aware that their chief advantage is how they are manufactured (mostly by various PRINTING processes, no silicon substrate needed).
But this HalfBakery idea stays. Fishbone it for being obsolete before ever being done, if you like, or bun it for audacity; I don't care. :)-- Vernon, Dec 03 2003 //Dare I say baked? // No, [kropotkin], you've still missed the point. I still see this idea as "create a LED array on a single chip.", which is different and would have many advantages (and some disadvantages) over OLED.-- Worldgineer, Dec 03 2003 Whoooooo! Infinite contrast screens!-- my-nep, Jul 22 2004 Think I'll [+] this as there's been too much talking and not enough voting. Vernon's got a good idea, done his research and no one has come up with a *reasonable* objection, though of course there are some serious obstacles to be overcome before market. The large silicon sheet would be my biggest worry - how big a sheet can we make now? Please tell me here Vernon as I just can't face the half-mile of links which will probably tell me somewhere. On the positive side, there is sufficient money in good displays for manufacturing faults to be overcome by overproduction. When plasma's were first manufactured, three out of four were destroyed off the assembly line due to pixel faults. As a result they cost a fortune, but they still sold...-- wagster, Jul 22 2004 [+] as it's a low power solution to flat screen technology, and apparently having so many plasma screens is causing a surge in demand for lectricity.
I like the idea of having a single pixel made of an array of diodes, would it be possible to just make loads of diodes in the 3 primary colours mix them up wash them over a substrate array and then calibrate?
How big is the diode that's inside an LED that I'd buy if I took all the plastic and metal connectors off?-- scubadooper, Jul 22 2004 wagster, the link you want is called "Making lots of flat silicon".
scubadooper, those pieces of silicon are too big for this idea. They are big because that lets them be brighter. When shrunk to phosphor-dot size, they only need to be as bright as phosphor dots.-- Vernon, Aug 04 2004 [-----] you can't just say baked and leave it at that, well actually I suppose you can, but it'd be nice if you gave us some detail and maybe a link to keep the others company as I think they might be getting lonely-- scubadooper, Nov 13 2004 I've definitely seen this before. It's baked. One use is in portable computers, where it uses less energy.-- -----, Mar 28 2005 I believe this is now baked (it's the year 2010). Google "OLED Display". My phone has one and it's almost too bring to look at in a dark room, even at the lowest brightness setting.-- cowtamer, Jan 03 2010 [cowtamer], the "O" in "OLED" stands for "organic". I was talking about inorganic LEDs here (they can have a longer lifespan).-- Vernon, Jan 04 2010 Where was I when this was first posted? Ah well. A couple of issues:
Using an entire wafer of silicon for a display is not ideal. Chip manufacturers have been chasing the ability to create larger wafers for reasons economies of scale: larger numbers of chips per wafer. Redundancy is built in at that level because of the incredibly tight requirements for purity. Modern chip facilities are probably the 'cleanest' places on Earth. Any defects, or operating faults in the LED are probably much more likely than a dead pixel in a standard digital flatscreen (plasma, LCD, etc).
I'm also skeptical of the switching on/off of the LED at the required resolution. Not that it will be tough to make; you could run the switching through the silicon nae bother - the problem will likely be heat generation. You know how hot modern processors get- this thing would basically be a giant hotplate!-- Jinbish, Jan 04 2010 random, halfbakery