h a l f b a k e r yThe embarrassing drunkard uncle of invention.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
You sit and face a flat glass plate. Let's assume a medium-size plate for a TV, say 25 inches (63cm) of diagonal measure. But we can also assume a 16:9 aspect ratio, so that we can describe a High-Definition TV.
The far side of this plate is covered with ordinary phosphor dots, just like an old-fashioned
Cathode-Ray-Tube TV. There is a cabinet holding the projection mechanism, but no vacuum or electron beam is involved here.
With the preceding information and an assumption of a fine "dot pitch" (about 60 dots per centimeter; see .15mm in link) we can compute the maximum resolution of this screen. I'm rounding down the figures I got, to 3072x1728 pixels.
Mounted horizontally a little ways behind the middle of the glass (maybe a foot, 30cm) is a spinning cylindrical mirror. The length of this mirror is a significant fraction of the width of the phosphor screen. In cross section this cylinder is actually a hexagon or an octagon or whatever-works in this particular design, and so the cylinder actually has a number of flat facets.
Just below the base of the glass face of this TV, and also located near the glass, is a "micromirror" device. (See DLP link.) This particular device is special for this Idea. Most micromirror arrays have a million or more tiny mirrors in them, arranged 1280x1024 for example (5:4 aspect ratio), or 1024x768 (4:3 aspect ratio). Here we want a single row of mirrors, just 3x3072 or 9216 of them, and that's all. This means they can be made a LOT less expensively than the typical million-mirror device.
Remember those 3072 horizontal pixels among the phosphors on the glass? That number applies three times, to red phosphors, green phosphors, and blue phosphors. So if we want to use the light from one micromirror to activate one phosphor, we need 3x3072 total mirrors, for all the dots in one row of pixels.
Next, since the micromirror device is tiny compared to the width of the glass plate, we need some lenses to properly expand and focus the light that is reflected from the micromirrors.
That light comes from a small mercury-arc bulb powered by an equivalent "HID" (High Intensity Discharge) ballast. This type of light is a rich source of ultraviolet.
OK, now, the light from the lamp is focussed toward the micromirror strip. The TV signal-processing electronics takes one complete scan-line of information, and activates all the micromirrors in parallel. They tilt to reflect light to the next lens -- or not. The number of times per second that they tilt controls the amount of light that is reflected to that second lens.
The second lens expands the light horizontally, but still sends it as a strip toward the spinning mirror. Depending on the momentary orientation of a long mirror-facet, the arriving light may be reflected toward the upper part of the overall glass screen, toward the lower part, or anywhere in-between. Note that because the light is widening as it travels, that's why the length of the spinning mirror need not be the same length as the width of the screen.
Due to the fact that the focal distance changes with respect to the angle that the light takes, from the spinning mirror to the phosphors on the glass screen, a final lens is probably also needed, to sharply focus the UV light upon the phosphors. This lens has to be as large as the glass plate, but can probably be of the stamped-out Fresnel variety (see link), and thus be fairly cheap.
Like an ordinary CRT-TV, the phophors glow when energized. They keep glowing for a long-enough fraction of a second that human persistence-of-vision can build a complete image while only one line of pixels at a time is illuminated by the UV light. As the spinning mirror finishes reflecting light to the base of the screen, the next facet rotates into position to reflect the next line of UV light to the top of the screen. Even if the "vertical refresh rate" is 100 times per second, a multifaceted spinning mirror is easily up to this task.
Finally, don't worry about sunburn from this TV; the ordinary glass of the screen is a good absorber/blocker of UV light. All in all, this should be a fairly affordable High-Definition TV.
About Dot Pitch
http://www.webopedi...RM/D/dot_pitch.html
As described in the main text [Vernon, Sep 07 2005]
DLP (Digital Light Processors)
http://www.dlp.com/Default.asp?bhcp=1
As described in the main text [Vernon, Sep 07 2005]
Some Fresnel Lens stuff
http://www.lanternr...m/misc/freslens.htm
As mentioned in the main text. The lens needed for this Idea can be quite flat. Also, the ridges are straight/horizontal and not circular, as is true for many Fresnel Lenses. [Vernon, Sep 07 2005]
Thermal imaging using just eight sensors - all done with mirrors
http://www.atra.mod...hoenix/L04-TITV.pdf
[coprocephalous, Sep 07 2005]
Digistar 3
http://www.es.com/p...digistar3-laser.asp
Replace the visible laser with UV and use a reactive screen, and you've got your idea. [Freefall, Sep 12 2005]
Grating Light Valve
https://en.wikipedi...Grating_light_valve
Mentioned in my anno [notexactly, Apr 19 2018]
[link]
|
| |
I'll have to read again in the morning. (Vernon, Are you, by any chance a descendant of Asimov?) |
|
| |
I mean this in the most respectful way. |
|
| |
{Zimmy], I've read most of Asimov's books, but am not related otherwise. |
|
| |
I've read about some of the tech you are talking about, I just am not in the right mind to understand the steps you've taken. |
|
| |
I didn't finish reading ... but sounds like you know what you are talking about :-) |
|
| |
Vernon, for goodness' sake put an executive summary in at the top. It makes it a lot easier to understand the relevance of all the minute details if you bother to tell me what you're trying to do with them all first. |
|
| |
This works in theory, but I can see a few issues with it. Firstly, one of the biggest costs with projectors is the bulb itself, which has a relatively short life of a few hundred to a few thousand hours, and costs several hundred pounds/dollars to replace. Also, for this to work properly you are relying on the spinning mirror being in exactly the right place; wear on the bearings would lead to vibration and consequent degradation of the image. Similarly, any vibration in the room would upset the picture to some degree. |
|
| |
Either you need something solid state, or you need something that corrects for its own errors. |
|
| |
// Also, for this to work properly you are relying on the spinning mirror being in exactly the right place; wear on the bearings would lead to vibration and consequent degradation of the image// Don't laser printers use spinning mirrors? Seems to work fine / long-term for them. |
|
| |
Having investigated, you're quite right, they use rotating mirrors and they do work long-term. It might be relevant to point out that the spinning mirror in a laser printer just diverts a single beam to scan across the width of a piece of paper, whereas this has to get over 9,000 seperate beams to hit a point target; if they're even 0.1mm off then they will light up a dot of the wrong colour. |
|
| |
The other thing that needs developing to make this viable is a DLP that can handle speeds 1728 times faster than current systems, as obviously having fewer mirrors to do the same job requires faster mirrors. That's just a matter of time and economics though. |
|
| |
[david scothern], I thought there were too many pieces to summarize this? Next, HID bulbs have been used for years in street lights and billboard illumination and industrial/warehouse illumination. They last for years because they do not have incandescent filaments; they use an electric arc through mercury vapor. The color of this light is not relevant because it gets absorbed by phosphors; it is not intended to be directly seen (or indirectly seen, after passing through a color-filter light valve, per ordinary projection TVs). |
|
| |
And, yes, I know that the spinning mirror is the weak piece of the design. But the more facets it has, the slower it can spin -- and it may be possible to mount it accurately enough using magnetic bearings (zero rotational friction). |
|
| |
Regarding the DLP mirrors, the alignment thing is a factory-adjustment problem. With the spinning mirror non-spinning, we want the line of DLP mirrors to illuminate a line of phosphor dots -- or not. I think a bigger problem is where does the reflected light go when you want a DLP mirror to NOT strike a particular pixel. It seems to me that each rotating mirror facet must be about as wide as the height of a row of phosphor dots, and that the rotating mirror should be behind a shroud, except for a slot through which light from the DLP mirror comes in and bounces out. Then the must-miss light hits the shroud instead of some other phosphor dot. Regarding mirror-moving speed, that's not quite so bad as you think. They already can move back-and-forth hundreds of times per second. An improvement factor of five or six may be sufficient. |
|
| |
I went to have a look as to why projectors suffer such short lamp life. What I found was that they can't use the lamp type you suggest as the colour that they put out is too far from white (strong bluish tinge) although I can't see why that couldn't be corrected for. |
|
| |
For my part, I look forward to LEDs being developed to the point where they last a long time while producing a true white light. I think that will revolutionise projectors by reducing their size, power consumption and running costs. |
|
| |
As for a summary, the point is that you leave out the detail. Just telling us that it was intended as a fairly affordable HDTV would have made a big difference. |
|
| |
HDTV standards offer refresh rates up to 60Hz. To support, say, 1600 lines at 60Hz, we're going to need mirrors operating at 100khz. Current DLP-based projectors operate at a maximum of about 100Hz, although that's not to say that the DLP itself isn't capable of operating at higher speeds. |
|
| |
[d_s] laser printers achieve very high resolution (2400dpi and up) and repeatability with their optical systems. However, the length of the optical paths in this idea will require a very rigid (read "heavy") casing to eliminate jitter.
The use of a linear mirror array is somewhat analogous to the mechanically scanned high resolution thermal imaging cameras, with the characteristic horizontal banding in devices like the TICM |
|
| |
It's the repeatability that is the interesting bit. It's fine while the whole system is rigid and operating on close tolerances, but after long use, surely it decays even in a laser printer? The bearings are going to wear; the question is do they wear fast enough for it to be a relevant problem? |
|
| |
I'm still not in the best of shape to understand. |
|
| |
Why Mercury light?(expensive) I thought laser worked better in this application. |
|
| |
[Zimmy], a mercury-arc light is, in a way, just a more powerful variant of an ordinary fluorescent light. The "more power" is the reason for the expense. But a projection TV doesn't need quite THAT much illumination power, as is used in street lights or billboards. So, the cost of a mercury-arc lamp specifically intended for projection TV use should be reasonable. |
|
| |
Regarding lasers, because this Idea requires ultraviolet light, that means you would need a UV laser. Those are I think, even more expensive than billboard lights! |
|
| |
Alignment is a problem, not at the factory, but after the set has been moved a few times. Big plastic cases are hard to make rigid (not "you can't" just "it might not be cheap"), and some HDTV's have had to be recalled due to misalignment, even with in the field adjustability. |
|
| |
Why UV light? To make the phosphor dots glow? |
|
| |
I'm going to have some acquaintances who work for a company developing an L-COS HD projection engine take a look at this idea. They might have some input about the optics involved. It doesn't sound simple. |
|
| |
[oxen crossing], yes, I'm pretty sure that UV can make phosphors glow just as well as electron-beam excitation. Heh, just walk into any of many nightlife spots that feature blacklights and phosphorescence. |
|
| |
Regarding maintaining of alignments, can't stiffness-of-chassis and vibration-isolation do the same job as sheer mass? So, how about fiberglass and that NASA stuff that's getting advertised in lots of mattress commercials? |
|
| |
Stiffness yes. It doesn't matter if the whole thing shakes as a unit, as long as the various components in the light path maintain their relative positions. However, large, stiff structures tend to be heavier than their more flexible brethren, obviously enough. |
|
| |
Designing a stiff structure would give better results than designing a heavy structure that doesn't vibrate simply because the vibrations it sees don't have sufficient energy to get it going. |
|
| |
I'm not sure what sort of vibration isolation you are proposing; isolating the optical components individually will cause problems as the optical path will be interfered with as the components move independently of each other; isolating the whole system from outside interference would be good up to a point, that point being susceptibility to vibration produced within the TV itself (eg from the motor spinning the mirror). |
|
| |
[david scothern], did you see my prior anno that mentioned using magnetic bearings for the rotating mirror (you can assume for the motor, too)? Wouldn't that just about eliminate that part as a source of vibrations?
Regarding vibration isolation, the substance I had in mind to use (without specifying exactly where) is that "Tempur-Pedic" stuff being advertised all over the various TV channels. (One commercial shows a glass of wine on the bed, and a few feet away somebody is jumping up and down on the same bed, and the glass isn't moving very much at all.) |
|
| |
I'm not at all convinced that vibration and physical stress
will be as much of a problem as you all make out.
Desktop projectors get banged, whacked and dropped on
the floor regularly (I know, I've done it) and maintain
alignment very well due to stiff design and lightweight
optical components. The mirror can be made fairly
lightweight by keeping it close to the DMD and very thin -
remember each face needs to be not much taller than
one row of pixels. |
|
| |
The DMD speed is the real weak point, Texas Instruments
(the sole manufacturer) seem to be fairly cagey about
switching times, but they are all designed around frame
refreshes, not line refreshes. They also claim that use
outside the visible spectrum may cause problems. |
|
| |
Interesting plan though. [+]\ |
|
| |
You've been reading the literature on the Digistar 3 system, haven't you? (see link) |
|
| |
It seems that the receiving & projecting methods for Television may change before the components of this type of system wear out anyway. I read 3-4 years ago that it would be best not to buy an expensive TV set until the technology that dominates emerges from the scrum. |
|
| |
It's possible, that the same thinking could be applied currently. The Technological Revolution is in Full Steam. Too bad we had to chop block the Soviet Union, as they made some bad ass High TEch MIG's & Rockets as well. (It may come out that their studies on the paranormal had some importance as well) |
|
| |
Perhaps, THE PROC (In name only) will re-fuel the Tech Revolution. |
|
| |
+ BTW. (The Pristine reading gave me Vertigo.) |
|
| |
I need Dr. Zarkov to explain this one to me...why not just go collect aluminum cans and sell them at the recycler until you have collected the $US 3,500.00 to buy a really nice, flat screen plasma at Circuit City. It would take less effort and the darn thing would probably work right out of the box. But, I'm voting affirmatively on this one just because of the bandwidth it took to describe it...probably some sort of record. |
|
| |
// this has to get over 9,000 seperate beams to hit a point target //
[2005-09-07] |
|
| |
When did "over 9000" become a meme? |
|
| |
// that they can't use the lamp type you suggest as the colour that
they put out is too far from white (strong bluish tinge) although I can't
see why that couldn't be corrected for. // |
|
| |
Lamp spectrum is mostly irrelevant in this application, because it's only
there to excite the phosphors. |
|
| |
// For my part, I look forward to LEDs being developed to the point
where they last a long time while producing a true white light. I think
that will revolutionise projectors by reducing their size, power
consumption and running costs. // |
|
| |
// a DLP that can handle speeds 1728 times faster than current
systems // |
|
| |
// we're going to need mirrors operating at 100khz // |
|
| |
Use grating light valves [link]. I've read they can switch light at up to
30 MHz (!) and they seem ideally suited for (rather, only suitable for)
monochromatic light, which is what we're trying to work with here. |
|
| |
// and it may be possible to mount it accurately enough using
magnetic bearings (zero rotational friction) // |
|
| |
What about air bearings like in hard drive motors? Those operate for a
long time with negligible vibration and friction. |
|
| |
// [various comments on stiffness]] // |
|
| |
What about active stiffness? |
|
| |