AVS Forum banner

1 - 20 of 48 Posts

·
Registered
Joined
·
6,752 Posts
Discussion Starter · #1 ·
Take an on-sale Optoma or Benq for 799-899 and triple that..oh right you'll also need some prism action (just like every cheap LCD that isn't a singlepanel), and different boards (which are still cheaper than tripling a the single dlp boards from three 1070's). Assembling shouldn't cost anything significantly different and R&D is already done. Brightness is increased in a much more efficient way than using multiple lamps, and there's no reason for it to cost over $2,500. The same price as a dual-lamp pj, with less heat, less moving glass parts, no RBE, less power draw, quieter operation, and cheaper upkeep. Now realize that anyone off the street (no manufacturers' discount..retail) can buy a dlpchip and controller for a lot less than 799. A 3chip should be more than possible at $1899-2000 while maintaining similar or better profit margins compared with budget singlechipDLPs.


This would allow options to both expand well past the 2000-2500lm maximum or use a lower powered, cooler running lamp in eco and still keep up 700-800 for your theater.


Why isn't this a thing?
 

·
Registered
Joined
·
6,752 Posts
Discussion Starter · #3 ·
Lol, true. But I'll sweeten this sales pitch. Imagine how many livingrooms, offices and schools you could invade with 5,400+ fullcolor lumens blasting out without a jet-engine fan. There'll always be a 40k market for a couple nudges of contrast and a cool spray of Runco black. ..plus I really want a dlp lightcannon that doesn't suck at primaries because most LCD drive me crazy. I wonder if there's someone big and important that takes calls and makes things happen.
 

·
Registered
Joined
·
1,375 Posts
I think part of it comes down to the law of diminishing returns.


With the same lamp, I don't think you'd get all that much extra brightness from a split-light-path in a 3-Chip DLP, when compared to a single-light-path via a color wheel (ie Single-Chip DLP). It might be a bit, but I don't think that the colored filters on a color wheel are all that much less efficient than the static colored filters used in 3-Chips. Sure there's a bit of light loss in the transition between segments; but I suspect those losses aren't particularly massive (actually one of BrilliantColor's purposes is to compensate for this - albeit at the cost of a bit of color accuracy)..


Other than increased cost due to the additional DLP chips, convergence might also be something of an issue in the case of 3-Chips as well; so sharpness might suffer (which is one of single-chip DLP's main strengths).


For the end-user, that'd leave the only truly major benefit being the lack of RBE. I think it's a bit of a tough sell for manufacturers for this reason; in that it's a lot of additional hardware (optical and DLP) with only one truly sell-able benefit. Couple that with the fact that most users don't notice RBE to begin with (especially on high-speed wheels); and it's not an easy proposition in such a competitive market...
 

·
Registered
Joined
·
7,072 Posts
I get the argument, but DLP really offers no product, other than the W7500 which has decent lens shift out there, and there is no compelling DLP product which costs more than the 7500 from BenQ. Considering that Epson delivers low cost, mid-cost, and higher cost models with the 2030, 3020, and 5030 models respectively, it is just a shock that DLP doesn't have a lot more that is trying to take over market share. RBE is a issue for more than a few people, a lot are just flat out worried. Others want the higher contrast which the 5030 (or AE8000) offers over the cheaper DLP models. There is a entire market segment which DLP fails to hit, which should be attainable and frankly, could be dominated, by the first DLP manufacturers which step up with $3,000ish 3-chip models with decent lens assemblies on them. Take the AE8000 or 5030 and change it to a 3-chip DLP model.


I am well aware that it isn't as easy with DLP as it is with LCD to get 3-chip to work, but it is just ridiculous how DLP completely fails to deliver lens setups like Epson and Panasonic have done for years. Even on the cheaper 8345 or AR100 models, you get a 2x zoom lens and phenomenal lens shift in a 3-chip design for under $1,500.


I certainly would like to offer more than one DLP model to the people I work with which is under $3,000 and has significant placement flexibility and a image which should match or rival the LCD competition. Maybe even give LCoS a run for the money at that point.
 

·
Registered
Joined
·
15,106 Posts
Yah but DLP has technological limitations to its lens shift in the manufacturing that LCD does not have. Benq kind of cheated a little even to get as much lens shift as the w6000/w7000/w7500. They can't really make it have more.


3 Chip DLP's would also lose their advantage in this market segment, as the big advantage of many of these DLP's is the sharpness, but low-priced 3-chip DLP's will not be as sharp, not nearly as sharp as single chip ones. Sure in the $10,000+ 3-chippers are really sharp, but that won't work in the sub $5k market.
 

·
Registered
Joined
·
6,752 Posts
Discussion Starter · #7 ·
Is the limit to lens-shift and zoom caused by the travel path of the dmd mirrors between on/off or is it something else?
 

·
Registered
Joined
·
15,106 Posts
I read about it in a TI whitepaper, but that was quite a few years ago.

It's an optical problem that causes an extraordinary loss of sharpness and loss of contrast, but I do not recall exactly where the optical issue was.


One reason the Benq's native on/off is so low is because of the way they designed it with the lens shift.
 

·
Guest
Joined
·
867 Posts
Three chip DLP design.

You are probably talking about using a high watt lamp or lamps to take full advantage of the higher light efficiency of 3 chip DLP, so a long glass integrator rod with optical coatings (less heat produced and higher efficiency and higher long-term reliability than a short hollow mirror tube used in most single chip DLPs). After the integrator rod probably a focusing lens with optical coatings followed by a fold back mirror then maybe some form of iris then another fold back mirror then another focusing lens with optical coatings to then go into the first of the three color splitting prism with optical coatings one for each of G R B and then a TIR prism with optical coatings and then the projector lens (which needs to be larger than with a single chip offset design) with optical coatings and probably an iris in the lens, It is also likely to have lens shift.


p20 of this link has a picture of the prisms used in three chip dlp
http://terahz.org/_pdf/Toshiyoshi%20-%20Projection%20Based%20Displays.pdf

The light path up to the prisms would I guess be similar to a telecentric single chip dlp
http://www.ti.com/lit/an/dlpa022/dlpa022.pdf

The followin link has a basic LCD/LCOS lightpath diagram for comparison. As you can see the prism is different to three chip dlp.
http://www.logiplayer.com/index.php/technology



In comparison to a single chip offset design (color wheel, short hollow mirror tube integrator, focusing lens, mirror, dlp chip, lens) far more components especially glass optical components with optical coatings that are going to be expensive and high quality to produce the best results and the projectors light path is also going to require far higher precision manufacture than a single chip dlp to produce best results.


An offset single chip dlp design is far simpler and has higher intrinsic contrast due to its offset design and is easier to manufacture. With high speed color wheels and VIDI lamp technology giving single chip dlp the capability to have adequate bit depth and color saturation. The main or only advantage to 3 Chip DLP these days seems to be light efficiency so what would be the point if you don't need the lumens.


What makes three chip DLP superior in the real world is probably the price point, that is single chip dlp projectors are usually built to a far lower price. Not that single chip dlp is inherently inferior, just that in comparison to three chip DLP they are usually poorly designed and poorly built using cheap components to be cheap to manufacture. They are built to be good enough, not as good as they can be.
 

·
Guest
Joined
·
867 Posts

Quote:
Originally Posted by Ftoast  /t/1522259/why-arent-there-any-inexpensive-3chip-dlps#post_24475309


Is the limit to lens-shift and zoom caused by the travel path of the dmd mirrors between on/off or is it something else?

Lens shift is not a problem with telecentric light path designed DLP projectors, it is difficult to implement in offset light path designed DLP projectors. Due to cheaper cost of manufacturea and higher intrinsic contrast most single chip DLP projectors are offset design. Telecentric requires more design and higher cost both in components and manufacturing precision to get as good or better results.
http://www.ti.com/lit/an/dlpa022/dlpa022.pdf
 

·
Registered
Joined
·
6,752 Posts
Discussion Starter · #11 ·
The simplified diagram of a LCoS makes it look like the panels are transmissive rather than reflective..I would have to believe a reflective design (as both are, unless I'm very wrong) could be the same or very nearly. Then again there aren't many inexpensive LCoS, though that's partially for other reasons than parts' cost. I also have to believe the usual markups for optical components, while beyond ridiculous to the public, aren't nearly so for any company already either kneedeep with their suppliers or manufacturing their own.


I realized my huge mistake for believing in the lumens/watt increase though. Once enough corners are cut to put out an inexpensive device, it would likely have no brightness advantage at all. If LED technology gets to a point where it can be kept cool without needing to strobe, using a 3chip configuration should practically triple the light output. That doesn't seem to be the direction it's headed though. Except the couple of hybrid theaters that need that extra power. Then again if they could keep them cool, they could also keep a brighter-strobed one just as cool and use a cheaper singlechip design that shouldn't suffer RBE anyway. Huh..


Thanks for giving me something new to read/learn. Hopefully any of my future rant-whine-posts will be better thoughtout for it.
 

·
Guest
Joined
·
867 Posts

·
Registered
Joined
·
6,752 Posts
Discussion Starter · #13 ·
There are some very cheap LCoS, but I suspect they are singlepanel or something just as deviously evil. I know Aaxa makes (or made) a little SD LED LCoS unit with claimed 1000:1 contrast for around $120. Ooh, that opens up another thought; a well-made single panel LCoS could have a 30% (single color) fill-ratio which would be very comparable to an inexpensive 3panel LCD, but with better contrast, smaller light path and less parts. I wonder how terrible 30-32% would look. ..maybe something for a different thread, but I personally don't mind the derail (what with it being my own train splitting).


Now that I look again, the way I was thinking 3chip dlps worked DOES nearly triple the light output (losses notwithstanding)! ..okay I'm back to wanting one again.

The filters redirecting and splitting the light spectrum rather than absorbing it as heat (like a colorwheel) paired with their constant on-state (rather than off-off-on)..is it really that simple?
 

·
Registered
Joined
·
7,072 Posts
Entry level QUALITY LCoS projectors are out there for under $3,000 from JVC. Sony is not far off. They all compete with the Epson 5030 and Panny AE8000... They all have lens shift, good zoom range, etc. So where is DLP? The excuses are deep, but that's what they've been doing now for years. Not just one or two years, but I asked why DLP didn't deliver lens shift for under $3,000 years ago, and they said it couldn't be done.


Frankly, it's all garbage.


There are no competing 3-chip designs because DLP manufacturers don't want to put that product out there. They are holding back on the design and engineering path which LCoS manufacturers have been walking down for years and have managed to create excellent product in for a fair price. We have seen single chip products which continue to push amazing '2x' color wheel speeds over a decade after 6x and 7x color wheel speeds were considered the gold standard for home theater use.


In my opinion, DLP manufacturers don't care. They are racing to the bottom and have no interest whatsoever in competing with the mid-tier models out there such as what Sony, JVC, Epson, and Panasonic deliver. On the low end, I think they have done a great job, and unless Epson sharpens up the 2030 and improves lag time, the use of DLP on the cheapest models makes sense.


But, if you want something different than the W7500 from BenQ, then you better not hold your breath on it... and if you want something better? Forget it at anything approaching a reasonable price.
 

·
Guest
Joined
·
867 Posts
Not all single chip DLP projector manufacturers are racing to the bottom. Some single chip DLP projectors are damn expensive, Projectordesign F35 and F32 spring to mind with upto 8,000:1 native on/off contrast depending on iris setting.


And there is nothing wrong with 2x colorwheel if people don't see rainbows. I don't see rainbows on a Projectiondesign F3 SXGA+ and that is only two speed, it might have something to do with the color wheel primaries or the light integrator or being a telecentric design so more even illumination, god knows, because I have seen rainbows on other DLP projectors with faster than 2x color wheels. The issue with faster color wheels is increased dithering, now less of an issue with VIDI lamp technology, long-term reliability as far as wearing out, and pitch of the sound of the color wheel spinning.
 

·
Registered
Joined
·
1,375 Posts

Quote:
Originally Posted by Ftoast  /t/1522259/why-arent-there-any-inexpensive-3chip-dlps#post_24477741



Now that I look again, the way I was thinking 3chip dlps worked DOES nearly triple the light output (losses notwithstanding)! ..okay I'm back to wanting one again.

The filters redirecting and splitting the light spectrum rather than absorbing it as heat (like a colorwheel) paired with their constant on-state (rather than off-off-on)..is it really that simple?

You would appear to be right! Assuming the 3-Chip uses a prism to separate the input light into primaries (and no color filters), the same lamp should yield much more brightness than when used in a 1-Chip equivalent. (The color wheel in a 1-Chip dissipates - ie wastes - any light produced by the lamp that differs in frequency to the filter (color-wheel segment) that's over the lamp at the time.)


So that means a serious leap in brightness coupled with zero RBE; at the expense of additional optics and a potential loss in sharpness due to convergence. At the right price, that's a trade-off many would make; especially for the extra brightness.


Also for lower-power lamp tech (LED comes to mind), this would hypothetically be a good way to compensate for the inherently lower brightness compared to traditional bulbs.


It's an interesting thought indeed, Ftoast...
 

·
Guest
Joined
·
867 Posts
If increased lumens is the main advantage why not just use a single chip projector with a higher power lamp or multiple lamps. A dual 300 watt single chip DLP projector is capable of a lot of lumens.


The leap in light efficiency with three chip dlp would I guess be just over three fold over a single chip RGB color wheel. It would be less of a leap from a RGBYCM color wheel to three chip as yellow and cyan segments boost light efficiency although at the expense of some accuracy, or for that matter a whie segment or using spoke time to peak white which boost brightness at the expense of color saturation.


If three chip DLP are similar to telecentric single chip DLP they may have lower inherent contrast. So to achieve the same or better contrast than a single chip DLP they might have to sacrifice lumens by using an iris or two.
 

·
Registered
Joined
·
1,375 Posts

Quote:
Originally Posted by dovercat  /t/1522259/why-arent-there-any-inexpensive-3chip-dlps#post_24480800


If increased lumens is the main advantage why not just use a single chip projector with a higher power lamp or multiple lamps. A dual 300 watt single chip DLP projector is capable of a lot of lumens.

Like the OP says: doubling lamps means even more heat, more cooling, and hence more noise. As well as even more optics for the additional light source. Also since lamps are inherently not 100% identical in their output, things like brightness uniformity may suffer (depending on the lamp). And of course since lamps and their housings are so expensive (one of the single most expensive items in any projector), doubling this cost is going to hurt the 'inexpensive 3chip DLPs' we're after

Quote:
Originally Posted by dovercat  /t/1522259/why-arent-there-any-inexpensive-3chip-dlps#post_24480800


The leap in light efficiency with three chip dlp would I guess be just over three fold over a single chip RGB color wheel. It would be less of a leap from a RGBYCM color wheel to three chip as yellow and cyan segments boost light efficiency although at the expense of some accuracy, or for that matter a whie segment or using spoke time to peak white which boost brightness at the expense of color saturation.

RGBYCM wheels have been found to cause quite a lot of color inaccuracy (white segment!) and once properly calibrated, most of the brightness gain from said segment is lost (see here ). Likewise for the remaining 5 segments, we're left with the exact same problem of a lot of light being dissipated; so even taking full advantage of said white segment (and the color-accuracy losses incurred), we still lose brightness for the remaining 5/6th of the time in a similar same manner to a 3-segment (RGB, or RGBRGB) wheel...


3Chip DLP's use just R, G, and B so color accuracy should be at least as a good a RGB wheel which doesn't have the above disadvantages.


So it's still a gain in brightness from the same lamp, with more potential brightness than even an RGBYCM single-chip; and potential perfect color accuracy.


I'm still keen!
 

·
Guest
Joined
·
867 Posts

Quote:
Originally Posted by kreeturez  /t/1522259/why-arent-there-any-inexpensive-3chip-dlps#post_24480866


Like the OP says: doubling lamps means even more heat, more cooling, and hence more noise. As well as even more optics for the additional light source. Also since lamps are inherently not 100% identical in their output, things like brightness uniformity may suffer (depending on the lamp). And of course since lamps and their housings are so expensive (one of the single most expensive items in any projector), doubling this cost is going to hurt the 'inexpensive 3chip DLPs' we're after
Doubling lamps does not oddly enough mean more heat and noise and less brightness uniformity in my limited experience. I own a dual lamp projector that out puts less heat than the single lamp projector I used to own and less noise and has superior brightness uniformity.


The differences that account for this are probably far supperior design of cooling, each lamp has its own small fan with air flow tunneled throw the lamp and there is also a large cooling fan below the lamp housing. The audible noise is less than with my previous single lamp projector on high lamp mode which on high mode was not in my opinion usable if anywhere near the viewer. The pitch of the noise may also have something to do with this as well as the dB level.


If you want no lamp noise and no lamp heat in the viewing room it is also possible to design a projector where the light source is seperate from the rest of the projector contected by fiber optic cable, there is at least one projector design that does this.


Less heat from my dual lamp projector maybe down to superior cooling meaning the lamps don't get as hot and as a side effect last longer, and may also be due to the projector using a glass integrator rod instead of a mirrored tude. Glass integrator rods are higher light efficient and result in less heat than mirrored tube ligh integrators.


Brightness uniformity and light efficiency is also far superior as it is an telecentric design single chip dlp not a offset design single chip dlp.


I don't see why dual lamp would cause less brightness uniformity even in a offset design as the lamps are before the light integrator so the light should be integrated prior to going through a focusing lens and on to the DLP chip.


Lamps and their housing don't need to be replaced often if the projector has adequate cooling and these days you can pay far lower than the manufacturers suggested retail price, and you can always buy the lamp without the lamp module.
Quote:
Originally Posted by kreeturez  /t/1522259/why-arent-there-any-inexpensive-3chip-dlps#post_24480866


RGBYCM wheels have been found to cause quite a lot of color inaccuracy (white segment!) and once properly calibrated, most of the brightness gain from said segment is lost (see here ). Likewise for the remaining 5 segments, we're left with the exact same problem of a lot of light being dissipated; so even taking full advantage of said white segment (and the color-accuracy losses incurred), we still lose brightness for the remaining 5/6th of the time in a similar same manner to a 3-segment (RGB, or RGBRGB) wheel...
According to Texas Instruments. A RGB color wheel has a light efficiency of 16.5%, a RGBYC color wheel is 24.7% at the expense of some color accuracy. Using spoke time can also be used to boost light efficiency at the expense of color saturation, and a white segment massively boosts lumens at the expense of color satuartion. A three chip DLP I guess is going to be just over 16.5 x 3 = 49.5% light efficiency, but most likely with a lower intrinsic contrast ratio than an offset single chip dlp projector. Using one or two irises to increase contrast ratio is going to drop that light efficiency. Yes it will be brighter using the same lamp, but at greater expense then just using a single chip dlp with a more powerful lamp or more than one lamp.
Quote:
Originally Posted by kreeturez  /t/1522259/why-arent-there-any-inexpensive-3chip-dlps#post_24480866


3Chip DLP's use just R, G, and B so color accuracy should be at least as a good a RGB wheel which doesn't have the above disadvantages.
Color accuracy just like on a RGB color wheel projector is going to depend on lamp used the color primaries used in the filters and video processing used. Three chip is not inherently better color accuracy than a single chip dlp, the only reason this is usually the case is because single chip dlp designs have chosen to trade color accuracy for lumen output. There used to be an advantage for 3 chip DLP as far as bit depth, but with VIDI lamp technology I think that is no longer the case. There may or may not be some sort of advantage in having the image made up of a constant RGB image rather than a sequential RGB image as far as perceived solidity of the image or color saturation, but I doubt it.
Quote:
Originally Posted by kreeturez  /t/1522259/why-arent-there-any-inexpensive-3chip-dlps#post_24480866


So it's still a gain in brightness from the same lamp, with more potential brightness than even an RGBYCM single-chip; and potential perfect color accuracy.
The gain in brightness is only if they have like for like lamps. A single dlp chip projector with a higher power lamp or multiple lamps even using a RGB color wheel or wheels can be birght enough for most or all uses and as like a three chip dlp it has plenty of lumens to begin with it can be designed to be both color accurate and high lumen output.
Quote:
Originally Posted by kreeturez  /t/1522259/why-arent-there-any-inexpensive-3chip-dlps#post_24480866


I'm still keen!
What about the inherent disadvantages of three chip DLP.

Less sharpness, due to alignment of three dlp chips and possibly due to far longer back focal length to the chips and due to the prisms.

Probably less intrinsic contrast due to stray light management presumably being more complex and the design I expect being telecentric.

Greater expense due to the three dlp chips, the color splitting prisms and TIR prism and the larger lens optics and tighter manufacturing tollerance bieng needed.


I would sooner buy an expensive projectordesign dual lamp single chip dlp than a even more expensive three chip dlp. I would sooner see single chip dlp projectors built as good as they can be at lower prices than cheaper three chip dlps built to a lower price point. How many single chip dlp projectors these days have decent native on/off contrast and decent lenses with high mtf and ansi contrast. DLP single chip is capable of at least 8,000:1 native on/off contrast and over a 1,000:1 ansi checker board contrast ratio.
 

·
Registered
Joined
·
6,752 Posts
Discussion Starter · #20 ·
Comparing two equally well made DLPs, the by-far largest source of heat (according to TI) is the lamp, and the pj with twice as many equal or higher watt lamps will be creating more heat. The singlechip designs also catch a lot of heat at the wheel where a 3chip would be splitting that light-energy and putting it on screen. A 3chip design could alternatively use a much lower wattage lamp for darkroom builds. Both should be running a lot cooler to achieve their competing (full color) brightness than a similarly bright, well designed singlechip. And lamp life would increase..and in the case of the brighter solution, 5000hours down the road you'll be paying for one lamp instead of two. If you need more..there's no reason a 3chip can't use multiple lamps 3X more efficiently.


Compared to an RGBRGB cw (or a very low wattage 3chip), an RGBYCM wheel gains its 50% brightness advantage at the cost of it's color brightness losing 50% and both its calibrated brightness AND contrast (because calibration will lower it's white level without lowering black floor)..you'll lose roughly 67% of its brightness. An RGBCYW will lose 73% of its brightness once calibrated! If you choose NOT to match your color brightness, your mismatch will be 3:1-4:1 and besides live-action content looking wrong, animated content looks horrible. The calibrated 3chip has a 200% advantage over the uncalibrated CMY, a 300% over the RGB, and a 600% advantage over the calibrated CMY!


For bright situations where it's competing against a dual(triple?)-lamp dlp, the contrast will be of little value and wideopen iris will allow the full advantage of having equal brightness with 1/2-1/3 as much lampheat and no colorwheel waste heat. Running cooler, it will last longer (maybe this is why they aren't manufactured..) and when it does need a new lamp it'll be at 1/2-1/3 the cost. And there's no good reason for two extra DMDs, one extra lightpath, and the same amount of build alignment that goes into every other 3chip device to increase the price beyond dual-lampDLP prices.


For darkroom solutions, the lamp wattage will be balanced with iris needs to result in a good mix of low heat/noise, longer lamp/device life, better ansi contrast and equal or better on/off and alignment/ fill-ratio than LCD, cleaner 3D than any non-DLP with more glasses options, lower lag-times than any non-DLP, no motion-blur, no RBE, and no reason to cost any more than a midrange LCoS. Might as well also toss in the no filter-replacement and no dust-blobs.


No alignment or cost issues that aren't already faced by every other technology, and a possible 3x brightness advantage to combat iris needs for excellent contrast (Sony has/had a dinosaur 3chip getting 8000:1). I don't see why it wouldn't have most of the placement flexibility of a LCoS too.
 
1 - 20 of 48 Posts
Top