AVS Forum banner
Status
Not open for further replies.
1 - 15 of 15 Posts

·
Registered
Joined
·
19 Posts
Discussion Starter · #1 ·
I just got my NEC Lt150 yesterday and yes it's great. However, like everyone else, there is the occasional rainbow that pops up. My question is if it's possible to modify the lt150's color wheel and motor for the newer models. It seems like a waste to spend $5k for another DLP projector just for the color wheel and motor. For this mod, you'd need to get a hold of a 6 segment color wheel and a high speed motor. This website sells used 3 segment color wheels and motors for $50: http://www.lasersurplus.com/optics.htm

Wouldn't the newer color wheel and motor be available for dealer repairs?

So now supposing you get a replacement 6 segment wheel and motor, you'll need to modify the timing of the mirrors on the DMD. This is where the mod may fail. From this website, a user describes how to modify the gamma linearity for his Proxima 4200 via it's aux port: http://www.visionscience.com/mail/cvnet/1998/0172.html

If it was possible to get the code to modify the timing of the micro mirrors for the Lt150 (from some service manual or something) the biggest problem could be the DMD itself. I'm not sure what type of DMD the LT150 uses and if the mirrors for that particular model could be adjusted for the 6 segment color wheel speed. So what do you think--is this feasible?
 

·
Registered
Joined
·
13 Posts
I imagine some sharp individual could make a nice profit replacing the color wheels on current DLP projectors with 6 segment wheels.


------------------

Best regards,

Alan
 

·
Registered
Joined
·
1,270 Posts
The complexity of replacing the color wheel would most likely make such a modification unfeasible. For starters, the DMD requires another set of chips to drive the electronics of the DMD. Converting an RGB signal in to field sequential color using pulse width modulation at 60 fields a second is a difficult challenge. Each individual mirror must be switched on an off literally thousands of times a second. The Driver board would need to be modified to allow for a six segment color wheel so that the proper signals are generated for the DMD.
 

·
Registered
Joined
·
848 Posts
tech_junkyard wrote:
Quote:
So what do you think--is this feasible?
Oh, goody. Someone asked for an opinion...


I'm a once-and-sometimes software, electrical and optical engineer (and manager). I say, "Yes. It can be done," but it will be fairly tricky.


(The LT150 also can work without the white stripe of the color wheel, so what is it you hope to gain from changing the color wheel, btw?)



JoeFloyd wrote:
Quote:
The complexity of replacing the color wheel would most likely make such a modification unfeasible. For starters, the DMD requires another set of chips to drive the electronics of the DMD. Converting an RGB signal in to field sequential color using pulse width modulation at 60 fields a second is a difficult challenge. Each individual mirror must be switched on an off literally thousands of times a second. The Driver board would need to be modified to allow for a six segment color wheel so that the proper signals are generated for the DMD.
I basically disagree, but based on limited knowledge.


Yes, there is significant complexity. However, RGB "conversion" to field sequential color is merely taking the R, G, and B values and applying them sequentially to the same DMD device. The projector already does this, although it also has a fourth "W" (white) period. And it already steps through the four color-fields at 240 Hz (i.e., 60 Hz for all four colors).


The pulse-width modulation is probably not done "literally thousands of times per second". It's more likely that a given mirror is turned on at the start of the color field, then turned off some number of "clock-ticks" later. The number of "clock-ticks" per color field must be very high (to provide for many steps from black to full-intensity), but the number of DMD on-off cycles per color-field could be (and probably is) as low as one.


I also am not sure that the driver board needs modification unless it is also responsible for timing as well as signal levels. But I don't know exactly what the driver board does.


Now, tech_junkyard, if you give me the schematics and code, I could tell you whether Joe or I am right.


Or we can sit back and wait for someone to do it.


And while we're at it, let's modify it to have a variable color-field rate. That way, higher refresh rates can further minimize rainbows, right?


-yogaman


[This message has been edited by yogaman (edited 08-09-2001).]
 

·
Registered
Joined
·
1,201 Posts
The DMD is driven by a custom ASIC which isn't exactly tweakable. The newer DMD's that can utilize the faster rotating 6 segment wheels also have the needed higher bandwidth data loading spec. As I understood it, this is not just a process tweak on the DMD.


So the short answer is rip out the old and stuff in the new.



------------------

plug in to play
 

·
Registered
Joined
·
1,587 Posts
Quote:
I'm fairly positive that you couldn't use a 6 colour wheel on

any XGA DLP projector. The DMD has to support it and the

XGA DMD doesn't. I could be wrong though.


Jeff



PeteB (TI) said on the av guest forum board that the Dmds that could take the 6 segment wheel are:

The new 1280x720 (0.8),

The new 848x600DM (0.7),

The 1280x1024 (1.1),

The current 1024x768 (0.9),

The yet to arrive 1280x1024 (0.9),

and the yet to arrive 800x600 (0.5).



One question that interests me is how many manufactures make color wheels and what are the primaries (color coordinates) being used on the current and future color wheels ?


As to the upgade of wheels I can not see why these sort of changes can not be done. If the timing is done from the wheel and software is changed this can not be a complete impossible task ? Maybe it is on some machines but I imagine that others could be ?


We have a lot of smart guys on the forum so who wants to donate their LT-150 ? Thank my lucky stars I fail to see these rainbow effects on my Dlp so its a mute point to me. But it will be very interesting to see if someone can hotrod a Dlp with changes such as this.


DavidW




[This message has been edited by David Wallis (edited 08-10-2001).]
 

·
Registered
Joined
·
4,351 Posts
The problem is that it is not software driving the color sequential output, its an ASIC chip. The only way would be to replace that chip with a new one, and also hope they are pin-compatible.


Andy K.



------------------

"L T One-Fifty

Did your light guide the wisemen,

to baby Jesus?"
 

·
Registered
Joined
·
1,587 Posts
Andy

No drama. Would these be supplied by TI do you think or are we talking custom spec by the manufacturer.I just do not imagine that the manufacturer changes much in the basic DMD/color wheel timing relationship.

I guess I am interested in what Ti supplies and what is left up to each manufacturer in this important timing relationship between the wheel/dmd and its electonic drive system.


DavidW
 

·
Registered
Joined
·
19 Posts
Discussion Starter · #10 ·
There's some interesting information here. It seems that the lt150's DMD does support the newer wheels. But the main problem would be that the timing is not software controlled but hardware controlled by an ASIC controller chip. So much for the "easy" mod. This leaves 2 scenarios(?):

1) the ASIC chips are pin compatible and a simple swap would work (very doubtful)

2) replace the entire driver board that contains the new ASIC chip.

Would this work? I'm no EE and am probably looking at the problem too simplistically. I think it'd be interesting to hear from someone with detailed service manual/schematics and access to these replacement parts. Of course this whole thing could be folly, but with many people with multiple lt150s it seems like something that could be investigated on a "spare" unit. And like Alan Ross said, imagine the profit that could be made!
 

·
Registered
Joined
·
3,653 Posts
The thing that tells me that this mod can't be done is that Seleco took the time and effort to change the colour wheel in the HT250 but didn't use the 6 colour wheel that they already use in the HT200DM.


Jeff
 

·
Registered
Joined
·
1,587 Posts
Quote:


The thing that tells me that this mod can't be done is that

Seleco took the time and effort to change the colour wheel in

the HT250 but didn't use the 6 colour wheel that they

already use in the HT200DM.


Jeff
Thats because it uses the Xga 0.7" Dmd which doesn't support the 6 segment color wheel. The Xga 0.9" Dmd doe's according to TI.


This sounds like a THUMPER mod.

He loves this stuff.


DavidW


[This message has been edited by David Wallis (edited 08-10-2001).]
 

·
Registered
Joined
·
1,270 Posts
Hi Yogaman,


There are a number of papers on the TI web site describing the DLP systems. One of particular interest is a paper IEEEIR.pdf that describes the reliability testing of the basic DMD designs. The normal switching period of an individual DMD is 200 microseconds which yields 5000 switching events per second. One neat little piece of trivia is that in their reliability testing they drove a DMD device at 50,000 cycles per second to accelerate aging. So, faster color wheels should not be a problem in terms of the actual DMD, but is a problem of how to integrate the new color wheel into the DMD driver electronics. There are some fairly tight timing tolerances in addition to the extra processing that would be required to decode each video frame into multiple sequential color frames that have each pixel pulse with modulated to achieve 24 bit color resolution.


The new color wheels that are Archimedes spirals will actually have all three colors illuminate the DMD simultaneously. The trick with this system is that the individual colors scroll across the DMD sequentially and the driver electronics must tracks what part of the DMD is being illuminated by which color. The advantage is that the color wheel can rotate much more slowly and the frequency of color transitions for a given pixel will be much higher than current single DMD DLP projectors. Unfortunately the transition from one color to the next on the surface of the DMD while being illuminated by spiral color wheel is not a 100% perfect transition. This indicates that the primary disadvantage is that it much more difficult to blend the individual color channels since there is an optical aspect to the equation that isn't there with field sequential color.
 

·
Registered
Joined
·
1,587 Posts
Quote:
Originally posted by David Wallis:

Pete

You mention increasing speed on the color

wheel via increased RGB area (double color

sections).

Is there a limitation on the reaction time of

the mirrors and life expectancy ?

Can the existing DMDs run with a double color

section or is there a limitation that requires a

new DMD along the lines of the new 480P

DMD ?


DavidW



David,


This is indepenent of DMD lifetime.

The data bandwidth into the DMD limits the

effective color

refreash rate, so new DMD's like the 480P and HD1

are

required. The have more rows available, thus the

ability to

load more data faster. The 0.9 XGA can do this as

well.


Pete


------------------

Peter J. Broas

Marketing Manager

Home Entertainment

DLP(tm) Products

Texas Instruments Incorporated

[email protected]

Text
 

·
Registered
Joined
·
848 Posts
JoeFloyd:


Quibbling with your numbers, the switching time listed in a ti whitepaper must be significantly less than 21 microseconds to achieve 256 relatively uniform 0levels per color in 5.33 milliseconds per color field for 60 Hz refresh. See Section 3.3.


And as that paper points out, the optical switching time is 2 microseconds.


I now understand, though, that the real problem is the driver chips.


I'm still looking for information on those driver chips. Peter, can you help?


-yogaman
 
1 - 15 of 15 Posts
Status
Not open for further replies.
Top