CurtPalme.com Home Theater Forum

[cmjohnson]

Nope, the Thomson focus yokes ARE inferior, the only reason they're in a Marquee is because Electrohome determined that they were "good enough" to meet the simulator resolution requirements of their military and commercial simulatior customers, and they did not feel it necessary to invest the money in re-engineering the platform to run K-D focus magnetics.

The difference between focus performance of the Thomson yokes and a G90 yoke in a modified machine is NOT SMALL.

I have done enough of these experiments to prove quite clearly, to my complete satisfaction, that the Thomson yokes are inferior,
and in fact they are the worst ones found in any projector you ever heard of. They'e just barely better than JUNK.

In actual fact a G90 focus yoke is PERFECT for a stock Marquee. But for ONE problem: The stock deflection/convergence yoke assembly is too long and doesn't allow the G90 focus yoke to be placed where it needs to be. It's too long.

But G90 deflection yokes are shorter and allow for proper focus yoke placement, which is obvious as that's how they're installed on a G90.

But there are electrical compatibility issues with the deflection yoke. A problem that I BELIEVE I'm just one more experiment away from solving, as I have already been able to run ONE G90 deflection yoke in a Marquee after changing a single component value on the HDM. With that G90 deflection yoke and focus yoke running on that tube, its performance was fully equal to the focus performance and stability of a G90's focus system. It was RIGHT, possibly the first time ever that a Marquee has focused CORRECTLY.

[mp20748]

[racerxnet]

But Mike,

Others are getting 65fl on a 110" screen with select neck cards. And your colors are horrid! Gasp, its not calibrated! Hope all is well and thanks for your efforts in moving the CRT community forward.

MAK

[racerxnet]

[gjaky]

[racerxnet]

[mp20748]

gjaky if you look back through the more than 15 years these mods have been on the table, you will find that I have been replacing those same three resistors, based on Scott at VDC recommendation, to include various other changes also being suggested by the more than 10 engineers who also contributed over the past 15 years or so. One of the reasons not many of the old-heads here are not excited every-time someone claims to have made a discovery, is because of the history involved with these projectors and my mods and others mods.

And if you were involved during that same time period, you would also know some of the history from the original design. to include what was going on when Scott was doing a re-design on the video chain of the projector. First let me inform you that Scott was formally an engineer for another CRT projector manufacturer before joining VDC. He also did a complete re-design of a VIM for the Marquee. I was also privileged to be invited to VDC and got to look at his design and what was in the plans. And has since had on-going contact concerning this and has done well under his advice and others who were very familiar wit the original design.

So, the focus fix you mention, sounds really good until you think about it only being a problem on the LUG tube. And we looked at that years ago, with things then moving to the video chain where the problem really is. It's just more obvious on the LUG because of its unique design. The resistor change helps, but its still minor in comparison to really solving the problem in the video chain.

I would just like to say for you and those who get excited about the changes get to search or inquire first on this forum and AVS as well, and there I'm sure you may find where some of this has already been on the table. An with that in mind, that the constant "our mods is better than yours" would simply stop. It boggles my mind how anyone can look at a screen shot and confidently state that what they are doing is better than the next person.


Oh, I forgot to mention and Tim and confirm that I've also been directly working with and for a former Electrohome engineer for the past 15 years.

[gjaky]

[cmjohnson]

Is it really THAT hard to keep this away from turning into a weenie measuring contest, guys?

I could be wrong but I think that those of us who are following BOTH of you gentlemen's work with interest
would like you both to at least PRETEND to exercise a modicum of mutual
professional respect and basic civility.

[racerxnet]

[mp20748]

I hope I haven't been a littler over reactive to Gjaky, and forgot that in spite of his affiliation with those who bash my work constantly on other forums (so I'm constantly told). I should be careful as he hasn't really come across as joining the nonsense.

And his willingness to share something that may be helpful and should have been appreciated, if useful to me or not.


I had also made some changes to the FGM board, but mine was mostly stripping noise from the power rails on the board, to tighten the focus by removing the LVPS switching pulses/noise (high frequency) from distorting the final stages (mostly) of the FGM. That required inductors (large) more than capacitors. I'll try and post a picture later on this.

The changes we made was a different resistor than what he suggested, but I understand what the changes he recommend would do.

The same applies really to the neck boards, in that they were not designed for 1920X1080P 60 or 70hz, and required changes to make them work better with the LUG's as well make them better overall. And that made for a cooler operating board while they are also far more capable of light output. So much so, that the Beam Limit let's you know you're driving the tubes too hard (never really worked before this). And still they maintain sharpness without blooming and are very stable. They can get warm, but when they do you're really driving them. I also provide taping to seal the tube housing for better fan flow to the neck boards. And as both who have received this version so far has indicated by one saying the upgrade from even a version I had sent back to me a little over a month or so earlier, that the later version was SUPER Sharp, and this is with LUG's. My other customer also spoke about the sharpness improvement over an earlier version and how much of an improvement the sharpness was over the previous version, and that his camera is not capturing the sharpness. So we're not seeing a sharpness issue at all. And I'm not expecting or seeing my camera to also capture what I see on the screen.

[mp20748]

I'll start posting a few shots to help in understanding what the real goal should be with mods.


Most are aware of and use the SMPTE pattern as a reference for proper resolution, or the Rule for when proper resolution is achieved. True that it can be used as a determiner, but it's far from being the absolute rule. It could be if all video info was at the same 100IRE level that the one pixel lines are, but since in an actual video image or scene, there is many other 'smaller' and less than 100IRE data that would require a very different video standard to properly reproduce. And this is when you see where i use the word "Speed" and it's its not really all about Speed and bandwidth only. There are several other elements that help make up the true High Definition image.

I'll start out using a shot that I posted on the previous page here (below). In that shot you'll see an image that nearly has almost perfect Clarity, from the foreground to the backgrounds. That means that even things in the background are as easy to discern and those things up front in the image. In order to make that happen, you'll need High Bandwidth and Fast video chain... OK, let me go a little deeper here...High Bandwidth or the required bandwidth would make for a perfect SMPTE pattern. But where "Speed" come in to play here, is when you'll need a video chain that's fast enough that the rise/fall rate is best performance. And that allows for an nearly to perfect pixel performance in an Analog Video chain. In short, it also means that the faster the square wave or Fine objects in the image have less rise/fall tracing/lag, which by-the-way is also a distortion, the Clearer things will be in the backgrounds. Or let me try another way of saying this... With a fast video chain, and that is what Gjaky sometimes also post when he post his supporting shots when he test a circuit. He shows the square or object of testing and also shows the get-there and recovery-timing (rise/fall) of the signal being used for testing. That allows you to see how well the signal will maintain it's original structure while going through that video chain or device.

And depending on how well the rise/fall performance, the better you'll be able to get those things in the background to have the same clarity as the stuff in the foreground. And that is when you really get into the best in HD. Unlike Standard Definition an High Definition Image literally means it's capable of also bringing out the detail in the backgrounds, whereas SD (Standard Definition) really only brings out the foreground as being detailed, because it lacks the ability to show the backgrounds being Properly Resolved.

Now before I forget, I've also mentioned the Absolute Black thing. This is also very important with an analog video chain, especially one that claims its HD (High Definition) ready or capable. In the video chain itself, Black is the equivalent of zero, or the lowest point in the video sweep. And the reason why Black is so important because Black (zero) is also the reference for everything of quality in the image. You'll also need a good zero or black reference point to get good colors. Ok, I'll leave that one alone for now and get back to Black being a reference for quality only for now.

So when you fire up your analog video chain, and if it's a True HD video chain, it has the awesome responsibility of maintaining that same 'black reference' that's on the source all the way through that video chain..and here is when it gets complicated. If the video chain is missing doing three things well, that black/zero reference will get lost in the video chain, and not only will you loose black reference, your colors and best background detail will also get lost, or will be diminished.

When that happens, you may see a sharp image (foreground), but if you're not able to see Clarity in the 'backgrounds' it's not really an HD image. And when an HD image is really good is when you'll able to see Equal or good Foreground Detail/Clarity over the entire image, and you would also be able to achieve "Absolute Black" - remember three things are required for this and that's two things beyond the SMPTE. Also when the three things are properly in place and when absolute black is really there, you'll also notice more and better pastel or subtle colors. Because without aa good black reference the finer or pastel, more complex colors will most likely appear muted.

Btw, this is all solved in my latest mods. And once I get the Green LUG installed, I'll crank things up here a bit in The Mod Factory and you'll be more able to see things on the screen here..


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

The below shot should do well in showing the clarity I spoke of here. But note that it could be so much better and a careful eye would show that the green is blooming and that limits things, yet it's still a good representation of an HD image.

[cmjohnson]

Chris Stephens had a great way of explaining the role of the entire video chain, all the way from the power supply forward to the CRT phosphor layer, which makes you really understand how dynamic, precise, and stable it has to be.

If you are displaying a 1080p- signal with a 72 Hz refresh rate, it works out that you need 180 MHz of clean video bandwidth.

Or, to put it another way, the system has to be able to respond to changes in video signal level and it only has about 5.55 NANOseconds (billionths of a second) to do it for a single pixel. It not only has to illuminate that pixel (spot) but it has to do it accurately, whether it's 0.00001 IRE or 100 IRE at whatever your target ft-l level is. It can't overshoot, ring, or clip. It has to hit the right output level every single time and it has 5.5 billionths of a second to do it, for every single pixel. And it has to also be able to hit a specific level and maintain it for some period of time without the voltage sagging or surging. Trace a full white scan line, go into the sync/retrace time, do it again, repeatedly, and never let that white scan line have any variance in it.

The power supply system, which is very much part and parcel of the video chain and NO less important than the neck card performance, has to be able to drive a steady full white screen with stability just as much as it has to be able to deliver the explosive instantaneous power required to light up a single pixel to full output in the middle of a dark field, without failing in any way to track the signal exactly and provide all the power it needs whether it's a full white field or just one pixel or anywhere in between.

Anyone who is smart enough to successfully modify the video chain and improve its factory spec performance is surely worthy of my respect.

[mp20748]

[cmjohnson]

All very true words, Mike. Image quality has not been the driving force in projection, other than meeting the usability goals of the end users. Which, for CRT projectors, were always military and commercial flight simulation systems.

Sony put SOME effort into the G90 in terms of getting a better picture. And I would say they succeeded, as the image out of a stock G90 IS better in some ways than the image out of a stock 9500LC. But it's not perfect, it still has some video noise to work out and of course you know this better than I do.

The G90, at a retali price of 38,000 dollars a unit, was produced in numbers only a bit above 1000 units. I am sure that Marquee production was something beyond 10,000, don't ask me how much more.

And then Barco actually built a specialty version 909, actually two of them, the Cine 9 and the CineMAX, which were actually
marketed to the home theater crowd and supposedly have fewer noise issues than any other CRT projector or so I have been given to believe. Price: I've heard something around 60K for a CineMAX. Not sure if that's true.

I'd be surprised if even 500 CineMax units were made. And I'd like to know what circuitry improvements they had over a standard 909.

Cost always drives the product's design. This holds true in all markets.

So it's fair to say that all products are compromised in their design and execution by cost constraints, to some level.

I

[jbmeyer13]

[mp20748]

[mp20748]

I had to get back into the Mod Factory for the past week or so, because I've been upgrading some boards I've sold going back over a year. This was a Free Upgrade that I had offered, because I wasn't pleased with a series of boards that I sold. So the upgrade was mention during the sale, and so far I'm finishing up the last of those boards. They meet my expectations well. Oh what a difference the proper Test Jig has made..

Now, this same set will need the LVPS also done to better compliment the video chain. If you notice in the shots below, even with the camera in the process. The low end and black is showing very well. You'll definitely need the LVPS also done to get this and much more.

And to add, I'm now also doing the Control Module. These both are substantial upgrades that easily show the difference after being done. But this may not be necessary for everyone. Only the very picky, because I don't include the LVPS and Control module in the cost for the video chain.




This shot was a quick-and-dirty, and that can be seen in the convergence and focus not being the best. But that will be getting better next month when I fire this baby up fully calibrated..


I can crank my contrast to 100 and things will remain tight with near to zero blooming...That's when you know you got the chain right..

Oh, from now on keep your eyes on RED in my shots and know that I I'm not using a RED C Element..

[mp20748]

Gotta get that Green LUG installed. The Green LCP presently in there is blurry again.

[mp20748]

The green LCP is getting worse. I took a shot to show what the convergence grids look like comparing the green LCP to the red LUG and the difference is very easy to see, however it did not turn out well.

Anyway, a few more showing the green out of control, but the good news is, I should it it out of there hopefully first part of next month..