Marquee Modification Performance enhancement thread

Hi,

The raster centering function in the help menus is completely unconnected to the convergence stuff. For horizontal position it adds a bias voltage to the horizontal deflection coils (using U7, U8 and U11 on the HDM). The vertical position works similarly by adding a bias voltage to the inputs of the vertical deflection amplifiers (U700, U800 and U900). It is possible, that the same circuitry is involved when moving the entire raster in the convergence menu. You could verify this by measuring the outputs of U8 on the HDM (Pin 8->RED, 7->GREEN or 1->BLUE) while using the controls (horizontal movement). If the output voltages don't change, then this raster movement is done entirely through convergence controls/amplifiers/coils...

Regards,
barclay66

I am quite confident that the HDM and VDM is used both when convergence is set in the center point and also during the installation procedure.
The service manual only referring as the CVA duties to process the ramp based signal those originated on the VDM, there is nothing there that could be related for horizontal centering. Yet this is where one can play, since the drive signals from the HDM centering circuits could be rewired to work on the convergence amplifiers if that would be desirable, I just don't know if it would gain any yield.

Of course raster centering functions are needed! Nobody is saying they aren't. But Kurt seems to think that you HAVE to use the Thomson yokes in a Marquee to center up the raster correctly and that is simply not the case.

The Thomson yokes ARE the problem. Remove that problem and you remove the need for the feature built into those yokes.

The raster centering feature of Thomson yokes is there to compensate for the lousy design of those yokes.

The other features in the projector are more than adequate for raster centering if you put in GOOD yokes.

I'm getting tired of repeating this. It's like people don't understand my words.

You dont have to use the yokes, but when running high bandwidt, and using the static convergence the HDM temperature seems to be in constant movement and shifting around the raster so that you get convergence error. If you 0 the static/ center convergence and do it manually you get a much colder HDM and more stable operation, and using the static convergense to its extremes wich is needet for pushing out raster ringing, it generates a shaky image wich blurr the image slightly, to the point improved focus ability dont make much sense, its actually very obvious if you try do screenshots, a bit of unstability in the deflection will bluur your screenshot as well.

The G90 dont operate like this, it has much more refined electric controles, wich i bet is much more of its focus ability than the focus yoke itself.

So in the end it would make more sense to buy some of those 100-300$ G90 projectors sold all the time and try put Gjakys marquee videochain in one of those if possible. I guess that would eliminate the poor ansi contrast and streaking the G90 have as well.

And i still think your wrong regarding the Marquee yokes, the mecanical centering is needet due to the bad electronics in the marquee wich create raster ringing to a degree that you need to offset the raster to the side, and the troublesome electronic raster centering that creates more problems than it solves when used.
So when you put in a fixed yoke you remove the posibility to mecanically compensate, and force you to introduce the issues created by the electrictronic controles.

I would very much like to know how you deal with the raster ringing today.?

And that's just what I'm working on. And, yes, I got credit...I just never made one penny off it. Not that I put forth any effort to do that anyway.

The fact that the Marquee CAN achieve high sharpness indicates that there's nothing wrong with the focus circuitry. However, one thing the G90 brings to the table is its active CPC system, which takes the concept of dynamic astig one step further. The 909 also has that additional feature, as I understand it (Never owned a Barco so I'm going on what I THINK I know there) and I believe that's the prime reason why corner sharpness is probably better than on a Marquee with those units.

Unfortunately, re-engineering the Marquee to add dynamic 6 pole magnetics would be a truly non-trivial project involving (probably) an entirely new CLM design at the very minimum, and all the software development that goes with it. WAY above my pay grade.

Oddly, I've never noticed streaking on any of my four working G90s. I can't verify any streaking problem exists.

When I have my G90 package in full test I'll be testing a stock Marquee side by side with the modded one. So it should be a pretty easy comparison to make. The candidate to be modded is also currently in stock form and their performance is definitely about equal. That makes for a good baseline for the testing.

Hmm, you must have missed the part where I said that my usual signal chain is direct from BD player to Moome card. So, whatever that pixel clock is. Or, I may have the HDQ in the loop. The only thing that really brings to the table for me is its ability to output 72 Hz and provide 11 point greyscale calibration, which I haven't done. My light meter isn't the type that works with the Lumagen calibration system and it would be quite a pain to use it to do the 11 point calibration manually. I'm not even sure it CAN be done.

As for light output, I keep it on the low side because bright enough is bright enough. Say 7 ft-l at most.
It's bright enough at that point that if the screen goes to a full white scene, it can make me squint. Tell me why I'd ever need it brighter than that. Screen is 1.0 gain flat white painted. Cheap but effective. Yes, I've seen more expensive screens in use. No, I don't see spending the money. The difference wasn't worth the price tag to me.

Will Chris ever finish a project he posts about?