CurtPalme.com Home Theater Forum

greg9518lc · Joined: 19 Apr 2016 Posts: 360

Gjaky seems to be on to something that could totally
revolutionize the CRT projector and provide more options for Marquee
and the other brands. Looking forward to the evolution of Moomee
Straight to the neck boards.

Tim in Phoenix · Joined: 21 Oct 2006 Posts: 4378 Location: Phoenix

Hello

Video to the neckboards maybe, but what about SYNC?

cmjohnson · Joined: 03 Apr 2006 Posts: 5180 Location: Buried under G90s

What about OSD and menus? We kind of NEED a visible menu system.

I'm all for new developments, as long as they don't break functionality that is definitely needed.

Tim in Phoenix · Joined: 21 Oct 2006 Posts: 4378 Location: Phoenix

CM makes a good point

Test patterns are helpful too.

gjaky · Joined: 05 Jun 2010 Posts: 2790 Location: Budapest, Hungary

Yes, VIM would be still needed for sync processing, and you need an interface you plug your input cable or moome card anyway, and for those the VIM is already perfect. The point is the video signal would fully jump through the VIM.

Functionality would not be breaked here either: you get OSD and test patterns as well with this solution too, however true you'd need to pull a new batch of wires to each neckboards (those carry vertical blanking, OSD/testpattern and contrast signals)

Originally on the VIM the video path starts with the video matrix (input switching), then a 2:1 divider, shortly after it is amplified with gain of 2 (logical isn't it?) this is where the OSD/test patterns are injected too, then goes the signal to contrast controlling, and goes towards the VNB on the RCA-coax cables, now on the cable between the VIM and VNB further loss is happening in gain (about -25%). So there is a lot of amplification happening on the VIM (that introduces noise, and lowers the bandwidth) yet it has an actual gain less than 1. Also take note that the Marquee was originally designed for 1V video input and not for 0.7V (that was the defacto standard).

Splitting the signal path into more separate blocks will only result in bandwidth limitation and more noise. This is eliminated here with putting the preamplifier right to the VNB, but I think this is the less important feature here. The big deal is IMHO how the OSD and test pattern injection to the input signal was done, it uses an approach that could not be done on the VIM anyway. And does not contain any series pass element (like switch or multiplexer) on the high frequency path at all, yet the best possible decoupling between the input and OSD source is maintained, so no noise and odd lines can leak into the signal path in "video mode".

There is also an experimental gamma corection (for blue) implemented on the board.
The problem with the original gamma correction is that it is a series element, ie. they put an extra opamp on the signal chain that makes the blue signal path "longer" because of this it has different bandwidth. This is actually even worse since they achieved the correction curve with diodes in the feedback network of that amplifier. This is not very nice because the diodes switch on/off rather slowly (4-5ns) and the amplifier they used is a current feedback amplifier, as such the bandwidth of the amplifier is inherently changing with the feedback resistor value, on the VIM however they change the feedback resistor in the first hand, so the frequency response of that amplifier is constantly changing in itself depending on the video content.

My gamma solution does not introduce series element on the blue signal path either, and has a response time less than 1ns!

Also this opens more oportunity on modifications of other projectors like the Barco 909. Since this board paired with a VNB offers an all-in-one video path, extracting the OSD signals, blanking and contrast signal would be much easier than interface the VNB alone instead of the Barco 909 neckboard...

Tim in Phoenix · Joined: 21 Oct 2006 Posts: 4378 Location: Phoenix

Okay

That sounds well thought out, but is the VIM video chain that awful?

gjaky · Joined: 05 Jun 2010 Posts: 2790 Location: Budapest, Hungary

greg9518lc · Joined: 19 Apr 2016 Posts: 360

racerxnet · Joined: 20 Jun 2007 Posts: 362 Location: Illinois

The questions to ask now are; When will they be available as a full retro kit, and how much. Do you need a beta tester?

Mak

gjaky · Joined: 05 Jun 2010 Posts: 2790 Location: Budapest, Hungary

Greg:

I don't know how the latest MP boards are gain wise, but compared to a standard VIM, this board offers more gain (assuming the beam limit is not activated on the VNB).
A modified Moome card never hurts, but I'm not sure about the neckboards yet -as for performance mods concerned. As in my XG the retrofitted VNB neckboards (with the father of this preamplifier board) are quite much stock and has no problem with 1080P 72Hz. This will need to be tested.

Mak:

The business model is not laid down yet, but I guess it will be available as soon as possible, I have no intention to sit on this unless it needs more improvement... As for price, whatever price is asked by Mike Parker for a similar mod, this will be cheaper than that Wink

At the moment I have enough PCB to build two more sets of these boards, but I'd wait the in machine tests results first, and yes, I already have a beta tester.

mp20748 · Posted: Fri Aug 12, 2016 9:58 am Post subject:

My latest mini board 02 VIM has only a high bandwidth switch (for internal OSD /Source selection) and the AD835 contrast (gain) amplifier in the entire video chain.

The mini board connects directly to the slot 2 connector from its source input. No relays, pedestal circuits, etc. It is a single source input that uses only slot 2 (BNC's not used) for direct coupling to the Moome.

Makes for an entirely direct coupled (no caps) video chain from the Moome to the neck boards.

gjaky · Joined: 05 Jun 2010 Posts: 2790 Location: Budapest, Hungary

greg9518lc · Joined: 19 Apr 2016 Posts: 360

cmjohnson · Joined: 03 Apr 2006 Posts: 5180 Location: Buried under G90s

A point to make about gamma correction:

Though to the best of my knowledge it has not been addressed in great detail, the truth is that all three CRT colors
have non-linear gamma curves but generally gamma is applied oniy to the blue channel as the blue phosphor's gamma curve
is substantially different from red and green, which are more linear.

But no color phosphor is linear.

Using the published gamma curves for each phosphor color, it would be interesting, and theoretically beneficial, to create
custom gamma profiles for each color and apply them all rather than only on blue.

The document that shows the phosphor gamma curves is available on Chris STephens' website (xymox1) if you don't have it already.

How to implement gamma in each channel? Designing that circuit is above my paygrade. Do it digitally, do it in analog, I leave that to the engineer.

But doing it digitally would allow you to make one gamma correction circuit and program it to the color of your choice. Red profile, green profile, or blue profile. Maybe even set that profile choice with 2 onboard DIP switches.

Tim in Phoenix · Joined: 21 Oct 2006 Posts: 4378 Location: Phoenix

Wow

That sounds like a lot of work for products with only a few dozens of potential customers.

cmjohnson · Joined: 03 Apr 2006 Posts: 5180 Location: Buried under G90s

Some people don't mind if the audience is small, they just like to solve the challenge and create something that they themselves can use.

In what little spare time I have left I'm working on a circuit that allows a complete stack of G90 magnetics to be run in a Marquee.

Actually it's TWO circuits, one is basically a broadband impedance matching transformer between the deflection yoke and the HDM, and the other is a 12 channel control circuit so you can use G90 electronic CPC coils in a Marquee.

I won't make a penny off it. I'll still make at least one set of each.

**kal** · Posted: Sat Aug 13, 2016 3:12 pm Post subject:

gjaky · Joined: 05 Jun 2010 Posts: 2790 Location: Budapest, Hungary

**kal** · Posted: Sun Aug 14, 2016 5:52 pm Post subject:

garyfritz · Joined: 08 Apr 2006 Posts: 12026 Location: Fort Collins, CO