CurtPalme.com Home Theater Forum

[cmjohnson]

Mike, your input has caused me to alter my direction. I WILL take your word for it, and am glad to NOT have to spend a lot of money on expensive parts that probably don't represent the best way to improve the circuit anyway.

I'd need to actually know the frequency characteristics of the noise on the DC rails to figure out proper filter circuit parameters for it,
but I figure your values are probably a really sensible place to start.

I have ready access to a rather impressive assortment of dirt cheap inductors ranging from tiny up to toroidals that are the size of a donut. Give me some recommended values and I'll go shopping.

[mp20748]

[Nashou66]

Chris, use Film caps in the HVPS.


These are 20uf 700 VDC clarity caps from Partsconnexion.







I talked to Chris Stevens and he said they would actually be better in that circuit than Wets. But back then they didn't have them small enough to fit. They were the size of a #10 ketchup can



I have wet tants on my CVA's and VDM's But you have to make sure they are in good working order. many Wets you find cheap are old and can fail. Id make a circuit to run some voltage through them for a few days before using them.

I have had a few leak out.

Chris said for some reason the best results he had with wets was the 6 or is it 8 6.3 volt caps on the HDM. He went with 1500 uf 6.3 volt variants there . I think originals are 3300uf .

And if you do use wet tantalums on the HVPS it is 120uf 100v x 5. That line is the 390 volt line from the LVPS.

The benefit he said was in the transient response from Black to White and white to black plus some noise filtering.

I have a pic of the wet tantalums floating around somewhere. He asked me not to post it but its all well known now.


he also said he wanted to do the LVPS in the same manner but the UL listing/ ratings wouldn't allow it and it had to pass those standards.


Nashou

[mp20748]

The 390 volt supply there is coming from the sub switching supply in the LVPS. It is the same supply that supplies all of the switching supplies in the LVPS (+/-85, +/-24, +/-15, +/-5, filament).

That section in the HVPS is more a decoupling network than a filter for the 390 to the HVPS. Its purpose is more to keep the HVPS isolated from the LVPS and more to prevent flyback noise from getting into the LVPS. No filter is really needed for the HVPS, because the 390 is already well filtered in the LVPS.

But for sure loading down those two caps with what's shown should no doubt help block anything from going either way. But still not sure why anyone would put very large values in a decoupling network, where its purpose seems to be for filtering a particular frequency area of concern.

[redfox001]

What you guys have not discussed yet is the image quality that comes with simplicity. Like simplicity in the video chain and I have to praise Mike for that. Take away all the processing all the scalers etc. But if you can not get it simple you have to do something else.

[tschaeikaei]

Hi, sorry for not reading your -very interesting- text before, i was in a hurry and had to do just the same you said you do: save my thoughts somewhere
I've read the whole thing now and there are still some questions. And some statements (i hope they're not offensive).
And yes, i've written some, about 50 lines about caps and stuff and since it was work, i'm not going to delete it.
But i'm glad you changed your mind about it and save the money


Resolution:

[cmjohnson]

Just a few things in general: I'm going to follow Mike's lead (and very generous offers of assistance) and pursue a more calculated program of filtering noise rather than just throwing expensive capacitors at a problem that I haven't even characterized.

Suddenly I find that I regret selling my excellent 400 MHz Tektronix 2465B oscilloscope. It would be useful for scoping out noise.

No problem, I have another that I can refurbish after recapping the power supply, and put it into service.

When redfox001 first said "vga" then I was really thinking of VGA resolution, which is just 640x480. But the general idea of extracting high resolution signals from a VGA connector via a pigtail or adapter is of course completely valid.

The catch to that is that I am not aware of any VGA source that goes beyond 2K x 2K. Nor any VGA equipped video card that goes beyond QXGA. (2048x1536). If such a beast exists, and you know about it, PLEASE give me the make and model number.

The rule I'm seeing is that all HD cards use DVI or HDMI, and if they have a VGA connector, it's limited to non-HD specific resolutions. Then UHD cards typically ONLY have HDMI.


Now as for Mike's point about limited good ground points on the Marquee chassis...

I expect to more or less totally disassemble the chassis in the course of the work.

There's no reason at all why I couldn't add some copper sheeting in strategic areas, thus greatly improving the quality of
the chassis as a ground point. At the very least, run copper tape or straps (screwed down) to all the factory ground points
in a star ground pattern.

I have some conductive shielding paint as well. That might be quite useful.

I wish there was an easy way to build an LVPS output filter system that could fit in the space between the HVPS and the LVPS,
but that'd be a substantial bit of engineering.

[redfox001]

He I have mods on the vertical, convergence and astig board (all boards on the back) and also on the focusboard, vim and neckboards. Together they lower the noise to not visible on the screen for me. Perhaps the noise originates from those boards and is travelling between them? I did nothing on lvps and hv. I think they deliver filtered power allready. Barco 909 also has thse inductor filters on astig, focus and convergence.

[Nashou66]

Here is one of Chris's Email's

Interesting guy I must say. His Engineering back ground is evident. But He is an eccentric chap.

[cmjohnson]

Yeah, Chris is a bit out there, but unquestionably he's brilliant and was thinking of things a decade ago that most of us haven't considered even today.

He looks at a Marquee the way a race car builder would look at any given brand new car off the lot that has some performance potential. "I think I can keep the radiator cap stock. The rest has to be replaced or modified if you want to win on race day."

I find myself wondering what we might have ended up with if the old quantum phosphor project had come to fruition. Theoretically, it could have resulted in flat panel CRTs suitable for projection use, with huge brightness ability and very high resolution capacity, with phosphor life that's orders of magnitude better than regular phosphors.

Today, we're starting to see the beginning of the usage of quantum dots. I wonder if there is any relationship between quantum dot principles and the quantum phosphor principles?


There was a time when I was thinking about how phosphors brown (burn) in a tube and it's oxidation between the binder chemicals and the chemicals in the phosphor, which occur as a function of time and temperature. It had occured to me that if there was a way to create a phosphor and binder system that had no oxidizers in it (Oxygen, chlorine, bromine, fluorine) then you would have a phosphor that can't burn as there'd be no way to complete the burning chemical reaction. That would have been a nice innovation. Permanent phosphor, never burning or browning.

I have found an HDCP removal device on amazon for just 25 bucks, and it's the HDMI 2.0, 4K capable version. So with THAT out of the way, all I would need would be a really kick-ass software engineer to contribute (for free, of course ) some programming time, to program up an FPGA to derive RGB or component signals from the datastream.

It may be a dead end road, but I'm enjoying the drive.

[cmjohnson]

Part of my plan (as it sits now) is to retrofit Barco magnetics into the Marquee chassis. This will be the full magnetics stack out of a Cine 9/909. Reason: Room on the tube neck. The Marquee deflection yoke and convergence coil take up more room on the tube neck than the Barco units do, and since I'll be running the same focus yoke you find on a Cine 9 (after any required mods, of course), and these are longer focus yokes, they're far more compatible with the Barco deflection and convergence yokes from a physical standpoint

I'll be able to put the focus yokes in their optimal position this way. This is going to be important so it's worth the trouble to try to adapt different deflection and convergence yokes to the Marquee chassis.

What I don't yet know is whether or not the deflection and convergence yokes will work properly with Marquee circuitry with only connector replacement and rewiring as needed.

Hopefully, they will.

[Nashou66]

Measure the indictance of both coils to see how close they are. Also the resistance. Remember what Chris said about the yolks:

Yokes are like speaker drivers. Woofers. Vert. Tweeters. Horiz. Lots of watts and pretty complex impedance. All that filter network stuff is really just like speaker crossover components. In fact, maybe it would be good to use better wire for the yoke runs.


Good luck

Nashou

[redfox001]

I know that those yokes have to be put further forward but I also had the 1292 yokes and did not try any more. If you can show how to do this I will put them on my Marquee too. I also foudn out that Mike modifies a resistor vallue on the focus board to get more adjustment range. Never tried it though. See if i can find it.

[cmjohnson]

I've started making notes and adding them to a file just to plan all this out! I can't rely on MY memory to retain everything, that's for sure.

I guess I should start by going on the hunt for a low cost MTF measurement rig. I will need hard data to determine the effect of any mods made, and as they'll be focused on reducing noise, as Chris Stephens pointed out, noise in the beam control circuits (deflection and focus, and convergence, too) will manifest in a larger spot size, affecting MTF readings. Noise in the video path is more obvious but noise in the deflection systems WILL have an effect on spot size, resolution, and MTF.

Ultimately, the CRT displays intensity and time. We want to quiet it down in BOTH domains.

I'm going to be looking at EVERYTHING. I seem to remember reading "the stig amps are noisy" somewhere, so yes, even
astig and convergence amps will be looked at for unwanted noise and quieted down if I can figure out how to do it.

[Nashou66]

[cmjohnson]

Thanks for finding that topic. I've bookmarked it for future reference.

For setup and adjustment of flare, astig, and triangularity, plus ultimate focus adjustments, I'm going to be using a
camera on a tripod, with its HDMI output feeding to a monitor, so I get a close-up view of the screen while I'm back at the projector. This method should be very good. For a camera I use my Nikon D800 or my D810, both are 36 megapixels with HDMI outputs.

I am looking forward to kicking off this project just as soon as the magnetics I bought from Redfox001 (Thanks, Red!) arrive.
I'll be picking up Projector no. 2 at the home of its former owner (a friend of mine) in the next day or two. Ironically, I helped him to buy it a few years ago. Loaned him the cash to get it and we took a road trip to pick it up.

[cmjohnson]

Parts are starting to come in. Projector 2 is here, finally, and I'd forgotten that it already has both red and green C elements so I have those to swap over to the HR tubes when the time comes for me to do that.

Once the Barco magnetics arrive, I'll start with a baseline evaluation of Projector 2 in its stock form. Then, rip it open, yank out the tube assemblies, put the HR assemblies in place with the stock Thomson magnetics, set it up, and get some basic performance data on it.

Then, swap to the Barco magnetics, performing all needed wiring adaptations to make it all work, and then set them up and see if things look any sharper.

Then begins what I expect to be a long process of iterative tweaking and adjusting. And circuit mods for noise reduction, one board at a time.

This won't be a weekend project. I expect it to divert some of my time on a regular basis for months to come.

The plan is also to fully upgrade Projector 2 to full Ultra standards from top to bottom. But an Ultra HR with additional performance mods.

[cmjohnson]

Magnetics are here, and I just scored a quantumdata test generator for 99 cents on ebay. Even after shipping it's quite a score!

This evening I want to start metering out the Barco magnetics and comparing their values to Marquee magnetics, and document them, of course, so as to have a good baseline for comparison.

If they're within 20 percent then I should be able to just swap the connectors and run them.

[cmjohnson]

I'm searching out as much data as possible on the Barco KF-3212 focus yokes so hopefully I won't have to open one up and disturb its internal alignment. Plus I don't want to break one.


Its extra connections are for an extra pair of 6 pole dynamic astig windings.

I believe i will be able to connect the Marquee astig circuit to EITHER the 4 pole windings or the 6 pole windings, but obviously, I won't be able to use both at once like a Cine 9 or 909 can.

That should be OK. I don't anticipate that making a large difference in what I can get from these tubes.

The idea of using the Barco deflection and convergence coils to allow for proper location on the tube neck for the Barco focus yoke seems to be workable, but I will have to modify the Barco yoke beyond simply changing the connectors.

The Marquee deflection yokes have a scan fail detection coil added to them, but it's glued in around the rim of the yoke and it appears to be a stand-alone coil system that can be lifted off the deflection yoke as an assembly and then glued down on the Barco deflection yoke. I'll just have to heat up the glue to extract it, then glue it back down on the new coil in a suitable location.

[cmjohnson]

I may not be right about the scan fail coil, I need to revisit the manual and read about the scan fail detection again.


The Barco deflection coils are more different than I expected. Their coil inductance values are roughly half that of the Thomson deflection yokes (regular Marquee variety) and the Thomson yokes are actually configured with two separate H coils which are connected either in series OR in parallel by relay switching on the HDM, depending on scan rate.

The Barco coils don't have this feature. While there are two H coils they are connected together in series, and that's hard wired. I will have to unsolder that connection and add wires to make the two coils indepenent and make them compatible with the HDM's system. But I am not sure yet if I'm going to need to modify the HDM to make it work right.

Fortunately I do have a "sacrificial lamb", an 8" LC tube assembly that can be used to test and validate all experimental changes made to any magnetics and circuits. If it gets a spot burn, nobody cares.

I came up with an idea just for testing the idea of what would happen if you were to actually have a way to run both the 4 pole and the 6 pole astig windings in these Cine 9 yokes.

Since my basic tests will be performed on a SINGLE CRT assembly on a test chassis, I'll test using the green channel. And I will be free to utilize the astig control off of either the red or blue circuits to drive the second astig winding on the Cine 9 yoke on the tube under evaluation.

Of course, that's not going to work with all three tubes installed, but I will be able to at least simulate what might happen if the Marquee could drive two astig systems.

I've looked at the Marquee's control architecture and I am rather disappointed that it wasnt engineered to control the individual boards via serial data over an I2S bus. If it did have control architecture like that, then actually adding a new secondary astig control system would be electrically and mechanically simple, and require modified firmware on the CLM. But since the CLM doesn't work like that, to add a secondary astig system to it would be a much more involved process that would still require new firmware and the addition of new hardware, AND require a significant physical mod to the CLM.