|
|
Post by radiohead314 on Mar 3, 2016 23:43:35 GMT -5
So to make things clear. Standard moome. Standard vim with trimpots. Standard new style neckboards. I did not expect this performance at all. Very surprised. I might still find some bad points when watching movie material or judging noise. If I overlooked something please say so. If I need to test something also say so. I want to know the truth on these and I am going to scope the hell out of them  |
|
nashou
Unmoderated Off Topic
Tech in Training.....
Posts: 1,239 
|
Post by nashou on Mar 4, 2016 13:12:56 GMT -5
Thats what I have been saying for years, the only thing bad on some of the new PJ's is the astig and CMM boards. Some might need work others might not.
Looks like you got a great specimen . A piece of electronics is only as good as all its parts. One out of spec part in a critical location can cause issues in the image.
I like the trim pots, especially if your going to use the same resolution all the time. If not then Its not too hard to have a second Vim for the other resolution and swap it out
f its really necessary.
Nashou
|
|
|
|
Post by radiohead314 on Mar 4, 2016 14:15:14 GMT -5
Ok perhaps with others the pots where adjusted wrong and without a scope it is difficult to get them right. Taking them off will result in a not so good bandwidth. Perhaps that is why they got a negative name.
|
|
|
|
Post by stridsvognen on Jun 7, 2017 9:03:33 GMT -5
The problem is that movie content is mixed resolution, and the trimpots make a very non linear responce between resolution and gain. Non of us sit looking at 1:1 on off vertical lines all night. So the peaking cant be trimmed for a specifik resolution like 1080P or 720P unless its a very gentle peaking/ trimming, like some places on the neckboards, the vim trimpots are verry agressive, try out the testpattern addet at different peaking settings, its made specifikly for that purpose. The peaking pots are put on the input buffer/ front end, but should be on the output end for a more gentle peaking, as agressive peaking before the bottle neck is a mess. If you look at peaking, you might be able to set it on a pattern at a given level, but if you change the level, you need to readjust the peaking, peaking dont work well at low levels. Everybody can make a miniboard mod on the vim, bypassing the input stage and old switcher, using the ADG1233, if you terminate the internal patterngenerator/ OSD right on that one, you will also be totally free of the CLM noise and vertical lines, even running 0-255 / 0-0,7V from a moome card, where you normaly will have to run the 16-235 setting to elevate the black reference, and then turn down brightness, and clip out some of the low level noise.  |
|
|
|
Post by gjaky on Jun 7, 2017 15:39:56 GMT -5
Everybody can make a miniboard mod on the vim, bypassing the input stage and old switcher, using the ADG1233, if you terminate the internal patterngenerator/ OSD right on that one, you will also be totally free of the CLM noise and vertical lines, even running 0-255 / 0-0,7V from a moome card, where you normaly will have to run the 16-235 setting to elevate the black reference, and then turn down brightness, and clip out some of the low level noise. I'd even say that it would be also possible to use the concept of my VNB-DB right on the VIM, this would also need a miniboard but would result in a signal path without switch. |
|
|
|
Post by stridsvognen on Jun 7, 2017 23:21:05 GMT -5
Everybody can make a miniboard mod on the vim, bypassing the input stage and old switcher, using the ADG1233, if you terminate the internal patterngenerator/ OSD right on that one, you will also be totally free of the CLM noise and vertical lines, even running 0-255 / 0-0,7V from a moome card, where you normaly will have to run the 16-235 setting to elevate the black reference, and then turn down brightness, and clip out some of the low level noise. I'd even say that it would be also possible to use the concept of my VNB-DB right on the VIM, this would also need a miniboard but would result in a signal path without switch. That should be the ultimate solution.. |
|
|
|
Post by gjaky on Jun 8, 2017 0:31:05 GMT -5
I'd even say that it would be also possible to use the concept of my VNB-DB right on the VIM, this would also need a miniboard but would result in a signal path without switch. That should be the ultimate solution.. While this would be easier to perform I think the VNB-DB still offer advantages those can't be made on a VIM solution. |
|
|
|
Post by thelasttubewarrior on Jan 17, 2018 23:32:59 GMT -5
Some of you may be interested to know that while my supply of new green LUG tubes is very, very short, I have a good number of brand new PT22 P43 fast phosphor greens available. While they will limit your light output somewhat, they're still brand new greens and if you are dealing with worn tubes with pentagonal patterns from the TOFT simulators they came from (L3 Communications TOFT sim based on their Simusphere platform) then these greens are certainly a lot better than the worn tubes.
I also have a very good supply of brand new or new condition red LUGs. I don't have many blues, unfortunately.
Contact me if you're interested.
|
|
|
|
Post by gjaky on Jan 18, 2018 2:21:55 GMT -5
For the P43 it's not just the light output, but they also have different spectra than "slow green" making it impossible to properly calibrate in critical applications (=HT).
|
|
|
|
Post by thelasttubewarrior on Jan 18, 2018 10:46:09 GMT -5
Nonetheless, I have them. I think anybody who loves movies could tolerate a not-perfectly-calibrated image out of a new green P43 tube if there simply aren't any new green "regular" phosphor tubes to be had.
|
|
|
|
Post by thelasttubewarrior on Jan 19, 2018 15:48:58 GMT -5
For anyone who was ever interested in the "Frankenyoke" experiment, to give better focus performance to a Marquee by substituting better magnetics, here is what I believe will be my final commentary on the project:
The best results to be had will be by making a small modification to the HDM and using G90 magnetics.
You will remove the Marquee stock deflection yoke and replace it with the G90 deflection yoke. (To include the convergence windings.)
You will remove the Marquee stock focus yoke and replace it with the G90 focus yoke.
This project assumes that we will be running the Marquee only on HD content as we will lose the series/parallel winding switching as the G90 yoke is not set up for that.
So, we will only be running at scan rate frequencies that are in the "high band".
While it may be possible to configure the G90 yoke to run in series or parallel like that of a Marquee, I have never put any effort into making that work. It is uncharted territory.
You will of course need to adapt the wiring harnesses/connectors. This is pretty straightforward. I'll pull my samples and show you the adjusted wiring.
Using cables taken from scrapped G90s I've made plug-and-play adapter wiring harnesses that allow me to use absolutely stock and totally unmodified G90s deflection yokes.
So, with the wiring taken care of, now there is but one thing left to do:
Go to the HDM and locate R175 under the deflection daughter board. It needs to be a larger value to reduce yoke current.
Its stock value is 357 ohms. We want to go larger. The correct value will probably be under 1K so if you have a small 1K trim pot available, you might try using that.
I am not going to define the exact value you need. I have not figured it out yet because at this moment I've tested with one G90 deflection yoke and two stock yokes, and I've tested with two G90 yokes and one stock yoke, but I have not yet gotten around to testing with three G90 deflection yokes.
The G90 yokes have lower impedance and require much less drive current.
The difference is such that in testing with 1 G90 yoke and 2 stock yokes on the projector, with the H. width dropped to zero, the grid rasters on the stock yoke equipped tubes are about half the width of the available raster area, but the raster on the tube with the G90 yoke is spilling off the sides of the tube.
Actually if the G90 yokes can be made to work perfectly, then the G90 focus yokes will also work perfectly with no modifications to the focus yokes or focus drive system aside from connector changes.
I have achieved perfect focus performance (equal to a G90) in these tests. I just haven't done a full-up three tube test yet.
The reason for using the G90 deflection yokes instead of the stock yokes: Two factors: Electromagnetic compatibility, and the fact that the G90 deflection yoke assembly is shorter than the stock Marquee deflection/convergence yoke assembly. This allows the G90 focus yoke to be placed correctly where it needs to be on the tube neck for best performance.
The fundamental reason for the entire Frankenyoke experiment is simply that Marquees have always run with the wrong magnetics ever since the changeover from Thomson tubes to Panasonic tubes back in the early 90s.
The two tube types have different requirements. Even the anode voltage and beam current profiles are dramatically different. And, focus yokes need to be engineered with beam current profiles in mind.
Electrohome was the one company that elected to use Panasonic tubes that did not also elect to use Kanto-Denshi magnetics, Kanto-Denshi being Panasonic's partner company for magnetics.
So, while the end result was still a focus system that was functional, it was never optimized for Panasonic tubes. Electrohome declined to invest the engineering time and cost into developing new magnetics that would work optimally in the Marquee.
AmPro also elected to use Thomson magnetics, but only the deflection yoke. Even AmPro used K-D focus yokes that were made for their application, but never with astig windings.
In theory, we could use Barco magnetics, too, (also made by K-D) if from a Panasonic tube equipped projector, but interfacing the Barco magnetics stack properly has always been problematic.
The best answer by far is G90 magnetics used as described.
|
|
|
|
Post by gjaky on Jan 19, 2018 16:27:50 GMT -5
Simply overriding the current limit resistor in the HDM is not the wisest idea when trying the G90 deflection yokes while they are in parallel mode.
The reason for this is that the retrace time is given by the product of coil inductances and the installed resonant capacitance (t_retrace=pi*sqrt(L*C)). First this is dangerous because the lower the retrace time the higher the retrace pulse is (assuming the same coil current), there is potential risk to exceed the maximal blocking voltage of the switching FETs (ie. destroying them), second the resulting super short retrace time will manifest in massive ringing.
Instead connect the G90 deflection H coil halves in series (it should be simple as they are wired out separately). Then the per tube inductance will be 180uH for the G90 yoke (series) opposed to the 135uH that would be the Marquee coil's inductance when in parallel. But this is a much better scenario to start with...
|
|
|
|
Post by thelasttubewarrior on Jan 19, 2018 22:44:06 GMT -5
An interesting thought. One I had not considered. (I'm a good technician, but not an EE.) I'm willing to risk an older HDM for this test in my originally described conditions but I can also just as easily reconfigure the yoke wiring to wire the coils in series. My original idea preserves the Marquee's series/parallel wiring switching, now that I dug out my samples and saw again how I did it.
I think that'd still require adjusting the value of R175 but probably not by as much. I'm sticking a 1K trim pot in there, preset to the default 357 ohms and marked at that position.
I'm hoping to do this test on my testbed projector over the weekend.
I do recall that I had my concerns about how hard the different yokes would drive the HDM. After using a larger value resistor for R175 (I forget what value I used) with the G90 test yokes, I did note that
the HDM heatsinks did in fact run cooler. I didn't see that as a bad thing.
|
|
|
|
Post by stridsvognen on Jan 20, 2018 9:08:26 GMT -5
Personally i find the frankenyokes a huge waste of time, as the yokes is not the main issue in the Marque, and by that i mean if you changes yokes you will need to do something to change the operation of the static convergence so it will operate stable all the way to its extreme outer settings.
As the Marquee works the biggest convergence drift factor is static convergence, and to optimize for minimum convergence drift you need the mecanical raster centering that comes with the original yokes.
For me the marquee is about image quality, and to reach optimum quality there is a number of compromises that needs to be concidered, and raster ringing is a huge factor, and on top of convergence drift introduced with static convergence, there is also unstability/ shaking, wich blurs/ reduces resolution.
I did spend a few days playing with a frankenyoke setup in a marquee once, a/ b tested it, and found absolutely no focus improvement, and the frankenyokes was removed in favor of the standard yokes to get stable operation, at identical focus.
But if someone can document improved focus capability and focus uniformity across the tube face, at a specifik peak light output ill be interested to travel there and see it for myself.
|
|
|
|
Post by thelasttubewarrior on Jan 20, 2018 10:28:44 GMT -5
You can try a lot of different focus yokes in a Marquee and get absolutely no improvement.
Or you can try one that's both magnetically compatible with the tubes AND electrically compatible with the drive circuits and then you will see the improvement.
To be clear about this, you won't necessarily see sharper focus than the stock Thomson yokes can deliver, but what you should see is better focus retention at high contrast levels.
Many people have noted that the G90 stays sharp at output levels where a Marquee's image has become a blur.
Why would this be? They both use the same kind of tubes. (Yes, a P19LNQ is a P19LUG. Identical tubes, but Sony insisted on a private part number for improved profitability as the "only source of replacement CRTs for your G90". The only difference is the label.
So if the TUBES are the same, then the improved focus retention has to be in the focus system.
Running G90 and stock focus yokes side by side in a Marquee, with the G90 deflection coil installed as well, the focus retention was tested and the evidence was very "clear". The tubes with stock magnetics blurred above a contrast level of 80. On the tube with the G90 magnetics, there was a trace of blur visible but not until contrast was above 95. Which is exactly what happens in a G90.
My very first, the original, Frankenyoke experiment used an AmPro focus yoke, specifically a K-D 2211, installed in a Marquee 8000 chassis, around the year 2004. No astig winding, but then again, an 8000 didn't have that function anyway. That particular model of focus yoke was actually designed to sit behind a Thomson deflection coil on a Panasonic tube. UNLIKE any other K-D focus yoke ever made other than a few prototypes. So its results were excellent. In fact, using my Sencore CM2125's pattern generator, I was able to achieve a resolved 1:1 test pattern on the tube face at a resolution of 2048x2048 in 4:3, equivalent to 2048x1536 in 16:9. However, the HD-8B lenses were not good enough to fully resolve this on the screen.
That was the only time I ever was able to achieve INCREASED resolution capacity by a focus yoke exchange. And, yes, focus retention at higher contrast levels was excellent as I recall.
All the other FY experiments met with limited or no success. The reason for that is now clear to me, but it was not clear then.
That reason varied from case to case but either coil values were incompatible or the yoke could not be placed in the "sweet spot" on the tube neck because the stock deflection/convergence yoke assembly is too long.
I have abandoned all other FY experiments and literally threw away all my samples and test articles recently. Except the G90 magnetics. I'm keeping those. I have every reason to believe they can run in a Marquee and give it the focus performance of a G90, excepting the slight improvements that come from electronically controlled CPC magnets which the Marquee doesn't have.
And, ironically, the mods required to run G90 magnetics are less involved than with any other yoke.
What gjaky says about retrace time, drive current adjustments, retrace pulse peak voltage, and yoke interaction with the system are all factors to take into account, but that puzzle is one that I believe can be solved as well.
|
|
