[jjasniew]

With the contrast down, looking into the lenses of my ECP 4100 I see that the raster is not centered vertically - in fact, it's bumped up against some metal clip thing at the top (bottom now - ceiling mounted) of all three tubes, which actually blocks some of the scan. (You can see the shadow...)


What is the adjustment to center the rasters vertically?


I'm afaid the answer is going to be "there is no verticle dc offset adjustment". Then if so, how do I fix the problem? It's as if 1/3 of each tubes phosphor isnt being used, because the raster's all jammed down at the bottom. Not that moving this would change the %, but I somehow think it best to have the raster at least somewhat centered both vertically and horizontally.


Thank-you for any help of suggestions.

 

[jcmccorm]

Have you checked to make sure that there is no vertical move used? (make sure "NO MOVE" is selected with the remote).


Cary

 

[Jay Wilson]

Cary's right. Other than setting to no move, there is no vertical adjustment in the ECP projector.

 

[M NEWMAN]

Have you done a hard reset yet? A lot of these PJ's came out of commercial installs where they were installed by idiots who used a LOT of convergence to fix install screwups. A hard reset might clear deliberately offset rasters.

 

[jjasniew]

I did the hard reset and it seemed to do the trick.


Now, I can proceed according to "All ECP owners doing toe-ins/raster centers, read me" thread and "Help converging ECP4500.....Curt???" thread, which I have printed out.


Anything else I should be looking at?


Thanks!

 

[Jay Wilson]

Hmmm, I thought we had done the hard reset in the last thread we were in...oh, well maybe not.


those two threads are the best for setup. The holy Focus Guy Kuo thread is pretty good too...but it's easier after you have a pretty good handle on the pj.

 

[jjasniew]

Ok, I did the big kabash reset on my projector.


Going through the "setup for new screen" menu, I notice that the Red tube isnt aligned with the Green tube vertically, for the rectangular target with the + presented. Horizontally, I can and did adjust it.


However, when it came 'round to the Blue tube, it was spot-on to the Green vertically!!! It was "off" similarly to the Red tube, prior to the reset!!


Of course, going through the convergence fields, the Red required a lot of manipulation. The Blue - hardly at all!


So the electronics is "working hard" with the Red tube, not doing a whole lot with the Blue.


Now, I've been told time and again that you cannot move the image vertically via some centering adjustment. Yet, I'd like to have that initial Green line-up for the Red tube be as good as the Blue. What do I do??


This would make the electronics do less work, hence less drift, more stability, etc. If the theory that starting with a physically well aligned unit and using as little electrical compensation is trully the way to go. And I believe it is.

 

[Jay Wilson]

Joe,

Cary and I were looking at this, and it isn't worth it. The initial convergence, that moves the entire image is global. The rest of the convergence is point and does make the elctronics work harder...at least for horizontal. Vertical should be pretty stable. Cary and I were trying to figure out a way to add a centering pot. He actually had a design working, and we researched it further. According to Cary (who did the measurements) this is what he found:

Quote:

I checked the output voltage on green centering (went home for lunch). This is the output of the opamp. Since this circuit, along with the DAC, allows the centering voltages to swing from -10V to +10V, I expected the green to be around -10V. It was closer to 0V! (close enough - about 0.7V). I checked the two outputs of the DAC that drive the opamp and the voltages were 5.38V and 0.345V. I went through the opamp circuit with the calculator and got an answer for the opamp output that is close to what I'm seeing. So, the output of the DAC was deliberately set up to make the green at 0V. Very interesting, but I guessed that this is where they wanted the green centering to be initialized, 0V, which makes sense. Then I checked blue. The outputs from the DAC that drive the opamp circuit were *way* different from green! They have been deliberately changed by the firmware! So, on a hunch, I went into the convergence routine and got to the part where you can act on the "total" red and blue crosshatch, moving it up and down across the whole raster. Guess what? The output from the DAC changes! So, these V centering outputs actually *are* being used by the firmware on the PJ and they are used for the overall (not zone) convergence. They don't use the convergence coil to do it though, the use the deflection coils for the overall stuff convergence movement. This is pretty cool. It also means that any mod we do isn't going to do anything unless we totally bypass using the DAC and just build a simple voltage divider between +12V and -12V to adjust the centering voltages. I'm not sure how much more stable, if at all, that would be than using the DAC. It looks like using the up/down overall convergence move accomplishes the exact same thing we're doing. Kinda makes this mod not very attractive in my opinion. It was a fun learning experience though.

 

[M NEWMAN]

Jay,


I told you guys to quit worrying about that vertical convergence issue now, didn't I?

 

[Jay Wilson]

WE DID! WE DID! Oh, you meant stop THE FIRST time you told us

 

[jjasniew]

Jay,


So you and Cary are "de-bugging" down to the component level? Do you guys have any ECP schematic sets you'd be willing to share with another fellow? (I'm an EE doing lab test automation for a large computer company)


So, let's say after the Hard Reset the DAQs which control the "total" crosshatch raster verticle position are set such that the op-amp outputs are near zero for all 3 colors. This assumption kinda makes sense as an overall starting point for convergence.


Therefore, if after my hard reset I find blue matches green (both ~ 0V) but red is off (-2V??) I might conclude that there's some fault in the red op-amp circuit. A leaky bypass cap, resistor gone out of tolerance, contact problems, bad chip etc.


Since I dont know where to look just yet, could Cary measure the 3 voltages out of the 3 op-amps - after the HR is done, to see if they are each initialized to zero? I'd expect that they would be - then you move them using the "total" crosshatch digital adjustment.


So I can find, perhaps, as an example, that when my Red DAQ outputs are initialized to something like 5.38 and 0.35 - and my op-amp output is not 0.7, then there's something analog circuit wise that I can fix.


Why not just comp with the digital adjustment? Well, if it's something I can fix...Also, I think I notice the red "zone" convergence adjustment running out of gas in some of the zones - barely making it. Maybe the zone adjustment would not have to work so hard, to comp for the other total adjustment, which would be better if electrically correct in the first place.

 

[Jay Wilson]

Joe, Cary is out of town until Thursday...I can't speak for whether he will measure all 3. I'm not sure he will want to doa HR...he just replaced a tube and spent all of a Saturday resetting the thing I am not an engineer, so most of what I do is based on others explanations. Cary just recently taught me the basics of a scope. If I could, I would do it for you as I am about to reset all my scheimflug adjustments and will probably start over afterwards.


As far as measurements, I have yet to have an ECP that has set all of the 3 tubes to the same vertical plane. I have ALWAYS had to adjust at least one of them somewhat vertically. As you can tell from Mike's answer above, he says we are wasting our time here. That the vertical adjustment in global convergence is adequate. After our little experimanet, I believe him and Cary and I are giving up on this. Normally it is not wildly off...just maybe an inch or so, and with the amount of global range, that 1" is negligable. And as Cary indicated, since it isn't using the convergence circuitry, but rather the deflection, it should be stable.


As for running out of range, that is why I am adjusting my lenses. I have that problem in one corner, which I am afraid may be cause by a bit of a mis-adjusted lens. They should be one inch from the crt mounting surface, and my red is 7/8" in one corner. That was set by a previous owner, and I have always just lived with one convergence line being at max in the one corner...now I want to fix that too. M Newman has indicated that this is so critical, that your ruler must measure down to 1/32" to be as accurate as possible.


I have the schematic, but it is paper. No electronic copy that I have found yet. I will be happy to look something up for you if you tell me what you need.

 

[jjasniew]

As far as schematics, I'd like to see the RGBHV input board RGB amplifier chain; at least one amp chain if they're all identicle.


I'd like to see the circuitry around the video drive where the three 10uf Rubycons are that KBK says mods effect as much as the whole amp chain mods in the RGBHV input module.


I measured the DC across the 100uf 25V coupling caps (got the value wrong in my last post) this morning with the projector running. Got around +5V across #2, +9V across #4 and couldnt get a DC measure across a #3. #1, the first cap in the chain, (input to output) wasnt oriented such that I could get a measurement. Who knows what they see at initial turn-on. It's tough to do these measurements with the PJ hanging from the cieling...


I couldnt get a resistance to ground (or anywhere else) from the (-) terminal of these caps. So I need the schematic to see what the loading is.

I assume they're 25V caps because the board has +/-12V going into it and maybe this circuit is symmetrical about ground - but who knows w/o a diagram.

 

[Jay Wilson]

Joe, You have a PM

 

[jjasniew]

It looks to me that not as much of the phosphor is being used as could be!


The area illuminated changes for the convergence crosshatch, depending on the presence of sync from the video. It also changes depending on the resolution and scan of the video input. There are equations in the literature that describe this effect, so I know that I can expect these changes.


With no sync input, I get an almost perfect 60X45 image of the convergence crosshatch. With the sync signal delivered, the horizontal drops to 54" and the verticle to 44".


I'm using a Viewsonic VB50HRTV as input to the projector. (I'll include my caveat here; "you have to start somewhere") 1024X768, 75hz. I have not yet explored the 800X600 nor the 60 hz settings much - other than to observe that the raster area changes...


My spouse joined me as I was watching a tape of "Eyes Wide Shut" and she commented that "those actresses all have such skinny hips". I failed to let her know that this was due, in part, to the projected image not quite filling the 60X45 frame in the horizontal dimension. (The movie was so exciting I fell asleep, it ended, and the VCR provided a blue screen which remained projected for an additional 3 hours @#$%^&*%)


So I notice when I adjust the horizontal using the digital control, the unit has to work with the sync some and the "sync" green led lights up on the panel. Is this what I need to do for the resolution/scan that I finially decide on, to get the correct proportion? Or, is there some other circuit adjustment for the width which is a better way to go?


Basically, here are some images which show where my raster is in the phosphor of the red tube; G&B are pretty much the same. What adjustments still need to be made, or would anyone here recommend, before running the unit much more?

 

[jjasniew]

Here's the lower right for the crosshatch -

 

[jjasniew]

This should give an idea of where the raster is relative to the left and top, for a 1024X768 image from the VB50 -

 

[jjasniew]

This should give an idea of where the raster is relative to the right and bottom, for a 1024X768 image from the VB50 -

 

[Jay Wilson]

Joe, I'm not sure I fully understand your questions...but...


the sunc light in on indicating that you are in a manual adjustment. If you push the button again, and go back to an automatic adjustmetn, the light should go off.


You are correct in that the different resolutions fed to the pj affect the settings. Always start with the highest resolution you plan to run, and then work your way back down through the lower resolutions. Do not start with a low resolution and work your way up expecting it to work. all settings will be way off.


also, after doing your geometry settings, simpy use the size controls to fit the image to the screen. the crosshatch will typically be larger than your screen surface by a varying degree (pending on a number of variables). there is a global width adjsutment on the h defl board under the green tube as well.


I didn't go back and read everything, but I am assuming that the referenced screen is what you should have based on your throw distance? I reember you had that skylight that kept you from movgin forward. Based on the dimensions above, I'll assume you throw is 81"? this also drastically affects teh amount of phosphor you use. How far back from teh screen is that green lens?


It actually looks somewhat ok based on your pictures. Hard to get a scale of reference from a pix, but mine is larger than yours on the horizontal side anyway...not sure what mine looks like vertically...never really cared. I definately don't have as much spacoing on the left side of the tube. But again, all a fuction of throew distance.

 

[Jay Wilson]

BTW, Don't try to setup the pj without a signal...