G2 is independent of HV. If HV were too high then the monitor would simply shut down.
Yes, but high brightness means more electrons hit the screen while the statistical amount of xray photons per electron stays constant =>more xray.
You may be on to something here. I'll check it out these days, unfortunately i haven't had time to mess with it yet. Could you give me a link to the service manual?
That would be horizontal output transistors. (HOTs) While it is true that the flyback transformer itself gets a bit hotter with higher frequencies, it's the horizontal output transistor (the big one which switches B+ into the flyback) which becomes more and more inefficient as frequency rises.
It's not, and this time it really has to do with the flyback. All transformers have a minimum operating frequency depending on core material, size of air gap and number of primary turns. If the frequency is too low, the core saturates. If the core would be allowed to saturate the HOT would blow up.
But the Nokia 417TV could do 25Hz (horizontal frequency 15.625kHz) as well as SVGA frequencies without problem. So far I know, other Nokia monitors (447 etc.) were based on the same hardware platform despite they didn't go that low, so it may be a marketing issue to avoid ugly looking modes and force people to buy new computers. I always wondered why multiscan monitors don't simply use separate oscillators for horizontal deflection and flyback input. So the flyback trafo could always run on its optimal frequency and waste less energy.
MAY THE SOFTWARE BE WITH YOU!
*============================================================================*
I CYBERYOGI Christian Oliver(=CO=) Windler I
I (teachmaster of LOGOLOGIE - the first cyberage-religion!) I
I ! I
*=============================ABANDON=THE=BRUTALITY==========================*
{http://weltenschule.de/e_index.html}
Yes, but high brightness means more electrons hit the screen while the statistical amount of xray photons per electron stays constant =>more xray.
Nope, X-ray emissions depend only on acceleration voltage. G2/brightness only adjust black level, the white level is fixed by the contrast control. The maximum white level is given by the EHT voltage. Attempting to exceed those limits given that EHT remains constant only results in lost highlight/shadow details.
I always wondered why multiscan monitors don't simply use separate oscillators for horizontal deflection and flyback input. So the flyback trafo could always run on its optimal frequency and waste less energy.
If they ran at different frequencies they would interfere with one another, the result being ugly distortions of the picture.
That link sent me to an ugly little exe file. Anyway, i just remembered i had it - just have to look up the correct section.
Sorry, the link seems to have changed. It opens only an EXE that connects to the internet and displays "install failure - please download again". But this one seems to work:
If they ran at different frequencies they would interfere with one another, the result being ugly distortions of the picture.
Vector monitors (for arcade, radar, EKG monitors etc.) generally run their HV trafo by a separate oscillator because they have no fixed horizontal deflection signal. I think it would work; may be that it just would make the power supply more expensive.
Back to our P1110. While that "drift correct" circuit does indeed look weird, i found an area that is a lot more interesting by far: G2_CONT. I see there a lot of high value resistors.
That is the area where the mods are usually done, but it seems to me they've been modding the wrong resistors. R458 (2.2Meg 1W) would be the prime suspect IMO. I am going to replace all high value resistors in that area, including the SMD ones, with 1% tolerance 1/4W ones, and see what that gives. The 2.2Meg 1W resistor will be replaced with a 3W.
Hi Th3_uN1Qu3, have you found a way to measure the G2 voltage on these monitors?
I have a 21" EIZO T966 (aperture grille) and I am trying to adjust it to specs and I need to measure the G2 voltage, the service manual says 700±10 VDC. So I attached the negative lead of the multimeter to the metal strips on the aquadag, and when I touch the positive lead to the G2 pin on the CRT neck-board, there is some kind of HV sounds from the CRT (like the monitor is turning off and discharges) and the screen dims a lot.
Now, I'm afraid to keep the multimeter touching long enough to get a reading, I don't want to damage something.
When I did this on a small 14" TV, I succesfully measured 370 VDC (and nothing much happened, screen dimmed a bit and some slight sizzling).
So I am not sure why this is causing this monitor to freak out, can you give me an idea what to do, am I grounding wrong or something, is this normal?
So I am not sure why this is causing this monitor to freak out, can you give me an idea what to do, am I grounding wrong or something, is this normal?
G2 is a high impedance signal and cannot be measured with a regular multimeter. Even on your TV, with the multimeter hooked up you were in effect reducing the G2 so now after you adjusted it it is too high.
You need to make a high impedance resistive divider for your meter, or an active probe. As a guess i'd say you need at least 100 Megohms for the side that you wire to G2, and 10Meg from the end of that resistor to ground. But that would still read wrong on your meter for x10 division if it only has 10Meg impedance (some cheap ones even have only 1Meg). I'd say 100Meg across the G2 (make sure it is a HV resistor, or better, use a string of 10 1% 10Meg ones), and 1Meg from there to ground. That would be a x100 divider so you would measure with your DMM set to 20 volts scale, 1 volt on the meter is 100 volts on G2. If your meter has only 1Meg input impedance (check the manual or measure it with another meter), then the resistor to ground needs to be 100k, and it would be a x1000 divider - ie you use the 2 volts scale.
I usually do G2 by eye though. Turn brightness to max, contrast to min, display a full black screen (which will be grey because of the high brightness) then increase G2 till the retrace lines start showing, then back off until they disappear. That's all.
As tubes age they need increased G2 to maintain full brightness anyway.
Hmm, I see. Well, for now I will adjust it by eye as you said, I am just trying to make sure everything is as close to specs as possible.
Are there any other voltages inside a CRT that cannot be measured the normal way?
I succesfully measured the "Main 85V" line at TP501 which is the first step in the manual, and it shows 85.0 spot on, so that is OK. I would still need to measure the HV (~27.0 kV, would need a probe like this - would that one also work for measuring the G2, I see it has high input impedance), and something called "H.DRIVE" between two pins but that is probably low volts.
The rest is just going through the service menu and adjusting various parameters like H. Hold, H. Sub Size, Linearity etc.
I would still need to measure the HV (~27.0 kV, would need a probe like this - would that one also work for measuring the G2, I see it has high input impedance)
Yes, that would work for measuring HV and also G2.
I now have examined my SGI GDM-5411 closer. First I opened the case and vacuumed a lot of tinder eh dust out to avoid a room fire. Then I measured some high megohm resistors on the CRT PCB; they seem to be ok.
- drift correct transistor (yes it does!)
So I checked the mysterious "drift correct" transistor Q406 between chassis GND and the GND input of the RGB cutof amp IC406 (actually a ceramic coated beige hybrid module running on 220V DC). In cold state without signal (only startup message) I measured 2.18V instead of the 1.7V rated in service manual. So I tried to pull it against GND with a resistor, which indeed lowers(!) the excessive brightness. However I had too pull it down to about 0.18V (using a potentiometer) to correct the black level (all in cold state), which seems to low to be normal, thus the transistor is unlikely the trouble source. (Short circuit current was only something like 0.98mA, which looks harmless, so it may be yet another suitable method to reduce overbrightness. I haven't checked how this affects colours.)
- WinDAS tips
So I installed WinDAS on my laptop IBM Thinkpad 760XD and connected my homemade 2 transistor RS232->TTL adapter. After measuring polarity of the monitor ECS port (GND is top, then +5V, then both data lines, cover can be pried out without opening the monitor), I connected it and too m surprise WinDAS started perfectly and found the monitor. It runs very stable on my laptop (166MHz Pentium MMX,106MB RAM,Win98SE) - likely because it was designed for such antique hardware. It yet never crashed nor lost contact to the monitor. The only bug I found is that the adjustment sliders are drawn twice (one upon each other) - the lower one is dead and only the smaller upper one works. After clicking it, you can also adjust the value with cursor up/down keys.
Unlike claimed elsewhere, the ECS communication works even when the monitor goes into standby. Important for this is only not to power the adapter by the ECS port +5V line. I instead used the +12V voltage from my RS232 (routed through a 100 Ohm resistor and pulled down with a 4.7V zener diode). Despite I discovered that the laptop anyway outputs only +5V instead of +12V and I have 2 LEDs (with 1kOhm resistor) burdening the data lines, my adapter works rock solid. So no matter where you plan to get your +5V from (dummies may even use 3x 1.5V batteries), to avoid trouble do not use the ECS port.
There are already instructions how to use WinDAS, so I won't fully explain it again here, but only note here some crucial details, those you may add to a FAQ.
IMPORTANT:
When using WinDAS, the absolutely first thing you must do is save the monitor's data to a file ("Save Data to File" in "File" menu) and make a write protected copy of it. (Copy additionally a backup on external media; USB stick works well, but even a diskette is sufficient to hold many versions of the only 10KB small file.) Do not accidentally select "Load Data to Set"; it may overwrite your monitor data with garbage. Especially do not accidentally select "MPU" - it would format your monitor EEPROM and reset everything to awful factory defaults.
WinDAS does not work offline, but like a stupid TV remote control most of its functions directly change the monitor's internal register data without asking and without any warnings! Thus unless you have made the backup, do not play or even test any other WinDAS functions while the monitor is connected, because this easily makes the monitor unusable until the backup file is uploaded back into the monitor.
To see what is in the file, select the menu "Help/Expert/View" and select your backup from the file requester. You can read but not edit register names and values in the file here. (This is the only function that behaves like a normal Windoze program. Stupid is only that you can not select it unless the monitor is connected.) To edit values, overwrite them with a hex editor (not Windows "Notepad" - you are not allowed to change the file size or add unexpected characters).
With my SGI GDM-5411 the G2 value showed up even as the 1st number on the list. It had been 171, so I set it lower because my monitor was way too bright and full of diagonal retrace lines. (A too low value can turn off the monitor picture, so using a 2nd PC (e.g. laptop) for WinDAS is a good idea.)
All higher WinDAS functions are only contained in so-called "procedures", those behave stubborn like menus in a novice's C64 BASIC program or old DOS .BAT file installer. Procedures are basically batch files those request you in a fixed order to use certain screen tests and make certain adjustments (either on screen using windows sliders or by physical adjustments inside the monitor) and measurements. For the screen tests you need an external program (use a different PC to avoid trouble) or signal generator, which is not part of WinDAS. But the most annoying thing is that you can not switch back to a previous step when you accidentally failed or skipped one, and since procedures (white balance...) can take several hours and only ask you at very few spots whether you want to redo something, any wrong button click can waste hours of work. :mad2: Especially be very careful with "CANCEL" buttons; they will abort the procedure without warning and void all previous steps by switching the monitor back to the previous state or even mess it up so you need to re-upload your backup file. Also some "READJUST" buttons are almost as bad as "CANCEL" since they send you far back in the loop and often reset previous adjustments to useless defaults instead of permitting you to tweak them further. So do not even try to begin a "procedure" when you only want to slightly tweak fairly well adjusted parameters; procedures generally first reset all related registers to useless defaults and tend to leave your monitor in this state when you abort them. Parameter names in description page often differ from the displayed adjustment slider name, and there may be other bugs in those descriptions.
IMPORTANT:
After successfully finishing any procedures, you must execute the "Final setting" procedure (before you quit WinDAS or unplug ECS cable) to save their results into the monitor's EEPROM. Else they stay only in RAM and will be lost once you power off or unplug the monitor, and also the monitor's menu will stay locked. After "Final setting" also do "Safe Data to File" to backup your work.
All available procedures are shown in a clickable list accessed through the "Adjustments" menu. For novices confusing is that their order changes because after use they are sent to the bottom of the page (likely to memorize which ones you have done), which at least with "Final setting" makes no sense since this one should be executed after finishing each lengthy procedure to avoid data loss.
- colour calibration trouble
I have gone through the white balance procedure to calibrate my SGI GDM-5411. As a colorimeter I used a SpyderTV (Spyder II hardware) and the great HCFR Colorimeter open source program. WinDAS requests you to show screen tests in 1280x1024@75Hz mode. I employed Nokia Monitor Test v1.0a and a homemade black and 30IRE grey PNG picture on IrfanView (on black desktop background) together with small HCFR window to see the xyY values in "Continuous measures" modes (those built in HCFR are fullscreen and can not be used simultaneously). With 6500k and 5000k mode I have set HCFR white point to D65 and D50 mode to see correct RGB % values; unfortunately for 9300k there is no matching (D93?) white point, so I selected sRGB. (In Windows I removed all colour profiles and gamma settings.)
The white balance procedure asks in a sequence for parameters of all 3 colour modes (9300k, 6500k, 5000k) and so is way too lengthy to be reasonable. (It took about 3h; why didn't they make 3 procedures of these?!) The first adjustment it asked for is the infamous G2 setting. Originally mine was 171 (way too bright with visible retrace), so I first tried 125 (grey looked still washed out) and in the 2nd try 114 (much better black) using the "Animated black" walking bar screen test of HCFR and black background from a DOS fullscreen text mode. In a dark room especially the visible scanlines of DOS and other low resolutions are well suited to see the black level. The darker bar in "Animated black" should be really hard to see. This way I got a nice black level, but unfortunately there is plenty of other trouble.
Before adjustment, my sRGB menu requested to set brightness 02 and contrast 00 (was ridiculously dark), now it wants 37 and 75, which looks less horrible. At least it already looks like a monitor again and not like "Who in this drive-in cinema has his xenon fog lights on?!"
With darker colours the Spyder II probe is infamous to output nonsense RGB values, but I first got rubbish partly up to 50IRE (yes, baffle removed and CRT mode). Now I still have at sRGB e.g. blue 140% at 0IRE, 120% at 10IRE, 110% at 20IRE. Much worse is that many requested Y (luminance) values are impossible to adjust even with the shown sliders at maximum. I reduced the IrfanView screen test window to 1/4 screen size to avoid the beam current limiter (so some modes indeed were bright enough to blur letters at max contrast) although according to WinDAS manual's signal generator description they seem to be supposed fullscreen. Especially with the last setting for each colour I got way less (e.g. 61 instead of the requested 90 cd/m²) with the CON_MAX_BR... slider (aka CMaxBCont in description) at 255 (maximum). My gammut triangle in CIE chart shows x0.561,y0.356 instead of x0.640,y0.330 (DeltaE 49.0) due to way too little red. Green has DeltaE 21.6, only blue is close. Strange is that red surfaces IMO look rather too bright (like neon colours, a tomato shines like a red traffic light) which may hint to poor linearity. In forums people claim that 6500k (sRGB) would be the best colour temperature (TV uses this) because it was modelled after sunlight. But with the luminous red IMO only 9300k looks like daylight and 6500k look really obnoxious.
Is really my CRT dead (worn out by overbrightness, or vice versa??), or is this the video amp or other electronic faults? :confused2
On dark grey screen test I see 2 symmetric crescent shaped smeared shapes (like ")(" ). Are these beam reflections inside the CRT or something worse? There is also occasionally slight colour impurity at the left side.
- dynamic convergence?
Despite I have tried hard to adjust convergence with the few available settings in WinDAS, especially the upper left corner has blue 2mm shifted to the left. Generally the top 3cm are slightly off. I hope I haven't messed up the magnet rings by vacuuming dust out (I cleaned everything with a soft paintbrush + thin vac hose), but I think it always was like this.
If I remember well SGI monitors were at that time considered absolute rocket science high end hardware, but the DCNV menu in WinDAS is greyed out. Has this monitor really no dynamic convergence, or needs WinDAS a special file update to handle this?
MAY THE SOFTWARE BE WITH YOU!
*============================================================================*
I CYBERYOGI Christian Oliver(=CO=) Windler I
I (teachmaster of LOGOLOGIE - the first cyberage-religion!) I
I ! I
*=============================ABANDON=THE=BRUTALITY==========================*
{http://weltenschule.de/e_index.html}
I hope someone comes up with a better solution than the Windas method. I've had to lower the G2 three times so far and after the last time the monitor takes over a minute to warm up enough to see a picture.
I have now selected in Win98SE the sRGB colour space and increased graphics card gamma to 1.10. This way the measured brightness curve (with HCFR + Spyder 2) indeed fits perfectly and gamma curve stays between 2.25 (dark) and 2.13 (bright). But saturation of the red channel at 50% is indeed 12.91% too high while the ends are ok. Also colour temperature above 30IRE is close to 6500K. I have set the monitor to sRGB now (recommended by all film and photo professionals), but IMO colours do not feel right at all. It is like watching photos on a yellowed old newspaper. The colour visually matches rather the 15W incandescent bulb in my bathroom than what they claim to be "day light". Or may my eyes simply have a different wavelength-to-colour translation curve than sold as "the truth" by such standards?
(CRT (or flu tube) light has a spiky band spectrum that is very different from the continuous spectrums of daylight. Phenomenons like red-green blindness prove that human eyes can differ very much, so there can never be a perfect standard for a simple colour model like RGB.)
I hope someone comes up with a better solution than the Windas method. I've had to lower the G2 three times so far and after the last time the monitor takes over a minute to warm up enough to see a picture.
When cold, my monitor is still too bright (and greenish) and it takes about 30 minutes to stabilize.
- May it be that the "drift correct" transistor is a better spot to modify the circuit?
When my monitor was cold, pulling its output to chassis GND immediately removed the overbrightness and retrace lines. I am not sure if with even more worn monitors it would be safe to pull the output line even below 0V chassis GND (like done with tube grids in many tube amps) or if this would kill the RGB cutt-off IC by latchup.
- Has anybody tried to heat individual parts to check which ones causes the bright warmup problem?
I had noticed that with my adjusted SGI GDM-5411 the red channel saturation was too high (+12.91% at 50%), which caused obnoxious glowing red. So some hours later I backed up the monitor EEPROM data with WinDAS (which resets the warmup timer :< ) and after re-warmup I did the infamous colour return on normal and sRGB mode. I measured again and noticed that the red saturation flaw now dropped to (still visible) 8.82%. Also other values have slightly improved. Only the colour temperature has slightly dropped and green level has slightly increased (also may be that colour return fixed just corrected a Spyder II inaccuracy), but generally the grey scale delta E and RGB levels have improved a bit.
So I backed up the EEPROM data again and compared both versions with my HxD hex editor. The bizarre thing is that they are almost identical. I only found a "2" instead of "0" at byte 397, which is REG value 48 "COLOR_MODE". It may be that this is just my toggle between normal and sRGB colour mode was indicated by it. Because a complex mechanism like colour return can not hold its data in a single byte, it obviously either stores its data elsewhere or does something completely different. Reports that overuse can damage or wear out the CRT hints to that it may be rather a CRT regeneration process based on controlled cathode overload instead of a simple measurement of RGB currents. But I think a CRT regeneration process would take longer than the few seconds.
So I conclude that Sony either considered the measurement data useless to be accessible through WinDAS or they again placed here one of their infamous anti-reverse-engineering features to prevent competitors from analyzing and copying the internal algorithm. Can you remember the complex copy protection tricks in their PlayStation 3, which first version allowed the user to install Linux but completely enshrouded the inner working of the 3D graphics by abusing one of the 8(?) CPU cores as a MMU, and how Sony rapidly cancelled the Linux feature once hackers attempted to find out ways to circumvent this? This attitude all reminds to the lyrics of an old German song: "Ich bin die fesche Lola, und zeig Euch was ich kann, doch an mein Pianola, da lass ich keinen ran..." (translated: "I am the sexy Lola, show proud what I can do..., but my player piano - no one can touch it!")
It may be that (e.g. to increase access speed) the colour return data is stored outside the normal EEPROM, e.g. directly inside the "RGB Drive" IC402 or "RGB Amp" IC 403. Also the "Input Select" IC401 looks suspicious because it has R_DET_IN, G_DET_IN, B_DET_IN each connected through a capacitor to the corresponding output, which looks like a measurement loop. If the algorithm simply measures RGB output currents and stores a curve to compensate, it may push up a scaling factors once a value in the curve reaches maximum (e.g. 15 in a 4 bit value) and runs out of range. And the algorithm may be a too simple linear approximation of something non-linear, that attempts to lower a statistical mean deviation without testing if dark values turn too bright, which makes it run into the infamous overbrightness (after the warranty is over).
- Has anybody reset or replaced one of these ICs and checked how this affects overbrightness?
Btw, have 100% confirmed that the bright screen before warmup is NOT a tube problem - i've had the monitor on all day and now after messing with WinDAS the screen was darker for a minute or so from when WinDAS shut it down to when it came back up - ie just like it should be. It probably traces to the same resistor people change to fix the G2 problem without using WinDAS. I'll try and take this beast apart next weekend and replace that resistor to see if the high initial brightness will be cured.
I have the exact issue on both my p991s. I always assumed it was normal because there is a 30 minute wait period before you can select 'Color Return' or 'Image Restore.' People have said that this has always been an issue since the monitors were brand new. If simply swapping a resistor fixes it I may do it as well.
If you find the proper resistor can you point out where you ordered it from? Initially when I got my Trinitrons I was going to do the resistor fix but I couldn't find anyone who had the correct resistor.
I don't order common stuff such as resistors because i have a bunch of parts shops nearby. So no help there, sorry. But depending on your location, Farnell, Digikey or Mouser should have damn near everything.
Anyway, i'm currently caught up with a lot of stuff. It'll have to wait a while (again).
The P1130 is here. It isn't as bad as the P1110 was in terms of brightness, retrace lines barely visible and display usable with brightness at zero. Unfortunately it's a lot worse when it comes to scratches in the anti-glare film - i definitely have to remove it on this one.
Hope all goes well when you start working on the P1110 and it's too bad you have to remove the anti-glare on the p1130.
I too am having some issues on my p991's as they both have some arcing on the high voltage somewhere. Hopefully there are no shorts inside the picture tubes. I'm gonna get some silicone and cover the joints on the flybacks and the anode connectors on the flybacks and the tubes and see they stop arcing. I have a Viewsonic G810 that does the same thing as well but it's much further along on the issue so I'll try this quick fix on the Viewsonic first. Hope I don't get zapped!
Does the picture change in any way when it arcs? If it doesn't, it's likely to be the main anode lead. If the picture changes brightness it could be the focus grid or G2, in which case you should pull the A board off the neck of the tube and blow the dust off it. If the picture changes color or goes very bright one color when the arcing happens then the short is inside the tube.
If it's the anode lead, it's unlikely to arc at the connector. Try covering the cable with some insulation.
When it arcs the picture blooms and gets slightly out of focus but only for a second. This has happened across all three monitors. On my viewsonic there was an arc so violent it shut the monitor off for a few seconds. Luckily it didn't die as i like the image it puts out so i stopped using it a few months ago. I fear the same thing might happen to my p991's if i don't do something soon.
After doing some cleaning i realized this has been sitting around for way too long, so i decided. The P1110 is going apart, NOW.
Fortunately all i had to do was remove 3 screws to get access to the A board, they didn't solder the shield in place like they do on other monitors. Kudos to Dell/Sony. There was a huge amount of dust bunnies so out with the vacuum cleaner. Much better... Now i'll start probing around and see what happens. I know where the fault likely is - i just need to confirm it.
M'kay, making progress... Somewhat surprisingly, all G2 related components tested fine. They'll still get replaced, especially those 10Meg resistors, but testing with my multimeter over time shows that G2 does not drift at all. I still do not know what causes the excessive brightness issue in the first place, but maybe i'm gonna find that out from the P1130.
I have found voltages in the same ballpark as CYBERYOGI did on the weird drift correct circuit on the cutoff amp. So at first sight it would appear there's something wrong with that. However, the cathode voltages don't really drift either. That leaves only one variable - the HV. If you think about it, it makes sense. The "warming up" period affects both brightness and picture size. The only voltage that can be responsible of this is the HV.
Taking a look at the D board schematic turns out we actually get a HV adjust pot - but since i do not have a HV probe it isn't a good idea messing around with that. I'm cheap - so i'll DIY myself a probe, and then if i confirm that it is indeed the HV which is drifting, we can go ahead and find a solution.
I'll be taking apart the P1130 too and see how things look like on that one.
Edit: The P1130 has a different internal construction and board layout, even though the parts are more or less the same. Also interesting to note is that the mysterious drift correct circuit isn't present on the P1130. We'll see what that gives us...
Edit 2: Got the P1130 aligned. Convergence ended up better than my eyes can notice - and this is no small feat to achieve. It's hands down the best monitor i've ever had so far. Looks crystal clear at 1600x1200 85Hz. There's some slight horizontal blurriness at the very sides but even with that it's better than the P1110.
And from what i've seen so far brightness doesn't drift at all on this P1130... I think i'm starting to connect the dots here. For optimal brightness range and picture clarity, the G2 on the P1130 ended up almost 100v higher than on the P1110, and even so i'm running brightness at 65 as opposed to 52. So, that extra brightness on the P1110 must be coming from somewhere... CYBERYOGI mentioned that messing with the drift correct circuit lowers brightness, so i'll go ahead and bypass the whole thing, readjust G2 and see where that gets me.
Edit 3: Bypassed the drift correct circuit on the P1110, upped the G2 using WinDAS to compensate. Guess what? I finally got a normal monitor! No more excessive brightness at startup, yippee!!!
The two monitors ended up with pretty much the same G2 voltage (albeit with slightly different WinDAS values), so clearly, something must be right in here. Also stuck my fingers in the advanced convergence adjustments on the side of the tube neck, just like i did for the P1130, and i got another clean display.
Slight horizontal blurriness is the same as with the P1130 so it's either the cable or the laptop video card - i borrowed my main computer to a friend for a while, and i remember the P1110 didn't have this kind of focus issues, so i'll report when i get my main rig back.
Bypassed the drift correct circuit on the P1110, upped the G2 using WinDAS to compensate. Guess what? I finally got a normal monitor! No more excessive brightness at startup, yippee!!!
I will try that with a p991 and see what happens. Unfortunately because my last computer's hard drive crashed I lost the original files for the monitors and I don't remember what the original G2 values were.
First I gotta locate that drift correct circuit. Do you have pics of where the drift circuit is on the p1110? Also what did you do to bypass the circuit?
The cutoff voltages of all cathodes will drift. Common drift is adjusted by the user by controlling the brightness. Different drift leads to a coloration of the black background level. In extreme cases vertical flyback lines will appear. Cutoff voltage can be adjusted with potentiometers, or there is automatic stabilisation. Still, the VG2 (screen) may need periodic adjustment too.
The RGB output amp transistors in some monitor/TV circuits apparently act as a high resistance current source, i.e. a worn out CRT draws less current by higher impedance, which makes the output voltage rise and even can make the transistors saturate and clip the signal (i.e. ugly streaks at the right side of bright image parts). So apparently a worn CRT can indeed turn brighter and not only darker.
What you have to do is connect pin 17 of the cutoff amp directly to ground. The transistor is the tiny SMD type, so a little piece of wire and a bit of solder is all you need.
If you can post a picture of the A board with the shield removed i can show you where it is. I'll post a pic of mine later.
@ CYBERYOGI: The cathodes emit more electrons as the voltage is lowered - as they go high impedance and draw less current and make the voltage rise, the picture goes darker, not brighter. In fact a common test when diagnosing TVs which are missing a color is to momentarily short each cathode to ground, and a very bright screen of each basic color should appear. If all three cathodes respond to this, then the problem is in the video amplifier, if not, the tube is gone.
I'll post a pic when I get home today. Just recently I have noticed something. Between the 2 monitors when they are cold my primary monitor isn't as bright as the secondary monitor.
Another observation: Normally when you wake the monitor from standby you hear the relay click at the exact second the light flickers from amber to green and everything turns on at the same time. Then you see an image about 6-7 seconds later. My secondary monitor, however, doesn't always do this. Instead, I see the light change from amber to green but I don't hear a relay click until about a second later. When the relay clicks then I see the power light, which is green right now, flicker and then I hear the tube fire up. When this happens when the monitor is cold it takes at least 60 seconds before I finally see an image. What is possibly causing this? Should I worry about it?
Some differences are normal depending on the age of the parts. One thing's for sure though: With the drift correct circuit removed, both monitors are going to start up at exactly the same brightness. Taking a few minutes to warm up and stabilize is normal, and it's inherent to all CRT displays.
With my P1110 for instance, if it's been really cold (i've been out all day today), black appears slightly green at startup and hence the picture appears brighter. This normalizes in less than 10 minutes - that's lots better than having brightness drift for 2 hours.
For the second monitor, you may have a problem with the power supply but it's likely something simple like a cold solder joint. I'd try and touch up the soldering, specifically all the joints around the CRT neck, all big semiconductors and relays.
Since you will have to up G2 quite a bit after eliminating the drift correct circuit, make sure you adjust it so optimal picture is at 50% brightness or below. I've talked about it sometime earlier, these monitors are wired to automatically lower the contrast when you up the brightness past 50%, to keep focus in check. Trouble is, this also shifts the color balance a bit, which is annoying. Better avoid it by setting G2 so that picture is bright enough before you reach 50% brightness in the menu.
Also i'm not sure anyone has mentioned this before: Always remove the ECS cable from the monitor when you're done with WinDAS. Keeping it in causes interference with the 5v rail which shows as slight wobbles in horizontal size.
Still works, but you can see jittering in text and small lines.
It's been disconnected for a year now.. Man it's been a long time since I switched to an LCD. Damn thing is still on my desk because I can't even safely lift it off. (Was for running classic games on, but I can't look at CRTs anymore, too much eye damage done, strain is too much now.)
Some differences are normal depending on the age of the parts. One thing's for sure though: With the drift correct circuit removed, both monitors are going to start up at exactly the same brightness. Taking a few minutes to warm up and stabilize is normal, and it's inherent to all CRT displays.
With my P1110 for instance, if it's been really cold (i've been out all day today), black appears slightly green at startup and hence the picture appears brighter. This normalizes in less than 10 minutes - that's lots better than having brightness drift for 2 hours.
For the second monitor, you may have a problem with the power supply but it's likely something simple like a cold solder joint. I'd try and touch up the soldering, specifically all the joints around the CRT neck, all big semiconductors and relays.
Since you will have to up G2 quite a bit after eliminating the drift correct circuit, make sure you adjust it so optimal picture is at 50% brightness or below. I've talked about it sometime earlier, these monitors are wired to automatically lower the contrast when you up the brightness past 50%, to keep focus in check. Trouble is, this also shifts the color balance a bit, which is annoying. Better avoid it by setting G2 so that picture is bright enough before you reach 50% brightness in the menu.
Also i'm not sure anyone has mentioned this before: Always remove the ECS cable from the monitor when you're done with WinDAS. Keeping it in causes interference with the 5v rail which shows as slight wobbles in horizontal size.
Maybe that's it. Other people on this thread have had this issue to the point where it would take 5 minutes before you finally saw a picture. I'll look into the power supply while I'm in there.
Still works, but you can see jittering in text and small lines.
Most likely dried up capacitors somewhere. Easy fix but tedious to locate the culprits. Have an IBM P275 with the same issue. I'll let you know when i get around to fixing that... And yes, if you add them up, i own three of those 21" beasts and counting. I plan to get rid of the IBM and the P1110 and get 2 more P1130s, because it's the best of the bunch. With three P1130s i should be good until i can afford a "proper" LCD (as in S-IPS) or maybe OLED... who knows. Certainly, i plan on keeping the tubes for at least 5 more years.
(Was for running classic games on, but I can't look at CRTs anymore, too much eye damage done, strain is too much now.)
That's funny - i find certain LCDs harder on the eyes than a well-adjusted tube, due to the high white point of some CCFL backlights. As long as the refresh rate is 85Hz or above, there's no eyestrain issue with CRT. And if it's an old DOS game where you can't force the refresh, there's the contrast control. Use it.
Now if you'll excuse me, i'll be back to my game... Hey look, the picture's so clear i can see the scanlines!
Comments
http://www.fixdevice.com/smanuals/file/39405.html But the Nokia 417TV could do 25Hz (horizontal frequency 15.625kHz) as well as SVGA frequencies without problem. So far I know, other Nokia monitors (447 etc.) were based on the same hardware platform despite they didn't go that low, so it may be a marketing issue to avoid ugly looking modes and force people to buy new computers. I always wondered why multiscan monitors don't simply use separate oscillators for horizontal deflection and flyback input. So the flyback trafo could always run on its optimal frequency and waste less energy.
MAY THE SOFTWARE BE WITH YOU! *============================================================================* I CYBERYOGI Christian Oliver(=CO=) Windler I I (teachmaster of LOGOLOGIE - the first cyberage-religion!) I I ! I *=============================ABANDON=THE=BRUTALITY==========================* {http://weltenschule.de/e_index.html}Nope, X-ray emissions depend only on acceleration voltage. G2/brightness only adjust black level, the white level is fixed by the contrast control. The maximum white level is given by the EHT voltage. Attempting to exceed those limits given that EHT remains constant only results in lost highlight/shadow details.
That link sent me to an ugly little exe file. Anyway, i just remembered i had it - just have to look up the correct section.
If they ran at different frequencies they would interfere with one another, the result being ugly distortions of the picture.
Sorry, the link seems to have changed. It opens only an EXE that connects to the internet and displays "install failure - please download again". But this one seems to work:
http://servicedownload.free.fr/moniteur/manual-gdm-5411-chassis%20g1.pdf
Vector monitors (for arcade, radar, EKG monitors etc.) generally run their HV trafo by a separate oscillator because they have no fixed horizontal deflection signal. I think it would work; may be that it just would make the power supply more expensive.
That is the area where the mods are usually done, but it seems to me they've been modding the wrong resistors. R458 (2.2Meg 1W) would be the prime suspect IMO. I am going to replace all high value resistors in that area, including the SMD ones, with 1% tolerance 1/4W ones, and see what that gives. The 2.2Meg 1W resistor will be replaced with a 3W.
I have a 21" EIZO T966 (aperture grille) and I am trying to adjust it to specs and I need to measure the G2 voltage, the service manual says 700±10 VDC. So I attached the negative lead of the multimeter to the metal strips on the aquadag, and when I touch the positive lead to the G2 pin on the CRT neck-board, there is some kind of HV sounds from the CRT (like the monitor is turning off and discharges) and the screen dims a lot.
Now, I'm afraid to keep the multimeter touching long enough to get a reading, I don't want to damage something.
When I did this on a small 14" TV, I succesfully measured 370 VDC (and nothing much happened, screen dimmed a bit and some slight sizzling).
So I am not sure why this is causing this monitor to freak out, can you give me an idea what to do, am I grounding wrong or something, is this normal?
G2 is a high impedance signal and cannot be measured with a regular multimeter. Even on your TV, with the multimeter hooked up you were in effect reducing the G2 so now after you adjusted it it is too high.
You need to make a high impedance resistive divider for your meter, or an active probe. As a guess i'd say you need at least 100 Megohms for the side that you wire to G2, and 10Meg from the end of that resistor to ground. But that would still read wrong on your meter for x10 division if it only has 10Meg impedance (some cheap ones even have only 1Meg). I'd say 100Meg across the G2 (make sure it is a HV resistor, or better, use a string of 10 1% 10Meg ones), and 1Meg from there to ground. That would be a x100 divider so you would measure with your DMM set to 20 volts scale, 1 volt on the meter is 100 volts on G2. If your meter has only 1Meg input impedance (check the manual or measure it with another meter), then the resistor to ground needs to be 100k, and it would be a x1000 divider - ie you use the 2 volts scale.
I usually do G2 by eye though. Turn brightness to max, contrast to min, display a full black screen (which will be grey because of the high brightness) then increase G2 till the retrace lines start showing, then back off until they disappear. That's all.
As tubes age they need increased G2 to maintain full brightness anyway.
Are there any other voltages inside a CRT that cannot be measured the normal way?
I succesfully measured the "Main 85V" line at TP501 which is the first step in the manual, and it shows 85.0 spot on, so that is OK. I would still need to measure the HV (~27.0 kV, would need a probe like this - would that one also work for measuring the G2, I see it has high input impedance), and something called "H.DRIVE" between two pins but that is probably low volts.
The rest is just going through the service menu and adjusting various parameters like H. Hold, H. Sub Size, Linearity etc.
Yes, that would work for measuring HV and also G2.
- drift correct transistor (yes it does!)
So I checked the mysterious "drift correct" transistor Q406 between chassis GND and the GND input of the RGB cutof amp IC406 (actually a ceramic coated beige hybrid module running on 220V DC). In cold state without signal (only startup message) I measured 2.18V instead of the 1.7V rated in service manual. So I tried to pull it against GND with a resistor, which indeed lowers(!) the excessive brightness. However I had too pull it down to about 0.18V (using a potentiometer) to correct the black level (all in cold state), which seems to low to be normal, thus the transistor is unlikely the trouble source. (Short circuit current was only something like 0.98mA, which looks harmless, so it may be yet another suitable method to reduce overbrightness. I haven't checked how this affects colours.)
- WinDAS tips
So I installed WinDAS on my laptop IBM Thinkpad 760XD and connected my homemade 2 transistor RS232->TTL adapter. After measuring polarity of the monitor ECS port (GND is top, then +5V, then both data lines, cover can be pried out without opening the monitor), I connected it and too m surprise WinDAS started perfectly and found the monitor. It runs very stable on my laptop (166MHz Pentium MMX,106MB RAM,Win98SE) - likely because it was designed for such antique hardware. It yet never crashed nor lost contact to the monitor. The only bug I found is that the adjustment sliders are drawn twice (one upon each other) - the lower one is dead and only the smaller upper one works. After clicking it, you can also adjust the value with cursor up/down keys.
Unlike claimed elsewhere, the ECS communication works even when the monitor goes into standby. Important for this is only not to power the adapter by the ECS port +5V line. I instead used the +12V voltage from my RS232 (routed through a 100 Ohm resistor and pulled down with a 4.7V zener diode). Despite I discovered that the laptop anyway outputs only +5V instead of +12V and I have 2 LEDs (with 1kOhm resistor) burdening the data lines, my adapter works rock solid. So no matter where you plan to get your +5V from (dummies may even use 3x 1.5V batteries), to avoid trouble do not use the ECS port.
There are already instructions how to use WinDAS, so I won't fully explain it again here, but only note here some crucial details, those you may add to a FAQ.
IMPORTANT:
When using WinDAS, the absolutely first thing you must do is save the monitor's data to a file ("Save Data to File" in "File" menu) and make a write protected copy of it. (Copy additionally a backup on external media; USB stick works well, but even a diskette is sufficient to hold many versions of the only 10KB small file.) Do not accidentally select "Load Data to Set"; it may overwrite your monitor data with garbage. Especially do not accidentally select "MPU" - it would format your monitor EEPROM and reset everything to awful factory defaults.
WinDAS does not work offline, but like a stupid TV remote control most of its functions directly change the monitor's internal register data without asking and without any warnings! Thus unless you have made the backup, do not play or even test any other WinDAS functions while the monitor is connected, because this easily makes the monitor unusable until the backup file is uploaded back into the monitor.
To see what is in the file, select the menu "Help/Expert/View" and select your backup from the file requester. You can read but not edit register names and values in the file here. (This is the only function that behaves like a normal Windoze program. Stupid is only that you can not select it unless the monitor is connected.) To edit values, overwrite them with a hex editor (not Windows "Notepad" - you are not allowed to change the file size or add unexpected characters).
With my SGI GDM-5411 the G2 value showed up even as the 1st number on the list. It had been 171, so I set it lower because my monitor was way too bright and full of diagonal retrace lines. (A too low value can turn off the monitor picture, so using a 2nd PC (e.g. laptop) for WinDAS is a good idea.)
All higher WinDAS functions are only contained in so-called "procedures", those behave stubborn like menus in a novice's C64 BASIC program or old DOS .BAT file installer. Procedures are basically batch files those request you in a fixed order to use certain screen tests and make certain adjustments (either on screen using windows sliders or by physical adjustments inside the monitor) and measurements. For the screen tests you need an external program (use a different PC to avoid trouble) or signal generator, which is not part of WinDAS. But the most annoying thing is that you can not switch back to a previous step when you accidentally failed or skipped one, and since procedures (white balance...) can take several hours and only ask you at very few spots whether you want to redo something, any wrong button click can waste hours of work.
IMPORTANT:
After successfully finishing any procedures, you must execute the "Final setting" procedure (before you quit WinDAS or unplug ECS cable) to save their results into the monitor's EEPROM. Else they stay only in RAM and will be lost once you power off or unplug the monitor, and also the monitor's menu will stay locked. After "Final setting" also do "Safe Data to File" to backup your work.
All available procedures are shown in a clickable list accessed through the "Adjustments" menu. For novices confusing is that their order changes because after use they are sent to the bottom of the page (likely to memorize which ones you have done), which at least with "Final setting" makes no sense since this one should be executed after finishing each lengthy procedure to avoid data loss.
- colour calibration trouble
I have gone through the white balance procedure to calibrate my SGI GDM-5411. As a colorimeter I used a SpyderTV (Spyder II hardware) and the great HCFR Colorimeter open source program. WinDAS requests you to show screen tests in 1280x1024@75Hz mode. I employed Nokia Monitor Test v1.0a and a homemade black and 30IRE grey PNG picture on IrfanView (on black desktop background) together with small HCFR window to see the xyY values in "Continuous measures" modes (those built in HCFR are fullscreen and can not be used simultaneously). With 6500k and 5000k mode I have set HCFR white point to D65 and D50 mode to see correct RGB % values; unfortunately for 9300k there is no matching (D93?) white point, so I selected sRGB. (In Windows I removed all colour profiles and gamma settings.)
The white balance procedure asks in a sequence for parameters of all 3 colour modes (9300k, 6500k, 5000k) and so is way too lengthy to be reasonable. (It took about 3h; why didn't they make 3 procedures of these?!) The first adjustment it asked for is the infamous G2 setting. Originally mine was 171 (way too bright with visible retrace), so I first tried 125 (grey looked still washed out) and in the 2nd try 114 (much better black) using the "Animated black" walking bar screen test of HCFR and black background from a DOS fullscreen text mode. In a dark room especially the visible scanlines of DOS and other low resolutions are well suited to see the black level. The darker bar in "Animated black" should be really hard to see. This way I got a nice black level, but unfortunately there is plenty of other trouble.
Before adjustment, my sRGB menu requested to set brightness 02 and contrast 00 (was ridiculously dark), now it wants 37 and 75, which looks less horrible. At least it already looks like a monitor again and not like "Who in this drive-in cinema has his xenon fog lights on?!"
With darker colours the Spyder II probe is infamous to output nonsense RGB values, but I first got rubbish partly up to 50IRE (yes, baffle removed and CRT mode). Now I still have at sRGB e.g. blue 140% at 0IRE, 120% at 10IRE, 110% at 20IRE. Much worse is that many requested Y (luminance) values are impossible to adjust even with the shown sliders at maximum. I reduced the IrfanView screen test window to 1/4 screen size to avoid the beam current limiter (so some modes indeed were bright enough to blur letters at max contrast) although according to WinDAS manual's signal generator description they seem to be supposed fullscreen. Especially with the last setting for each colour I got way less (e.g. 61 instead of the requested 90 cd/m²) with the CON_MAX_BR... slider (aka CMaxBCont in description) at 255 (maximum). My gammut triangle in CIE chart shows x0.561,y0.356 instead of x0.640,y0.330 (DeltaE 49.0) due to way too little red. Green has DeltaE 21.6, only blue is close. Strange is that red surfaces IMO look rather too bright (like neon colours, a tomato shines like a red traffic light) which may hint to poor linearity. In forums people claim that 6500k (sRGB) would be the best colour temperature (TV uses this) because it was modelled after sunlight. But with the luminous red IMO only 9300k looks like daylight and 6500k look really obnoxious.
Is really my CRT dead (worn out by overbrightness, or vice versa??), or is this the video amp or other electronic faults? :confused2
On dark grey screen test I see 2 symmetric crescent shaped smeared shapes (like ")(" ). Are these beam reflections inside the CRT or something worse? There is also occasionally slight colour impurity at the left side.
- dynamic convergence?
Despite I have tried hard to adjust convergence with the few available settings in WinDAS, especially the upper left corner has blue 2mm shifted to the left. Generally the top 3cm are slightly off. I hope I haven't messed up the magnet rings by vacuuming dust out (I cleaned everything with a soft paintbrush + thin vac hose), but I think it always was like this.
If I remember well SGI monitors were at that time considered absolute rocket science high end hardware, but the DCNV menu in WinDAS is greyed out. Has this monitor really no dynamic convergence, or needs WinDAS a special file update to handle this?
MAY THE SOFTWARE BE WITH YOU! *============================================================================* I CYBERYOGI Christian Oliver(=CO=) Windler I I (teachmaster of LOGOLOGIE - the first cyberage-religion!) I I ! I *=============================ABANDON=THE=BRUTALITY==========================* {http://weltenschule.de/e_index.html}(CRT (or flu tube) light has a spiky band spectrum that is very different from the continuous spectrums of daylight. Phenomenons like red-green blindness prove that human eyes can differ very much, so there can never be a perfect standard for a simple colour model like RGB.)
When cold, my monitor is still too bright (and greenish) and it takes about 30 minutes to stabilize.
- May it be that the "drift correct" transistor is a better spot to modify the circuit?
When my monitor was cold, pulling its output to chassis GND immediately removed the overbrightness and retrace lines. I am not sure if with even more worn monitors it would be safe to pull the output line even below 0V chassis GND (like done with tube grids in many tube amps) or if this would kill the RGB cutt-off IC by latchup.
- Has anybody tried to heat individual parts to check which ones causes the bright warmup problem?
I had noticed that with my adjusted SGI GDM-5411 the red channel saturation was too high (+12.91% at 50%), which caused obnoxious glowing red. So some hours later I backed up the monitor EEPROM data with WinDAS (which resets the warmup timer :< ) and after re-warmup I did the infamous colour return on normal and sRGB mode. I measured again and noticed that the red saturation flaw now dropped to (still visible) 8.82%. Also other values have slightly improved. Only the colour temperature has slightly dropped and green level has slightly increased (also may be that colour return fixed just corrected a Spyder II inaccuracy), but generally the grey scale delta E and RGB levels have improved a bit.
So I backed up the EEPROM data again and compared both versions with my HxD hex editor. The bizarre thing is that they are almost identical. I only found a "2" instead of "0" at byte 397, which is REG value 48 "COLOR_MODE". It may be that this is just my toggle between normal and sRGB colour mode was indicated by it. Because a complex mechanism like colour return can not hold its data in a single byte, it obviously either stores its data elsewhere or does something completely different. Reports that overuse can damage or wear out the CRT hints to that it may be rather a CRT regeneration process based on controlled cathode overload instead of a simple measurement of RGB currents. But I think a CRT regeneration process would take longer than the few seconds.
So I conclude that Sony either considered the measurement data useless to be accessible through WinDAS or they again placed here one of their infamous anti-reverse-engineering features to prevent competitors from analyzing and copying the internal algorithm. Can you remember the complex copy protection tricks in their PlayStation 3, which first version allowed the user to install Linux but completely enshrouded the inner working of the 3D graphics by abusing one of the 8(?) CPU cores as a MMU, and how Sony rapidly cancelled the Linux feature once hackers attempted to find out ways to circumvent this? This attitude all reminds to the lyrics of an old German song: "Ich bin die fesche Lola, und zeig Euch was ich kann, doch an mein Pianola, da lass ich keinen ran..." (translated: "I am the sexy Lola, show proud what I can do..., but my player piano - no one can touch it!")
It may be that (e.g. to increase access speed) the colour return data is stored outside the normal EEPROM, e.g. directly inside the "RGB Drive" IC402 or "RGB Amp" IC 403. Also the "Input Select" IC401 looks suspicious because it has R_DET_IN, G_DET_IN, B_DET_IN each connected through a capacitor to the corresponding output, which looks like a measurement loop. If the algorithm simply measures RGB output currents and stores a curve to compensate, it may push up a scaling factors once a value in the curve reaches maximum (e.g. 15 in a 4 bit value) and runs out of range. And the algorithm may be a too simple linear approximation of something non-linear, that attempts to lower a statistical mean deviation without testing if dark values turn too bright, which makes it run into the infamous overbrightness (after the warranty is over).
- Has anybody reset or replaced one of these ICs and checked how this affects overbrightness?
I have the exact issue on both my p991s. I always assumed it was normal because there is a 30 minute wait period before you can select 'Color Return' or 'Image Restore.' People have said that this has always been an issue since the monitors were brand new. If simply swapping a resistor fixes it I may do it as well.
If you find the proper resistor can you point out where you ordered it from? Initially when I got my Trinitrons I was going to do the resistor fix but I couldn't find anyone who had the correct resistor.
Anyway, i'm currently caught up with a lot of stuff. It'll have to wait a while (again).
Need some help nothing to big.
I have the 24' Sony GDM FW9012
I did a big no no by adjusting parameters in my original dat file and now i don't have the stock .dat file.
Where did the stock file go you ask? Of course i saved it but by mistake i modified it and now i need the stock file.
If anyone has it and could help me it would be grately appreciated.
I do have a P1130 coming my way, so i'll have an usable monitor while i tinker with the other one... stay tuned, lots of goodies coming up.
I too am having some issues on my p991's as they both have some arcing on the high voltage somewhere. Hopefully there are no shorts inside the picture tubes. I'm gonna get some silicone and cover the joints on the flybacks and the anode connectors on the flybacks and the tubes and see they stop arcing. I have a Viewsonic G810 that does the same thing as well but it's much further along on the issue so I'll try this quick fix on the Viewsonic first. Hope I don't get zapped!
If it's the anode lead, it's unlikely to arc at the connector. Try covering the cable with some insulation.
Fortunately all i had to do was remove 3 screws to get access to the A board, they didn't solder the shield in place like they do on other monitors. Kudos to Dell/Sony. There was a huge amount of dust bunnies so out with the vacuum cleaner. Much better... Now i'll start probing around and see what happens. I know where the fault likely is - i just need to confirm it.
I have found voltages in the same ballpark as CYBERYOGI did on the weird drift correct circuit on the cutoff amp. So at first sight it would appear there's something wrong with that. However, the cathode voltages don't really drift either. That leaves only one variable - the HV. If you think about it, it makes sense. The "warming up" period affects both brightness and picture size. The only voltage that can be responsible of this is the HV.
Taking a look at the D board schematic turns out we actually get a HV adjust pot - but since i do not have a HV probe it isn't a good idea messing around with that. I'm cheap - so i'll DIY myself a probe, and then if i confirm that it is indeed the HV which is drifting, we can go ahead and find a solution.
I'll be taking apart the P1130 too and see how things look like on that one.
Edit: The P1130 has a different internal construction and board layout, even though the parts are more or less the same. Also interesting to note is that the mysterious drift correct circuit isn't present on the P1130. We'll see what that gives us...
Edit 2: Got the P1130 aligned. Convergence ended up better than my eyes can notice - and this is no small feat to achieve. It's hands down the best monitor i've ever had so far. Looks crystal clear at 1600x1200 85Hz. There's some slight horizontal blurriness at the very sides but even with that it's better than the P1110.
And from what i've seen so far brightness doesn't drift at all on this P1130... I think i'm starting to connect the dots here. For optimal brightness range and picture clarity, the G2 on the P1130 ended up almost 100v higher than on the P1110, and even so i'm running brightness at 65 as opposed to 52. So, that extra brightness on the P1110 must be coming from somewhere... CYBERYOGI mentioned that messing with the drift correct circuit lowers brightness, so i'll go ahead and bypass the whole thing, readjust G2 and see where that gets me.
Edit 3: Bypassed the drift correct circuit on the P1110, upped the G2 using WinDAS to compensate. Guess what? I finally got a normal monitor! No more excessive brightness at startup, yippee!!!
The two monitors ended up with pretty much the same G2 voltage (albeit with slightly different WinDAS values), so clearly, something must be right in here.
Slight horizontal blurriness is the same as with the P1130 so it's either the cable or the laptop video card - i borrowed my main computer to a friend for a while, and i remember the P1110 didn't have this kind of focus issues, so i'll report when i get my main rig back.
First I gotta locate that drift correct circuit. Do you have pics of where the drift circuit is on the p1110? Also what did you do to bypass the circuit?
But I found in this FAQ some interesting info about CRT aging:
http://www.repairfaq.org/sam/crtfaq.htm#crteec The RGB output amp transistors in some monitor/TV circuits apparently act as a high resistance current source, i.e. a worn out CRT draws less current by higher impedance, which makes the output voltage rise and even can make the transistors saturate and clip the signal (i.e. ugly streaks at the right side of bright image parts). So apparently a worn CRT can indeed turn brighter and not only darker.
What you have to do is connect pin 17 of the cutoff amp directly to ground. The transistor is the tiny SMD type, so a little piece of wire and a bit of solder is all you need.
If you can post a picture of the A board with the shield removed i can show you where it is. I'll post a pic of mine later.
@ CYBERYOGI: The cathodes emit more electrons as the voltage is lowered - as they go high impedance and draw less current and make the voltage rise, the picture goes darker, not brighter. In fact a common test when diagnosing TVs which are missing a color is to momentarily short each cathode to ground, and a very bright screen of each basic color should appear. If all three cathodes respond to this, then the problem is in the video amplifier, if not, the tube is gone.
Another observation: Normally when you wake the monitor from standby you hear the relay click at the exact second the light flickers from amber to green and everything turns on at the same time. Then you see an image about 6-7 seconds later. My secondary monitor, however, doesn't always do this. Instead, I see the light change from amber to green but I don't hear a relay click until about a second later. When the relay clicks then I see the power light, which is green right now, flicker and then I hear the tube fire up. When this happens when the monitor is cold it takes at least 60 seconds before I finally see an image. What is possibly causing this? Should I worry about it?
With my P1110 for instance, if it's been really cold (i've been out all day today), black appears slightly green at startup and hence the picture appears brighter. This normalizes in less than 10 minutes - that's lots better than having brightness drift for 2 hours.
For the second monitor, you may have a problem with the power supply but it's likely something simple like a cold solder joint. I'd try and touch up the soldering, specifically all the joints around the CRT neck, all big semiconductors and relays.
Since you will have to up G2 quite a bit after eliminating the drift correct circuit, make sure you adjust it so optimal picture is at 50% brightness or below. I've talked about it sometime earlier, these monitors are wired to automatically lower the contrast when you up the brightness past 50%, to keep focus in check. Trouble is, this also shifts the color balance a bit, which is annoying. Better avoid it by setting G2 so that picture is bright enough before you reach 50% brightness in the menu.
Also i'm not sure anyone has mentioned this before: Always remove the ECS cable from the monitor when you're done with WinDAS. Keeping it in causes interference with the 5v rail which shows as slight wobbles in horizontal size.
Still works, but you can see jittering in text and small lines.
It's been disconnected for a year now.. Man it's been a long time since I switched to an LCD. Damn thing is still on my desk because I can't even safely lift it off. (Was for running classic games on, but I can't look at CRTs anymore, too much eye damage done, strain is too much now.)
Now if you'll excuse me, i'll be back to my game... Hey look, the picture's so clear i can see the scanlines!