CPD1604S Vertical Collapse

[vol.2]

So after a little while using this monitor, I have realized that the HV regulation seems funny. I noticed it was bouncing in and out a little bit on the sides when it's cold, and sometimes.

As a test, I use the HV Regulation Pattern, which is a white box that flashes inside the middle of the screen. With this I discovered that the raster is stretching pretty significantly in the horizontal direction during flashes.

I began looking for a problem in the Horizontal, and subsequently discovered there's some problem with the 160V source. I see 155V on the 160V line. There must something blown up in there?

There are a few diodes and such between them, but I am wondering if anyone see anything as a smoking gun?

[old_tv_nut]

You need to look for some voltage that changes when the width jumps. A steady 155 volts may just be due to component tolerances and not something failing.

[Alex KL-1]

+1 for watching voltages.
I can add, if PSU caps are good, I fear about the pulse caps in the HOT. Some monitors have switching for different resolutions (adding more cap for lower resolutions=lines), for tuning the flyback, and perhaps one can be "added" at these flash times. But, the image will be dim if this occurs (at the moment).
Also, if image srhinks at that time, can be a failing pulse cap. This can indeed produce/induce a high output voltage from flyback and produces a flash somewhere (depending about the severity of the failing cap).
Investigate it only if voltages are all stable at the flash times.

[vol.2]

Originally Posted by old_tv_nut

You need to look for some voltage that changes when the width jumps. A steady 155 volts may just be due to component tolerances and not something failing.

Originally Posted by Alex KL-1

+1 for watching voltages.

Okay, I took readings at the 60V chopped voltage with feeds Q403 and to the collector of Q402, which I believe imparts the horizontal drive signal.

I also watched the 155V (160V) during the flash test.

Voltage stayed fairly steady at 155V during flash test. Dipped down maybe 200mV at most.

Both waveforms (Q402,403) stayed completely steady during the flash test. No movement whatsoever.

What was more interesting is that the waveform #2 which is supposed to be 660Vpk is more like 1.1kVpk.

This seems excessive. Although someone else has told me that 1kV is probably normal for a color monitor, even though the SM says 660kV. I don't know what to think about it.

I had to use a 100X probe for the readings on my scope, but both readings were taken with the same scope and probe on the same settings, so unless the scope or probe is somehow going to start misbehaving at a higher voltage, it's accurate (or service manual is wrong).

Originally Posted by Alex KL-1

I can add, if PSU caps are good, I fear about the pulse caps in the HOT. Some monitors have switching for different resolutions (adding more cap for lower resolutions=lines), for tuning the flyback, and perhaps one can be "added" at these flash times. But, the image will be dim if this occurs (at the moment).
Also, if image srhinks at that time, can be a failing pulse cap. This can indeed produce/induce a high output voltage from flyback and produces a flash somewhere (depending about the severity of the failing cap).
Investigate it only if voltages are all stable at the flash times.

I measured all the capacitors in the path of the horizontal yoke and it's related circuitry (basically about 15 or so large film caps between the collector of H OUT transistor and the H Yoke coils). All capacitors are perfect, one cap is 10% out, and the rest are within 1%.

[vol.2]

After getting it back together, I see a slight reduction of the horizontal breathing during the flash test. I also saw that the Pincushion settings had changed a small but noticeable amount.

I believe that I must have stumbled on a poor solder joint or something that was causing a little bit of issues with the Horizontal, and reflowing a lot of that section helped.

It's still not "perfect" though. I'm seeing two remaining flaws.

1) the top of the raster is still a little pinched inward under optimal settings, on the upper left-hand corner. not sure how to fix it, but the "Pin Up" control is all the way down to get it to the best possible setting. I would assume that the default setting for the Pin Up potentiometer is center, but I have to turn it all the way down to achieve the best outcome. This seems suspect to me, and possibly indicative of another issue. I traced the signal from the Pin Up control, and it runs to Q201 on the D board. This is labeled V RAMP. The collector of Q201 measures 11V steady instead of 11.4V, so I just kind of let it pass, but maybe it's worth checking the other voltages? I have it closed up, so somewhat reluctant to open it again for a long shot.



2) the brightest whites don't get as bright as would expect them to before I start to get color smearing. This only seems to happen with the red, and that seems to be par for the course with 90s VGA monitors. I understand that red takes the most juice to drive or something, so it tends to crap out before the other colors.
However, it's happening right at the edge of what the SM states this monitor can handle, and I've generally seen it be okay with some headroom.

SM states it can handle 96-106 Nits, and I think the last time I did a white balance, I got it about 100ish Nits before it started to smear red towards the right of the whitest whites. It's not exactly dim, but it seems a little skimpy.

So my question to all here is: can I do anything to help with the red smearing, or is that just a function of the phosphors and nothing I can do? I'm new enough to this, and I've heard enough conflicting information about it to not conclusively know if it's fixable or not.

For example, could upping the value (or lowering the value) of a cap somewhere tighten up the color resolution at higher voltages?

[zeno]

The red bleeding is probably the CRT is weak. If it gets better with warm up & looks better with the brite & contrast reduced, CRT is most likely the suspect.
Quick test: Usually there is a low ohm resistor ( typ 1 ohm, 1W +-) in series with
the filament. Jump it out & see if it improves. That will increase the filament
from 6.3 to @ 8V & heat up the cathodes more. There is a very small risk
to this, its a last ditch way to get more life.

73 Zeno
LFOD !

[vol.2]

Originally Posted by zeno

The red bleeding is probably the CRT is weak. If it gets better with warm up & looks better with the brite & contrast reduced, CRT is most likely the suspect.

It doesn't change with warm-up, and brightness has zero effect on it. I only see this when I push the contrast up high. There's a certain point at which if I turn the contrast knob up, the white solid areas ghost red to their right. I can manage at least ~103 Nits, which is already in line with what the manual says it can do (it says to lock the contrast down to 96 Nits +\- 10).

If I turn up the contrast to just before it starts to bleed, and I crank the brightness, it does not bleed; it goes right on getting more washed out until I can see retrace lines.

I have read in some places that this effect can appear when you have too much length to you cables or use an extension and is thus a bandwidth issue. I don't have an extension though, this set has a built-in cable.

[old_tv_nut]

If you turn up contrast just to where it does bleed, and then turn brightness way down, does it go away?

I don't know, but it's possible it has a beam current limiter that works to turn down contrast as you turn up brightness, and that's why the brightness can't cause smearing. Or it could be something else peculiar to this chassis that I can't think of.

Anyway, the red smearing is most often an indication that the red video output stage is saturating (going to the lowest possible output voltage), often due to the setup voltages required with a tired tube.

[zeno]

On the red problem it may be bad filtering on the @200 V supply for the
video outs. This is a common problem on all brands ! Symptoms are:
Left side normal, right side too brite.
Streaking.
Shading.
Jail bars.
Pix too bright.
Retrace lines in some or all of pix.
Looks like bad CRT
+ more I cant remember !
Poor px, just dont look right.Any combination of above.

Usually its a 4.7mfd or 10 mfd on the @200 V.
G-2 setting must be right also. Be sure you can extinguish or nearly so
the raster & have NO retrace lines.
Now back to the @200 V. On most the neg goes to GND. On some
the neg goes to the main B+ then there is another 'lytic neg to GRD.
That can cause width & jail bars, plus all the above.

Time for a bier
Zeno
LFOD !

[vol.2]

Originally Posted by zeno

On the red problem it may be bad filtering on the @200 V supply for the
video outs. This is a common problem on all brands ! Symptoms are:
Left side normal, right side too brite.
Streaking.
Shading.
Jail bars.
Pix too bright.
Retrace lines in some or all of pix.
Looks like bad CRT
+ more I cant remember !
Poor px, just dont look right.Any combination of above.

Usually its a 4.7mfd or 10 mfd on the @200 V.
G-2 setting must be right also. Be sure you can extinguish or nearly so
the raster & have NO retrace lines.
Now back to the @200 V. On most the neg goes to GND. On some
the neg goes to the main B+ then there is another 'lytic neg to GRD.
That can cause width & jail bars, plus all the above.

I tested C421 (10uf on the 200V) when I did the whole set and it looked good, but maybe it's a sneaky failure?

In any case it was one of few that wasn't replaced. I guess I can try that.

One interesting thing I found is that the far right side of the raster has bright fringing on the edges when the G2 is on the edge of making the raster glow, and it seems that things get weird if I push the brightness past that point. It's a really fine line where I can just barely get the brightness up past a dead black background and the image gets a lot worse and there's weird (and irregular) bright fringing on the right edge.

If I keep the G2 at right around dead black, the fringing goes away and the picture cleans up a lot (meaning I can get more headroom out of the contrast with a sharper picture). After I got it to that point, I was able to use a colorimeter and software to create a color profile that gamma corrects the monitor, so now the white point is correct and the picture is ramping the brightness correctly across the board. Looks a whole lot better now than it has at all so far.

Originally Posted by old_tv_nut

If you turn up contrast just to where it does bleed, and then turn brightness way down, does it go away?

No, the brightness doesn't seem to have an effect on the red ghosting/bleeding.

I don't know, but it's possible it has a beam current limiter that works to turn down contrast as you turn up brightness, and that's why the brightness can't cause smearing. Or it could be something else peculiar to this chassis that I can't think of.

That could be true. It's got an automatic contrast control knob that I am supposed to set in the white balance procedure. It wants me to start off with everything super maxxed out and then use the contrast limiter to adjust it down. But this doesn't seem to work out when I follow what they want me to do. It seems to look much better when I ignore the brightness limiter and just turn the G2 down and the sub contrast down until everything is reasonable.

]Anyway, the red smearing is most often an indication that the red video output stage is saturating (going to the lowest possible output voltage), often due to the setup voltages required with a tired tube.

It might be that, but also I might just be hitting some limits and it's a combination of things. As it stands now, it is plenty bright to use. I'm mostly now just concerned with the weird bright fringing on the right side of the screen. It might be worth replacing that 10uf cap on the 200V that measured good just in case.

[Alex KL-1]

Try to measure the DC voltages of the CRT cathodes. This can indirectly *sometimes* make a clue. If DC reading of the red cathode differs too much from others, something are ocurring. For example, a too low red cat. voltage can make the output stage saturating (on peaks/high luminance). More easy on monitors, since it have less dynamic range than a regular TV since it operates with lower voltages due to larger bandwidth (for not over dissipating).

[vol.2]

Originally Posted by Alex KL-1

Try to measure the DC voltages of the CRT cathodes. This can indirectly *sometimes* make a clue. If DC reading of the red cathode differs too much from others, something are ocurring. For example, a too low red cat. voltage can make the output stage saturating (on peaks/high luminance). More easy on monitors, since it have less dynamic range than a regular TV since it operates with lower voltages due to larger bandwidth (for not over dissipating).

Okay. I have done a number of different things.

I checked the cathodes, and they have a +50VDC offset. Not entirely sure why, but I think it might have something to do with the multisync nature of the monitor in that it operates the cathodes at a floating voltage. Perhaps it has something to do with the way that it overlays the video on top of the raster. Not sure, it's a little above my head at the moment.

But for whatever reason, it does float the voltage and it relies on the G2 to G1 to raise the pedestal of the image to the right level.

That means I could potentially push the whole voltage level down a bit and get more headroom.

I also bought a slightly higher bandwidth version of the video amp IC (it can do 100MHz vs 70MHz), and tried it out.

There were a couple takeaways from that. First, I had to install a green drive pot to match the red and blue because the new video amp has slightly different gain characteristics that don't work well with the existing circuit. I just matched the values of the other two and that seems to be okay.

Secondly, I started messing with the 33 ohm resistors on the TE pins of the video amp inputs. I understand that this is a HF compensation method and that it will alter how fast the video amp can function. However, changing these values also seems to have an effect on the clipping of the video amps.

You can see them on the right side of the video amp IC:



Raising the values of the red and blue channels, I am able to get more total brightness out of the raster without clipping from the video amp. I think there's something besides the total swing of the video signal on the amp that is contributing to the clipping. It seems like the TE pins are effecting the biasing of the IC as well.

The internal schematic of the Video Amp IC:



Unfortunately, the new amp doesn't seem to play nicely with the current circuit in other ways as it seems to have a different DC offset, and it ends up too dim when I adjust the G2 for proper black levels. Basically, I think it ends up with the pedestal too high up, and then the G2 has to be lowered a lot to compensate, which lowers the G1 G2 relationship and the overall luminance goes down.

I think after that, I realized something interesting the service manual. The color balance procedure has you max out the bias controls to start, and then to pull them down to balance. That means that you are effectively pulling the pedestal all the way down and then selectively pulling guns up to reduce their effect on the background. This IIRC is not how it's done in other monitors and TVs I've worked on; typically, they have you start in the middle or at the bottom and then work your way up. I have to assume that they are considering zero volts as the brightest possible bright, and this monitor expects that to be between 45-50V. Black levels are supposed to be about 90V and it's supposed to swing around 45V in the middle.


In any case, I'd love to figure out how to redesign the bias network to use the VPS10 (100MHz part) in this monitor. From my testing, it was much, much crisper of an image (at a similar level of brightness to the old video amp), even though it turned out dimmer with the present circuit.

This is the bias network circuit if anyone has any suggestions. I guess I would assume I would have to mess with these resistors in the voltage divider for the 180V to the bias controls. The new video amp looks like it maxes out at like 106V instead of 90V or something, so I'd need to lower the DC offset of the video signal by about that much. I'd still have the bias controls to do the fine tune, but it's apparently too far out.

[Alex KL-1]

Originally Posted by vol.2

Okay. I have done a number of different things.

I checked the cathodes, and they have a +50VDC offset. Not entirely sure why, but I think it might have something to do with the multisync nature of the monitor in that it operates the cathodes at a floating voltage. Perhaps it has something to do with the way that it overlays the video on top of the raster. Not sure, it's a little above my head at the moment.

But for whatever reason, it does float the voltage and it relies on the G2 to G1 to raise the pedestal of the image to the right level.

That means I could potentially push the whole voltage level down a bit and get more headroom.

I also bought a slightly higher bandwidth version of the video amp IC (it can do 100MHz vs 70MHz), and tried it out.

There were a couple takeaways from that. First, I had to install a green drive pot to match the red and blue because the new video amp has slightly different gain characteristics that don't work well with the existing circuit. I just matched the values of the other two and that seems to be okay.

Secondly, I started messing with the 33 ohm resistors on the TE pins of the video amp inputs. I understand that this is a HF compensation method and that it will alter how fast the video amp can function. However, changing these values also seems to have an effect on the clipping of the video amps.

You can see them on the right side of the video amp IC:



Raising the values of the red and blue channels, I am able to get more total brightness out of the raster without clipping from the video amp. I think there's something besides the total swing of the video signal on the amp that is contributing to the clipping. It seems like the TE pins are effecting the biasing of the IC as well.

The internal schematic of the Video Amp IC:



Unfortunately, the new amp doesn't seem to play nicely with the current circuit in other ways as it seems to have a different DC offset, and it ends up too dim when I adjust the G2 for proper black levels. Basically, I think it ends up with the pedestal too high up, and then the G2 has to be lowered a lot to compensate, which lowers the G1 G2 relationship and the overall luminance goes down.

I think after that, I realized something interesting the service manual. The color balance procedure has you max out the bias controls to start, and then to pull them down to balance. That means that you are effectively pulling the pedestal all the way down and then selectively pulling guns up to reduce their effect on the background. This IIRC is not how it's done in other monitors and TVs I've worked on; typically, they have you start in the middle or at the bottom and then work your way up. I have to assume that they are considering zero volts as the brightest possible bright, and this monitor expects that to be between 45-50V. Black levels are supposed to be about 90V and it's supposed to swing around 45V in the middle.


In any case, I'd love to figure out how to redesign the bias network to use the VPS10 (100MHz part) in this monitor. From my testing, it was much, much crisper of an image (at a similar level of brightness to the old video amp), even though it turned out dimmer with the present circuit.

This is the bias network circuit if anyone has any suggestions. I guess I would assume I would have to mess with these resistors in the voltage divider for the 180V to the bias controls. The new video amp looks like it maxes out at like 106V instead of 90V or something, so I'd need to lower the DC offset of the video signal by about that much. I'd still have the bias controls to do the fine tune, but it's apparently too far out.

Perhaps the VPS10 have less gain, setted internally. If is possible to raise the gain without altering the bias (perhaps a divider from the local supply to ground). But not entirely promising, since the IC have internal emitter R. Better is to provide more gain in the preamp.

About cathodes, ia relatively common to have this type of pedesta clamping, direct to the cathodes; the DC restoration being here and even in ultra hi-end and low end CRT monitors. What interesting, is like a return to the 1940's or suchlike (a simple diode DC restorer in the CRT section).

[vol.2]

Originally Posted by Alex KL-1

Perhaps the VPS10 have less gain, setted internally. If is possible to raise the gain without altering the bias (perhaps a divider from the local supply to ground). But not entirely promising, since the IC have internal emitter R. Better is to provide more gain in the preamp.

I took a couple of scope readings of the red cathode as a baseline, while viewing a color bars.

I started with VPS07T installed and got it as bright as possible with true black pedestal and no smearing caused by saturation of the amp.

The VPS07T was showing about 58Vpkpk with black slightly over 100V, and the whites at 45V.

Then I installed VPS10 and took the same measurement first without changing anything.

The VPS10 was showing about 48Vpkpk with the black at about 96V and the whites at about 50Vpkpk.

So at the same settings, the gain of the VPS10 swings a little lower, resulting in a lower voltage for black and higher peak voltage.

Then, I color balanced the monitor with VPS10 installed, and I was able to increase the swing all the way to 60Vpkpk, but to get it there, I have to peak the manual brightness control and lower the G2 such that the pedestal is still below the level of the VPS07T. It's not a huge difference, voltage-wise, but it seems to make a big difference in the overall brightness and the color linearity.

The final scope reading with the VPS10 tuned the best I can get it is about 97V for black, and 40V for whites. Any higher contrast than that and the colors start to clip and I get smearing.

So it seems that the black level (the pedestal) is about 6V lower for the VPS10 than the VPS07T, and that's what I'd have to find a way to shift.

Currently, the only thing I can do to shift the whole DC offset range is to lower the bias controls, but if I do that, then image smears at a much lower drive level. I think the output circuit is somehow designed to have the bias controls at the top of their travel, and the G2 is supposed to be set with the bias as high as possible. The Service Manual instructions state to do this, and it seems to be correct.

Aside from the Drive, Amp, and Bias sections, there is the preamp IC, but I don't have a datasheet for it, nor can I find one anywhere on the net. Its a CXA1209P, a Sony IC.

Perhaps the answer is to alter values to the blanking pins or something?

[Alex KL-1]

The bright pin have a fixed divider. Perhaps the IC have a pedestal clamp, independent of the fact of it being clamped at the CRT? Perhaps it have, for having more defined levels to the vide out.
So, *perhaps* altering the bright pin voltage the pedestal will alter. Is worth investigating.