If it works fine for a while then problem starts I would check for cracked solderjoints/reflow solderjoints and then check the capacitors.
Even if the capacitors looks fine they can be marginal and needs to be replaced. If you got a i circuit capacitor tester that will help alot!
CRTs can be quite sensitive with bad/marginal capacitors.

But first, check solderjoints. Thats easier. 😀

Also, many CRTs have service manuals with troubleshooting pointers.

CRTs can be dangerous, but just double check everything and take it slow. Safety first! Never work on it plugged in unless you REALLY need to.

If you can, also test with another video input source!

So I had a buddy over this weekend, and we took a crack at this monitor. The good news is, after taking off the cover, there really wasn't anything else that need be done. It's a single PCB, almost fully exposed, with all the soldering on the exposed side exclusively.

After spending an hour looking for cracked solder joints, and trying to test the capacitors in circuit, he noticed a heat damaged portion of the PCB around two transistors. One of them is definitely defective, with a short between the base and emitter. I'm lead to believe this is bad. The other doesn't seem obviously broken, but with the heat damage on the pcb around it, I ordered a replacement for that as well just to be safe. Gonna take a few weeks for the parts to get here, so we'll see how it goes then.

Also quite fortuitous, these transistors are on the outside of the PCB, easily accessible. So I'm not anticipating any difficulties when the replacements arrive.

Once you do, you will have blown components when that transistors shorted out, resistors and other items. Real reason of this happening usually are degraded capacitors and solder joints. Need ESR meter to check the health of electrolytic capacitors and close visual inspection of solder joints. Also be careful of board cracked, rare but does happen.

Need multimeter and ESR tester.

Cheers,

The capacitors that are needed to keep transistors alive are typically high-voltage film capacitors, not electrolytics. You definitely want to check the capacitors around the broken transistors for their capacity (you need to do so out of circuit), but you don't necessarily need to measure their ESR. While degraded electrolytics often are best identified by ESR measurement, for film capacitors capacitance measurement is the better diagnostic.

Be aware of fake short reading, though. In some monitors, there are transistors driven using a transformer with its secondary between emitter and base. The transformer secondary is a short circuit for DC, so you might see a short measuring in-circuit although the transistor is fine.

Yes, electrolytic and solder joints are usually the two main causes of killing transistors or IC then other items burn up in the shorted circuit area next. The other one that kills the horizontal scan transistor is faults in horizontal deflection and driver circuits: bad solder joints, dried up electrolytic capacitor or blown flyback transformer, the first three main causes of faults. Bad deflection yoke coil can happen but not as often.

Second most common fault is vertical IC blown by solder joints or their kick (return to top scan kicker), electrolytic capacitor direct connected to the vertical deflection IC and same capacitor is powered via a diode from a power circuit that also powers the vertical IC as well. When vertical IC goes, it will take out a fusible resistors in the power circuit supplying the vertical circuit, and usually also takes out the diode as well in the power circuit (from power supply or from the flyback transformer's secondary windings.

I often see people try to replace something and fail to account for all the issues to be checked or replaced and replacement blows again.

I was electronic repair tech back in the day in trail end of CRT era and beginning of LCD era from 2003 thru 2010. I know this.

One of my inherited 17" monitor in 2000 or so, was infamous Mag model with a design is considered flawed, that kills the horizontal transistor and one more transistor due to control circuits degraded by a electroytic capacitor. That thing kept doing that few times till I had best grade capacitors replaced stopped doing that since, because I was using bog standard caps which is not low ESR to begin with. Then it lived to old age, CRT go dim with age for so many years later.

Cheers,

mkarcher wrote on 2022-10-23, 21:57:

The capacitors that are needed to keep transistors alive are typically high-voltage film capacitors, not electrolytics. You definitely want to check the capacitors around the broken transistors for their capacity (you need to do so out of circuit), but you don't necessarily need to measure their ESR. While degraded electrolytics often are best identified by ESR measurement, for film capacitors capacitance measurement is the better diagnostic.

Be aware of fake short reading, though. In some monitors, there are transistors driven using a transformer with its secondary between emitter and base. The transformer secondary is a short circuit for DC, so you might see a short measuring in-circuit although the transistor is fine.

So, it's been a while. I took your advice, and looked closer after removing the transistors. The transistor was not actually short, but as you describe. Or at least close. It had a 4 ohm resistor between base and emitter, which was enough for my multimeter to beep. I removed that and tested that it was indeed 4 ohms and supposed to be 4 0hms. Even so, heat damage in the area was evident, I had already ordered replacement parts for 25 cents a pop, so I went ahead and replaced them anyways. Cleaned up the old thermal compound, replaced it with something fresh, and things seem better?

I'm learning my way through this, and I'm assuming, as the poster after you mentioned, there is likely something up stream of the transistors causing problems. So when I ordered replacements, I ordered a bunch. I figured why not, might as well make the shipping worth it, right?

I left the monitor on at max brightness for a good chunk of time. I might have seen it dim again, but I'm not 100% sure? Before it was impossible to miss, got bad quickly, and was staying bad longer and longer. This time around I'm barely certain I actually saw anything happen.

This is probably a stupid question, but is it at all possible that the thermal compound between the transistor and the heat spreader had dried out so much, they were overheating and drifting out of spec? Is it remotely possible replacing the thermal compound on the other transistors attached to the head spreader could fully resolve the issue?

I'm not turning up anything new on visual inspection, so if the problems persist, I guess I'll be removing capacitors to test out of circuit.

Namrok wrote on 2022-11-14, 04:17:

So, it's been a while. I took your advice, and looked closer after removing the transistors. The transistor was not actually short, but as you describe. Or at least close. It had a 4 ohm resistor between base and emitter, which was enough for my multimeter to beep. I removed that and tested that it was indeed 4 ohms and supposed to be 4 0hms. Even so, heat damage in the area was evident, I had already ordered replacement parts for 25 cents a pop, so I went ahead and replaced them anyways. Cleaned up the old thermal compound, replaced it with something fresh, and things seem better?

It's normal for some components to run hot in a CRT monitor. If a transistor has 4ohms "pull down" resistor between base and emitter, this resistor is going to drain like 200 milliamps when the transistor is turned on, so the transistor itself likely gets at least as much base current, which is an indication that it carries several amps collector current. Probably it's the main drive transistor for the horizontal transformer. It's mounted on a big heat sink for a reason. It's normal that the PCB gets darker in the hot area, it does not necessarily indicate "damage". On the other hand, the hotter components run, the faster they wear, so inspecting the areas of the PCB that get hot is generally a good idea.

Namrok wrote on 2022-11-14, 04:17:

I left the monitor on at max brightness for a good chunk of time. I might have seen it dim again, but I'm not 100% sure? Before it was impossible to miss, got bad quickly, and was staying bad longer and longer. This time around I'm barely certain I actually saw anything happen.

This is probably a stupid question, but is it at all possible that the thermal compound between the transistor and the heat spreader had dried out so much, they were overheating and drifting out of spec?

The hotter a transistor gets, the less good it gets at switching. A worn transistor or a transistor with insufficient cooling can gradually operate worse as the monitor warms up. So it's indeed possible that bad thermal coupling between the transistor and the heat sink is a cause of the picture getting dim. On the other hand, thermal compound doesn't need to stay wet after the stuff is assembled. The main point of thermal compound is to make sure there is no air between the transistor and the heat sink, because air is a terrible heat conductor. If the transistor is firmly mounted to the heat sink with no air inbetween, no air will get in - unless you get mechanical stress moving the transistor. Rough transportation of a monitor might cause the transistor to "tear off" the dried heatsink compound at get some air in, although typically the mounting of the transistor to the heat sink is firm enough that it will stay firmly attached to the heat sink.

Namrok wrote on 2022-11-14, 04:17:

This is probably a stupid question, but is it at all possible that the thermal compound between the transistor and the heat spreader had dried out so much, they were overheating and drifting out of spec? Is it remotely possible replacing the thermal compound on the other transistors attached to the head spreader could fully resolve the issue?

Possibly yes. Replacing the thermal compound won't do any damage, but the smaller the transistors get, the better they stay attached to the heat sink, and the less the chance that the thermal coupling gets bad over time. In case the issue re-appears, you definitely need to check the capacitors near that transistor - not only the electrolytics, but also the film capacitors. At the end of every scanline, a hefty voltage spike occurs at the horizontal output transformer. This voltage spike is required to quickly re-magnetize the horizontal deflection and something that can not be avoided. The voltage spike is caused by a certain amount of energy getting dumped into film capacitors. If those capacitors degrade and lose capacity, the same amount of energy generates a higher voltage. The transistor can only withstand a certain maximum voltage and will quickly degrade and get hot if that voltage is exceeded. So if the symptom re-appears shortly after replacing the transistor, you really need to check those capacitors, because bad caps can quickly kill the replacement transistor.

Ok, well, I spent some time on it today and the problems began to re-emerge more obviously again. So I'll be removing the film capacitors and testing them. Any advice on picking suitable replacements? Or is anything that matches the footprint, capacitance, voltage, and anything else printed on the packaging good to go?

Namrok wrote on 2022-11-14, 21:10:

Or is anything that matches the footprint, capacitance, voltage, and anything else printed on the packaging good to go?

Basically, voltage and capacitance are the primary important parameters. Slightly higher design voltage doesn't matter. As the charging current during the horizontal retrace can be quite high, "impulse type" capacitors that are meant for short current spikes will likely help endurance - but when you pick a capacitor with a similar footprint and volume, you are likely getting capacitors rated for enough impulse current. Don't pick capacitors that are significantly mechanically smaller than the ones currently inside the monitor, though. Film capacitors did not get that much better over the last 30 years. If the capacitor is smaller, it is less robust, and robustness is important.

Ok, sounds good.

Thanks for all the help and advice in this thread. A lot of it's going over my head, so I appreciate the patience shown.

Ok, I removed the film caps that were most directly traced to the transistors in question. Measuring their capacitance out of circuit seemed to check out? But I don't know how that relates to their voltage rating, which if I understood you correctly could be the failing specification in this instance? I attached a picture of them.

So for example, the one labeled MPPS472G, I believe I found it's datasheet. Can't seem to find anywhere selling it though. I found one on mouser that looks to be the closest match, right down to the flame retardant epoxy material. But there were many more, and cheaper versions, that appear to match the specifications just as well, but come in a "self-extinguishing solvent resistant plastic case". Both the datasheets on mouser list the capacitor as being appropriate for deflection circuits in televisions. How hung up should I be about the material of the packaging if the specifications, so far as I can discern them, seem appropriate?

For the 5100pF capacitor, I can only find 4700pf or 5600pf at 630V. I'm assuming when in doubt, go higher? Alternately, I found a 5100pf at 800V? Is stepping up to a higher voltage preferable to messing with the capacitance in this application?

Namrok wrote on 2022-11-20, 19:16:

Ok, I removed the film caps that were most directly traced to the transistors in question. Measuring their capacitance out of circuit seemed to check out? But I don't know how that relates to their voltage rating, which if I understood you correctly could be the failing specification in this instance? I attached a picture of them.

It's missing capacity that can kill the horizontal output transistor. You measured the caps and their capacity is fine. If they fail under full voltage, they start losing capacity. That's something you would have seen on your tester, even at low voltage. So I conclude that these caps are most likely not the reason why your monitor fails.

Namrok wrote on 2022-11-20, 19:16:

I found one on mouser that looks to be the closest match, right down to the flame retardant epoxy material. But there were many more, and cheaper versions, that appear to match the specifications just as well, but come in a "self-extinguishing solvent resistant plastic case". Both the datasheets on mouser list the capacitor as being appropriate for deflection circuits in televisions. How hung up should I be about the material of the packaging if the specifications, so far as I can discern them, seem appropriate?

Capacity is the most important rating, with impulse handling next. Most capacitor datasheets list the impulse handling capability as dV/dt. In case of the Vishay cap, the specification indicates 10kV/µs as acceptable, whereas the KEMET datasheet only specifies 2.5kV/µs. That's why I wouldn't substitute the Vishay cap with the KEMET cap without proper research in the specific circuit. Let's do the research: A typical voltage to which this cap is charged during each retrace is 1.2kV. The highest horizontal frequency of that monitor is 70kHz, which means a retrace every 14µs. The retrace can be estimated to take around 10% of the scan period, so around 1.4µs. The capacitor is charged to 1.2kV and discharged again during that time, so a charging time of 0.7µs is a good assumption. To estimate the slope, I will round it down, because the charging rate isn't constant over time. At 1.2kV target voltage and 0.5µs charging time, you get a voltage rise of 2.4kV/µs, which is just in spec with the KEMET one, so if I were to replace the capacitor, I would buy one class higher, especially, as this calculation includes some hand-waving and gross estimation of values. A standard 15kHz television is far less demanding on the capacitor than a high-resolution computer monitor.

Namrok wrote on 2022-11-20, 19:16:

For the 5100pF capacitor, I can only find 4700pf or 5600pf at 630V. I'm assuming when in doubt, go higher? Alternately, I found a 5100pf at 800V? Is stepping up to a higher voltage preferable to messing with the capacitance in this application?

800V instead of 630V is not a problem. 4.7nF vs. 5.1nF is a problem. The monitor manufacturer has chosen the bigger value for a reason, and the value will affect picture geometry. Substituting a cap with one with a different capacitance can cause the range of the digital adjustment to be insufficient to get good picture geometry.

But, as I said, these caps are most likely not the issue, so I recommend to solder them back to the monitor.

You also said in the initial post that geometry is not affected while the problem appears. This should have pointed us directly away from these caps and the horizontal deflection transistor. If something fails in this region, you will get geometry issues like sudden picture size changes. If brightness is wrong, but picture size stays constant, either the "SCREEN / G2" potentiometer at the horizontal transformer or something on the neck board (like a "sub brightness" pot) should be among the primary suspects.

Would it have just been a coincidence then that after I replaced the suspect transistors, it seemed to work better for an hour or two before the problem manifest again?

Namrok wrote on 2022-11-20, 21:58:

Would it have just been a coincidence then that after I replaced the suspect transistors, it seemed to work better for an hour or two before the problem manifest again?

While you replaced the transistor, you had the monitor cool down completely. If the problem is heat related, maybe due to mechanical stress, having the monitor cooled down to ambient temperature completely might be an important factor. It might also be coincidence. I am confident that brightness problems rooting from actual issues in the horizontal deflection are always accompanied by geometry issues. Good and stable geometry is an indicator for a flawlessly working horizontal deflection circuit.

So, I found these potentiometers on the neck. They were labelled Yhc, Delta V and Yv. I googled a little and I don't think these are it? I found a giant solder joint at the very top of the neck labeled G2, with a thick green wire that goes straight to the flyback transformer I think? It vanishes underneath it without an exposed solder connection. Then on the flyback I found 3 more knobs labelled Screen, Focus, and then I can't see the top one.

So between the G2 solder joint and the Screen knob directly on the flyback, I think I found what you are talking about? I don't think I'm going to be cracking open the flyback transformer though. Well, maybe if I work up a lot of courage and someone tells me it's not as bad as it looks.

I guess I could try reflowing that big fat solder joint labelled G2? Or resolder it? Just on the off chance something about it is screwy and the heat is bringing it out?

Namrok wrote on 2022-11-21, 18:43:

So, I found these potentiometers on the neck. They were labelled Yhc, Delta V and Yv. I googled a little and I don't think these are it?

No, that's not the pots I was talking about. The pots you found seem to be for geometry, focus or convergence adjustments. Don't touch them unless you have specific problem. The "neck board" is not a PCB mounted along the neck of the tube, but the board that is plugged to the very end of the tube. Unless the monitor is fully digitally controlled, you should find internal pots to set the base level brightness there.

Namrok wrote on 2022-11-21, 18:43:

I found a giant solder joint at the very top of the neck labeled G2, with a thick green wire that goes straight to the flyback transformer I think? It vanishes underneath it without an exposed solder connection. Then on the flyback I found 3 more knobs labelled Screen, Focus, and then I can't see the top one.

The join is likely OK. The flyback transformer outputs a voltage (several 100 volts) for a grid in the CRT. This grid is called G2 (second grid) or screening grid. The knob on the flyback labelled "Screen" directly affects the voltage at the G2 solder joint. The G2 voltage is one of the primary factors setting the general picture brightness. I meant to suggest that you use an isolated screwdriver to slightly turn G2 and check whether you see sudden changes in brightness. If you do, the brightness issues are caused by the G2/Screen pot inside the flyback.

Namrok wrote on 2022-11-21, 18:43:

So between the G2 solder joint and the Screen knob directly on the flyback, I think I found what you are talking about? I don't think I'm going to be cracking open the flyback transformer though. Well, maybe if I work up a lot of courage and someone tells me it's not as bad as it looks.

No. Don't crack open the flyback transformer. The only thing you are supposed to do is careful adjustment of the screen knob. If touching/moving the screen knob doesn't fix the brightness problem, the transformer is not the source of your issue.

Yeah, I didn't see any pots on the board at the very end of the tube. So I'm clear, it sounds like I'm supposed to be slightly turning that screen adjustment knob with an insulated screwdriver while the monitor is on? And if the screen brightness seems to have sudden disproportionate jumps giving it a light adjustment, it could be the problem? I'm assuming I should wait until the CRT warms up to some degree and the problems begin manifesting before I do this?

You mention touching/moving the screen knob possibly fixing the fluctuating brightness. What would be the mechanism at work in that scenario?

I may try to resolder the capacitors I removed back on, and attempt this over the weekend. Thank you for all your guidance.

Namrok wrote on 2022-11-21, 21:33:

You mention touching/moving the screen knob possibly fixing the fluctuating brightness. What would be the mechanism at work in that scenario?

The screen pot might be slightly corroded or dirty. Moving it will scratch away the corrosion or dirt and ensure proper contact.

Well, I put the same capacitors back where I found them. Turned the monitor back on in my shop, gave the screen pot a slight twist left and right, and waited. And waited. And waited. Haven't caught it dimming yet, but it's probably high 50s in my shop, plus the cover is off. So if the problem is heat related, it's not getting warm enough. Gonna give it another hour or so, then it's time to clear off my workbench for other projects. Just have to see how it fairs inside under normal conditions.