I'm guessing the real cause is an aged electrolytic capacitor. The electrolyte can dry out over time causing a change in capacitance. Power supplies are the most common failure in electronics and electrolytic capacitors are a common reason for power supply failure.
There are electrolytic capacitors near where he was heating, and the capacitance of electrolytics can have a strong temperature dependence. He probably managed to heat one of the electrolytic capacitors, which happened to change its capacitance in the correct direction to make the circuit work.
Chances are the monitor would work reliably if all the electrolytic capacitors were replaced.
Edit:
I'm guessing the problem is C208. Section 10 of the LDO data sheet, linked in the article, talks about how stability is dependent on the output capacitor (C208). C208 has probably dried out, reducing its capacitance and making the LDO unstable. Heating was enough to make the circuit stable (for a while).
Further edit:
Predating my comment, "Gordonjcp" also called out C208.
> Power supplies are the most common failure in electronics
In over 35 years of troubleshooting gear and systems, my experience is that 90% (no exaggeration) of problems are bad cables.
It seems even worse, these days, with high-speed serial cables, running on razor-thin margins, and often with embedded ICs.
That's why, when the IT geek comes to your desk, they just rip out all your cables, and replace them with new ones, out of the shrink-wrap. They'll toss out a hundred dollars' worth of perfectly good cables, because they know the deal. They could waste an hour, trying to troubleshoot a problem caused by an intermittent USB C cable.
Also network cables. After spending days troubleshooting weird network issues a few too many times, I just toss the entire box of miscellaneous network cables in the attic every year or two.
But capacitors are definitely the second most common issue.
Way back when I was a staff engineer at Motorola, we'd often have network problems from workstations caused by marginal cables. At first, the IT people would come over and replace the cable, and toss the bad/old one in the person's cubicle trash can.
But I noticed over time, they all started adding one extra step: they'd cut the cable in half before tossing it in the trash.
Before the cable cutting, engineers, being engineers, would fish the "bad" cable out of the trash as soon as the IT person left ("it works almost all the time...") and use it for the next lab build-out. And then integration tests would fail intermittently, etc.
The capacitor plague was real. So many electronics I have from that era have failed due to bad caps. They are almost always cheap bottom-tier components with brand names and packaging suspiciously similar to the more reputable ones. (I’ve also read about inferior components that have been re-sleeved but I’ve never come across these. I don’t doubt they exist, though.)
We had a 500W subwoofer amp just die on us one day. Since the replacement was going to be several hundred dollars, I figured I’d try disassembling it to see if I could find a problem.
Lo and behold, I saw three 1000uF caps that had leaked, one of which had a clear bulge on the top. So I ordered a bunch of replacements off Mouser, bought myself a soldering iron, and replaced all three. Worked like a charm!
I’ll never understand why even high-end equipment manufacturers wind up using crappy knockoff capacitors in their stuff. It seems like it’s just a failure waiting to happen. I guess they get to make good money on the support and service?
It's such a widespread phenomenon that there's an entire site (with good repair forums) about electrolytic capacitors causing failures: https://badcaps.net/
For those unaware, the great capacitor plague of the 2000s has an interesting backstory: it is believed to be the result of corporate espionage gone wrong. Somewhere in the chain of theft, then transfer, then use by a competitor, a misappropriated formula for capacitor electrolyte was altered. Faulty capacitors ended up in all sorts of electronics, including the particular Abit VP6 motherboard which failed and led to the creation of badcaps.net.
An old trick I learned from a tech who repaired CRT televisions was to test components by spraying them with an aerosol. If something was close to failing, the cold propellant would push it over the edge and you could target individual capacitors to identify which ones need replacement.
A few years ago, I got ahold of an old Samsung Syncmaster LCD that didn't work anymore. After replacing all electrolytic caps for around 2-3$, it worked like new again.
Same here - I've had my SyncMaster 226BW for roughly 15 years now and it's still a great monitor, but a few years back I had to replace one of the capacitors after the screen would turn "on" but had nothing on the display. Did the same thing with a Dell monitor a few months back that I use as a second display. I am guessing that bad caps are the single biggest point of failure on monitors.
There's also a YT video showing troubleshooting on an LCD TV down to a bad multilayer ceramic cap. Used a microohm-meter to identify the shorted cap without replacing all of them.
I had a computer that could only be booted with a hair dryer. I did not disassemble any parts, so I never got to know which exact part was failing, but a good 5-10 minutes of pre-heating with the dryer allowed it to boot.
A few years ago, my neighbor's PC wouldn't boot. Unlike similar situations I'd had with my own hardware, the power-supply wasn't dead -- in fact, at the back, there was a green blinking light. I went online and found people suggesting that, in such a case, blowing a hair-dryer at the power-supply might fix things. I tried it and...sure enough, the PC then booted.
Cool (or hot, as the case may be). Can I ask how you stumbled upon this solution? It's not one I would immediately reach for, especially the 5 - 10 minutes bit.
I saved a monitor once that had a bad cap in the power supply— very satisfying and straightforward fix, like a $2 part and fifteen minutes of taking it apart and soldering.
When visiting China as a teenager in the early 90s, my brother and I decided to invest our hard-saved cash in a Micro Genius. This was a rip-off of a Super Famicom which I'd seen a Malaysian school friend play back home in Australia.
There's a photo of us smiling at the counter of a department store, handing over money. We bought a couple of multi-game cartridges. 190-in-1 and 27-in-1 or something.
We tested it in a Hong Kong hotel room, and briefly played a few games. Then imagine our dismay when we eventually got back to Australia and the thing wouldn't reliably load a game. We were blowing on cartridges and all that.
One day, we gave it a shot up in our non-A/C, second-storey bedroom. It was a 40 deg C day, so absolutely cooking upstairs. The console worked! The games loaded! We got to play an assorted of games we'd been eagerly waiting on.
We eventually decided it must be the heat and on the next day I can remember us taking it in turns with a hairdryer trying to warm the console or cartridge to get a game to start while the other person played. It might let us play for several minutes and then fail. Unfortunately, this trick didn't last for long and then the console was surpassed and the games no longer kept our interest.
30 years later, I still have the useless boxed console in my garage and can't bear to throw it out.
This is a really cute story. It might be really nice to fix it up and play it with your brother at Christmas (or other similar holiday) as a bonding fun time.
I felt so responsible as the older brother, convincing him that we should buy it together. Of course, the expense is trivial now and just a story we laugh over. He was 9 then and 40 now. He lives interstate and we tend to bond over Fortnite when it comes to games these days, playing as a squad with my 9yo son.
For my kids, I'm not sure the Micro Genius era of games holds up to the sandbox style of say Goat Simulator or constant chaos of Fortnite. Bit of a shame. In hindsight, I should've walked them through a bit of history of gaming in sequence.
Problem is, even as new, it's still the worst of about a dozen consoles or devices in the house capable of letting my kids play video games. They play PS4 or the 360, and occasionally an original Xbox for a particular party game. There are older consoles than those that get completely ignored but would be superior to the Micro Genius!
But you reminded me that we did take it to an electronics store for a quote to try and fix it, and I think the quote was more than we paid for it. Also a bit pointless with the pace new consoles were being released!
> A common issue with these types of components is that the quiescent current which ensures for the internal circuitry to work properly depends on the ambient temperature. If it’s too low and the regular doesn’t have enough supply current left the output appears to be dead.
Which honestly makes very little sense to me, and also reads a bit like a terminology tombola. A quick googling did not turn up more material on the idea that voltage regulators depend on the temperature like that, and it would be surprising (generally electronics performs better when cooled).
I would expect the problem to be due to a bad solder joint, which would explain why heating it helps since it might make the solder flow a little bit back into making connection (although hair dryer temperatures at 200°F/93°C) are too low to properly reflow solder). Or it might just make components and/or solder expand enough to make contact (which is kind of the same thing but different).
In the olden days we used to use hair driers and cans of "freezer spray" - ozone-depleting freon in a can, now replaced with more eco-friendly stuff - to heat and cool components to see which ones were temperature-sensitive. Quite often you'd get a fault that would only show up when the set was cold, or only after it was thoroughly warm.
I'm wondering if perhaps C206/C207/C208 in the LDO circuit that decouple the output might have gone a bit leaky and warming them up causes them to act more like capacitors and less like resistors to ground. If they're SMD multilayer ceramics that would be a pretty common failure.
The part is a LDO (low-dropout regulator). You can read about these here. See also the section about the quiescent current. One more thing: what made you believe that the given explanation is wrong?
> quiescent current which ensures for the internal circuitry to work properly depends on the ambient temperature.
This statement seems to get the relationship between quiescent current and the operation of internal circuitry backwards.
Quiescent current doesn’t power the internal circuitry. Quiescent current is a measure of current consumed by the internal circuitry while it’s idle.
The temperature relationship exists because the circuitry consumes more power when it hot. But there isn’t some temperature dependent magically quiescent current provider that must work correctly for the rest of circuitry to operate. Just like there isn’t a “standby power provider” inside of a TV to allow it to remain in standby. The standby power is just a measure of the power consumed while in standby.
I really do understand what LDO means, and at least a decent amount of how they work.
I still don't think it makes sense; it's not as if you have the heat the device to be hot enough to draw the quiescent current, it's the other way around. As the internal temperature rises, the efficiency of the components degrade, things start to "leak" more current and the quiescent current draw goes up.
I think it makes sense when you read it carefully. What tripped me up is that quiescent current is usually only seen as a bad thing. It not drawing enough current also seems more like a symptom than the cause.
On my first pass I was also a bit confused whether he is talking about the regulator's internal circuitry or something else: It's not obvious that LDO stands for low-dropout regulator, so at first it felt like he's misusing "quiescent current" to refer to some kind of "standby current" (comparable to an ATX power supply +5V Standby line) that the regulator has to supply to some other circuitry that then powers on the regulator to supply the rest of the device.
It's not uncommon for solder joints to get damaged due to heat stress. It happens to BGA compontents too when not properly cooled. The hair dryer may or may not have provided enough heat to fix a small crack (I didn't look up the temperature a hair dryer provides).
Did kind-of the same with my Philips TV a while back. Still going strong.
"Hair dryer" is such an imprecise term in a use case like this. One can find 600 watt units all the way through about 2300 watts. A typical general-purpose heat gun one might use to strip paint or shrink some heat-shrink plastic is usually between 1500 watts and 1800 watts.
Put a thermocouple in some hairdriers and you'll find some get well over 400 Celcius (750F!).
They just rely on the fact air has a low thermal mass, and it's easy to just keep it slightly further from your skin if necessary - the air quickly cools with distance as more room air mixes in.
There is a huge difference in whether one can actually reach that temperature, how quickly, and how much airflow it provides across the heating element and on target. A heat gun is a much more consistent tool for the uses for which it's designed.
A heat gun is also designed to put pretty consistent heat an inch or so away from the nozzle, whereas hair dryers often have what appear to be left-over jet engines for fans, for when you need to dry someone's hair from ten feet away.
I agree, it makes no sense. I am a real hardware electronics engineer and I don't understand his explanation. I believe the LDO could fail in a temperature dependent way. I do not believe the explanation.
I am not an expert, just a hobbyist. It's hard to tell anything without measurements. Either the regulators are half dead (factory spec for operating temperature is -40 to +85 C°) and should be replaced, or thermal expansion causes a cracked solder joint to touch again, or some other component is half dead and needs a bit of warmth to work properly.
>A quick googling did not turn up more material on the idea that voltage regulators depend on the temperature like that, and it would be surprising (generally electronics performs better when cooled).
Not a hardware expert either, but Wikipedia points at this TI doc[0] which claims ambient temperature is necessary for the quiescent current. There's no mechanism described there, though.
These have ground pins (as opposed to floating regulators), so they draw whatever bias current they need (which mostly depends on temperature, input voltage, output current and the lottery). The explanation is bogus, and doesn't explain how heating once would help with a problem caused by too low ambient temperature during operation anyway.
> ambient temperature is necessary for the quiescent current
No it does not say that. It says ambient temperature is a factor contributing to quiescent current:
"The value of quiescent current is mostly determined by the series pass element, topologies, ambient temperature, etc."
In an LDO you normally want as little as possible quiescent current when idle.
You certainly wouldn't design stand by operation to be dependent on temperature.
If it turns out to be so with time, it's an aging problem.
Old thread now. But often a minimum load is needed for an LDO, particularly on older parts, to work correctly. Such that these are given in datasheets. This is continuously required so is effectively quiescent current. I think this is what the OP was trying to get at.
But failed capacitor reulting in instability seems much more likely.
You are right that a small quiescent current is necessary for correct idle operation. But much of the discussion here is about the misconception that a certain ambient temperature is necessary for correct operation by design.
By the way it is in fact imaginable that aging has caused the quiescent current to become too small at low temperature.
There are several possible temperature dependent fault mechanisms caused by aging, so I wouldn't make a guess in this case.
Yes, to me the text read as if the author put some relevant words into a rotating drum and then picked them out. The result is a random-seeming sentence, which is my attempt at explaining how it read, to me. I was not aiming for snark, apologies if that's how it came across.
This observation isn't directly applicable to this story in particular, as this case the fix did require some in depth troubleshooting knowledge of the subject.
That being said, I'm often struck by how often something can be fixed by just opening it up and looking at it, even if you know next to nothing about the internals.
Two examples, both car-related:
- I used to drive an old Ford Ranger, and one day it started running like crap. Horrible acceleration, engine running rough. I made an appointment with a mechanic, but the day before my appointment I thought "What the heck, I might as well look at it." I popped the hood and immediately noticed that the air filter housing was cracked in half. Patched it up with duct tape, and it was good as "new".
- One of my wheels started making a godawful constant squealing. I couldn't drive 10 yards without turning heads. I brought it into the mechanic, where they took the wheel off and promptly a pebble fell out of the brake calipers. Had I just jacked it up and taken the wheel off myself, I would have saved a trip to the mechanic.
I would never guess that a piece of even consumer hardware was designed with such low tolerances as to develop such issues.
Is this some kind of a trend that I'm not up to date with?
I was surprised to find out that my laptop fans started first getting noisy and then rattling after less than two years since purchase. I searched around and apparently the tight tolerances combined with low quality of the bearings eventually produce this effect.
This is especially audible if I let them heat up - it appears that thermal expansion is enough for the blades to get too close to the housing.
I ordered a set of new ones and appropriate tools, but I can't imagine doing this every two years. My previous laptop lasted around seven, after which both the battery and the power socket gave out.
> Is this some kind of a trend that I'm not up to date with?
It's called planned obsolescence and it's part of the factory-to-landfill pipeline. It's not exactly new.
Do we know how to make a long lasting laptop fan? Yes. Would nearly every consumer pay $0.25 more for a laptop with a longer lasting fan? Yes. Can you buy laptops with high quality fans? Yes, but seemingly only by dumb luck.
By the time you figure out that a product has a high failure part the company will no longer be manufacturing it and therefore reviews won't be relevant (granting relative immunity to bad reviews). And when every brand is doing it, there's no way for "free market" competition to sort it out. It's a race to the bottom. (3. 2. 1. Cue "The morality of protecting share holders eclipses the morality of ripping consumers off.")
I only buy used laptops now. The significant reduction in price is a reduction in risk. Also a used product has had "burn in" time to weed out the lemons. The engineer calculated xx% of fail-early laptops often aren't the ones being resold.
I'm bitter. I'm cynical. Despite being aware of my mind's ability to find patterns to confirm my biases... I'm really struggling to be excited about new products. I'm spent like nuclear fuel; I'm toxic. They say knowing is half the battle... not in psychology. Doesn't help me a damned bit.
Today's sponsor is Better Help. I should just stop now.
> They say knowing is half the battle... not in psychology.
I've felt this so many times that I gave it a name: "the falling physicist problem".
A physicist falling without a parachute from the very top of the troposphere knows, that his terminal velocity is around 50m/s and was reached via gravity pulling him towards the ground.
Nevertheless he's going to go splat the same way anyone else would, because sometimes knowing is just not enough.
Or you buy a laptop brand that's specifically targeted for long-lasting professional use only, like Panasonic Toughbooks or Thinkpad P series workstations. So all the problems associated with consumer race-to-the-bottom type stuff is avoided. And if not the 5 year+ next day on site warranty would cover it.
Buying older hardware is also one of the few ways to have it be documented enough to be able to remove all of the crapware baked in by the manufacturers. Every laptop newer than the Thinkpad X230 is basically dead to me.
He's talking about a 12 years-old screen that probably saw daily use. Something failing at that point is more than expected.
As others have noted, it's in fact probably bad capacitors, which is a really common issue for electronics of that era. I also encountered that several times, it's a quick fix if you know how to solder new ones, and you can find such capacitors for cheap (like 1$ cheap last time I had to look, though finding that price for a single one is hard, and much less when bought in bulk).
>I would never guess that a piece of even consumer hardware was designed with such low tolerances as to develop such issues.
It wasnt, the explanation in the blog is nonsense. 10 years is a good lifespan for capacitors working in hot environment, and that is what failed. Electrolytic Capacitors are perishable, they age even when not used.
>I ordered a set of new ones and appropriate tools, but I can't imagine doing this every two years. My previous laptop lasted around seven, after which both the battery and the power socket gave out.
then use better quality replacement mechanical part. People arent surprised when servicing cars, why different expectations with modern electronics?
> Electrolytic Capacitors are perishable, they age even when not used.
But not that perishable. I have some old electronics, along with fully analog devices (guitar effects) and they, along with their power supplies(which get hot) still work.
> People arent surprised when servicing cars, why different expectations with modern electronics?
Because they have orders of magnitude less moving parts - if any.at all. Is it unreasonable to expect something that has one moving part to not fail after two years?
Actually, I wouldn't want a car exhibiting mechanical problems after such a short period.
Old gear didnt depend on low ESR (simplified resistance at high frequencies). Every time you turn on a piece of electronic with linear power supply electrolytic caps will slowly reform and start working better and better with time. Modern gear is DC-DC converters all the way down. Aged cap means no supply voltage at all (doesnt turn on), or crazy ripple (crashes, produces funny sound or visible artifacts).
>Actually, I wouldn't want a car exhibiting mechanical problems after such a short period.
Nobody would, thats what scheduled services are for.
If this is the case, you could consider bearingless fans, which should prevent them from wearing out. The only issue is you may not be able to find them in the size requirement you need.
Bad caps was a very common problem in the old days, I remember I replaced the ones on my motherboard at least twice.
The classical symptom was that heating the motherboard will allow the PC to start.
I fixed a 2008 macbook with a GPU issue (wouldn't boot past BIOS) by turning it on and letting it run full-tilt under a blanket. Eventually it just resoldered itself.
It didnt. 8600M had a design defect in microbumps connecting die to package. Thermal cycling in high heat scenarios (Apple is famous fo cooking components at the margin of T-junction) softened improperly selected underfill and broke microbumps connecting die to package. Repeated heating up can again soften it to release stresses and temporarily reconnected broken traces. It never fixes the main issue of broken chip.
>On July 2, 2009, the date being ironically a year after the notorious 8-K that publicly kicked off bumpgate, the company put up a job listing for a “DIRECTOR OF PACKAGE TECHNOLOGY”.
Can confirm that this saved our old Samsung lcd television.
It took longer and longer to turn on. Some guy on YouTube used this hair dryer trick. So I did just that, blow the air inside the TV from below through the panels, and like magic it works again. Life hack!
Sandwich one of those silly cheapo USB mug warmers into the case. Plug it in briefly for a little warmth just to get it going. Like a choke on an old-fashioned car!
Not to be a worry wart, but a long time ago we got a batch of monitors with a defective board. A capacitor overheated, melted some sort of glue on an adjacent component, which in turn dripped on a power supply component, shorted out and started a fire.
A couple of days later, it happened again… and we ended up getting all hands on deck to find those monitors.
Reminds me of the "myth" of putting a Radeon GPU into the oven on low heat for a while if it was broken. I tried it when mine stopped working and indeed it fixed it.
Also another Radeon card was identical as a more powerful one, except that two pins (maybe wrong word) were not connected. I drew the connection using a pencil directly on the board and it worked as well, saving around 100 Euros.
It's over a decade ago, so details may be slightly wrong. But still interesting how low tech solutions worked on such complicated machinery.
What you did was reflow the solder. This isn’t an uncommon technique for amateur/small batch repairs or assembly of surface mount circuit boards. At the maker space I use, there’s a toaster oven dedicated to this task.
That being said, you don’t want to use that oven for food purposes anymore. Lots of toxic chemicals will off-gas in the reflow process.
ETA this reminds me of the Xbox 360 “red ring of death” fiasco. One DIY repair technique was to wrap the entire Xbox in towels blocking all the ventilation. The theory was the resulting overheating would reflow the failing BGA solder joints. I don’t know if this really worked or was anecdotal but it was one I remember seeing a lot.
> you don’t want to use that oven for food purposes anymore. Lots of toxic chemicals will off-gas in the reflow process.
Just run the oven at full heat for an hour and you are fine - the same process that off-gassed the chemicals in the first place, will also deplete them from the oven when you run it later. (Keep the vent on, or ventilate the kitchen.)
You can also stick a fume sucker[0] in the (cold) oven for a hour.
But with either approach, make sure nothing dripped and landed on the bottom of the oven chamber (or worse, the corner seams or the space under the chamber), since that can stick around and keep outgassing for months or years.
0: A length of dryer hose with a salvaged PC cooling fan on one end and the other out a window - usually used to ventilate solder fumes.
I love the last picture with their desk covered in stickers. I thought we were supposed to keep the stickers in a box forever until we find the "perfect" use for them...
Hair dryers can be quite useful. If you have a small engine that won't start like a lawnmower, heating it with a hair dryer will often allow it to crank.
My old Benq monitor was blinking for the fist 10-20 minutes when I turn it on and I used similar trick. I knew nothing about electronics and used common sense. I would direct hair dryer towards air vents of the monitor and wait for 30 seconds.
I got bored after a month repeating the same thing every day and sold it unrepaired.
I wonder if there could be another issue. I remember we had a hardware guy once where I worked who took apart a broken cable modem prototype, found that one of the traces was thinly broken. Traces are the lines on a printed circuit board, Vias the circles. He used an emory cloth to strip the protective green coating, then a small narrow heat gun to melt the trace slightly and bring it together.
If OP had a broken trace in there, then heat might have fixed it. Then again, OP's reasoning is probably better as I am not a hardware guy.
Kinda reminded me of the time I used listerine on my MacBook screen and it solved the gross reflective coating issue that apple wouldn’t fix for me. Crazy part is it worked
i wonder if this would help my issue, my 165hz monitor is not usable anymore at 165hz it flickers, has weird grainy image and what looks like scan lines. Had to use it at 144hz for a while as it didn't happen there but after some time i had to go to 120hz and now 120hz is starting to degrade also : (
I had a similar problem with some small exotic parts on my Hyundai L90D+. After being able to source them, I replaced the parts, and my monitor came back to life, briefly.
After 30 minutes or so, another part of the board has fried, so I just replaced the monitor, quite sadly.
A lot of theory and weird explanations (weird for an EE), but no verification like actually measuring VEN. hair dryer = dried capacitors, not LDO. Heating up temporarily rejuvenates dead caps.
I used to throw the graphics card from my Dell XPS 1710 into the oven whenever it stopped working. That worked about 10 times over the space of a few years!
Active LCD shutter glasses for 3D movies were much nicer than 3D gaming that way. Sadly, so little content came out for those setups that the fad passed.
There are electrolytic capacitors near where he was heating, and the capacitance of electrolytics can have a strong temperature dependence. He probably managed to heat one of the electrolytic capacitors, which happened to change its capacitance in the correct direction to make the circuit work.
Chances are the monitor would work reliably if all the electrolytic capacitors were replaced.
Edit:
I'm guessing the problem is C208. Section 10 of the LDO data sheet, linked in the article, talks about how stability is dependent on the output capacitor (C208). C208 has probably dried out, reducing its capacitance and making the LDO unstable. Heating was enough to make the circuit stable (for a while).
Further edit:
Predating my comment, "Gordonjcp" also called out C208.