Comment by femto | Hacker News Reader

Comment by femto | Hacker News Reader

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.

Yes, network cables.

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.

Cutting the cable is important. I’ve learned to do that, by habit.

There's a good overview of the problems with certain Chinese electrolytic capacitors here: https://en.wikipedia.org/wiki/Capacitor_plague

Also, here's a related article describing how a Chinese failure at industrial espionage created a worldwide capacitor problem: https://www.theguardian.com/technology/blog/2010/jun/29/dell...

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?

no matter the brand, being hot while charged wears capacitors out.

drop the operating temperature of a capacitor by 10°C and you extend the life by 10x. The inverse is also true.

Crappy caps will die more quickly than top of the line capacitors, but they'll all die eventually.

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.

My experience is that this practice brings things back into spec while still cold.

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.

Replacing electrolytics that'd gone bad was my entry point to electrical engineering.

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.

More examples: https://www.google.com/search?q=power+supply+blinking+light+...

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.

Reply on news.ycombinator.com