Faulty 120W charger analysis (Anker GAN Prime) [video] | Hacker News Reader

Faulty 120W charger analysis (Anker GAN Prime) [video]

youtube.com

78 pointszdw71 comments

I think this justifies my habit of buying Anker chargers. They're expensive but seem decently engineered. If they were cost-cutting, they wouldn't have 2 temperature sensors, input filtering, properly rated & branded components, etc. Even if this one failed, it seems to have failed safe.

Also the transformer manufactured from a multi-layer PCB is pretty wild [about 20 mins into the video], I've never seen that before.

Indeed, this is way different than the cheap aliexpress stuff bigclive usually tears down, where 1mm track clearance separates you from death by mains voltage.

You do get what you pay for.

I would be interested to know - has anyone ever died when a consumer charger becomes faulty and connects the AC to the DC side?

Lots of cases where cheap chargers cause house fires. Whether that's specifically because the AC & DC sides were connected would be hard to tell after the fact. https://www.bbc.co.uk/search?q=house+fire+charger&d=HOMEPAGE...

We have no idea if they were cheap.

Too many people assume that just paying more will get them quality, that's far from guaranteed, and opens the way to so many scams.

DiodeGoneWild also has quite some of those: https://www.youtube.com/playlist?list=PLfUGOPyKWwiriWhjmdtvj...

The worst has to be the charger without any isolation between mains and output: https://youtu.be/qTpVhNCot3Y?t=81

Edit: just to make it obvious, check the upload date for the second link

You often do get cheap when you pay cheap, but paying more doesn't guarantee that you don't get the same cheap thing with a different brand on it.

Unfortunately at least two of the popular known-to-be-good brands, Aukey and RavPower, misbehaved in ways that got them banned from Amazon (IIRC fake reviews). IIRC Anker also had some scandal, recall or exploding product or other reason why I stopped trusting them (in addition to the Eufy scandal, where they released a supposedly privacy-friendly/local-only/e2ee camera that was everything but that, then lied about it for months when caught).

On the other hand, the guy said "bloody complicated", and that's not a good sign.

Complicated designs are neither a good nor bad sign. Sometimes complicated designs are just the result of solving complicated problems. And USB-C is anything but simple.

If you have ten types of charger, all of them working fine, and one of them is 10x as complicated as the others, then that's at least a bit suspicious, since more components means more opportunities to fail.

Could also mean more insurance. In case nothing happens, insurance complexity seems overprotection, but when something happens, nothing is too much for people to get back whatever value is lost.

Insurance can also break and take down the rest of the circuit.

Only if the circuit is designed by amateurs.

Firstly you can have extra components (like the temperature sensors in this Anker supply) that monitor other components, and will shut down the power supply in case of an anomalous condition. Secondly for most circuits, most of the time, you can simply leave out components and the circuit will continue to work, at least on the happy path. For example take a typical PC motherboard, snip off some of the capacitors, and chances are that it'll still work. Big Clive has examples of very cost-reduced circuits where components are simply not mounted - there are even pads on the PCB for them, they appear in the reference design, but the manufacturer didn't include them, and the device will appear to work until it suddenly doesn't.

You are assuming they use the temperature sensors as protection, and not as a way to drive the circuit to its upper limits.

You clearly have no idea what kind of engineering goes into something like this. All of your arguments are pointless and uninformed. You have no idea at all what you're talking about.

Let me just quote the guy from the video: "An interesting design, but bloody complex. With so many components the probability of some failing is adding up."

And using less components so fewer things can break is actually a thing in engineering contexts.

The GP literally provided you with examples of where additional components can provide resilience while still allowing the circuit to function should they fail. And yet you still reply with the reductive argument that there is some magical one-to-one correlation between component count and probability of failure.

And totally ignoring that there are many chargers out there which work totally fine with a much smaller BOM.

Nobody has ignored that fact.

What we disagree on is whether those additional components add value to the quality of the overall product.

Then again the expensive equipments do not survive the error conditions (like gridspikes and blackouts either). You either guard at the gate or your fortress blows up with suboptimal results either way. Single point of failure filter design it is for economic reasons ?

the video is a total joke for people with any clue. yes, Anker throw the kitchen sink at problems, but not for quality. their only metric is reviews. so now warming up is much more important than working after two months, etc.

the cherry is the top port receiving 5v that will then go to the negotiated c ports

i don't know why you took it as an endorsement

If anyone is interested in bonus GAN content. Electrarc240 recently did a very accessible teardown of a similar GAN power supply, while explaining all the bits and pieces as he goes along and comparing to an old non-GAN style supply.

https://youtu.be/vgmqUhvQlww

Very interesting video even though I only did a little bit of electrical engineering. These videos where they don’t wear gloves or seemingly have no protection really stresses me out though.

The protection is that it's not plugged in.

True in this case, but this isn't universal. Some devices with high voltage caps will absolutely still kill you even if they've been unplugged for days. This is why people can and do die while messing about with unplugged microwaves in particular.

This is true, but people handling those devices generally do protect themselves. In this case it wasn't necessary, so no need to stress!

Should be quite unlikely though. They are meant to have resistors to drain the caps. Can also short the cap with a screwdriver if you want to be extra sure.

People should not be messing with microwaves period.

There aren’t gloves you can wear at this scale that would offer any protection and allow you to do anything.

The only protection you need is for it to be unplugged and if it has large caps you want to make sure those are discharged.

I don’t mean protection from electricity.

He’s sawing into plastic and electric components and has no cut resistant gloves or a vacuum to suck away dust. At one point he also had a soldering iron quite close to bear fingers. Parts where he’s just prying things off with pliers wouldn’t need maximum dexterity with both hands.

I wouldn’t use gloves to do any of those things you’ve described. They would be a hindrance, and the risk just isn’t that bad.

There are many different philosophies. My neighbor is 84 years old and worked for the US Forest Service for 30 years doing timber management and fire prevention / suppression. He has thousands of hours of experience using chainsaws, much of that on trees that are 100+ feet tall. Even in his 80s, he prefers to use a chainsaw wearing nothing other than shorts and a t-shirt - no chaps, no long pants, no socks, and no shoes.

His argument is that safety equipment makes people complacent, especially when they are experienced enough to move quickly. If you have bare feet / legs, you are much more aware of the dangers and hence more careful. He still has all his limbs and the fingers / toes at the ends of them. Too many of his former coworkers do not. He is nearly deaf, so using hearing protection might have been better.

That's all anecdotal at best, but experts often disagree on the safest approach for themselves vs. others. And what's best for experts is often not what's best for novices, and vice versa.

My Anker GaN chargers seem finnicky with cables, resulting in drop-outs and flapping phone charging on my Pixel 6 Pro. Anker customer service were happy to send me replacements but they weren't any better.

I tried Ugreen GaN chargers and they seem to work much better. And a couple of my Anker GaN chargers have failed now, not long out of warranty. Since then, I've stopped buying Anker. I just can't get them to reliably charge all my devices, whereas Ugreen chargers seem to work better.

Conversely, it could also be the case that the Anker chargers are noticing too much resistance on the cable (I've had a number of nice and poor quality USB-C cables break in strange ways which indicate some portion of the wiring is damaged, like cables that only charge one orientation, or only negotiate data one way, or ...), and degrading to avoid risk while still working.

You'd need a tester with some fine detail to be able to tell the difference prior to a problem, though...

Ironically, it was Anker chargers with Anker cables that worked the worst (all were new), and a new Ugreen charger with the cable it came with that worked the best.

I got the impression that the Ugreen charger was more tolerant somehow - perhaps delaying longer before considering a previous PD negotiation terminated. I don't have the ability to tell for sure though.

I genuinely thought the audio was playing backwards...

The accent took some getting used to for me. It sounds Italian, but not quite.

Czechia

His website: https://danyk.cz/index_en.html

Ahh very interesting, thank you. I know other Czech people but they don't draw the vowels out this much , I wonder if it's a local thing or just specific to this person.

I really like the accent! At first I found it extremely odd but it's grown on me so much over time it's honestly the main reason I watch his videos now.

It seems really unusual to me too. I've got Czech relatives and have visited several times as a tourist but seem to have missed out on hearing this accent somehow (probably as I've barely left Prague).

It sounds normal + imitation of 'American YouTuber' to me: the drawn out vowels being common in many US accents anyway, and then the 'let's fiind oooouuuutt' sort of YouTuber affectation that's itself an imitation of a certain sort of TV presenter I suppose - Jeremy Clarkson for example.

To my native Czech ear, this sounds like the English accent of a native Czech speaker who learnt English primarily by reading, back when English audio wasn't readily available.

I'm impressed that someone with such electronics knowledge has opted for pliers over hot air for removing chips. I suppose it worked though...

He’s clearly an engineer and not a technician, and even jokes about his ‘sophisticated technique’.

He does use a hair dryer later on.

I have bad experiences with two of my last multi-port GAN chargers. One Anker was DoA and then I got a Baseus which burnt within one year. Small powerful multi-port chargers sound great in theory, but I will stick to 'normal' chargers for now.

What is this accent? It sounds kind of Indian but also not at all... Iranian?

Czech.

His website: https://danyk.cz/index_en.html

Strange. I've never heard a Slavic person talk like that.

This accent is common among Czech people who learnt English by reading without much exposure to English audio. Some of Danyk's apparent pronunciation idiosyncrasies used to be common among English teachers in Czechia. I presume they learnt English pronunciation from written descriptions of it. It is fairly rare today thanks to internet streaming. Danyk's blog is older than YouTube.

That PCB transformer design is interesting.

Was this done to save costs? I can't imagine any other reason.

The opposite. Planar (PCB-based) transformers are more expensive to make, but allow much tighter control over where wires and insulation go.

A properly designed one will be significantly more compact for the same power rating (compared to wire-wound one).

They are quite common in high-wattage high-reliability power supplies like modular PSUs in rack-mount servers.

I suppose it uses blind/buried vias then, to explain the costs.

Blind vias, yes. Also - compared to average 4/6-layer PCB - thicker copper layers, thicker metalization of vias, more layers, thicker board overall.

This video started to get comical as he started breaking away components.

Nandor up skilled!

On a serious note isn't all that pink stuff a fire risk due to insulating high wattage components?

I doubt it's flammable but it's still eventually burnable and insulating everything else.

Going by the comments on the YouTube video: It seems to be for hear dissipation, so quite the opposite of insulation.

Whats the point of this product over just using a regular USB block or the charger your laptop came with?

Some examples:

- the one that came with your laptop no longer works

- you use your laptop at home and office, and don't want to carry the charger back and forth every day

- you travel a lot and want something smaller/lighter than your stock charger that can charge your laptop and phone at the same time, both at high speed

Not sure how this hits as a hn story given the large amount of YouTube electronics from agile greenhouses out there doing full teardowns to wearing through fabrication processes, but ok...

It’s not a particularly in depth failure analysis beyond identifying the failed component but he shows knowledge of what the components are and what they should be doing, and I like his understated humor.

In this world of AI can we not get a summary?

In this case, I think the interesting part is the video itself / the process. He breaks down a very inaccessible product and explains how it all works.

fwiw YouTube has an experimental “Ask about this video” button. This is the verbatim output:

Here's a summary:

• The charger was donated by a viewer and stopped working despite seemingly good construction. 0:02

• It has two USB-C ports (100W each) and one USB-A port (12V, 22.5W). 0:25

• When plugged in, none of the ports output voltage. 1:00

• The outer box is filled with resin, making it difficult to open. 1:10

• Open, the charger reveals several circuit boards in a sandwich-like design. 2:57

• A shorted rectifier bridge (3:49) and a shorted capacitor on the USB-C output control circuitry (7:20) were initially suspected. 3:49, 7:20

• Further investigation with a thermal camera pinpointed a specific communication/voltage conversion chip as the source of the short. 13:11

• This chip controls the gate of the MOSFETs used in the synchronous rectifiers and DC-DC converters. 14:00

• The main transformer is a hybrid flyback design with a unique AC input and half-bridge rectifier. 17:15

• The short was traced to the capacitor's pins on the same side of the board, suggesting a potential multi-layer board issue. 12:19, 12:41

• The video ends with a cross-section view of the transformer and measurements of some donated 18650 battery cells. 18:46, 20:35

Essentially, the video demonstrates a complex charging circuit and troubleshooting process, highlighting the potential for multiple failure points in high-power chargers.

Where?

Certainly!

I shall describe the UI in English: good luck

First, tap on the dooblydoo to expand it.

Find the powder blue/violet, four-pointed star that represents Google Gemini.

Tap the down-chevron to expand summary.

Also take note of various “explain” lozenges outlined in multicolor. These are quick-action queries for Gemini.

Have fun!

[The production style exactly reminds me of predecessor “Play With Junk” who undertakes massive teardowns of spectacular enterprise-class Unix machines. https://youtube.com/@playwithjunk?si=4-hu0iIA7bGovnGe]

As always with new Google products: available only in certain countries...

At Google, something as trivial as changing the website theme is enough to trigger legal action so it’s not surprising they roll stuff out carefully.

the short was actually traced to one of the two PD controllers, the other stuff seems kinda right

This video dives deep into a faulty 120W GaN charger, exploring its intricate design and uncovering the source of the failure. The video details the process of disassembling the charger, revealing its complex circuitry and identifying various components like the GaN transistors and the printed transformer. The investigation concludes with a surprising discovery about the cause of the malfunction.

Reply on news.ycombinator.com