debauer's liquid metal mod causing resistors to fall off ?

[haihane]

give this video a watch:



one thing really caught my ears: that the liquid metal contained Gallium (What?!).
isn't gallium the thing that eats through coca cola cans, so i thought to myself, and wtf is it doing there in liquid metal used as TIM (at this point i'm not sure what to call that, so i'm just going to call it TIM).


here's what i think after watching the video:
- liquid metal used as meant to be a replacement for thermal compound, isn't designed to have direct or near direct contact with PCB components. i realize too i'm being the annoying captain hindsight of southpark here; and that debauer, like many others (just youtube up "liquid metal on GPU" to see what i mean) didn't do it out of malice, but merely he was sharing something which worked, not realizing the full consequences of his actions.

- a little part of me feel he's pulling an "equifax" here (the equifax breach scandal and how they seemingly shirked responsibility). perhaps accusing him of doing an equifax is rather harsh here. to be honest, i had hoped for an apology here towards his viewers, that he need not be fully responsible for resistors or whatever falling off the pcb where he applied liquid TIM on, but an apology regardless would have been nice as admission of "sorry, i didn't know that would happen.". he sounded aversive (but i do still appreciate that he did try to offer free repair to that one dude who contacted him).

- continuing on the above. i realize too, that in his line of work (extreme overclocking), for most folks, is like stepping into unknown territory of science. you try and test out new methods for the end prestige of recognition and hall of fames. something in me tells me that what i saw here is not much different when medicine is first explored (anyone remembered that bloodletting was a common practise for several ailments?). with that in mind, i still find it hard to fully blame him for this F-up. he couldn't have predicted that his mod 5 months back would have caused parts to fall off (though i did somehow suspect, that when he first did this 5 months prior, he could be thinking "is this thing safe? the liquid TIM touching the resistor like this?", and perhaps he thought too. ah wthell, #YOLO.


- this is also serving as indirect advertising for him, from a complete stranger (me), that he needed data from anyone who ever did liquid metal repasting on their GPUs. if you repaste / re-TIM yours, maybe try sending him your findings what happens now.

- perhaps my last question: i'm curious, and would highly appreciate it if i get an answer for this. what does ASUS use as material to solder their components to pcb? i know that ASUS probably won't condone own repaste because of the possibility of this happening, and any warranty would likely by mercilessly voided if found. i can't help but wonder about ASUS' opinion about this though 😕


tl;dr: i'm still a bit salty. a sorry for the F-up would have been nice from him. i want to know what y'all think about this.

[xeromist]

I disagree with your assessment. As far as I know he has no direct affiliation with the companies producing liquid metal. So it's not like he produced the product that caused the damage. Grizzly may be his sponsor but we don't even know if that's the brand that ate the shunts. Correct me if I'm wrong.

Further, it appears his choice of card and liquid metal didn't interact so how would you expect him to know that someone else's would? Years ago I went skydiving and it was fun. If someone I told was later injured in a skydiving accident I would feel bad for them but it would not be my responsibility. Bad stuff happens in the world and you don't always need to find someone to blame. He's not being evasive because he didn't do anything wrong. He's just being his usual analytical self and trying to figure out the cause so it doesn't happen to anyone else.

Your equifax comparison is pretty messed up. You're talking about a company that collected highly personal information on millions of people without their consent and then leaked it due to negligent security practices. The victims will be at risk for the rest of their lives. That is nothing like enthusiasts sharing inherently risky mods. Anyone attempting a mod is taking a deliberate risk with their hardware.

[haihane]

xeromist wrote:
I disagree with your assessment. As far as I know he has no direct affiliation with the companies producing liquid metal. So it's not like he produced the product that caused the damage. Grizzly may be his sponsor but we don't even know if that's the brand that ate the shunts. Correct me if I'm wrong.

hmm.

Further, it appears his choice of card and liquid metal didn't interact so how would you expect him to know that someone else's would? Years ago I went skydiving and it was fun. If someone I told was later injured in a skydiving accident I would feel bad for them but it would not be my responsibility. Bad stuff happens in the world and you don't always need to find someone to blame. He's not being evasive because he didn't do anything wrong. He's just being his usual analytical self and trying to figure out the cause so it doesn't happen to anyone else.

i'm trying to make sense of this (also, your skydiving analogy almost did sway me).

but, in your analogy, you went skydiving, you didn't tell him where to go and/or what to do for extreme thrills. if debauer were to just tell his folks that overclocking-modding is fun, yet someone blew their card's capacitor or something; that wouldn't be his fault. and i would agree with you. perhaps the skydiving analogy didn't hit that close to home.

intead, when i watched his video (and to be honest, if he didn't make this video, i wouldn't have known about it. i appreciate him coming out for it), i had the impression that he was taking something (liquid metal) and applying it to a component (something close to pcb), i've never used liquid metal before. was there an instructions guide on how the liquid metal should and shouldn't be used? i.e.: use it only on heatspreader, don't put it in mouth, don't eat it. don't put it anywhere else other than on top of cpu. avoid direct contact with PCB tracers/ solder joints ?


my gut told me: "this shiny solder, if my liquid metal touched this stuff, won't it be dangerous?". granted and admittedly, i didn't have the guts to do half what he does for a living. i'm still trying to make sense of my saltiness.


if i were to put it into analogy: (i was having a hard time finding a compatible analogy. finally remembered it.)
Crazyrussianhacker once made a video on how to make homemade air conditioner using dry ice. the end result worked. but with consequences he immediately didn't think about. i remembered reading about this, and he did receive flak like hell. people started pointing out that the original idea came from another youtuber (i forgot his name), and the youtuber used normal ice.

in your opinion, would crazyrussianhacker be responsible if someone were to suffer mild carbon dioxide poisoning following his how-to-s ?

granted, the scale of harm would be completely on a different magnitude (one being a mere GPU, fixable with a resolder-debauer, at worst just a voided warranty. and the other: potential poisoning, loss of life),

help me solve this cognitive dissonance. if one is okay whilst the other is not, what's the benchmark used? the scale of damage? if yes, i would agree with you on this.
my head has this weird way of thinking, that i like to break each individual problems to its base constituents, then analyzing what ifs. then i run this thought experiment in my head: if i were to not consider the consequences (not because it's not important, but temporarily ignoring it to satisfy the reasoning, what makes one action okay while the other not-okay in its basic form), what separates debauer from crazyrussianhacker in terms of actions alone?

Your equifax comparison is pretty messed up. You're talking about a company that collected highly personal information on millions of people without their consent and then leaked it due to negligent security practices. The victims will be at risk for the rest of their lives. That is nothing like enthusiasts sharing inherently risky mods. Anyone attempting a mod is taking a deliberate risk with their hardware.

True.
i apologize to the victims and to debauer for being terribly insensitive. this is purely my mistake for taking such a comparison.

[xeromist]

haihane wrote:
what separates debauer from crazyrussianhacker in terms of actions alone? .

You are not alone in thinking this. After I read your reply I also thought of the homemade air conditioner vs this.

The difference to me is that PCB modding is inherently risky and I'd say most people doing this mod would understand that overriding Nvidia's power targeting could have negative consequences up to and including complete bricking of the card. Also DerBauer has a niche audience so his tools and advice are not intended for a general audience.

Contrast that with the Russian who is making entertainment videos for everyone and consider that most people wouldn't think an air conditioner to be inherently risky.

The Russian's response was different too. When Thunderfoot called him out for endangering people the Russian filed a takedown on the video. He basically blew it off and tried to cover things up. Even now, the original dry ice video is still up, unaltered. The only difference is that the title now suggests water ice but doesn't explain why CO2 ice is bad. The video still uses dry ice and says dry ice is better than water ice. Meanwhile DerBauer has been warning people of the potential risk of liquid metal since at least the 1080Ti power mod video in May.

[haihane]

xeromist wrote:
The difference to me is that PCB modding is inherently risky and I'd say most people doing this mod would understand that overriding Nvidia's power targeting could have negative consequences up to and including complete bricking of the card. Also DerBauer has a niche audience so his tools and advice are not intended for a general audience.

So, Target audience matters. True. (i hadn't considered this, and i'm grateful for you pointing it out)
i was tempted to ask Louis Rossman (the guy who schooled Linus when Linus baked his GTX 780TI(?)) about his opinion during one of his PCB board repair livestreams. what he thinks about putting liquid metal on it.

edit #2 changelog. i can't be that lazy.
*(?) = doubly checked, it's a 780TI. ref here. and here.

Contrast that with the Russian who is making entertainment videos for everyone and consider that most people wouldn't think an air conditioner to be inherently risky.

The Russian's response was different too. When Thunderfoot called him out for endangering people the Russian filed a takedown on the video. He basically blew it off and tried to cover things up. Even now, the original dry ice video is still up, unaltered. The only difference is that the title now suggests water ice but doesn't explain why CO2 ice is bad. The video still uses dry ice and says dry ice is better than water ice. Meanwhile DerBauer has been warning people of the potential risk of liquid metal since at least the 1080Ti power mod video in May.

and i missed the part where he already warned the inherent danger of doing so. Intent Matters (how one deals with one's own negligence).

learnt two things today: Target audience and Intent (how one deals with it)


i'm pleasantly surprised you know about Phil Mason (Thunderf00t) too.


i'm thoroughly satisfied with the answers. thanks all of you.

[haihane]

i wonder why i was salty over falling component(s) in the first place, when it didn't happen to me (even).

i seem to be unable to get it out of my head that debauer was using unorthodox methods to achieve his means, that if he had burned his own GC, that's okay. but if he put it up on youtube showing the how-tos, with promises of greater reward (OC potential), and somebody else's kidneys resistor start failing, i would feel inclined to ***** on behalf of him.

that even if his own resistor didn't fail, but mine did, following his advice, he'd owe me a sorry.
was i wrong thinking this way? (and why?)


i'm kinda shameless when asking this kind of stuff. i don't mind if you were to even judge that i'm messed up. i'd rather be corrected and find out what's wrong with me (my way of thinking).
btw. TIL huh? redditor detected! hello!

[Menthol]

One has to take responsibility for ones own actions, even if he stated you must do this or else, it is still your responsibility if you do it

[haihane]

Menthol wrote:
One has to take responsibility for ones own actions, even if he stated you must do this or else, it is still your responsibility if you do it

!!!
You, Sir, Are Right!

damn, the answer was this all along.
Thanks!


(I'll go back to sulking at one corner to reflect on this).

[Nate152]

Hi guys !

debauer reminds me of a mad scientist that likes to experiment, unfortunately it did not go well. You have to look up to him for trying it though, at least we know to not do what he did.

[xeromist]

This is my favorite kind of thread. We can have a reasoned debate and I get to learn someone else's thought process. Thanks Haihane!

(although I admit that me saying someone else's opinion is "messed up" wasn't very constructive of me. I plead temporary insanity, being an equifax victim myself)