All right, guys. I guess welcome back to Hardware Unboxed. This is not a video we were planning on making. It's the early hours of the morning. I was doing some testing for another video and I suspect my 5090 is on the cusp of melting. I just took the cable out and it burnt my hand. It was extremely hot. We put the thermal gun on it and we were seeing it. We basically, I put it back in, run a test, and we saw the temperatures getting up over 150°. So, I've just chucked the Thermal Grizzly Wire View Pro on. We're about to fire up a game, put some load on it. I want to see what's going on. I thought we might as well film this, make a video out of it, because I've been running many of these cards with many different power supplies and cables, and I've never ever had a problem. Of course, it is a fairly widely documented issue that the 12V high power connectors can melt. I think mine is getting to the point where maybe it's melting. This is a relatively new test system. I think I've only had this, what, this has been about two or three months. Balon, this one's been in use for... Balon's behind the camera, which is why I'm sort of looking at you guys asking that bit weird. Anyway, let's fire up a game. I'm loading up Doom: The Dark Ages because that has a good looped demo mode where we can just run over and over again. If you want to just watch this, Balon. So, this is loading into the game at the moment. Interestingly, I'm only seeing load on two wires. You seeing that there's a massive load imbalance which we weren't able to see, and that's not drawing much power. So, what are we at? 160 something watts. 170 now. Yeah. So, there you go. So, the Wire View Pro has detected the problem. Why only two pins are sending current, I don't know. So, we haven't even gotten to the game yet. We're loading into the menu. So, there's a connection issue either at this end or the power supply end. The power supply end we haven't touched since we installed it. But there you go. That's why we're getting the overheating issues. If we didn't have this, we would have never known. So, that actually works. That's cool. There you go. So, pin one, 22 amps. Pin six, 0 amps. Why is it 0 amps? What has gone wrong? I'm going to do something dodgy. Let's stand this up. I just want to see if the power cord, because this is one of those side-facing power supplies. So, I've got the cord here. It's not hot at the power supply end, which, given it was hot at this end, would suggest that the resistance issue is at this end. So, this cable here, when it was plugged into the graphics card and when it's plugged into this, there's a problem here. And I think part of the cable has already melted from the previous test. So, we'll shut the system down. So, the Wire View Pro all looks good. It's already quite warm. Okay. So, I don't know how easily you're going to be able to see it, but that cable there looks melted at the tip of my fingernail. There's like a shiny bubble on that. So, that must have been the cable that had the 20 amps going through it. And this end, which apparently has zero for some reason, looks good. Now, the connector at this end's okay, but we know there's a problem. Can't grab the cable that easily to push it back. But this one looks like it has receded slightly. We're talking only a very small difference compared to the other cables. Or those two cables, I think, are stuck together. But have they melted together or were they actually bonded together when manufactured? I think they were pre-bonded, which makes the pin slipping back even harder to believe. There we go. I've got them apart. So, they hadn't melted together. That pin's pushed all the way in. It won't go further in. Well, let's just show the temperature it was getting up to. So, all of those cables should be in correctly. I don't know why this has just started playing up now. It was fine yesterday when I was using it. I took it out yesterday. It was cool as a cucumber when I took it out. So, it was the underside though that was getting hot on the thermal gun, wasn't it? So, I have to lay the case back down. So, we've removed the Wire View Pro. So, this is, I guess, out of the box. This is Nvidia's safety features on display. So, we'll fire Doom back up. So, as I said, we're firing Doom back up without the Wire View Pro installed, which I obviously wouldn't recommend. And at this point, now that we know there is an issue, it would be wise to replace that cable with a brand new 12V high-powered cable. I don't know why that one looks fine visually. I can't spot anything wrong with it. It was working. I didn't change anything. Now it's not working. And the Wire View Pro told us there were pins or wires that are no longer sending any current. So it's not even like a slight load imbalance. It's just for some reason of the six wires, at least two of them just don't send anything anymore. Though when we were getting the bars, it looked like only two of the six wires were actually sending, well, conveying current. So not really sure why that is. But I'll put this on here so we can see the actual temperature. I'll just get this to re-calibrate. Give it a second. Okay. So, it looks like there isn't much room for Balon, so he's banging into things. This is just calibrating at the moment. It looks like that wire... So, see, that is the second wire is the problem wire, the one that I bent out. But if we try and bend that back, it's not like it's being pulled out. Like, I don't... Yeah, I can't imagine what's gone wrong there. Like, it's not like the wire would have snapped or broken without any physical damage to the wire itself. But that is clearly where the problem lay. Put a bit of load back on it for a brief moment, a brief period of time. Sorry, I'm bumping that all around everywhere on you. I try to hold it steady. Okay, so we're now under load. So, you can see the temperature is just skyrocketing up as we're trying to pull 20-something amps through that wire. We have seen it get as hot as about 170° on the thermal imaging. So I think it's pretty conclusive at this point that this is a failure. This will go up in smoke if we leave it. So I don't think there's much point doing that. We've seen the result of what happens when people just, you know, don't detect the issue like I have by touching the cable and burning their thumb. Not a great way to do it. I don't recommend. So yeah, this is definitely going to go up in smoke within the next half an hour if we leave it. And as I said, there were already signs of the cable starting to melt. It has started blistering and bubbling in spots. So, I'm going to shut it down. I'm going to change this cable. I'll let it cool down before I grab it, though. And we'll see if a new cable solves the problem, which, you know, I strongly suspect it will. Actually, we'll just do one more temperature update before I shut it down. We'll see how hot it's already gotten just while I was talking. So, we'll just calibrate this. Give it a second, guys. So the test has started again, as you can see by the temperature, which is sharply rising. So you can see how, without monitoring this, how quickly it would exceed 200° because we've seen it get to... it got to 170 within minutes. From 150 to 170, we were talking about, you know, two minutes and it was there. So once the problem occurs, you do need to monitor it quickly and basically stop using it. Okay. So I think you guys get the idea. The cable is running way hotter than it should be. So, let's go and replace it.
Okay, guys. I have found the real culprit. So, when I went to replace the cable from the power supply, I could not get it out. It was just jammed in there. I was basically hanging off this thing trying to get it out. Had to use a pair of pliers in the end. And it has completely melted into the power supply. So, the reason why the Wire View Pro was detecting no current on some of the pins is because the pins failed at the connection within the power supply. So, it was perfectly seated in the power supply. There was no user error, nothing like that. It was in there nice and snug. It just... it was a failure on the PSU end. So, the power supply unit has melted the cable at that end, which then started to cause issues at the other end. And perhaps the million-dollar question at this point, not that I'm blaming the power supply because we've seen this happen with all different models of 5090s or just graphics cards that use the 12V high power connector. We've seen it with different power supplies and stuff like that. So I don't want to throw be quiet! under the bus here, but it is their Dark Power 13 power supply. So super duper expensive high-end power supply and it has failed. It's fallen victim to the 12V high power connector. Bit of a shame. So yeah, just to recap, we can see clear bubbling. Oh, hopefully you guys can clearly see it. I can clearly see it. This pin here, or wire rather, has started to bubble up as the insulation has gotten too hot. So it started to melt and that was the problem that was caused from this end. So, the damage here was me, that's not melting, that's me with a pair of pliers trying to yank it out because I couldn't get it out. It was... Balon joked, 'Oh, maybe it's melted in there.' I said, 'I don't think so at this end.' But Balon was right. I was wrong. The failure occurred. You can see these ones completely gone. I mean, there's... I don't know if I'm holding that the right angle for you, but huge damage on that connector. Multiple points of failure. And in the end, thanks to the Wire View Pro, we know that only two pins weren't melted. And you know, full disclaimer, we do advertise the Wire View Pro. Thermal Grizzly advertise with Hardware Unboxed, but I just had it here because of that reason and was able to validate its usefulness. It does what it's advertised to do. It detects the failure. An alarm sounded. So, that was really cool. We actually got to test that. I suppose the good news in all of this is that my 12V high power connector on my ASUS Astral graphics card is perfectly fine. There is no damage at all on that. So, I suspect when we change the power supply, this card will continue to work just fine. I mean, until we potentially run into the problem again. Hopefully we don't. But as you guys know, it's a problem that can pop up. But my, what is this, like $8,000 Australian dollars? So my $8,000 Australian graphics card isn't broken. Have to admit I was a little bit nervous about that because I do use it for testing. And that's exactly what I was doing when this whole situation occurred. And yeah, I've sort of burnt my thumb a bit, but that'll heal. My graphics card wouldn't have healed without some sort of intervention from a repair shop. So there's not much more to say on this one. Bit of a scuffed pop-up video that we weren't planning, early morning thing, but yeah. So, I'm not really sure what else to say about this other than it is a problem, which you guys already know. It can happen and it's finally happened to us. Daniel Owen recently had a cable melt on his, I think it was a Founders Edition card. We've just had it happen on the Astral card. And it's really annoying that had I not pulled it out, that would have led to catastrophic failure, as in damaged graphics card. The power supply is already damaged well beyond my repair, that would have to be sent back under warranty. But yeah, had we not pulled the cable out to change graphics cards, it would have melted in both ends. So, it would have been melted to the graphics card and melted the power supply, and all of those parts would have needed to be sent back in under warranty. And sending a 5090 in for warranty at this point in time is scary stuff given, you know, availability, pricing, and all that. I don't know how we would go with all of that. But anyway, I don't know what else to say for this video. As I said, this is just off the cuff. I haven't... This was in no way prepared. It's like a behind-the-scenes video. It's not our usual pretty well polished, thought-out videos. It's just here's the thing that we think's happening. And sure enough, it did happen. So, thanks for watching. I guess I'm your host, Steve, and I'll see you again next time.