Discussion Thermaltake PSU defects? My rant and your thoughts?

[DeatM47]

Alright, So this is gonna be kind of a rant, But I wanted to share a very concerning experience that I had with a Thermaltake Toughpower Grand RGB 750 Watt PSU, This supply I got from my current workplace at my local high school where the system it was in was used for visual ASL teachings and used motion cameras to teach students ASL using their own hand movements, Needless to say it is a higher end system, Intel I7-8700k, 16gb DDR4 2666, And a Quadro P2000 on an Asus Mini ITX H370 Board. It came in to our tech office about 3 months ago completely dead, No power or POST, So when I took it apart after trying to test it I discovered that the modular SATA cable going to a 1TB HDD had a wire (a ground wire) that had completely melted through the jacket and had clearly gotten very hot, Now this was concerning because that to my understanding means there is either a short or something drawing insane amounts of current, (Ultimately I believe the former to be true) So we ended up ordering completely new parts and rebuilding it with a corsair supply and it has worked great since. I ended up taking all the old parts that we replaced (Board, PSU, HDD and the SSD) back home to test all of it, Where I discovered that inside the dead PSU there were multiple leaking caps, Which I am quite sure they have shorted. (I plan to gut this thing and test it with A multimeter, Will update when I get to it.)I ended up going down the capacitor plague rabbit hole because these capacitors that had leaked were Rubycon ZLH Series caps, The descendant of one of the same series that had an incorrect formula that was smuggled to China and Taiwan, (ZL and ZA) Looking at these capacitors you can clearly see that the printing on them is very low quality (Images below) compared the other Rubycon caps inside that are YXG series that have not leaked, And brand new Rubycon YXF caps I have for a project. So are there still fake formula caps being made now with ripped off Rubycon branding 20 years later? And to top all of this off, the included Molex to Floppy power adapter that I was using with a Windows 98 PC I was messing with was assembled backwards, and had 12V on what should have been the 5V pin, Smoking the floppy drive it was plugged in to. (Image below) This is very concerning and whether this is just me or super rare I do not think I am going to buy Thermaltake for my personal projects now. I would like to hear everyone's thoughts on this and maybe share similar experiences if you all have had any. Mad stuff!

				Images in numerical order 1-2: Leaked suspected fake caps with blurry print 3: suspected genuine cap with sharp print 4: Fried floppy drive IC from bad Molex to floppy adapter 5: Overhead of PSU 6: Pinout of defective included Thermaltake Molex to floppy 7-8: Melted ground wire on SATA power cable 9: Brand new genuine ordered from Digi-Key Rubycon YXF caps for example

 

[JohnBonhamsGhost]

of course there's always the chance that someone will receive a unit that may have some manufacturing flaws that somehow passed inspection.
there's also the chance that this system was severely mistreated before it came into your hands.

the Toughpower Grand RGB series' Gold units are pretty highly rated but Thermaltake does produce some very low quality units too.
i would personally never choose a Thermaltake power supply, but i also would never state that everyone should avoid them.

 

[4745454b]

My own personal thoughts. First, there is more that can go wrong with a PSU than bad caps. Look around the innards to see if there are other melted/broken parts. That may provide a clue as to what happened. Second, if those are bad/wrong caps, that might not be TT's fault. I don't remember which OEM makes that model, but even they might have bought parts they thought were correct only to be fooled. If it was the bad caps that caused the issue I'm not sure where the blame/fault lays.

I personally would run some of TTs units, after doing research as to which are good to use. Bad units can happen to anyone, even seasonic. All you can do is look at research and buy the best you can. Failures can and do happen however.

 

[DeatM47]

I can say for sure this computer was rarely powered off in it's life, It was run 24/7 only stopping when it had issues or needed updates. Hard life especially if it was the lucky one that got the bad egg supply. The PC came from some European company that does hardware consulting for global schools with special needs kids, And when I was working on it they did a poor job of cable managing so it likely got hot as well, Cooler master ITX case with the PSU mount above the CPU, That could be a factor as well.

 

[DeatM47]

All aluminum heatsink on an I7-8700k, IMO a recipe for disaster.

 

[DeatM47]

4745454b, I feel like what you said about the OEM is likely, Digging in a little deeper now I agree that whoever made this likely was fooled into buying fake parts, Someone in the supply chain for components, A Chinese broker or distributor or whoever probably found "Rubycon" caps for a cheaper price than other distributors and took the gamble and flowed down the line to whatever unfortunate products got them installed. Testing a bit with my meter those 2 caps that leaked are reading like they are nearly shorted, like almost acting as resistor, I am going to pull them out and put them in a tester and see what is up.

 

[DeatM47]

I also see zero other components that appear broken, nor do I see any heat marks, solder bridges or signs of shorts, tested all diodes and the full bridge rectifier, fuse all good. Very odd that that cable melted but nothing in the supply itself did.

 

[Juular]

I highly doubt there are any fake 'Japanese' capacitors, let alone that Sirfa would use them but all 'Japanese' capacitors are made in China now, i guess it's possible that there are some QA issues. What would need to happen to that wire on SATA cable tho, no idea, a single HDD shouldn't draw nearly enough current to do this, are you sure it was only powering the HDD ?

 

[DeatM47]

I highly doubt there are any fake 'Japanese' capacitors, let alone that Sirfa would use them but all 'Japanese' capacitors are made in China now, i guess it's possible that there are some QA issues. What would need to happen to that wire on SATA cable tho, no idea, a single HDD shouldn't draw nearly enough current to do this, are you sure it was only powering the HDD ?

The SATA power cable melted up to the first connector where it was connected to the HDD, This was a mini ITX case where there was not room for much else, As for the caps I still believe them to be fakes, Look up "Fake Rubycon capacitors" (or any other fake name brand cap) and you will see that a lot of times sketchy companies basically rewrap low quality caps as "Japanese brand" caps and try to swindle lesser distributors in Asia to buy them as genuines, This seems more prevalent on cheap aliexpress audio gear but who knows what these scammers have moved into, Fake parts still exist for sure, look up how various asian salvage companies rip old used IC's off broken or scrap electronics and repackage them as new by cleaning the leads and re-etching the chip part numbers. If the distributors or anyone else in the purchase line don't check their product then it makes sense they could end up in a product like this, On the contrary it could be Rubycon QC issue but the last time I ever recall seeing a genuine Rubycon cap leaking was on a Dell motherboard from a 2001 optiplex. Who knows, But I am still trying to figure out why that ground cable melted, At this point I am more curious than anything. Amazingly the ASUS motherboard and the 8700K survived, But that hard drive did not.

 

[DeatM47]

And I must admit there is likely no way to prove these 2 caps are fake unless you had a brand new sample from the original batch and chemically compared it to a genuine cap of the same series, But to my eye the printing on the plastic wrap is not consistent with new Rubycon caps, Or ones from even 20 years ago, Which all have sharp readable print with little blur, Now maybe if the bad caps heated up it could have warped it, But they are 105 Degree rated caps that even if they got that hot should not have experienced physical damage unless exposed for a long time to such conditions, ZLH series according to Rubycon specs are 6000-10000 Hours life under 105 Degree Celsius conditions.

Chart I made of a bunch of Rubycon caps in my collection comparing wrap print quality between the suspects and other Rubycon caps from different time periods.

 

[DeatM47]

I am not gonna shun Thermaltake too hard for one sample out of many products and millions of units. Still I feel it is an interesting example of what can happen in the industry and how it can sometimes affect a product. Everything else about that supply suggests it is of good quality, It has a very heavy duty looking design, And the fact they used Rubycon caps at all instead of lesser known ones is still a step in the right direction. Also considering it is a 150 USD supply my personal PSU bias leans toward something from Corsair of the same price, But they don't have RGB if that is your thing.

 

[Juular]

"Fake Rubycon capacitors"

Fake capacitors going through retail and small distributors is one thing, fake capacitors being used by a big OEM is a completely different story. Besides, the sharpness of the text on the label doesn't define the quality of the capacitor, as i said, they're now being manufactured in China too, they could've grabbed just whatever labeling equipment they got there at the partner factory, since that has nothing to do with the quality of the product itself. But yeah, them dying like that definitely says something about either the quality of the caps themselves or the Sirfa's design.

PS: About that Berg connector, i'm not completely sure since that's just the first entry in the google query but the pinout seems right. Your diagram is probably load-side. Nevermind, there are both versions on the net lol. Wikipedia says that the 5V (red) is supposed to be right by the key so in that case Thermaltake pinout is indeed incorrect. Says something about Thermaltake QA.

 

[Juular]

Everything else about that supply suggests it is of good quality, It has a very heavy duty looking design

No really all that heavy duty, it's a budget oriented OEM and fairly budget platform. 150 bucks ? Absolutely not. More like 70-80$, maybe 90$ if you really need RGB. Thermaltake GF1 would've been a better choice but it looks like they've swapped the internals for pretty much the same thing as here. Just Thermaltake being Thermaltake. Corsair would've been a way better choice, they have a very strong QA/QC, at least while JonnyGURU is still working his butt off there. So <Mod Edit> can happen with them too, just as with everyone else, but way more unlikely.

 

[DeatM47]

No really all that heavy duty, it's a budget oriented OEM and fairly budget platform. 150 bucks ? Absolutely not. More like 70-80$, maybe 90$ if you really need RGB. Thermaltake GF1 would've been a better choice but it looks like they've swapped the internals for pretty much the same thing as here. Just Thermaltake being Thermaltake. Corsair would've been a way better choice, they have a very strong QA/QC, at least while JonnyGURU is still working his butt off there. So <Mod Edit> can happen with them too, just as with everyone else, but way more unlikely.

My bad on the pricing I had accidentally looked up the 850W version due to google auto-correcting. And about the Floppy connector, comparing it to a 1998 Gateway PC's PSU Floppy connector (Which is where the floppy came from) And the image again, It is actually correct, And also correct for another floppy I had, So somehow that picture is wrong, Maybe It was twisted and it isn't visible? No clue, But the chip that roasted on that floppy had a motor driver in it, And I recall reading something after I had an 80gb Seagate IDE hard drive motor driver IC blow that when those old driver chips wear out they have a tendency to blow up, (which was my original theory before having a brain fart and misreading the pinout) So that was likely just my sad old Panasonic floppy biting the bullet, Nothing to do with the PSU or cable QC, My mistake.

 

[DeatM47]

(image of a hard drive motor driver IC that melted, similar to what happened to me and what likely happened to the floppy drive)