#5520151

nizzen

2018-02-15 10:27

Thank you for the info 🙂 I have EK threadripper block on my 1950x. Love from Norway

#5520202

David Lake

2018-02-15 13:39

nowhting?

#5520249

schmidtbag

2018-02-15 16:11

I don't get it - what exactly did they do wrong?

#5520254

Denial

2018-02-15 16:21

schmidtbag:

I don't get it - what exactly did they do wrong?

Their original threadripper CPU blocks barely covered all the cores, the fin area wasn't large enough (you can see the difference in the pictures of the article). The performance was significantly worse than competitor blocks. https://images.hardocp.com/images/articles/151460117490s2ycebe7_1_1.png I don't know why they designed it that way in the first place but I think they are handling it well.

#5520255

schmidtbag

2018-02-15 16:25

Denial:

Their original threadripper CPU blocks barely covered all the cores, the fin area wasn't large enough. The performance was significantly worse than competitor blocks. I don't know why they designed it that way in the first place but I think they are handling it well.

I'm surprised it makes that much of a difference. But considering the product wasn't faulty (just simply "sub-par") I would definitely agree they are handling it very well.

#5520264

Embra

2018-02-15 16:46

It was faulty(inferior) in that it did not do what those buying it expected. But yes, they are handling it as they should.

#5520283

Agent-A01

2018-02-15 18:27

schmidtbag:

I'm surprised it makes that much of a difference.

Even HSF designed for threadripper compared to normal versions on those Noctua heatsinks made a significant difference; normal versions didn't cover all the cores.

#5520313

RealNC

2018-02-15 20:28

So wait, they design them and then never test them? How did this get to market?

#5520316

schmidtbag

2018-02-15 20:36

RealNC:

So wait, they design them and then never test them? How did this get to market?

I'm sure they did. It's not like their design didn't work, it just wasn't as efficient as it could be. I guess EK just underestimated the importance of taking advantage of TR's entire surface area. After all, there are higher-wattage CPUs than the 1950X with significantly smaller dies, so if I were them I'd also assume "meh, we probably don't have to cover the whole thing".

#5520318

Embra

2018-02-15 20:40

They may have felt rushed to get something out, which by no means is a correct thing to do if so. They are trying to make amends, hopefully not too many were troubled by the fault.

#5520427

__hollywood|meo

2018-02-16 06:07

they were indeed rushing to push out a TR block & did not revamp the channels within...just sort of crammed the existing AM4 design inside & hoped the block would dissipate enough heat by virtue of the blocks mass, it seems over the years, EK has been intermittently plagued with QC problems & some questionable cost-saving measures that give their customers the short end of the stick. that being said, their official responses have improved massively. giving out replacement plates & partial refunds is a very solid move towards banking good will.

#5520439

KissSh0t

2018-02-16 07:56

schmidtbag:

I don't get it - what exactly did they do wrong?

This shows basically the internals of the cooler which lets you understand why it performs so badly for Threadripper. [youtube=nw-TqUjhdmI]

#5520484

Fender178

2018-02-16 11:54

I'm glad that they admitted their mistakes and they are fixing it at no additional cost to the customer. Nice job guys.

#5520664

__hollywood|meo

2018-02-17 06:03

yikes. the guys @[H] are absolutely eviscerating EK over this. i dont recall them reacting like this when the EK AIO units were leaking, or had defective pumps...maybe my memory is just fuzzy. anyway seems like they feel the PR response is less "good will" & more "lip service"