Intel Recalls Boxed Xeon E-2274G Processors Due to Inadequate Stock Cooler Effectiveness

[EarthDog]

Oops, lol. Should have stuck with the copper slug version. Oh well.

Man there are a shitload of Intel haters here...lololol. it's like they took your first born or something.

 

[bonehead123]

Man there are a shitload of Intel haters here...lololol. it's like they took your first born or something.

Maybe, but there are also a LOT of "el-cheapo crap" haters here, regardless of who dumps it on us

I run intel cpu's in all my rigs (6 at home, 400+ at work), but I would neva, eva, eva even think about using one of their p.o.s. coolers....

 

[newtekie1]

I run intel cpu's in all my rigs (6 at home, 400+ at work), but I would neva, eva, eva even think about using one of their p.o.s. coolers....

I use their stock cooler on ever CPU that isn't overclocked. I haven't really found a reason not to use their cooler. And actually, I find their fans seems to last forever, while I find myself replacing fans after a year or so when I use cheaper after market coolers(looking at you here Hyper 212).

 

[efikkan]

I would prefer if by defaul all CPUs were without boxed coolers. If some manufacturer wanted to bundle one, they could make a special edition. Of course, beside TDP they should publish also some upper power requirement for CPU, e.g. PPT or PL1 or PL2 or whatever.

I'm glad Intel stopped shipping boxed coolers with their K-models, and I hope they will stop doing so with the rest. Boxed coolers will always be the bare minimum, and most custom builders will throw them in the trash anyway, so you're basically paying for trash, even if it's just a few dollars.

I've seen OEMs like Dell put these in machines without any case fan, and even on 65W TDP CPUs they get loud and eventually throttle during sustained load or in some cases even crash the computers.

These coolers are only sufficient if you don't put any sustained load on the CPU, like just light surfing and office work may be fine, but not anything that puts real load on the CPU.

 

[Keullo-e]

For Core 2 the stock coolers do their job, but like back in the day i7-4790K throttled with the stock cooler for example... there's a reason why they stopped shipping the K CPUs without one.

 

[Apocalypsee]

Yup Intel boxed cooler getting crappier by the day. Look at boxed cooler during LGA775/1155/1156 days and compare to new cooler. You get why people annoyed by this. For the record I tried using boxed cooler with i5 4670K, after 2-3 application the plastic clip holding the CPU to the motherboard already loose thus the heatsink don't make proper contact with the CPU.

 

[MelonGx]

LMFAO Intel are fooled by the faking TDP made by themselves.

 

[R-T-B]

Oops, lol. Should have stuck with the copper slug version. Oh well.

Intel has some ok stock coolers, certainly designs capable of cooling this anyhow. Anyone remember the old lga1366 tower one?

I find their fans seems to last forever

I mean, the fans are Nidecs, so yeah. They are actually pretty darn solid.

Look at boxed cooler during LGA775/1155/1156 days and compare to new cooler.

I am unsure what your point here is, this same design has been around SINCE that era virtually unchanged.

The only varations in consumer land are the copper slug version and the weird briefly made lga1366 tower fan.

 

[Apocalypsee]

I am unsure what your point here is, this same design has been around SINCE that era virtually unchanged.

The only varations in consumer land are the copper slug version and the weird briefly made lga1366 tower fan.

I don't mention design changes. Of course I meant the copper slug, and thicker aluminium. And their plastic clip legs are more sturdier than newer cooler. They made their processor more 'cutting edge' but cooler is getting worse. You see how THIN those cooler are?

 

[Aquinus]

I am unsure what your point here is, this same design has been around SINCE that era virtually unchanged.

The only varations in consumer land are the copper slug version and the weird briefly made lga1366 tower fan.

I think the point is that even when I bought a Core 2 Duo E6600 with a 65-watt TDP, it came with the stock cooler that had the copper slug. The Xeon E-2274G has a higher TDP (83w) with a worse cooler (no copper slug.) Intel just flat out screwed the pooch on this one.

 

[1000t]

I'm glad Intel stopped shipping boxed coolers with their K-models, and I hope they will stop doing so with the rest. Boxed coolers will always be the bare minimum, and most custom builders will throw them in the trash anyway, so you're basically paying for trash, even if it's just a few dollars.

I've seen OEMs like Dell put these in machines without any case fan, and even on 65W TDP CPUs they get loud and eventually throttle during sustained load or in some cases even crash the computers.

These coolers are only sufficient if you don't put any sustained load on the CPU, like just light surfing and office work may be fine, but not anything that puts real load on the CPU.

For me it does not make much sense why boxed coolers exist. It's a separate part that has to be assembled by the user, and for many it's just a waste of money. If some manufacturer feels like it has play with psychological marketing or business tricks, I'd rather see a coupon for cooler instead. I've got that RGB stock cooler from AMD, but I've replaced it shortly. While it's not the worst thing in universe, I've only tried it because of some positivity about it. For me, it would be better to get maybe a €10 coupon or what could be manufacturing costs instead.

 

[hat]

I don't mention design changes. Of course I meant the copper slug, and thicker aluminium. And their plastic clip legs are more sturdier than newer cooler. They made their processor more 'cutting edge' but cooler is getting worse. You see how THIN those cooler are?

I remember seeing half height coolers for the first time when 45nm Core 2 chips arrived. They worked, but that was back before turbo when the clock speed on the box was the clock speed you got and TDP was mostly accurate. It's nothing like today where you have a "65w" i9-9900 that draws God only knows what during turbo.

For me it does not make much sense why boxed coolers exist. It's a separate part that has to be assembled by the user, and for many it's just a waste of money. If some manufacturer feels like it has play with psychological marketing or business tricks, I'd rather see a coupon for cooler instead. I've got that RGB stock cooler from AMD, but I've replaced it shortly. While it's not the worst thing in universe, I've only tried it because of some positivity about it. For me, it would be better to get maybe a €10 coupon or what could be manufacturing costs instead.

Many system builders, unlike you or me, don't necessarily care about having the best cooling. They just want to throw a computer together that will work, and call it a day. Ever built a computer for somebody else? Unless it's my computer, I'm not buying a fancy aftermarket cooler and spending a week fine tuning the thing to get the most out of it. Neither I, or more importantly, the person I'm building the computer for, has the time for that... nor do they care. They just want a computer that works and doesn't suck. They get stock coolers and no overclocking. I think Intel is on the right path here. They give you the plain, uninteresting cooler for the plain, uninteresting processors and they leave it out on the K series processors... the ones that are likely to be under a better cooler anyway.

 

[efikkan]

For me it does not make much sense why boxed coolers exist. It's a separate part that has to be assembled by the user, and for many it's just a waste of money.

Agree, it's also increasing the shipping volume of the product by at least 5x, for something that most will discard anyway.
I don't mind them offering a separate cooler though, like they did back in the Sandy Bridge-E days, but at least then people get an option.

I remember seeing half height coolers for the first time when 45nm Core 2 chips arrived. They worked, but that was back before turbo when the clock speed on the box was the clock speed you got and TDP was mostly accurate. It's nothing like today where you have a "65w" i9-9900 that draws God only knows what during turbo.

Intel's TDP is "accurate". According to spec, it allows burst speeds beyond the TDP for up to 28 seconds, but not sustained loads over TDP.
So the 65W i9-9900 will have short bursts over 65W, but will throttle and maintain a sustained power usage of ~65W, unless you disable power limiting in the motherboard BIOS settings.
But when the Intel stock cooler tries to cool it at 65W, it will be loud! Loud enough to annoy colleges in an office environment.

Many system builders, unlike you or me, don't necessarily care about having the best cooling. They just want to throw a computer together that will work, and call it a day.

Starting with Haswell, and even more so with Skylake and Zen 1/2, CPUs have been boosting much more aggressively pushing up towards the TDP in a way only heavily multithreaded workloads used to do. So the need for proper cooling have increased over time.

Even for a value gaming build, I wouldn't recommend anything lower than a cooler in the class of BeQuiet Pure Rock or Cooler Master Hyper 212 EVO. Even the stock coolers provided with Ryzen 3000 are loud and hot, so any build with a 6-core or more from either Intel or AMD deserves a better cooler, and downdraft coolers are inefficient in most computer cases. Finally, you should always have some thermal headroom, as there will always be some buildup of dust in computers.

 

[hat]

Intel TDP is measured at base clock. That means the 65w i9-9900 will either pull around 65w or produce 65w worth of heat (forget which it is) while under load... at 3.1GHz. Max turbo goes all the way up to 5GHz (not sure what all core turbo is). While turbo is active, that 65w TDP goes flying out the window. Of course, it's the same way with AMD.

On dust buildup, yeah, that happens, however it is the user's responsibility to either clean the computer once in a while, or take it somewhere to be cleaned. Sure, most users don't do this... but that's nobody's problem but their own.

On downdraft coolers, I note that these coolers shouldn't pose an issue in any case with adequate airflow. These coolers also have the added benefit of moving some air over VRMs and such.

 

[efikkan]

Intel TDP is measured at base clock. That means the 65w i9-9900 will either pull around 65w or produce 65w worth of heat (forget which it is) while under load... at 3.1GHz. Max turbo goes all the way up to 5GHz (not sure what all core turbo is). While turbo is active, that 65w TDP goes flying out the window. Of course, it's the same way with AMD.

No, let's end this nonsense right now.
As per Intel spec (chapter 5).

Power Limit 1 (PL1): A threshold for average power that will not exceed - recommend to set to equal TDP power. PL1 should not be set higher than thermal solution cooling limits

Sustained power (even with Turbo) should be steady at PL1 (TDP rating).


AMD on Zen have a different approach, where TDP is not tied directly to average sustained power, like Intel's is.

On downdraft coolers, I note that these coolers shouldn't pose an issue in any case with adequate airflow. These coolers also have the added benefit of moving some air over VRMs and such.

Downdraft coolers will in most cases recycle some of the air. These really only work well in special cases or open test benches.
Tower coolers in a case with proper intake and exhaust fans and positive pressure will not pull in all the air flowing by, and will not dump its hot air down on the hot VRMs.

 

[newtekie1]

I am unsure what your point here is, this same design has been around SINCE that era virtually unchanged.

The only varations in consumer land are the copper slug version and the weird briefly made lga1366 tower fan.

There have been major changes since the 775 era. In fact, back in the 775 days there were at least 4 different stock coolers. There was the thin all aluminum cooler, the thick all aluminum cooler, the thick aluminum with half copper slug, and the thick aluminum with full copper slug. All of these had curved fins to aid in airflow.

The current cooler, AFAIK, is just a thin all aluminum cooler with straight fins.

Edit: It seems Intel actually currently has two different stock coolers that they provide with their CPUs. The lower TDP CPUs seem to come with a stock cooler with straight fins, while the higher TPD chips seem to come with a cooler with curved fins.

LMFAO Intel are fooled by the faking TDP made by themselves.

The big problem that Intel has put themselves in is that the higher you push the clock speed, the more inefficient the processor gets. People expect the processors to run at the boost clocks these days. If the processors drop to their base clock, they call that throttling these days. Whether or not that is the correct way to think is a totally different topic.

So Intel is pushing these high boost clocks and the processors get really inefficient and there is no way the stock cooler can handle them. The 2274G is a 4-Core part, but the boost clock is 4.9GHz, and that is where the stock cooler is going to fail. On the other hand I have an i5-9400 that under full Aida64 stress test load will maintain it's all core boost under load with the stock cooler at only hits 75°C. But it's all core boost is only 3.9GHz. At at that 3.9GHz, HWInfo says the CPU is only using ~55W. I know that power reading isn't totally accurate, but it gives an idea. But anyone that overclocks will tell you that the heat output rises exponentially as clock speeds go up(and the voltages increases go up to make those clock speeds possible).

 

[AusWolf]

Would it hurt them to replace the heatsink with one that has the copper core in it? I'd never return a processor just to get it back without the cooler.

 

[newtekie1]

Would it hurt them to replace the heatsink with one that has the copper core in it? I'd never return a processor just to get it back without the cooler.

It doesn't even have to have a copper core in it, just make it taller. If they doubled the thickness I'm betting it would be fine. Copper costs about 3x as much as aluminium. So doubling the amount of aluminum, by doubling the height of the cooler, would probably be much cheaper than adding a copper slug.

 

[hat]

No, let's end this nonsense right now.
As per Intel spec (chapter 5).

Sustained power (even with Turbo) should be steady at PL1 (TDP rating).
View attachment 136716View attachment 136717

AMD on Zen have a different approach, where TDP is not tied directly to average sustained power, like Intel's is.


Downdraft coolers will in most cases recycle some of the air. These really only work well in special cases or open test benches.
Tower coolers in a case with proper intake and exhaust fans and positive pressure will not pull in all the air flowing by, and will not dump its hot air down on the hot VRMs.

Thermal Design Power (TDP) represents the average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload. Refer to Datasheet for thermal solution requirements.

Note, it says Base Frequency, which in the case of the 65w i9 9900, is 3.1GHz. Anyone who's looked at a review for a recent Intel chip knows those TDP figures are well under what the chip in question is capable of pulling, at least if it's a high end chip like this one.

 

[Chrispy_]

its things like this that may have caused AMD to develop the wraith cooler.

Whilst the original Ryzen 1000-series coolers (Wraith Stealth, Wraith Spire, Wraith Prism) were all good for their respective TDP ratings, AMD seem to be cutting corners now.

The 3000-series Wraith Stealth now uses a nastier 7-blade fan than the original version of the Stealth. It is much noisier because of the blade angles. I'm sure it moves more air, allowing AMD to bundle a the cheaper/lighter stealth cooler with chips that would otherwise have needed a Spire cooler instead. The high-quality 5-blade Coolermaster fan on the original Stealth and Spire was QUIET. The new one is audible over typical case fans at almost any RPM.

The 3000-series Wraith Spire now comes without a copper core, and I've seen listings on ebay for non-copper Wraith Spire coolers with both the high-quality, quiet 5-blade fan, as well as the newer, noisy 7-blade fan. The loss of the copper core is pretty bad, but added to the worse fan is a kick in the teeth. I believe the LED-equipped Spire on the 3700X is still using a copper core and quieter 5-blade fan.

The Wraith Prism MAX that I recieved with each and every 3900X I ordered seems to have a nasty fan compared to the original Wraith Prism. Perhaps they've changed the fan OEM but whilst the 2700X and 1800X version of the Wraith MAX had balanced, reasonably quiet fans, most of the new 3900X coolers seem to have no fan balancing and vibrate because of the imbalance. Given that the 3900X will use around 140W with PBO enabled, and with Zen2's rush-to-idle behaviour, the Wraith Prism MAX constantly revs up and down making both a drone from the fan blades at higher RPM and humming loudly because of the imbalance.

So there are now at least a dozen variants of the "three" coolers AMD offer. Take a look on eBay as people sell unused or lightly-used boxed coolers from all three Ryzen generations. Between the copper slug, 5/7 blade fan, no-LED/white-ring/RGB there are probably closer to 20 permutations of AMD Ryzen cooler now, and only a few of them are equal to the well-reviewed models that were extensively tested all over the web at Ryzen's launch.

If AMD/Intel aren't going to bother putting effort into noise levels any more, I'd rather they just didn't bundle the coolers at all. AMD and Intel have no problem putting labels on CPU boxes explaining that there is no integrated graphics and you'll need to use a discrete GPU, so they should have no problem putting a sticker on the box and the IHS warning people to install a cooler. Joe Average may not realise that he needs to buy a cooler before he sees the CPU but he can at least READ a warning label.

 

[Keullo-e]

Heh, maybe they sold those crappy coolers to Asus, you know those Dual coolers in Asus' cards..

 

[DanTampa]

The author incorrectly refers to the Intel CPU cooler in question as a "E973708-003". As far as I know, that is not a valid part number. There is however, a "E97378-003". I have been building workstations and entry level servers around the E-2176G boxed CPU's and the cooler that comes in the box is a "E97378-003". Each and every one of them has had a copper core. The E-2176G is a 80w TDP processor and the E-2274G (subject of the recall) is a 83w TDP.

I believe the CPU cooler the author meant to refer to is the "E97379-003", as can be seen if you look at the blue label in the pictures he provides. The heat sink of that model does not have a copper core, and that is the model that was typically used for CPU's up to 65w. As as I stated, my E-2176G boxed CPU's have always included the correct "E97378-003" copper core CPU cooler which I believe is used for CPU's up to 95w. Since no other boxed E-21XX or E-22XX series CPU's were recalled...I agree with the author's assumption...I suspect the recall came as a result of some "E97379-003" CPU coolers mistakenly being packaged with some of the E-2274G boxed processors.

 

[Keullo-e]

Copper core, these days? Without a pic I can't believe, last time I've seen one was with Pentium G3258 (the unlocked, you know).

 

[INSTG8R]

I mean how long can they push those almost substandard coolers for so many any years with zero improvement while pushing their CPU TDP and thermals up all the while before they became truly “inadequate“? They were always just barely adequate to begin with. I’ve got an old Opteron 170 cooler from the SKT 939 era that’s just about the same as the current Wraith Prism, heat pipes and all. Intels aluminum design hasn’t changed or improved in a decade, even getting the cooper core one has become rare...