Scan & Intel CPU Warranty Issues

[stroberaver]

My advice would be to test the CPU using the retail heat sink, and if there is a problem, put in a warranty claim. Once you have established that the CPU is fine, go ahead and void the warranty by using a better heat sink.

Thanks for that tip. I'm waiting on Scan for an E8400, and I've got my fingers crossed that it will be from a batch that doesn't have these heat problems. If it turns out to be running too hot or has the cores reporting inconsistent temperatures, I fully expect it to be RMAed. One of the reasons for choosing it was the fact it was supposed to run cooler (and thus, quieter) than the old 65nm chips. I could do without the hassle of installing it with the stock HSF first to test it, so I'm looking forward to some clarification from Scan on this issue as well.

[Supershanks]

Tests run using Prime95, CPU-Z v1.43, CoreTemp v0.96.1 & Abit Guru
Bios was loaded with Default settings initially, the only amendment being to Vcore to match the Vid.
Prime was run for 3 hours on each CPU.
The respective VID for each CPU was taken from coretemp & Set using Abit GURU.

E8400
Pre Adjustment to 1.1225V

Q6600
Vid was correctly set in default bios.


Test Results:-

E8400 At Stock speed using Stock cooler
Load__Avge_Max__MIN
Core0 56.7__59___48C
Core1 56.1__58___49C
Avge. 56.4__58.5_48.5C

Idle__Avge_Max__MIN
Core0 39.8__46___38C
Core1 43.0__43___43C See Note *
Avge. 41.4__44.5_40.5C

Delta 15.0__14.0__8C

* Core1 on idle appears to be locked on 43C this is confirmed from the CPU-Z Dumps. The same result occured on the earlier water cooled test.

Q6600 At Stock speed using Stock cooler
Load__Avge_Max__MIN
Core0 60.3__62___49C
Core1 57.6__59___46C
Core2 60.4__62___49C
Core3 57.3__59___47C
Avge 58.9__60.5_47.75C

Idle__Avge_Max__MIN
Core0 35.7__39___33C
Core1 34.2__40___31C
Core2 37.2__43___35C
Core3 33.9__40___32C
Avge. 35.3__40.5_32.8C

Delta 23.6_20___14.9C

I have the CoreTemp files & CPU-Z Dumps if anyone would like them. I just pulled the logs into Excel & Calculated the stats

I will add the CPU-Z Dumps & The temps from them later if anyone is interested. They bear out the coretemp results, using the methodology decrined in post#13

Coretemp only reads the CPU register see directly access the data from within the Intel CPU register that holds the digital thermal sensors (DTS ) for the respective core. The absolute temp is purely TjMax (105C on 8400) - DTS value. I have verified this myself using the method described. (cpu-z dump). Now that basically is so simple it puts the onus firmly with the CPU. This leaves us with the following posabilities.

Conclusion
I'll leave you to make your own. , but to me the E8400 or my E8400 doesn't live up to the advertised claims ogf being cooler, it's 2 cores are barely that cooler than the 4 cores of the Q6600. Please also bear in mind when looking at the results the motherboard was on a desk, not in a HTPC case.

[wesleyaldred]

Hello Everyone

I have spoken directly to the Intel rep who luckilly is here in person at Scan today. Firstly Intel WILL NOT void the warranty on the CPU for use of aftermarket air or water cooling, providing the use of such has not directly contributed to the failure, however, they will not take into account any results or temperatures produced whilst aftermarket cooling is in use.

Specifically any readings provided to Intel for warranty purposes MUST have been obtained whilst setup in a system using the supplied heat, sink & fan with the retail CPU.

Hope this clarifies the situation for everyone.

Wesley Aldred
Returns Manager

[pauldarkside]

Hope this clarifies the situation for everyone.

Thanks very much for straightening this out - I'd subscribed to this thread as it greatly affects my purchases. Only accepting temperature measurements using the stock heatsink/fan does make sense in regards to a direct comparison to their readings. However, you'd think the water cooled setup would highlight the fact that something is definately "odd" with those chips.

[w3bbo]

Hello Everyone

I have spoken directly to the Intel rep who luckilly is here in person at Scan today. Firstly Intel WILL NOT void the warranty on the CPU for use of aftermarket air or water cooling, providing the use of such has not directly contributed to the failure, however, they will not take into account any results or temperatures produced whilst aftermarket cooling is in use.

Specifically any readings provided to Intel for warranty purposes MUST have been obtained whilst setup in a system using the supplied heat, sink & fan with the retail CPU.

Hope this clarifies the situation for everyone.

Wesley Aldred
Returns Manager

Well they can't prove afaik that any aftermarket cooling was used so it would be hard for them not to honour the warranty. In any case its good to know we wouldn't have to lie to them about what cooler/cooling was used as you have confirmed it has no bearing on the matter apart from temps reported. Thanks for clearing that up.

[Clunk]

Hello Everyone

I have spoken directly to the Intel rep who luckilly is here in person at Scan today. Firstly Intel WILL NOT void the warranty on the CPU for use of aftermarket air or water cooling, providing the use of such has not directly contributed to the failure, however, they will not take into account any results or temperatures produced whilst aftermarket cooling is in use.

Specifically any readings provided to Intel for warranty purposes MUST have been obtained whilst setup in a system using the supplied heat, sink & fan with the retail CPU.

Hope this clarifies the situation for everyone.

Wesley Aldred
Returns Manager

Thanks for the reply Wesley, but what about the temp issue that supershanks has described above using the stock cooler?

I'm very keen to know how a 45nm dual core can be 3c less than a 65nm quad core at the same room temp, using the stock Intel coolers, and yet claim to be "Eco-Friendly. Faster. ‘Cooler’." Also, see here for the same "cooler" phrase used to describe the CPUs.

So, going off the above information, that 45nm runs "cooler" than 65nm, mustn't it be true that if the "hotter" running 65nm quad core is only 3c more than the new 45nm dual core, that Intel's claims aren't true in this case, or that the chip is simply faulty?

[wesleyaldred]

Supershank should contact Intel directly to discuss this.

Wesley

[Clunk]

So, are you still saying that the CPUs with the heat issues can't be returned to scan for a refund?

Edit: With all due respect, it should be Scan that is talking to Intel on the customers behalf. There are lot of people watching what happens here.

[scruffy]

While you've got Intel there, can you ask them if they've released a batch of faulty CPUs that read 15 degrees different between cores across multiple motherboard vendors on up-to-date BIOSs please?

Can you also ask them why they've provided no input into the well documented problem too?

Asus, Gigabyte, Abit etc.. are all affected on the P35, G31, 650 and 680 chipsets. I find it hard to believe that every single chipset would exhibit the same fault of random and sticking temperatures.

Abit in particular have been toting 45nm support proudly across their flagship range, when the CPU is finally released in it's retail form they withdraw that from their website and release beta BIOSs to counter 127-240 degree readings at BIOS/uGuru level. What did Intel change from the ES to the retail chip to make such a monumental error in reading of temps?

I'd love to hear their official word on this, while I'm torn between the board and the CPU at fault, recent evidence suggests this is an Intel owned foul up.

[wesleyaldred]

Yes, if Intel verify that the temperatures are outside of their threshold for that CPU running with the stock cooler then the item can be returned to Scan within 28 days from purchase for a refund.

Obviously as per my earlier post supershank will need to speak to Intel to verify this.

Wesley

[wesleyaldred]

Scruffy unfortunately I cannot as he left hours ago.

Wesley

[scruffy]

Yes, if Intel verify that the temperatures are outside of their threshold for that CPU running with the stock cooler then the item can be returned to Scan within 28 days from purchase for a refund.

Obviously as per my earlier post supershank will need to speak to Intel to verify this.

Wesley

I think the real question really here is not will they honour specifically his return, but more over the returns of the countless other people out there in Internetland who are having the same problems. I presume you as the returns manager met with Intel to discuss the finer points of what you can RMA for a reason? Is there something wider spread perhaps about to happen?

Scruffy unfortunately I cannot as he left hours ago.

Wesley

Got an email address for him?

[wesleyaldred]

I did not discuss any persons specific case, the information i received from Intel was general information which will be applicable to everyone.

[scruffy]

I did not discuss any persons specific case, the information i received from Intel was general information which will be applicable to everyone.

I think the real question really here is not will they honour specifically his return, but more over the returns of the countless other people out there in Internetland who are having the same problems.

I appreciate you can't/won't discuss individuals cases, however I'd be more interested to see what their stance is on the widely reported issues with their latest 45nm releases.

[wesleyaldred]

Sorry but i have no idea, certainly nothing was said to myself on this.

[Supershanks]

Specifically any readings provided to Intel for warranty purposes MUST have been obtained whilst setup in a system using the supplied heat, sink & fan with the retail CPU.

Hope this clarifies the situation for everyone.

Wesley Aldred
Returns Manager

I specifically ran the test yesterday with the E8400 using the stoick cooler. I also compared irty to my q6600 go , using the stock cooler.

Supershank should contact Intel directly to discuss this.

do you want to provide a contact please ??