Bizarre: dying CPU or Power Supply or other?

[Noel 4,504]

I think thats a great idea Noel - good luck - I will have to try that someday also but now now  :wink:

 

I did it tonight, applied a square pattern by making an X first, then filling in the spaces to make a square.  Temps are maybe a little lower but not much, and Core1 still is about 3-5F higher than other cores, so not much change.  Realtemp only displayed 4 cores for some reason, even though it ID'd my 3930K.  Probably put a little more volume than I should have, but I didn't see any evidence that much if any oozed out--I had to use a dental exam mirror and bright light to inspect.  

[martin-w 3,238]

Looks to me that none of these methods solve the problem of getting full contact.   The X method gave the biggest coverage which makes some sense.  The spread method creates bubbles.  The thicker line is similar to the blob method.  After looking at this video, then looking at this layout of 6 cores in a 3960X CPU, it strikes me the best approach is essentially a square blob--after all, these are square mating surfaces, not round ones.  All of these methods potentially skimp a little on the corners & memory controller of the die:

 

 

I think I will for my own satisfaction try a bubble-free square shaped blob.  I will test this out w/ a piece of glass on a hard surface to see how big a square will cover this surface area.  I won't outline a square, I will try a filled in small square.

 

 

 

 All of these methods potentially skimp a little on the corners & memory controller of the die:

 

 

No they don't!

 

 

 

As I said previously, although it seems to be ignore Martin time...

 

You seem to be missing the point Noel. Your image above is of the "die". The IHS is "much" larger. The die is "small", and in the very centre of the heat spreader. The vast majority of the heat is in the centre of the IHS! Covering the far corners, or worrying about "full" coverage of the IHS is not necessary. It makes very little difference.

 

A  blob in the middle, and then relying on the pressure of your cooler to spread it is sufficient. Pretty much no gains will be made by obsessing about covering every last square millimetre of the IHS!

 

 

 and Core1 still is about 3-5F higher than other cores

 

As I said... one core will always be hotter. It's been like this on the last three CPU's I've owned. It's not an issue, it's normal! I've yet to hear of an Intel CPU that doesn't do this.

 

To sum up... one core is always the primary core, and will run a little hotter, dependant on the software being run. Again... normal on a quad core.

 

temp difference in one core can also be related to the inaccuracy of the temperature diodes in the die itself. They tend to be inaccurate, especially when far from TJ Max.

 

In fact, just a few degrees is very good, many chips have one core much hotter than that.

 

As for the merits of a UPS... for the vast majority it isn't required. A reasonable quality power strip like the one I linked to, with surge protection is enough, plus the protections built in to all modern PC PSU's!

 

However... if you live in a region with pretty crap mains power, with frequent outages, spikes brown outs and all manner of silliness... then by all means a UPS would be a useful addition to your kit.

 

Something tells me that once some peoples minds are made up, nothing will convince them otherwise though, and they will spend their money through fear. so be it. Personal choice.

[Gary A 8]

As mentioned I have my primary surge protection between the AVR

 

Hi Noel, I was responding at the same time you were, missed your post.  Also responding for the benefit of the masses that may be reading and actually be interested in power protection, even for my own edification.

 

 

 

risk for damage seems to be minimal

 

 

Well....depends on how you wish to define minimal.  If your personal choice is to spin the wheel or otherwise console your choice for protection or no protection, then you may wish to define as minimal. :smile:

 

For those more statistically inclined I offer a: NEMA  study indicating 31% of respondents indicated equipment damage.  That study consistent with a German Engineering study HERE.   A white paper concluding that power protection was prudent HERE.  Coupled with many North American utility providers recommending protection on their websites like HERE  HERE  HERE and so-on.

 

I suspect the failure rates are actually higher than 31% and that many failures are not reported or not adequately diagnosed as attributable to a power issue.   Example, MB fails and user simply buys new or returns for warranty and no cause of failure is ever determined.

 

Of course what any one person chooses to do or not to do is their business and of little concern to me.  That is an open statement not directed to you Noel specifically.  As I said before, most of my electronics are not protected (doesn't make it right) and certainly doesn't make the best advice to others to be that it isn't needed.

[Noel 4,504]

 

 

The vast majority of the heat is in the centre of the IHS! Covering the far corners, or worrying about "full" coverage of the IHS is not necessary. It makes very little difference.

 

Yes, makes sense, very little difference.  OTOH, as long as you aren't slathering on a tablespoon what's the harm using a small square versus a small circle?  No harm, and very likely very slightly better heat transfer!   What I haven't seen yet is a photo of how much smaller the die is than the IHS.  Was the IHS designed so much larger than the die to protect from seeping interface media?

[Noel 4,504]

For those more statistically inclined I offer a: NEMA  study indicating 31% of respondents indicated equipment damage.  That study consistent with a German Engineering study HERE.   A white paper concluding that power protection was prudent HERE.  Coupled with many North American utility providers recommending protection on their websites like HERE  HERE  HERE and so-on.

 

I suspect the failure rates are actually higher than 31% and that many failures are not reported or not adequately diagnosed as attributable to a power issue.   Example, MB fails and user simply buys new or returns for warranty and no cause of failure is ever determined.

 

Of course what any one person chooses to do or not to do is their business and of little concern to me.  That is an open statement not directed to you Noel specifically.  As I said before, most of my electronics are not protected (doesn't make it right) and certainly doesn't make the best advice to others to be that it isn't needed.

 

I think the whole thing is still way overblown and I base that on loads of personal experience and the casual testimony of friends and neighbors who have zillions of appliances, PCs, and other devices sporting microcircuitry that simply rarely have problems.   I don't have whole house surge protection.  The white paper mentions tales of woe even w/ whole house surge protection.  My close friend had all of his electronics and then some zapped when his mountain home was stuck by lightening--surprise, not!  Insurance replaced it all.  I have been lazy about putting anything but an adaptor to go from 2 outlets to 6 and that adaptor in one living room outlet, God help me, I've had a digital Cable TV box, surround audio system, and a now 9 y/o Sony LCD plugged into it the whole time with no failures.  Again, where I work--80 desktops, never heard of one dying yet--they're replaced for other reasons than hardware failure.

 

I think a decent surge protector is good enough for all practical purposes until you're dealing w/ mission critical stuff which might warrant highest quality protection.   Our local power supply must be pretty decent.  Since I have 3 midi workstations, a drum machine, an 8 channel mixer, a stereo amp, two PCs, and a small LCD TV in my playroom it's worth the $54 to provide whatever AVR really offers, if anything.

 

So what killed, apparently, the master boot sector on my SSD?  I'll never know--and in the end, it's now been wiped and reloaded but now I learned by P3D license won't reactivate and they will get back to me in '2 business days', meanwhile I have a whole weekend I could have played around w/ setting up 2.5 with, but can't!  The bastages!   

[martin-w 3,238]

Yes, makes sense, very little difference.  OTOH, as long as you aren't slathering on a tablespoon what's the harm using a small square versus a small circle?  No harm, and very likely very slightly better heat transfer!   What I haven't seen yet is a photo of how much smaller the die is than the IHS.  Was the IHS designed so much larger than the die to protect from seeping interface media?

 

I was going to say take a look at the Ivy Bridge delliding guide in this forum Noel.  However, the images aren't available it seems.

 

Your CPU is soldered though so no one will have peeked.

 

 

 

 

This will give you a rough idea...

 

 

 

Nothing to do with seeping TIM as far as I know. Just offers more stability in terms of cooler mounting I suspect. The IHS covers most of the PCB. The IHS is exactly that, a heat spreader, thus the surface area is large.

 

 

I think the whole thing is still way overblown and I base that on loads of personal experience and the casual testimony of friends and neighbors who have zillions of appliances, PCs, and other devices sporting microcircuitry that simply rarely have problems.

 

 

Yes that's absolutely correct. It seems to me that a UPS is more of a US obsession than here in the UK.

 

I suspect the failure rates are actually higher than 31% and that many failures are not reported or not adequately diagnosed as attributable to a power issue.   Example, MB fails and user simply buys new or returns for warranty and no cause of failure is ever determined.

 

 

Hi Gary,

 

That statement above means nothing, it's not factually based at all. 

 

The problem here is that all of these quoted percentages are actually guesses. Same for electrostatic discharge and other failures. The point is we have no idea how many systems are damaged through mains voltage irregularities, or indeed through other causes like ESD. It's impossible to determine such a thing with any accuracy at all. Did my component randomly fail in my PC because there was a mains voltage fluctuation? Did the component in my PC fail because there was an electrostatic discharge when I cleaned the dust out of my PC? Did that same component fail due to a manufacturing defect that only reared it's ugly head months or years down the line. Did that component fail due to wear and tear?

 

You get the picture. All we have is estimates, guesses, nothing definitive.

 

As far as mains voltage fluctuation and the merits of a UPS... generalising and quoting random percentages is silly. It doesn't take into consideration regional variations in mains power quality for one thing. Person A in a small Californian town might have a dodgy mains supply, but person B that lives in Birmingham England might have an awesomely stable supply.

 

 

or not adequately diagnosed as attributable to a power issue

 

 

Do you see how suspect that statement is? You could say that about absolutely anything, any PC or hardware failure. You could apply it to ESD, you could apply it to manufacturing defect, you could apply it to anything.

 

I could go way over the top and use the same logic as you and apply it to ESD ... and point out that ESD can damage a PC component at a much lower level than could be felt or seen by a human being, thus you would never know you had zapped your PC. I could quote research that "estimates" the number of PC components that fail due to ESD. I could quote invented figure like the previous posters 80 - 90%, or your guess at greater than 31%. But it means nothing.

 

We don't know hoe many components fail due to ESD just as we don't know how many components fail due to voltage irregularities. But what we do now is that those that live in a region with reasonable mains voltage rarely have an issue. If you don't, buy a UPS, if you do, don't waste your money.

[Noel 4,504]

UPS:  useful if you need to be certain you don't lose data before it's saved.  Example:  one of my hats at work is I run a diabetic retinal screening program.  Since it's very useful not to lose retinal images captured on patients sitting in front of a retinal camera before the images are saved, we use a UPS on the camera's PC.  For home PC use?  Never matters here.

 

'Automatic Voltage Regulation':  I got sucked into buying this device thru the discussions above, but I'm thinking it verges on totally irrelevant for my application.  At least I didn't buy a UPS w/ AVR which gets up to the $140 on up cost.  I may connect my 55" LCD TV and a few other goodies to it as if Gary's review was correct it has some surge protection built in.  Or, I may return it and turn a $54 probable mistake into a $10 one.

 

Gary's cited white paper states,

 

'Short-term (0.5 cycle to one minute) rms voltage variations between 0.1-0.9 p.u. are voltage sags...Data confirms that voltage sags are by far the most common type of disturbance. Although sags vary according to depth and duration, it can be stated that in general they do not cause damage...'    and, 

 

'For the first seven years of our power quality program, monitoring and investigations focused only on commercial/industrial customers. Damage claims due to power quality events were practically nonexistent.' 

 

'We conclude that low-level low-energy transient voltage surges are fairly common in every electrical environment due to capacitive and inductive load switching. Although somewhat inconclusive they do not appear to be a big concern on a day-to-day basis. Transients from utility capacitors rarely show up in a facility. The most common source of damaging transient surges in Oklahoma County is lightning, with an average of nearly 18,000 strikes per year (flash density of 8-16/km/year). Yet there are still unknowns [one of Martin's main points]. ' 

[johnman 7]

I'm NOT going to Read or Write a 50 Page White Paper about all the variables that were Discussed Here - in 5 Pages ! - I know a UPS or Battery Backup Works - and I Work It - And I am Very Happy that I Spent $75 - and know My Equipment is Protected - Has been Protected - and Will be Protected for Years to come - You got Sucked into NOTHING ! - Johnman B)

[Noel 4,504]

I'm NOT going to Read or Write a 50 Page White Paper about all the variables that were Discussed Here - in 5 Pages ! - I know a UPS or Battery Backup Works - and I Work It - And I am Very Happy that I Spent $75 - and know My Equipment is Protected - Has been Protected - and Will be Protected for Years to come - You got Sucked into NOTHING ! - Johnman B)

Who knows Johnman but the $54 AVR device from APC arrived and is installed, and justified or not, it does give me a modicum of peace of mind.  The bloody thing is heavy and also gave me the opportunity to shield a few other devices from transients and less than sterling power regulation.  I think for those who like the concept of 'conditioned power' it's a good alternative to a UPS which costs quite a bit more to get the 'conditioned power' piece.

 

SOOOO happy it was a corrupted SSD MBR or something.  Anyway, at the moment all seems fine, back to rock stable.   Thank you all for your input on this let's hope it sticks a while, at least until P3D 64-bit w/ full DirectX 12 support comes along.  Still hoping there is something meaningful to come from DX12.  I have version 3.1 running and I can say w/o a doubt--can't really tell much of a difference beyond slight changes in the GUI and the ATC window.  Hopefully the VAS management issue will make it worth having over 2.3.  Still waiting to get REX working which should help restore things visually.

[johnman 7]

Hey Noel ,

Glad You have things back on the Rails - Hope You are able to Enjoy this all Absorbing Hobby for Years in the Future - Remember - Even though it is Frowned Upon  - When Mentioned in these Hallowed Forms - Fun is the Underlying Result of US Being Here - Among US - Johnman B)