New overheating tests with photos and screen shots.

[philipma1957]

3 screens shots before and after cooling tweak. 2 photos before and after cooling tweak.


this is a 10 dollar tweak ugly but truly function first.

on its side is best for cooling.


but it could be turned on top of the cooling fan.


I am going to run 3 cpu tests. with max fan speed.

[philipma1957]

mods my apologies for double post but I want to show many screen shots. well a few minutes in and all is good with proper venting and cooling.


I am about 10 minutes in at this time my temps were 211 to 216 now 190 or less. I failed at the long run (100 reps) of this cpu test at the 1 hour mark and the 2 hour mark I then said stop test and check ram.

I still believed the problem was heat more then anything else. I will post this test until 1 hour or until it fails.

At one hour I will turn the mini down to be less ugly but it will still be on the laptop cooler.

It may go higher then the 185-190 it has been doing on its side. It may fail. .


If it still works at the 2 hour mark i will put the black cover on the mini and rest it on the laptop cooler.

all spots would be considered good venting or even great venting.

My guess is this will pass the side test and the sitting on the laptop cooler without a black cover.

I think it will over heat with the black cover on sitting on the laptop cooler.


18 minute marked all maxed temps are good. I will post back in 20 minutes or so. well between the 17 minute mark and the 37 minute

mark cpu test got stuck it was maxed at 100% on all 8 cpus but stopped counting higher. so I have become pretty sure that the test is flawed. and yesterdays and the other days failures were test related not heat. i also have been able to see a 25 to 30 degree heat improvement with the machine on its side. I will stop all cpu tests and now do prime95 tests. this is the cpu test without the skin or gui.

link


http://www.mersenne.org/freesoft/

so far it is running with no faults at 190 to 200 c . If I can run it for 1 hour with mini on the side I will try to run it for 1 hour with mini on the laptop cooler.

I am putting off hand brake tests until tonight.

[dasx]

I guess you're using manual set of the fans so could you try leaving it Auto and see if the bottomless mini needs less rpms in order to keep temps down?

Thanks!


P.S: I'll do my own tests as soon as I get my spare bottom cover and have time to mod it.

[philipma1957]

yes and here are my tests with prime 95 as


cpu test is not good.


use prime 95 at this link


http://mersenne.org/freesoft/

I am about 35-38 minutes into prime 95 and my high number is 203f all 99 to 100 percent.

in 20 minutes I will flip the mini in to normal position and see how much the temps rise.

I am 1 hour in.. since all is good I left the black plate off and flipped the mini.

[dasx]

Ohh, Prime95 is available on Mac? Sweet. I still remember it from my times overclocking PC based machines.

Post your results as soon as you've tested it with the auto setting for the fans.

[philipma1957]

Originally Posted by dasx

Ohh, Prime95 is available on Mac? Sweet. I still remember it from my times overclocking PC based machines.

Post your results as soon as you've tested it with the auto setting for the fans.

yes I will . here are the two photos of the first cooling spot and the mini simply flipped to normal with no black plate. now in under ten minutes temps jumped up a lot. look at screen shots at 1:51 flipped up and 2:01 flipped down.


down if they don't stabilize real soon this machine will overheat. sitting on an 8 inch fan with no black plate and house temp is 79f.

I am running this same test on a 2010 core 2 duo mac mini wall mounted as i type I am going to check its temps.

it is running at 170f and the fans are only 3600rpm.

but it is wall mounted on its side.

now some would say why not wall mount my 2012 quad,

because My wall mounted mini needs a dvd player.


I am now 2 hours in.

here is an interesting tidbit my cpu has had possible hardware errors multiple times even with cooler temps. worker 1 has had more then 1 possible error while worker 8 has been fine.


So my cpu may be a loser


workers 1 2 and 7 have had possible errors.

that is 3 of 8 cores so all my tests may show I just have a bad cpu.

i did quite a few ram tests but I think I am going to put the oem ram back in the mini and re test everything.


one thing is true so far no throttle 2 hours of tests at 995 to 100% with the aid in cooling.


I will put the black plate on right not to see if it gets real hot.

went up a few more degrees still avoided throttle highest temp is 214 to 216 .

but it is sitting on a laptop cooler .

One thing this testing has shown to me is my cpu or my ram or both cpu test and prime95 have a problem.

the possible rounding errors on 3 of 8 cores is not very good. 30 minutes into the test wit the black plate on and I have stopped

at 219f for a max still no throttle. yesterday and the day before I never got past 217 with out the machine locking up or throttle. this was with cpu test not prime95 test. still 219c is about 104c and i am sitting on a laptop fan! I will run this for a half hour more then i will test with oem ram. I would like to know if the ram the cpu or both programs are wrong.


I just put the stock ram back this time 4 cores are failing on the prime95 test.


I am thinking the cpu is defective. which is not good for me . but may not effect other users as much.

I still wonder if the black screen issue goes hand in hand with this. I am going on intel's forum for a while. see what I find.

[philipma1957]

cpu is bad. here is a 1 core or 1 worker test done with oem ram instant error, I set up an appointment at my apple store on monday.


to sum it up

mini fails cpu test


mini fails prime95 test

mini has black screen issue

usb3 has disconnects

usb3 has beach balls used 3 usb cases with 2 different drives

lacies t-bolt has a stutter issue with raid0 or raid1 not jbod same gear works on 2011 mini


used 3 sets of ram for cpu test and prime95 and got the same fails in fact the oem ram had worse results then the kingston.

I purchased at jr.com they don't take returns of opened and used computers so off to apple store on monday.

[All Taken]

Appreciate you took a lot of time doing this but how is this relevant to anyone? You had a faulty processor.

[dasx]

Originally Posted by All Taken

Appreciate you took a lot of time doing this but how is this relevant to anyone? You had a faulty processor.

He's testing also temperature fluctuations depending on different circumstances. Leaving aside the CPU being faulty that is of a great use, at least for me.

----------

Originally Posted by philipma1957

cpu is bad. here is a 1 core or 1 worker test done with oem ram instant error, I set up an appointment at my apple store on monday.


to sum it up

mini fails cpu test


mini fails prime95 test

mini has black screen issue

usb3 has disconnects

usb3 has beach balls used 3 usb cases with 2 different drives

lacies t-bolt has a stutter issue with raid0 or raid1 not jbod same gear works on 2011 mini


used 3 sets of ram for cpu test and prime95 and got the same fails in fact the oem ram had worse results then the kingston.

I purchased at jr.com they don't take returns of opened and used computers so off to apple store on monday.

Definitely so bad machine you got there. I hope this was just a one time thing with QC!

[All Taken]

I see that, however, given that the op has a faulty processor which is what has been used to give these results....

To me they simply tell me that these are the temperature results for a faulty QC in a mac mini. Really not of use to anyone, if the op had performed operating tests on a functional QC then it might be of use as a gauge.

[philipma1957]

Originally Posted by All Taken

I see that, however, given that the op has a faulty processor which is what has been used to give these results....

To me they simply tell me that these are the temperature results for a faulty QC in a mac mini. Really not of use to anyone, if the op had performed operating tests on a functional QC then it might be of use as a gauge.

yes you are correct. but if you have a quad you now have a link to test with prime95 .

you also know how to cool the machine down ; side mount is better then flat mount. the cpu broken or not still ran cooler with side mount, also when i get the replacement don't you think I am going to test it? darn right I am.

[atMac]

Originally Posted by All Taken

I see that, however, given that the op has a faulty processor which is what has been used to give these results....

To me they simply tell me that these are the temperature results for a faulty QC in a mac mini. Really not of use to anyone, if the op had performed operating tests on a functional QC then it might be of use as a gauge.

It's also possible that it was the high temps for so long that made the CPU fail. If that is the case we may see a few failures in the future. A core failure doesn't impact the heat much, if anything a non-working core will show lower temps implying that others could be running even hotter.

[cyclotron451]

It's very relevant to test heating/overheating parameters in the 2012 Mac mini, I've recently suffered a GPU meltdown in an early 2011 quad i7 MBP. I'd been using the laptop happily for a year, then I started to do some GPU assisted computing and the thing now needs a €496 new motherboard! MBP was 2 months out of warranty, no repair offered by apple, according to the genius it's not a series fault. Maybe the combined CPU/GPU MBP heat sink is a bit iffy, however I'm now doing the calculations on my 2010 macPro. It was enough to convince me to go for the lowest TDP CPU in my mini, hence just basic 2012 2.5 i5 mini, and wait a few years to do "cluster supercomputing" at home. I also decided to get the mini over the new anorexic iMac. Out of 6 Qi7 MBP's at work (very small sample) I know of two that have "overheating problems" including my fatal meltdown. I hope the mini statistics prove much better than 30% with problems, and it is VERY bad to discover a 'problem' outside of the warranty period! /rant

[philipma1957]

Originally Posted by atMac

It's also possible that it was the high temps for so long that made the CPU fail. If that is the case we may see a few failures in the future. A core failure doesn't impact the heat much, if anything a non-working core will show lower temps implying that others could be running even hotter.

yeah but all told less then 5 or 6 hours at temps that anyone running handbrake over night will get.

in fact we had posters say they ran handbrake for 48 hours at all 8 cores for 100% with no overheating .

plus the machine always had decent air circulation .

if you are correct and you run handbrake on this overnight in a Home theater case good luck.

that is assuming that my testing did it. btw this is the standard cpu test for many pc users do for stability. i also did this test on my mac pro 2010 with a 6 core i dropped in ran it for 3 days 0 errors.

I just finished this test on my 2010 2.4 c2d 20 hours no errors.

here is my guess and hope;

" I just got a crappy cpu".


it had some usb3 beach ball issues.

has the black screen issue.

both before any testing of the cpu.

[Snowcake]

Originally Posted by philipma1957

at 219f for a max still no throttle. yesterday and the day before I never got past 217 with out the machine locking up or throttle. this was with cpu test not prime95 test. still 219c is about 104c and i am sitting on a laptop fan! I will run this for a half hour more then i will test with oem ram. I would like to know if the ram the cpu or both programs are wrong.

So your cpu did throttle? How did you know it was throttling?
What for monitoring program were you using to monitor the speed/mhz/ghz for the cpu?

[xlii]

What app is your mini running that is reporting the temperatures inside the mini? I haven't run prime95 since 2004.

[philipma1957]

Originally Posted by Snowcake

So your cpu did throttle? How did you know it was throttling?
What for monitoring program were you using to monitor the speed/mhz/ghz for the cpu?

i went by the istat numbers on the top bar .


the 100% use on the cpu dropped.

the watts on the cpu dropped.

and temps dropped

Since bios is hard to get at in a mac I

inferred that :
1)temp drop
2) cpu use drop
3) watts dropped means throttle.

Originally Posted by xlii

What app is your mini running that is reporting the temperatures inside the mini? I haven't run prime95 since 2004.

istat menus 3



at least the 2 core i5 is good so far.

[Snowcake]

Originally Posted by philipma1957

i went by the istat numbers on the top bar .


the 100% use on the cpu dropped.

the watts on the cpu dropped.

and temps dropped

Since bios is hard to get at in a mac I

inferred that :
1)temp drop
2) cpu use drop
3) watts dropped means throttle.

Well that is doubtfull. Read this link http://forums.macrumors.com/showthre...1484787&page=4

There is a great realtime mhz monitoring tool MSR tools. you can download it here http://answers.yahoo.com/question/in...6003600AAN7vW7

[philipma1957]

Originally Posted by Snowcake

Well that is doubtfull. Read this link http://forums.macrumors.com/showthre...1484787&page=4

There is a great realtime mhz monitoring tool MSR tools. you can download it here http://answers.yahoo.com/question/in...6003600AAN7vW7

thanks I will give it a try.

[Snowcake]

Originally Posted by philipma1957

thanks I will give it a try.

I hope it works because i can´t test it since i don´t have a mac anymore.