Hello dear community,

 

first a small introduction from mine: 

I am happily using a Lenovo Thinkpad W530 built in 2012 successfully for years now.

 

Originally it had an i7-3740QM CPU and Quadro K2000M GPU and I just upgraded to a i7-3940XM (not OC).

 

Unfortunately I did not measure the temperature of my old CPU before.

 

But the i7-3940XM I built in recently seems running on quite high temperature - especially running it in my Lenovo Mini Dock Series 3 Plus.

 

When I did the upgrade I cleaned the heatsink and vent and used the Kryonaut thermal paste

 

So after searching and reading a lot in this forum, I still have some questions regarding the i7-3940XM.

 

1. Is the information correct, that the W530 with i7-3740QM uses the same heatsink and vent as the W530 that was sold with i7-3940XM CPU by stock? 

Or could/should I upgrade it?

 

2. What is the normal and healthy operating temperature for the i7-3940XM in the W530?

The cores seem to differ a bit, but on idle some cores are on 42-46° C, two cores are always a bit higher on temperature with around 50-52° C on idle.

 

Is that normal? Also the difference between the cores?

 

When I run Intel XTU and using CPU stress test, the temperature goes up to 100° C within 1-2 minutes and the CPU throttles.

Normal and safe on this setup?

 

On "normal" usage with browsers, some software usage etc. it runs between 70-90° C.

 

3. The i7-3940XM is not OC via BIOS or similar - but Intel XTU CPU test measures 3.2 Ghz as base clock (although Intel lists it with just 3 Ghz base clock).

Is that normal?

 

4. The Intel XTU shows a maximum TDP of around 67w, but the CPU specification of Intel says 55w is TDP.

Is that normal as well?

 

5. If anyone has experience with either the upgrade I did or even has a W530 with stock CPU i7-3940XM what is your experience with the CPU and temperatures on idle and different CPU load levels?

 

Thanks in advance for any answer and sorry if some questions are annoying - I really researched a lot to answer this, but could not find much information and user experience, especially regarding the Thinkpad W530.

Is it possible that you put on too much paste? More is not netter. 100C might be OK if it happens quickly after a stress test starts only if the fan quickly speeds up and temperatures quickly fall. I think all the W530 fans are the same

@ 3940XM wrote:

Hello dear community,

 

first a small introduction from mine: 

I am happily using a Lenovo Thinkpad W530 built in 2012 successfully for years now.

 

Originally it had an i7-3740QM CPU and Quadro K2000M GPU and I just upgraded to a i7-3940XM (not OC).

 

Unfortunately I did not measure the temperature of my old CPU before.

 

 

2. What is the normal and healthy operating temperature for the i7-3940XM in the W530?

The cores seem to differ a bit, but on idle some cores are on 42-46° C, two cores are always a bit higher on temperature with around 50-52° C on idle.

 

Is that normal? Also the difference between the cores?

 

Yes, since intel will clock the CPU differently depending on active nr of threads. So for example CPU 1 and 3 can get higher. It could also be a result of a uneven thermal paste application or uneven heat sink.

 

When I run Intel XTU and using CPU stress test, the temperature goes up to 100° C within 1-2 minutes and the CPU throttles.

Normal and safe on this setup?

 

On "normal" usage with browsers, some software usage etc. it runs between 70-90° C.

 

3. The i7-3940XM is not OC via BIOS or similar - but Intel XTU CPU test measures 3.2 Ghz as base clock (although Intel lists it with just 3 Ghz base clock).

Is that normal?

 

4. The Intel XTU shows a maximum TDP of around 67w, but the CPU specification of Intel says 55w is TDP.

Is that normal as well?

 

If XTU allows your CPU to reach 67W that is really high and will of course generate heat. If you want to reduce this you can manually reduce it I believe. It is an embedded motherboard value.

I think the important part is how the machine feels during normal usage. I would not panic about thermal throttling during XTU. Take any new machine like P53 it will be the same.

 

For example using 3940XM on T530 with Intel igpu the max TDP will be 45W, using T530 with Nvidia GPU the max TDP is 55W.

 

As a first step I would probably try to repaste it and be careful that the heat sinks makes even contact. You can fish it out just popping the keyboard with some agility so only a 10 minute job.

 

First thank you very much for your replies! On XTU Stress Test it reaches 100C-102C within 1-2 minutes when the W530 is in the Dock Series 3 Mini (which is quite known for blocking the airflow on the side and the back of the W530 by design) and without the dock it takes 2-3 minutes to 100C and then keeps between 85-100C. So thanks for your hint regarding the thermal paste - I unfortunately indeed did not spread it, but did it like Intel recommends on some of their pages by applying kind of a "rice corn" big and shaped piece of paste on the CPU - I also thought it might be too much. But might it have this impact? I saw some tests with different methods of applying and it seemed even too less or too much of thermal paste just made a difference of a few degree. But I will try this and report later if it worked. So I assume from your answers at least it is not a normal temperature right now?! Btw. I looked it up in XTU - the Power Boost wattage is 55w and the so called "short extra power boost" is shown with 67.27w. I tried to undervolt the CPU or disable the boost in XTU - but it seems not to react to that changes. But as I heard OC is not possible with the original BIOS and also not with tools I think this is expected.

Did xtu CPU stress tests on three machines, 

 

P50, stock paste, i7-6820HQ, 40x multiplier overclock, thermal throttled after one/two minutes. On lap.

P51, repasted, E3-1535v6, undervolt -100mv, no thermal throttling, on Dock, 

T430, i7-3630QM, unknown paste, no undervolt, never gets more than 40W durung xtu CPU stress test, no thermal throttle

So your numbers are not unreasonable, but I would still repaste once to be sure.

 

Thanks for your super valuable input!

I am hesitating with repasting for now as I looked up a lot of tests and benchmarks and they all seem to show that even very bad pasting (uneven, too less, too much etc.) only seemed to make differences of a few degrees (maximum 5 degrees C in all the tests I have seen).
Reason seems to be that when too much paste is used, it will simply overflow from the surface where the heatsink contacts the CPU when the heatsink is tightened and so the thickness will automatically adjust and be okay in the end.
As I used no liquid metal and the CPU surface to be cooled has enough space on it's sides on this model it will also not spill into the socket or something. But I did not use *that* too much thermal paste anyway, I am sure.

When I reopen the case I would also - again - polish and clean the heatsink much more, as I did before on the CPU change.

I also read people using the W530 with i7-3840QM and TDP 45w reach these numbers. And I simply put in an i7 with TDP of 55w - producing more heat, of course.

Still open for hints and tricks to really improve CPU cooling in general and I also want to see what maximum temperatures will be reached on "normal usage" as I use the W530 as a developer mainly.

If you have advices for more "real life" tests instead of the XTU stress test please also let me know - thank you! :-)