Thinkpad T410 normal operating temperatures

[TPFanatic]

So this integrated gfx T410 has the 2.66 ghz i5 that overclocks itself to 2.93 ghz with a TDP of 35 watts, and it seems to be running hot. Right now it's barely doing more than Windows Update and hanging between 80-90C at only 25-50% usage @2.93 ghz. With Prime95 running it clocks around 2.8 and 2.53 and sometimes drops to 1.1 ghz around 95C to keep from overheating. The fan is running at 4500 RPM (equivalent to Fan 7) in Bios mode.

If I disable the Windows Update service temps quickly drop to sub 45C at low fan. Sounds like a good idle to me?

I feel I applied too little thermal paste and will probably try again. What temperatures should I be achieving under load with a T410?

Edit: It has the single heatpipe version of the heatsink.

[UMPC2024]

I think this post might help http://forum.thinkpads.com/viewtopic.php?f=45&t=121297. It has a similar description.

I haven't gotten to measuring my T410 that I bought used and never cleaned out the heat sink yet, but under load (docked with multiple programs, playing PayDay 2 on the lowest settings, things of that nature...) the left hinge would get hot enough to be uncomfortable to touch for more than a second. T410 generally runs hot though (like hotter than T420 and T430).

[TPFanatic]

Thanks, so I'll consider 75-85C as my target temp range.

Considering the laptop idles down to 41C doing nothing and only heats up (dramatically so) with CPU usage, what should this suggest? That the heatsink sucks? Thinking I should buy the dual heatpipe version, or not care. I just don't want the fan to burn out since I've read some stories about that....

[TPFanatic]

I redid the paste with a smaller dot and the cooling is still, eh, not what I'd like, but I'll chalk it up to these being hot CPUs, the heatsink being the single pipe version, and my not having the best thermal paste. I'd upgrade the latter two if I really cared so much but I don't.........

[thinkpadcollection]

If you have this integrated GPU T410, get 2 pipe heatsink, and tell seller to pack heatsink carefully. These heatsinks is too soft especially the metal fan housing and pipes. The fit is very tight in these T410 and I had a fight getting that right. I had two of these 2 and 3 pipes heatsinks for T410 arrive damaged. All are panasonic fan maker as well "[M]" logo.

Heat is very good running ubuntu 16.10. I use ceramique 2 or arctic silver 5 (my go to thermal grease tubes.)

Cheers, thinkpadcollection

[jaspen-meyer]

From what I've read, generic paste is 5-12C worse than the best thermal pastes.

[TPFanatic]

Great. As soon as I have money to spend I'm replacing the cheap crap out of all of my computers.

[ajkula66]

Great. As soon as I have money to spend I'm replacing the cheap crap out of all of my computers.

You don't need to spend a fortune on thermal paste, IMO. Arctic Silver 5 is superior to stock paste by many margins and you likely need less than $15 worth of it to re-paste all of your systems.

My $0.02 only...

[zephxiii]

I just reapplied thermal paste on my T410 with Nvidia and single heatpipe (for CPU) cooler and before i was at 91C running Prime95, after new Artic Silver it was running at 81C.

CPU idle temp went down from 51C to 38-41C.

The big change however was the drop in GPU temps that went from 78C to 59C during Prime95...which i know was not loading the GPU but GPU temp was really high and being a problem.

[TPFanatic]

I just reapplied thermal paste on my T410 with Nvidia and single heatpipe (for CPU) cooler and before i was at 91C running Prime95, after new Artic Silver it was running at 81C.

CPU idle temp went down from 51C to 38-41C.

The big change however was the drop in GPU temps that went from 78C to 59C during Prime95...which i know was not loading the GPU but GPU temp was really high and being a problem.

I actually did the same thing today but without as much of an improvement. My guess is I used too much. Congratulations and t hanks for the reference.


Just wondering does anyone know which die on the CPU chip is the actual CPU and which one is the igpu?

[RealBlackStuff]

AFAIK the small square is the CPU and the big one the GPU.

For thermal pastes there are new kids on the block!
The top pastes now come from Thermal Grizzly, top-flavor is Kryonaut, followed by Hydronaut.
Another advantage they have: no more silly 200-hour "burn-in" waiting time as with AS5.
In the list below with 48 tested pastes, AS5 comes only 28th.
http://overclocking.guide/thermal-paste ... gen-ln2/6/

[TPFanatic]

Thanks, I'll remember for when I buy paste next.

[kfzhu1229]

AFAIK the small square is the CPU and the big one the GPU.

For thermal pastes there are new kids on the block!
The top pastes now come from Thermal Grizzly, top-flavor is Kryonaut, followed by Hydronaut.
Another advantage they have: no more silly 200-hour "burn-in" waiting time as with AS5.
In the list below with 48 tested pastes, AS5 comes only 28th.
http://overclocking.guide/thermal-paste ... gen-ln2/6/

This is really NOT something you can buy at local stores in Canada unfortunately... Then on eBay how can you even tell if they are genuine or refills

[RealBlackStuff]

I bought my 7.8 grams Hydronaut on eBay: http://www.ebay.com/itm/331809847541
It comes in a resealable black bag, like this: http://www.ebay.com/itm/331958473007

[kfzhu1229]

Does it make a big difference in a T43 with ATI X300 and Pentium M 780 compared to AS5? It currently has brand new AS5 and it throttles on 1080p playback if without fan64 or undervolting and I have another T43 that has AS5 instead of thermal pad to cool X300.

[RealBlackStuff]

I only got my paste last Saturday.
First use was this morning in my wife's Lenovo G560.
I had last refreshed its P6200 with AS5 in July 2016, but there was still a fair bit of heat blowing out all the time.
Today I replaced the P6200 CPU (2.13GHz) with an i3-380M (2.53GHz) and used the new Hydronaut.
I'll have to check later today how it behaves with Hydronaut.
(TPF does not work on this laptop).

The Windows Experience Index went up quite a bit: (even though I don't give a rat's a$$ about that).
Old / New
4.7 / 6.5 Processor
4.7 / 5.5 Memory
3.2 / 4.2 Graphics
3.0 / 4.9 Gaming graphics
7.8 / 7.8 Primary hard disk (SSD in this case)

http://cpuboss.com/cpus/Intel-Pentium-P ... re-i3-380M

[RealBlackStuff]

UPDATE
Before the CPU and paste swap, there was hot air coming out of the vent permanently.
After the swap, even under full load, there is hardly a breeze, which is no more then lukewarm, if that.
The TDP is the same for both CPUs (35W).
In other words, that Thermal Grizzly Hydronaut works fantastic! (and no, I'm not paid or anything to say so )

[kfzhu1229]

Does that mean the heat transfer is almost 100% to the heatsink or the heat is trapped inside the CPU. A dried thermal compound will also result in cold air coming out

[RealBlackStuff]

A dried thermal compound will also result in cold air coming out

What did I just do on Nov. 28?

I've probably done 7-800 thermal paste refreshers...

[jaspen-meyer]

Hydronaut...
http://cpuboss.com/cpus/Intel-Pentium-P ... re-i3-380M

Those tests show Hydronaut performing 1 degree C better than AS5, measured on a "160 Watt continuous load (which) simulate(s) an overclocked Intel Core i7 Haswell CPU on high load."

On a thinkpad pushing 20 Watts what will the improvement be, a few tenths of a degree?

[axur-delmeria]

To be honest, I'm not sure if the results for desktop CPU tests are applicable to laptops, especially since:

1. Laptop CPUs don't have a heatspreader, resulting in a smaller contact area.
2. The clamping force of laptop heatsinks is much weaker compared to desktop heatsinks.

That's why I also look at thermal paste tests on GPUs. Like laptop CPUs, GPUs don't have heatspreaders, and their heatsinks don't apply too much clamping force (at least, I think so).

[kfzhu1229]

One of the reasons laptop heatsinks do not have that much clamping forces is because you are risking to break the glass die.

[axur-delmeria]

One of the reasons laptop heatsinks do not have that much clamping forces is because you are risking to break the glass die.

I know that.

What I'm trying to say is that the results of the thermal paste comparison tests on desktop CPUs cannot be directly applied to laptops because of the different conditions (bare silicon die and lower clamping force). I take the results of GPU thermal paste tests as something like a "second opinion".


.

[kfzhu1229]

Unfortunately that Hydronaut is twice the price on eBay before shipping vs AS5 which I typically get from Canada Computers for CAD $9

[TPFanatic]

I've finally bought a T410 dual heatpipe hsf to play with and my preliminary test results suggest the dualpipe HSF gets almost 10 degrees cooler temps than the monopipe under load, both cores running at 100% utilization w/ Prime95, @2.66Ghz (i5 540m), measured with HWMonitor.

Monopipe under load: 10 minutes in 84C @ max fan.
Dualpipe under load: 10 minutes in 75C @ max fan. (slowly rose to 78C at lower fan speed, before spinning up to max and cooling to 75C)

Neither test pushed the CPU to downclocking temps.

I did not clean off the "old" thermal paste from the die before attaching the new HSF, so the results may not be optimal. I am seeing funny symptoms after putting on the new HSF (Core 2 idling 5C higher than Core 1) so I'd best repaste it. Idle temps were pretty close with both HSFs, around 40C. So if anyone was wondering how much better the dualpipe fan is, it's at least 10 degrees cooler.

Another thing to consider when getting the dualpipe HSF, the dualpipe fan is different from the monopipe's fan. The dualpipe HSF's fan is labelled "Panasonic", the blades can be removed from the motor, and the housing is unique. The monopipe's fan is unlabelled and the blades are captive, so they're very difficult to lubricate. The fan housings cannot be swapped.

I don't know if the dualpipe fan is more reliable than the monopipe's fan when considering T410's notoriety for failing fans. All of my T410's have come with the monopipe HSF stock, including my 256mb Nvidia model, with 2 of them complaining about the fan and my last one being worrisome enough to warrant this replacement. I'm not sure why there are 2 different HSF types (4 in fact, since each has an Nvidia version with an extra noodle-sized pipe) and both appear in the HMM with some confusion: the HMM identifies the monopipe HSF as the Nvidia heatsink, and the dualpipe as the IGPU heatsink, which is incorrect, since both mono and dualpipe HSFs have Nvidia versions.

I'm going to sell this computer so I won't be watching to see if the Panasonic fan fails.

Thoughts on either fan type's reliability?

[jaspen-meyer]

As for the subtopic: Thermal Grizzly Hydronaut
I've used it on 10 machines and give it's a big thumbs up. It came in the same sealed black bag RBS mentioned.
After the first use I was skeptical -- the first application was a mess -- the paste was more note very thick -- it was runny like thick motor oil. I guess it was just the top-most layer of paste...all other applications were artical-silver-5 like consistency.

I run my machine's, including my T410, with the fan OFF if temp is < 70C. The Hydronaut has lowered temps.

[kfzhu1229]

I run my machine's, including my T410, with the fan OFF if temp is < 70C. The Hydronaut has lowered temps.

Are you sure in this way that the GPU won't desolder? For something like A30/p, the stock setting was to turn on the fan when the CPU reaches 80C which I have now set it into turning on the fan when the CPU reaches 45C and with the GPU connected to the heatsink via thermal sponge from a dead A31p motherboard. With this, the only thing that burns your lap are the ram sticks (2x512mb PC133) which runs VERY hot, sometimes hotter than the CPU!

[TPFanatic]

I run my machine's, including my T410, with the fan OFF if temp is < 70C. The Hydronaut has lowered temps.

Are you sure in this way that the GPU won't desolder? For something like A30/p, the stock setting was to turn on the fan when the CPU reaches 80C which I have now set it into turning on the fan when the CPU reaches 45C and with the GPU connected to the heatsink via thermal sponge from a dead A31p motherboard. With this, the only thing that burns your lap are the ram sticks (2x512mb PC133) which runs VERY hot, sometimes hotter than the CPU!

Yikes. I haven't used A31/p on my lap enough to know about that. But I do use this T500 here. The ram placement is the same on the Z60-thru-T500 era of 14" and 15" Thinkpads, strategically placed under the touchpad, and ranging between pleasantly warm to blazing hot. Now unless you reach under on purpose, you won't really feel it in normal usage, unless you're a touchpad user. And it gets HOT in my T500.

[kfzhu1229]

But I do use this T500 here. The ram placement is the same on the Z60-thru-T500 era of 14" and 15" Thinkpads, strategically placed under the touchpad, and ranging between pleasantly warm to blazing hot. Now unless you reach under on purpose, you won't really feel it in normal usage, unless you're a touchpad user. And it gets HOT in my T500.

I guess that is the reason why in T510/20/30 series, they reverted back to the way how T4x did: one stick beneath the keyboard and one stick in the base, I guess in this way, it neither heats up your lap nor become too unreachable to be upgraded.
Btw I've seen T430u as well, even though it is an ultrabook, unlike most other ultrabooks, it still gives you option to upgrade RAM (both slots beneath the keyboard), communication cards, battery and storage VERY easily!

Yikes. I haven't used A31/p on my lap enough to know about that.

Well A31/p series does heat up A LOT. Just when the fan comes on at around 55C, the air can burn your hand if you put your hand inside the holes.
However, my A30 and A30p don't even heat up even close to that! In fact, it runs MUCH cooler than my T43p with a lower CPU and GPU TDP even with Pentium III 1.2ghz! My A30p can settle down at 39C - if it is not plugged in and I don't know why since neither I allow speedstep to lower the clock speed on my A30p nor it performs differently with vs. without power. Either way, the CPU only peaks at around 70C while my T43p's CPU will still peak up to 80C after undervolting and its GPU can still reach in the 90s!
When running with battery and doing light office work or presentation, my A30p most likely doesn't even feel more than moderate warm (36C) except for the RAM sticks (The worst of all for that, is that the model for the pair of sticks is Infineon HYS64V64220GBDL-7.5 which even has a unibody metal heatsink on it and heats up like a Pentium 4) and the inverter.

[cynicx]

First post don't know if this is helpful to anyone or not.

Upgraded my T420 CPU from a i5-2520m to a i7-2640m and used some Thermal Grizzly Kryonaut I had laying around.

I wanted to see the speed difference by encoding a movie via handbrake but I also recorded max temps. Both are 35 TDP CPU's so temps are going to mostly reflect the thermal paste.

2520m with OEM paste.
Encode time 20 minutes
Max fan 4633RPM
CPU 0 - 97c
CPU 1 - 97c

2640m with Kryonaut
Encode time ~15 minutes
Max fan 3633RPM
CPU 0 - 83c
CPU 1 - 84c

Recorded via psensor in Linux.

Since I was curious about time more than anything else I wasn't paying attention to see if the 2520m was throttling, which could account for the 25% difference in encode times. Regardless there was a very significant difference in temps. I've had lower temps with Kryonaut over Arctic Silver in the past so I just stuck with it.

I don't see a way to upload screenshots directly to this forum however if anyone is interested (or wants proof) I can host them somewhere.