Thinkpads Forum

Just got my W510 today, sweet machine...hot too.
I've been running some memory and CPU stress (prime95) and integrity tests for the past 2 hours.
Noticed that the CPU will slow down to around 1.2GHz when it gets to 95 degrees C.
I'm guessing that this is normal, however, question is why have such a fast CPU if it's going to get too hot and clock down?

At idle, Core Temp will show that the CPU is still at around 52-55 C.

Is there something wrong with my W510's heatsink&fan? I'm starting to think that I've wasted the money on a CPU that can't run cool enough to run at full speed for, say, 45 minutes to an hour.

Can someone please enlighten me?

It is normal for the CPU to clock down at that temperature, but my opinion is that it is not normal for the CPU to get to that hot. I would place a call to Lenovo service.

Here's an update, with a bit more details.

Software:
Prime95 v25.11 build 2: In-place large FFTs (8 worker threads, maximum heat)
CPU-Z (for reading clockspeed)
TPFC 0.62: Smart mode

With the three programs running, TPFC gets the cpu fan up to around 4360 - 4460 rpm (smart mode). The CPU stays around 90-91 degrees C. If, however, I then choose BIOS mode from TPFC, the cpu fan will drop down to a maximum of around 4000 - 4100 rpm. I guess that's why my cpu was previously able to get as hot as 97 degrees C (and it decided to drop the clockspeed).

With TPFC running in smart mode, I'm able to run Prime95 at the above mentioned settings for over 30 minutes now, and without the CPU clocking down.

So, I guess it comes down to the BIOS not spinning the CPU fan at anything faster than 4100 rpm. (Already have the latest BIOS as of last night 28th Sep 2010).

Yes, 90*C is also too hot.

I think so as well. Will try to get Lenovo to fix it.

However, can I take this opportunity to ask what other W510 users are getting if running the above test (Prime95 & TPFC)? Would like to find out the general acceptable temp is.

My W510 reaches the mid 80's when running Prime95, this is a 720QM CPU though and should run cooler than yours.
In the HMM i have found that two different coolers are available for the W510, an AVC cooler, witch I have, and a more expensive one that I placed an order for yesterday.
When I receive it I can post some details, I hope it's an improvement.

I have the same problem now. I believe this was not happening from the beginning, just noticed recently while exporting a lot of pictures out of Adobe Lightroom.
If I run Prime95, I get following situations quickly before and after overheat:





When clocked down, the machine is in agony - single window redraw takes 10 seconds, mouse movement is jerky etc.

I've contacted the service company, let's see.

Aleq, looking at your screen shots, the lows and highs are very similar to mine. My lows are around 52 to 54, highs are around 88 to 91 (all while having set CPU fan control to Smart in TPFC). I'm wondering if this temperature range is normal for higher specs cpus.

Adda, and looking at your numbers, I'm starting to think maybe the 720, 820 and 920 are meant to be in this temperature range. There's definitely a pattern here. Wish there's more samples to compare with.

Just got the cooler, they sent me the wrong one, identical to the original one
But the old one had some solder on the CPU cooler pad, bad quality control.
I retested the temperatures with the new cooler, the GPU gets a lot hotter than it used to, but there have been several BIOS updates since last time I tested.
When running Prime95 with 6 threads and FurMark in extreme burning mode my temperatures reach 90 for the CPU and 87 for the GPU, when I got my W510 temps where 86 for the CPU and 78 for the GPU when doing the same test.
I have found that performance has increased with BIOS updates, it doesn't show in 3DMark but WEI for the GPU is 6.5 it was at 6.4 when I first got my W510.
When playing games it seems more smooth than it did a while ago, it could be that they have increased GPU performance by changing things in the BIOS resulting in higher temps.
When doing real things and not pushing my W510 in the extreme as these test do, it doesn't get nearly as hot.

Adda,

According to the HMM

http://www-307.ibm.com/pc/support/site. ... 74250.html

there are three different coolers for the T510 / W510 (see thermal modules on pages 104 and 105). Did not your W510 ship with the workstation model on page 105?

My W510 shipped with the FRU: 60Y5493 AVC cooler for workstation models, but a more expensive FRU: 60Y5494 is listed in the HMM as well.

Edit: look at page 159 No. 20 in HMM.

What does the AVC stand for?

Besides the difference in price, do you know what the difference is between the two coolers?

Also, could you tell which cooler you had without opening the machine? When I do the parts lookup, I cannot find the cooler listed.

When I run Prime95 with 6 threads (instead of 8 ) with In-Place (max heat) mode, together with FurMark at 1280x1024 non-fullscreen at the same time, the CPU will go from 2GHz and clock down to 266MHz, then back to 2GHz...every 3 to 4 minutes in an attempt to cool off the CPU and GPU. Max temp at this point for the CPU is at 91c, with GPU at 91c as well.

jketzetera wrote:What does the AVC stand for?

Besides the difference in price, do you know what the difference is between the two coolers?

Also, could you tell which cooler you had without opening the machine? When I do the parts lookup, I cannot find the cooler listed.

AVC is the manufacturer of the cooler, by removing the keyboard you might be able to see AVC printed on the CPU cooler head if it's the AVC model.
The more expensive cooler is most likely made by another vendor.

I suspect that the non AVC cooler might be used in 920XM models since it's a 55w CPU, while the 720QM and 820QM are 45w, but this is guess work.
Or maybe the non AVC cooler got introduced as a fix for the heat problems so many W510's have, but I'm just guessing.
I was hoping to find out, but they sent me the AVC one, I think it's because the other one was not in stock, and to Lenovo a cooler is a cooler I guess.

But it all smells like the non AVC cooler was introduced later, since it's not in stock, and there must be a reason for this.

My heatsink & fan unit is also a 60Y5493...wondering if I should get a service engineer to replace it with a 60Y5494 instead.

Right, after extensive testing I can say that not all W510 coolers are created equal.
I found that after a long run of Prime95 x64 with 6 threads and FurMark in extreme burning mode, the CPU would throttle down to 933MHz when it reached 90-91 degrees, stay at that low speed for a while and clock it self back up to 1.6GHz untill the temperature reached 90-91 degrees again.
I took my W510 apart at saw that the pressure pattern in the cooler paste on the cooling pad was not even, not by a long shot, witch explained why some cores got hotter than others.
So since I had two coolers I could experiment with bending the heatpipes so that the cooler would fit against the GPU and CPU in an even manner, with little risk.
I got to install and remove the cooler many times before I got a good pressure pattern in the paste on both CPU and GPU, but now the CPU never exceeds 87 degrees and the GPU 84 degrees, and it doesn't throttle at all, heat problem solved.

Beware though that taking a W510 apart is no trivial task if you haven't done it before, the palmrest has to be unhooked in a certain way to avoid damage to some of it's claspers, and the cooler is a pain to remove.
Also when putting it back together it's important to do it in the right order or you will damage the palmrest, I know, I did the first time I took my W510 apart, had to order a new one.

I'm still curious about the 60Y5494 cooler though.

There have been reports in a few forums (notebookreview.com, thinkpads.com, lenovo.com) that W510's are suffering from overheating. If the overheating is due to the same reason as discovered by you, it is quite alarming, as that would indicate that Lenovo has serious desgin / QA issues with the W510 product line.

jketzetera wrote:There have been reports in a few forums (notebookreview.com, thinkpads.com, lenovo.com) that W510's are suffering from overheating. If the overheating is due to the same reason as discovered by you, it is quite alarming, as that would indicate that Lenovo has serious desgin / QA issues with the W510 product line.

It should be simple to figure out, look at Aleq's CoreTemp shots, how can you explain such a big difference in core temperature if it's not due to a poorly fitted cooler?
After all the adjusting I have done to mine, there is only a 2 degree difference in core temperature when all cores are under load Aleq's CPU as a 13 degree difference!
I'm sure that if he removes the cooler in his W510 he'll find the the cooler puts pressure along one of the edges of the CPU core and not on the middle, one of the core corners might even be untouched (yes it's that bad).

But to be sure more people with heating problems has to take a look for them selves.

I can also confirm that the pressure from the heatsink applied onto the surface of the CPU and GPU does in fact make a substantial difference to the temperature.

As a matter of fact, after trying to bend the heat pipes (ever so slightly), the CPU can now run all 8 threads in Prime95 max heat mode at between 84 to 86 degrees c. However, I've made the matter worse for the GPU...and it will now experiencing heat induced speed throttling...even to the point of shutting down the laptop all by itself when running 6 threads in prime95 and furmark together.

I'm going to get the Lenovo engineer to deal with it...as it's seems like a QA issue.

Good job Chatbox, now it's confirmed!
If you are up for it you can fix the GPU cooling pad yourself, the mounting bracket that holds the GPU cooler down, can be unscrewed and turned over, use a plier to bend the legs that hold down the cooler to increase pressure, if this is not enough, you can bend the heatpipe as well.

I've been trying to do it for 3 hours now... Bend...test...bend....test....

It's coming to a point where it's too time consuming and tedious. I purchased the on-site warranty a week ago, so going to give them a call and hope some super talented engineer can deal with it. I honestly can't be arsed anymore.

I'm sorry to hear that, I think I used about two hours adjusting, but I will admit it requires a fair bit of luck to get it just right.
I hope the technician will sort it out for you and not just replace your cooler with another faulty one, fingers crossed.

For those fiddling with your heatsinks and fans... are you reapplying the thermal compounds and thermal pads? plus applying them properly.

Also normal operating temps are a little subjective but for these types of systems it's understandable to run 5-10 degrees hotter than a normal laptop w/o a high end GPU and high CPUs. When I got mine with a 720QM it ran mid to high 40s idle, I changed my CPU to a 940XM and it runs at low 50s to mid 50s at idle. Normal operating temps are around 60s and up to 70/72. Gaming normal is anywhere from 70-85 depending on the game. But fair to say my avg gaming temp is around 78. All temps in degrees celcius.

Comparing this to my old Gateway dual core with a dedicated ATI card these temps are a dream. Comparing to a Dell XPS with ATI 4670, slightly better graphics card than ours temps are about the same, especially when it comes to gaming. The overall system on the dural core though was slightly less but oh well.

As far as overheating issues, it all depends on what you consider overheating... I've heard people complaining about their notebooks hitting 80C when playing agame, while temperature wise it is hot, it's normal for notebooks.

I use Arctic Silver thermal paste, and for the final cooler install I applied fresh paste.
Also the thermal pads on the GPU cooler head got squeezed a bit out of shape, but by peeling it off and hanging it by the fingers for a short time makes is contract back in to shape and ready to be re-applied.

Overheating = so much heat that the CPU and/or the GPU throttles to avoid damage and/or heat induced crashes.

As for the person having problems with the GPU, making it worse, you need to make sure you put the thermal pads on properly onto the 4 memory modules of the graphics card. Whatever you use to monitor the GPU temp will not register temps just on the memory, and that will cause your system to shut down. So for me one time the pads weren't touching properly and when trying to run final fantasy XIV benchmark it would shutdown the system. I thought it was something else and underclocked the GPU and it would at least run at lowest memory clock cycle. So I knew that was the problem and reapplied the thermal pad and now I have the GPU overclocked.

Running at 690Mhz core and 945Mhz memory clock stable for FF XIV. Technically 700 was stable but at very very extreme situations it would crash and clock down.

Overheating, to me, means the systems will self-protect itself by speed throttling and / or system shutdown.

Yes, reapplying thermal paste.

Under normal, interactive use, the laptop works fine. But it's not able to handle the load from artificial tests/benchmark programs...that's an issue I would like to address.

playersnoopy wrote:For those fiddling with your heatsinks and fans... are you reapplying the thermal compounds and thermal pads? plus applying them properly.

Also normal operating temps are a little subjective but for these types of systems it's understandable to run 5-10 degrees hotter than a normal laptop w/o a high end GPU and high CPUs. When I got mine with a 720QM it ran mid to high 40s idle, I changed my CPU to a 940XM and it runs at low 50s to mid 50s at idle. Normal operating temps are around 60s and up to 70/72. Gaming normal is anywhere from 70-85 depending on the game. But fair to say my avg gaming temp is around 78. All temps in degrees celcius.

Comparing this to my old Gateway dual core with a dedicated ATI card these temps are a dream. Comparing to a Dell XPS with ATI 4670, slightly better graphics card than ours temps are about the same, especially when it comes to gaming. The overall system on the dural core though was slightly less but oh well.

As far as overheating issues, it all depends on what you consider overheating... I've heard people complaining about their notebooks hitting 80C when playing agame, while temperature wise it is hot, it's normal for notebooks.

Chatbox,

I am very curious if you get Lenovo to admit that there was a problem with the cooler and get a service technician to fix it for you.

New Update.

I just found out that I've been conned The i7 920XM in my W510 (bought from eBay...I should have know...it was just a really good deal though), is actually an Engineering Sample (i7 920XM ES, Q3B8), and is not a Lenovo part.

Now, the tech is extremely nice and understand my situation and said that he can replace my current 920XM with a 720QM, slower, but will be an official Lenovo part and will have warranty. I'm inclined to go with this replacement, having support and stability is pretty important to me. However, I would like to hear what some of you might suggest. Go with a free replacement, supported with warranty, but slower CPU...or stick with the current 920xm and try to burn a hole through the laptop chassis?

As to my current i7 920XM and the stock heatsink & fan: I've managed to get it to run for 30minutes without throttling....but man, it's HOT (Prime95 8 threads, max heat). See image below.



Also got the GPU down to 84-86 c. Technique: Like one of the previous posters said, it's all about pressure!!! Now, what I did differently though is, rather than trying to bend, twist...etc to get even temp on all 4 cores, basically just have the entire CPU heatsink contact lowered by about 2 to 3 mm by bending the pipe AND not to worry about getting it leveled. The total increase in pressure is proportionally much great than that supplied by the 4 springs at each of the corners. So what I'm saying is, when it comes to the pressure, don't rely on the four springs to apply the pressure, but instead, use the copper pipes, bend them so that the entire heatsink is pressing against the CPU surface even without the screws. This will ensure the CPU surface will get REALLY decent pressure overall. Then, of course, you'll still need to use the screws to secure the heatsink itself.

I might also add, my typical usage of the laptop isn't for gaming, CAD or anything graphics intensive. Nor do I use it for any mathematical research or number crunching. However, I do use the laptop for running multiple VMs and application servers, test environments, software integration POCs...

So, from a usage perspective, it's not going to "heat" intensive like Prime 95, but I really do like the 8 threads capability as well as the high speed of turbo boost when needed.

Another update.

I bit the bullet and went with the replacement of having a slower CPU. Really going to miss the speed.

This time, same test, the 4 cores are around 79-81 degrees C, which got me thinking...given that nothing else got changed, maybe 93-94 C was "normal". Having said that, I just couldn't risk having other nearby components getting that hot, 90+ c is just not healthy for a laptop.

As for the heatsink & fan, didn't replace it, so I can't tell what the differences are, if any. The tech guy did say that, from the internal notes, both thermal units are supposed to function identically and there's no thermal/cpu requirement that say they should use one over the other.