Overheating mSATA SSD

Hey guys, has anyone else experienced a similar problem? I just stared using a new 850 EVO 1TB mSATA SSD and I noticed it's temperatures are wayyy too high. Even idling it stays north of 50C and it's even creeped into the 60s during normal use.

I've tried the various techniques such as those presented here:
http://www.abxzone.com/forums/showthread.php?t=120968
and here:
http://forum.crucial.com/t5/Crucial-SSDs/m500-msata-high-temperatures-70C/td-p/152140




Is there anything else I can do to lower my temperatures, or am I out of luck and just got a defective drive?

 

It's all about airflow, or lack of it inside. When my 850 EVO hits 50c and it will do so even siting idle, I ramp up CPU fan to max and that cools it off. If you want to keep it cool, get laptop cooling pad, just make sure fan blows air inside close to where msata drive is . I'm planning on permanent fix, I just need to find fan small enough to fit inside, even little air moving around will keep it cool. And no, I don't think your drive is defective.

 

Looks like the Max temp according to the spec is 70c , how long does it stay at 52c, it might only be for a short time while your fan speeds up.

http://www.samsung.com/global/busin...SSD/global/html/ssd850evo/specifications.html

John.

 

A very valid question indeed, for several reasons.

While SSD is not going to melt at 50-something degrees, I'd be willing to bet it will start throttling and that data throughput will suffer. That's the reason for having a temp sensor in the first place, regardless of BS that manufacturers might throw upon us.

 

Try the 30 day trial software below, to monitor your ssd temp, hopefully it will work with an ssd.

https://www.argotronic.com/en/screenshots.php

John.

 

I consider an SSD idle temperature in the 50's C to be quite reasonable and it won't go lower unless it is located where there is cooling airflow or has a thermal pad to conduct heat to the notebook casing. The SSD idle temperature largely depends on proximity to heat-generating components (particularly CPU / GPU) but also the fan operating rules. You can use the Sensors tab in HWiNFO to monitor the temperatures of the key components in your system and, possibly, the fan speed to improve your understanding of the factors affecting the SSD temperature. It could be informative to see the maximum temperatures when you run an SSD performance benchmark. Do they approach / exceed 70C?

John

 

Idle temp's at 50C? That is defective to me, whether it is within 'design specs' or not. More Samsung junk... now new and improved with 3D heating tech...

Anything that heats up a platform without an increase in performance is not a contender for my systems (which is one reason why discrete GPU's are worthless to me for most of my workstations mobile and desktops).

If the indicated temp is 50C+ at idle then the controller is obviously throttling (throttling starts at anything over 30C with an SSD controller). I would at least try another example before writing it off completely.

Btw, are you using Win7 or Vista? You might want to try a newer O/S too.

 

First we need to find out the ambient temperature inside the computer at the SSD location. That could easily be over 50C in a notebook where the fan rules are designed to keep noise to the minimum and the SSD is near the CPU. The SSD won't be cooler than the temperature of its surroundings. The 840 EVO mSATA SSD in my Dell E7440 idles at around 40C while the CPU idles at around 50C. However, the SSD's location is on the opposite side of the notebook from the CPU.

Nor do I believe that throttling will onset as low as 30C. If the design maximum SSD temperature is 70C then 60C or 65C are more plausible thresholds.

John

 

Hey everyone, thanks for the responses. To help you help me I'd like to give you all some more information.

First of all my laptop is a Lenovo Ideapad y400, pictured to the left. Below is a full disassembly video for my laptop. At around 1:09 you can see the position of the mSATA slot. It's in the bottom right of the notebook (looking at it from the view point), to the right of the RAM slots and below the fan/copper heatsink.



With that out of the way, I decided to run CrystalDiskMark and see how far I could raise the temperatures. The first time I ran it, the drive started at 62C and climbed to 70C. Immediately after that, I reset HWMonitor and ran CrystalDiskMark once more. This let me see what the maximum CPU and GPU temperatures were during the benchmark, and raised the mSATA's temperature even further, to 74C.

Below I've colour coded the results.

• The maximum temperature of my CPU was 63C in the package and one core. It's circled in purple.
• The maximum temperatures of my GPUs were 43C and 46C respectively. They are circled in green.
• The maximum temperature of my HDD was 37C. It is circled in yellow.
• The maximum temperature of my mSATA in question was 74C. It is circled in red.

Somehow the drive wasn't throttled even at that temperature, but it is no doubt bad for it's longevity. When I took the screenshot, roughly 20 minutes or so after the benchmarking, the drive's temperature dropped to 59C. That's still far too hot for comfort. As with the maximum temperatures, the minimums from 20 or so minutes later are recorded in the screenshot.

Also regarding how long it stays at 50C+, that's it's constant idling temperature. The drive can be as low as 29C immediately after booting, but it immediately begins its climb to 50C+ and only takes a few minutes to get there. Blasting the fans doesn't help much, as the video shows they are not positioned to move any air around the mSATA.

 

My take would be to forget about mSATA altogether and just go for a full-size SSD in the main bay.

 

I don't think there's anything wrong with your SSD, it's just that mSATA location on the Y400 is not ideal since it's in a corner with no airflow. I would consider modding the bottom cover so it can intake some fresh air. Take a dremel and extend the air hole over the RAM slots to twice its length, ending just before the rubble foot, then cover it with a not-so-fine wire mesh. Additionally, consider getting a cooling pad such as a Notepal U3/U3+ with movable fans and put one fan directly underneath the drive.

 

Hmmm that just may work. I'll use the laptop without the back cover for a while and see if that helps. Also have you ever used an mSATA in your y500 octiceps? I'd assume it would have the same issue if you did.

 

No I never put a real mSATA SSD in. Had the stock caching SSD (SanDisk U100 16GB) for a while but it didn't run hot at all, probably because it was a low-performance drive and wasn't heavily used.

 

Ah I see. I actually have a 16GB mSATA lying around that I used as a cache drive for a short period as well. Like yourself I had no overheating issues with it.

 

If the ambient temp is 50C that is a problem in itself.

But the throttling at 30C has been reported - sorry, I can't look for those posts now.

 

Hey guys, just a bit of an update. Removing the laptop's back cover does next to nothing. The idle temperature immediately begins it's climb to 50C+ and a single run of CrystalDiskMark is enough to push the temperature up to nearly 70C. I'm starting to suspect something about this drive is off.

 

50C+ is not something that can be fixed - it may be the motherboard, the proximity of other chips and/or power phases and of course it could be just bad design (in both chassis and mSATA SSD) too.

See:
http://forum.notebookreview.com/threads/overheating-msata-ssd.777486/#post-10028285


Suggest you try another sample and if it's the same - try another brand. If the trend continues - dump your expectations of mSATA SSD's (at least with your current system.

 

Well, that's not good news at all, but you don't need me to tell you that...

If the drive were mine, I'd be returning it for a refund.

Given what octiceps has written in this thread, my take would be - once again - to forget about mSATA and just go for a 2.5" SSD in the main bay.

My $0.02 only...

 

Taking the bottom cover out doesn't improve airflow by itself, as a matter of fact it may make it worse, for example CPU fan will draw portion of air from inside of laptop, cooling msata a little bit, but without cover more air will be drawn from bottom and less from msata area. If you want to see results, take the bottom off and then run some small external fan blowing underneath towards msata area, you should see drastic change in temp. I did. As far as I know most msata drives have similar power draw, in a range of few watts and replacing the drive with other make won't do much if you don't do anything about airflow.

 

Not much you can do about it here. The mSATA form factor really doesn't allow the designer to do much in terms of passive or active cooling. Although the SSD is running within spec, sometimes its just unavoidable with a given design to use a particularly hot running part.

For instance, my Sierra Wireless MC7700 idles, not even sync'ed to the LTE network, at 45degC. Its firmware throttles at 60degC. Nothing I can do about it other than accept the limitation during heavy transmission activity, or buy a newer radio based on newer technology (ie: a process shrink of the Qualcomm Gobi chipset it uses!).

 

Too hot. I think I've had my Intel 525 60GB mSATA max out at 43 degC and my crucial m4 128gb max out around 45 degC.

 

My 850 evo 1Tb also overheats to ~60°C under heavy load. (in Latitude E7450) Idle should be somewhere around 45 though.

 

For comparison, I've just run CrystalDiskMark on the 840 EVO mSATA in my E6440 and the temperature peaked at 64°C (starting from 37°C at idle in room temperature of 22°C). I'm not not concerned by that peak temperature under worse case conditions.

John

 

Some mSATA drives tend to run cool, others really hot - disregarding their placement. It is an important factor that should be mentioned in all reviews, but often is ignored. This ignorance itself is even greater problem than drive temps, IMHO - because it wastes people's time on investigating the issue.

 

It increased by 27C and peaked by 42C over ambient in a 14" notebook? Crazy (to think this is normal)!!!

Even a HDD from yesteryear would have me concerned for it's mechanical stability at over 35C idle temps. But those drives never even hit 50C for me in their harshest use.

With an SSD at 30C or above, the drive throttles itself and starts stuttering/pausing at double that temp and a some do that at lower temps too.

I guess this is why I don't run mSATA SSD's in any of my notebooks - but the NUC's, with the thermal pad for the mSATA drive and the small active fan makes them run almost like normal (2.5") SSD's, most of the time.

The newest mSATA drives with higher peak performance get the hottest and deliver the least sustained performance because of this throttling. Buying these for notebook use is the opposite to increasing performance vs. using a quality 2.5" SSD.

 

Apples and oranges. HHD-s have moving parts, which might be displaced if the temps are out of spec, killing-wearing off the drive. SSD-s are all circuitry, having another safe range for operational temperature. No one raised a concern yet, that his GPU or CPU exceeds 60C under heavy load...

BTW in the first generation of 7200rpm notebook drives were quite hot too...

Anyway, throttling, damage to nearby plastic parts may be a real issue,

 

The SSD location in the Dell E7440 / E7450 is away from the hot components but, as a result of this, also doesn't get any significant cooling airflow.

Samsung's datasheet for my 840 EVO mSATA says

so to be within that range when running an SSD performance benchmark gives me no cause for concern. Such benchmarks are not representative of my everyday workload. However, this issue flags up the potential problem of heat management for M.2 SSDs with faster than SATA 3 interfaces. Perhaps that's one reason why they are not yet in abundance.

John

 

Means throttling shouldn't be an issue at 60C.

M.2 btw is more problematic than mSATA: mSATA drives have 2 screws acting as a heatpipe, m.2 drives have only one.

 

Apples and oranges to you, but to me a storage subsystem is a storage subsystem. Yes, we have finally achieved storage performance that is finally better than HDD's. But for it to be better overall, the temps need to be similar or lower too.

But storage subsystems have other issues (like data retention, for example) vs. cpu/gpu overheating issues. A safe range over which throttling is allowed and even expected is not a properly designed storage subsystem (or notebook chassis, for that matter).

All first gen 7200RPM HDD's and even my desktop 2.5" VRaptors at 10K RPM never exceeded their maximum temps - ever; nor did they throttle themselves to below their rated spec's. Even if they (the 7200RPM drives) ran a few degrees warmer than previous slow/clunky drives of their time. Believing this is normal is allowing manufacturers to sell us inferior subsystems with performance we'll never see or use at prices that indicate the opposite.

When I first discovered this in my systems, I returned as many of the dozens of mSATA drives as I could. Not because I could not find a use for them elsewhere, but because it was the best way I knew how to give the manufacturers a clear and loud message that overheating piles of crap were not to be rewarded (at least not by me).

If these were the first gen products, I might be inclined to forgive them. Instead, this issue gets worse with each new gen of mSATA drives. M.2 is useless for exactly that reason (80C+...) and gives performance figures less than single digit percentages of what is stated on their spec's.

Unless you can compare a system that isn't throttling to one that is, they still seem fast (enough). But that is just not good enough for me to give real money for.

John Ratsey and tamas970,

If these were properly designed and installed in a chassis that took this excessive heat output into account (like current NUC's do...), at 60C it should not throttle at all, agreed. But both of these statements are not true for any current notebook I've worked with or any mSATA/M.2 SSD I've touched. They start throttling at well below 60C - ime, specifically at about half that; 30C, and well below the 70C point most current manufacturers state for the upper limit.

When a mere benchmark is run for a few seconds and hits temps of 42C+ above ambient (and ambient is 22C...), this is a major design flaw.

I do know why this is happening; more power/heat gives higher performance spec's and that is what they're trying to give us. But the implementation leaves a lot to be desired. And the communication to chassis manufacturers is non existent.

At least Intel listened (eventually) with their NUC's. A proper cooling pad is now standard on all new models for years now. Where is the notebook makers response to this?

I go out of my way to not have a discrete gpu inside my notebooks because they simply cause the processor to overheat and throttle sooner (and I get no use from the gpu itself...). Why would I want another source of heat from my storage subsystem?

One can be content and believe the provided spec's. But actual use proves otherwise.

Those high temps can be hit even on a NUC simply copying files from a server for 5 minutes or less - if it didn't have the thermal pad in place - and it still had the fan moving air. When it did get that hot, the system would literally crawl and even start stuttering and freezing/pausing. And long before that happened, the slowing down of the system was obvious (even if I didn't know what was going on initially).

High performance parts need a properly designed and accounted for cooling system. Nothing in the notebook space addresses that as far as I've seen for mSATA/M.2 drives.

If/when we get chassis' built properly for these types of drives, it will be another huge jump in real world performance.

Until then, even using the drives normally (for my light uses) in any notebook I've tried them in, I was paying for 2013/14 SSD's and getting 2005 heated chassis' and 1999 and worse HDD performance out of them.

I'm glad that your notebooks, mSATA and workflow combinations don't show the limits of these drives so easily. But for me. This is a tech that is dead for any current notebooks I use.

P.S. As for the two screws holding down mSATA drives acting as a heat sink... lol... in most systems those are simply inserts into the plastic chassis. No benefits at all for cooling the simply over the top heat output of almost any mSATA I know.

 

I've seen an mSATA SSD idling at 50C, and getting 76C under benchmarks. I've also seen mSATA SSD idling at ambient temp inside the notebook chassis (slightly lower than room temp) and getting mid-40C under very same benchmarks. It really depends on the model. The fact that some mSATA SSDs are bad indeed doesn't automagically make them all bad. Choose wisely.

 

Were these differences between different examples of the same series or different brands? Which did you find idled near ambient?

Yeah we shouldn't paint all mSATA's with the same brush, but life is not about beating your head against the same wall continuously.

Share your info!

 

tilleroftheearth, I can't disclose the hot drive's brand and model; the cool one has a review published, link in my signature. The drives do use different NAND obviously, but I wouldn't make any assumptions unless it's proven that the whole product line is subject to heat issues, not just single, possibly faulty, unit. And these two are the extreme cases; most drives are caught in between.

Reading other users' reports, it appears to be common problem for drives of different brands. We need more data to determine what causes the problem exactly, are there any software ways to overcome it and which parts should we avoid.

 

I don't know if the NAND or the controller is to blame. Controllers nowadays are quite beefy ARM-like chips, they might also take a fair share of the power consumption...

 

tamas970, it may theoretically be poorly written firmware alone - we don't know yet, need (a lot of) data on different drives.

 

That would be the best news, because firmware can always be corrected later.

 

Are you going to make me turn sigs on again? Come on...

Why is the other brand make/model so secretive? Just curious...

 

It appears, that the 2.5" 850 EVO-s can also heat up quite a bit. Interestingly the same 2TB 850 PRO runs cooler!

I want a 850 PRO in mSATA! 10 years warranty would be another benefit pls Samsung, shrink the PRO-line to mSATA and M.2...

 

Maybe a bit late but I have an mSATA Samsung on my MSI notebook and when using the RAPID option of the SSD at startup/boot, the temp at start rises from some 30C to +50C in a few seconds. Because I mostly use the build-in fan manually, the temp stays below 45C. If I don't run the fan it climbs to 50C-55C.

What worries me the most at this time is the rapid temp rise at startup. Maybe I should disable the RAPID mode.

 

Does disabling RAPID mode actually make the temps lower?

In either case, I would recommend you don't use your system with CRAPID in any event. The system is slower, overall, not faster.

 

Actually according to the Samsung Magician Tool, the RAPID mode should increase overall performance if certain requirements are met. I don't see, except at boottime maybe 3C, a temp. drop without RAPID mode enabled

 

RAPID has proven to not increase overall performance in any system I've tried it in. In client's systems in which I've disabled it, the ones that noticed (not all do...) asked how did I make the system snappier (note: by disabling CRAPID).

SM will, of course, say it increases performance when CRAPID is enabled (it's called marketing).

The temp issue you're having is a function of the mSATA SSD you have and the notebook's chassis cooling design (i.e. it has none for the mSATA drive).

 

I won't get side-tracked by the merits / demits of RAPID mode. Nor would I be unduly concerned by the temperature reaching 55C during the boot process which involves some intensive SSD activity. I would be more interested in the maximum temperature reached during some intensive SSD writing. Does the temperature get anywhere near to 70C during the CrystalDiskMark benchark? If so, then you may start to encounter SSD performance throttling in order to control the temperature.

The SSD temperature may also be influenced by proximity to heat-generating components. Is it near the CPU or dGPU or in a cool corner of the notebook.

John

 

The temp in fact stays below 55C. But allthough I have a game MSI notebook I rarely use my full CPU and never use the notebook for gaming. In fact my average CPU use (i7) rarely goes above 15 to 20%.