fanless build - concerned about component lifetime

Got a shopping cart of parts that you want opinions on? Get advice from members on your planned or existing system (or upgrade).

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
FrankL
Posts: 72
Joined: Sun Sep 28, 2003 10:28 am

fanless build - concerned about component lifetime

Post by FrankL » Mon Dec 02, 2013 1:40 pm

Hi,

I'm planning to build a low noise or fanless mini-itx desktop machine. So far I'm tending towards the following choice for case/motherboard/cpu/hdd:

Akasa Euler + Asus Q87T + Core i3-4130T + Samsung 840 EVO SSD

While I have no doubt that this fanless case can cool the CPU adequately, I'm slightly concerned about the influence of a much higher than normal case temperature on the motherboard components (such as capacitors, etc.) and the HDD.

I tend to try and build systems that do not require maintenance for extended periods, e.g. up to 10 years. I imagine the trapped air in the sealed case cause a much higher ambient temperature when the machine is in operation... would this reduce the lifetime of the components to less than, say, 8 years?
Asus does not provide any operating temperature in their specifications for the motherboard; I have troubles finding information on what these things are made to handle while still lasting a normal lifetime.

Should I instead opt for a case with ventilation holes and add a small fan in order to not worry about component failure?

edh
Posts: 1621
Joined: Mon Mar 17, 2008 1:49 pm
Location: UK

Re: fanless build - concerned about component lifetime

Post by edh » Mon Dec 02, 2013 2:13 pm

You'll be fine. Most of the components on motherboards will deal with much higher temps, components like capcitors are used in a range of electronic devices, and when was the last time you saw a TV or a video recorder have a fan? The SSD will be no worry at all, they are very tough things.

I have been thinking about this recently and think that computers are lasting TOO LONG. We have a 13 year old computer that hasn't been used for a few years and just seems to have stayed because there was space for it. It works but is no use and I doubt anyone else would want it. It would have been better to have been quieter in operation, failed 3 years ago and then it could have been binned!

HFat
Posts: 1753
Joined: Thu Jul 03, 2008 4:27 am
Location: Switzerland

Re: fanless build - concerned about component lifetime

Post by HFat » Mon Dec 02, 2013 3:00 pm

Different components handle high temperatures differently. Read the documentation. Talk to your suppliers. If there's no relevant documentation and your suppliers aren't talking, be wary!
This sort of gear is less failure-prone than mobile devices or hard drives of course but there are no easy answers or guarantees. I built fanless computers according to vendor guidelines and haven't seen a failure. Others ignored them and haven't seen a failure either. But SPCR ignored them, pushed the envelope... and Mike says he's had several failures.

If you look, you'll notice differences like this SSD's performance is throttled when its temperature rises while that SSD ignores its temperature. Amazingly, there are people here who figure such throttling is a reason to avoid the product! I know which product I'll prefer in a situation without decent airflow...

You can implement your own throttling in software by the way. You might as well buy a cheaper CPU while you're at it.

xan_user
*Lifetime Patron*
Posts: 2269
Joined: Sun May 21, 2006 9:09 am
Location: Northern California.

Re: fanless build - concerned about component lifetime

Post by xan_user » Mon Dec 02, 2013 6:47 pm

whats the highest ambient temps the PC will experience? aka are you in Finland with minimal house heating or Belize with out AC? also how long do you own a PC before retiring it?

FrankL
Posts: 72
Joined: Sun Sep 28, 2003 10:28 am

Re: fanless build - concerned about component lifetime

Post by FrankL » Mon Dec 02, 2013 9:33 pm

The machine will be in an environment where the maximum room temp will likely be in the 25-32C range.

I think the PC should be build to last 8 to 10 years in this particular case.

about the advice to get a more expensive CPU: I've seen people saying you can achieve a similar power envelope with software limiting of a regular Haswell. However, this won't be a PC for my private use (building it for someone else), so I wish to use something that is going to have low power envelope guaranteed, without needing to run special programs (thinking of apps like RMClock that I use on my laptop), and which will work reliably within specs with the low envelope for years on end.

xan_user
*Lifetime Patron*
Posts: 2269
Joined: Sun May 21, 2006 9:09 am
Location: Northern California.

Re: fanless build - concerned about component lifetime

Post by xan_user » Mon Dec 02, 2013 11:20 pm

operating range 0-70c -samsung 840evo

also,
Samsung’s Dynamic Thermal Guard Algorithm in 840 EVO’s firmware monitors the temperature of the SSD, preventing it from overheating. When the SSD is exposed to hot weather or extreme temperatures, the algorithm controls the SSD’s power supply to keep it cool. -samsung
so youre good there.
afaik the asus has similar protection capabilities, especially being a thin board.

id bet my money the build will cope just fine for many years, unless he's folding with it 24/7 (or bit mining ).

FrankL
Posts: 72
Joined: Sun Sep 28, 2003 10:28 am

Re: fanless build - concerned about component lifetime

Post by FrankL » Tue Dec 03, 2013 2:47 am

The Samsung 840 EVO being rated for operating till 70C seems good, but the following temperature observations make me wary:

http://www.silentpcreview.com/article1321-page4.html

under heavy stress, the case temp goes all the way to 62C with 22C ambient temp.

Admittedly that's with a CPU with a 55W TDP (I'm targeting CPU a 35W TDP), but it's also without a motherboard connector plate in place, allowing for a certain amount of cool ambient air to enter the case (and hot air to leave).

the other thing that worries me is that Asus does not give any temperature range for their products. The closest I could find was this:
http://support.asus.com/FAQ/Detail.aspx ... 1&m=P5Q-EM

which lists to a 65C max operating temperature range. 62C then suddenly is awfully close to the upper range limit...

the PC is not going to be used for heavy processing right now, but who knows how windows behaves in 5 years from now. I remember well the episodes of Windows Update in XP pegging the CPU at 100% for half an hour on end (funnily enough I read there recently was yet another outbreak of exactly this behaviour).

Vicotnik
*Lifetime Patron*
Posts: 1831
Joined: Thu Feb 13, 2003 6:53 am
Location: Sweden

Re: fanless build - concerned about component lifetime

Post by Vicotnik » Tue Dec 03, 2013 3:48 am

Even if a software bug would put the CPU at full load it wouldn't come close to Prime95+Furmark. I wouldn't worry. Don't use a passive system for any intense full load 24/7 stuff and you're fine.

HFat
Posts: 1753
Joined: Thu Jul 03, 2008 4:27 am
Location: Switzerland

Re: fanless build - concerned about component lifetime

Post by HFat » Tue Dec 03, 2013 3:49 am

FrankL wrote:Admittedly that's with a CPU with a 55W TDP (I'm targeting CPU a 35W TDP)
Don't take comfort from that!
TDP ratings aren't to be used like that. I don't think the actual heat outpit of those CPUs would be all that different. Hopefully the T part packs a more efficient GPU and operates at lower voltages but typically within the same generation the faster CPU heats up more. The 4130T is only barely faster so that's not a real concern but if you want lower temperatures, I must insist: do get a slower and cheaper CPU!

FrankL
Posts: 72
Joined: Sun Sep 28, 2003 10:28 am

Re: fanless build - concerned about component lifetime

Post by FrankL » Tue Dec 03, 2013 4:12 am

HFat wrote:
FrankL wrote:Admittedly that's with a CPU with a 55W TDP (I'm targeting CPU a 35W TDP)
Don't take comfort from that!
TDP ratings aren't to be used like that. I don't think the actual heat outpit of those CPUs would be all that different. Hopefully the T part packs a more efficient GPU and operates at lower voltages but typically within the same generation the faster CPU heats up more. The 4130T is only barely faster so that's not a real concern but if you want lower temperatures, I must insist: do get a slower and cheaper CPU!
ok so let me get this right. Are you saying that a slower CPU such as a Pentium G2010 uses less power in practice?

Or when you say slower do you literally mean lower clockspeed? Because the one I listed is already near the low-end of the spectrum that Intel offers. Incidentally, the even lower clocked Pentium models carry the same TDP or lower, so there seems to be a strong correlation between the indicated TDP and the clockspeed, which is basically what you're saying?

Personally I get the impression that the TDP figures for the different processor products in the Intel ARK database are representing the different processor 'bins'. If a CPU doesn't meet requirements of a certain TDP bin, it gets placed in the TDP bin one above. This does not mean that all CPUs with the same TDP use the same power, as it depends on the individual CPU's characteristics. However, it should indicate the maximum power draw you can expect from a CPU (in case you get unlucky and get a CPU at the upper TDP limit inside its bin). And maybe later into a production ramp, production of CPUs that end up in a low TDP bin might outstrip demand for them (as they are more expensive), which may let Intel sell them as a higher TDP bin part. This is all just a hypothesis and wildly speculative, mind you...

it'd be interesting to read a good analysis of the real meaning of TDP on Intel parts. If my hypothesis is right, the indicated TDP might not exclude that you can get a cheaper chip with a higher TDP rating using equal or less power than a low TDP rated part, but it definitely does not guarantee it! The Euler case review, where the G2120 with 55W TDP uses only 23W, but a i5-2500K with 95W rated consumes 61W (without stressing he CPU), show a power draw at different percentage of the rated TDP.

HFat
Posts: 1753
Joined: Thu Jul 03, 2008 4:27 am
Location: Switzerland

Re: fanless build - concerned about component lifetime

Post by HFat » Tue Dec 03, 2013 4:53 am

FrankL wrote:ok so let me get this right. Are you saying that a slower CPU such as a Pentium G2010 uses less power in practice?
At full load, yes.
Except we're comparing an Ivy to an Haswell which needlessly complicates things.
If you want an Ivy, there are slower ones (Celerons). If you want an Haswell, the most reasonable thing to do would be to buy the cheapest one (60$?) and underclock it in BIOS.. if you want lower load temperatures and unless you actually need an i3 feature that is!
FrankL wrote:Personally I get the impression that the TDP figures for the different processor products in the Intel ARK database are representing the different processor 'bins'.
TDP figures apply to a whole line and not to individual CPUs. That should be obvious if you look at a table with all the CPUs. To the extent they're based on any single model's performance, they must be based on the fastest model which gets a particular TDP rating.

We don't know how much binning really goes on behind the scenes but what we know is that Intel enjoys such a tech lead that they really don't need to bin. It might be cost-effective for them to bin. Or it might be cost-effective for them to randomly cripple batches of CPUs for sale at lower price points.
And we know that if they're binning a small number of CPUs, they're binning the most expensive ones and not the ones we're talking about.

CA_Steve
Moderator
Posts: 7650
Joined: Thu Oct 06, 2005 4:36 am
Location: St. Louis, MO

Re: fanless build - concerned about component lifetime

Post by CA_Steve » Tue Dec 03, 2013 9:09 am

It's pretty hard to run a CPU at Prime95 levels of utilization with normal applications. I think you'll be fine as you've specc'ed the system.

Post Reply