Haswell comes to NUC

Want to talk about one of the articles in SPCR? Here's the forum for you.
CA_Steve
Moderator
Posts: 7650
Joined: Thu Oct 06, 2005 4:36 am
Location: St. Louis, MO

Re: Haswell comes to NUC

Post by CA_Steve » Tue Dec 02, 2014 3:17 pm

hmsrolst wrote: The issue is not with the CPU overheating; it's with the wireless adapter mounted beneath the mSATA SSD overheating. If you dig around on the web, you'll find that there are complaints about the wireless adapter misbehaving as a result of overheating. The way that Intel has chosen to prevent that is better ventilation through higher fan speeds.
The speculation from a year ago was that the heat from the wireless card was causing the SSD to behave poorly. Either way, it's a bad layout and Intel chose to solve it with a bandaid. Here's hoping they figure out a better solution for Broadwell. (and yes, I'm also in the camp that avoids wireless when there's a wire available.)

pngai
Posts: 1
Joined: Sat Dec 13, 2014 1:30 am

Re: Haswell comes to NUC

Post by pngai » Sat Dec 13, 2014 1:46 am

I agree that the version which takes a 2.5" drive (D54250WYKH) looks clunky but the ability to install both an mSATA SSD and the 2.5" drive seems like a good system for using Intel's SSD caching system, SRT (Smart Response Technology).

However, my NUC's BIOS only offers IDE or AHCI modes, not RAID which SRT requires.

Has anyone been able to use SRT with this Haswell NUC?

hmsrolst
*Lifetime Patron*
Posts: 427
Joined: Sun Aug 11, 2002 3:26 pm
Location: Arlington, VA USA

Re: Haswell comes to NUC

Post by hmsrolst » Sun Dec 21, 2014 12:54 pm

Now having had some real world experience, I'm providing an update for any of you considering buying an Intel NUC. First, there are ongoing issues, some of which I've worked around, but others may not be able to. Second, I'm deeply suspicious that the NUC that Intel sent Mike was intentionally set differently to make it appear to be quieter than it was in actual practice.

On the first, when the minimum duty cycle is set at 25% the fan runs for the most part at about 2200 rpm (very quietly). But when, for example, streaming audio, it constantly ramps up to 3300 briefly and then back down. For an HTPC this isn't a problem because the music drowns it out, but as a desktop it would be unbearable. In addition, the wireless adapter does fine for awhile, but every few hours requires resetting. It could be the adapter, but I had no problem with it in my Tesla NUC (which I upgraded to AC). I also have a wired LAN back-up for my HTPC through my power lines which isn't very fast but works fine to stream music. So I basically don't need the wireless (but might if I watched more video). So I'd definitely not recommend the Intel NUC unless you have hard-wired capability.

On the second issue, when I looked back at the thread that I linked in my earlier post, I noticed the date of the initial post--December 8, 2013. This was followed by more complaints of the same kind within the month, and on the 4th of January, someone identified the source of the noise and posted about the minimum duty cycle being set at 40%. Mike's review was posted one day before the original post, December 7, 2013. I don't know how long Mike had his test NUC before his review came out, or how long the original poster with the problem had his, but I'm very suspicious that Intel was sending out these NUCs set at 40% and intentionally sent one to Mike set at 18%. Of course, it's possible that they only identified the problem later and changed the stock setting after they had sent Mike the test device. However, given the close proximity in time, shouldn't they have notified Mike that what they sent him was not set at what became the stock setting? So basically, Intel benefited from Mike's review about the quiet of the unit, knowing that it wasn't true at stock settings. Unless I'm overly interpreting, it's pretty sleazy.

I've also now had a lot of every day experience with the Haswell NUC in a Tesla case, as it's my primary work machine. It works fine. No problems with the wireless adapter. So perhaps it's a matter of getting what you pay for.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Haswell comes to NUC

Post by MikeC » Sun Dec 21, 2014 1:17 pm

hmsrolst --

All your speculations could be right... but OTOH, Intel is a big company & in such, the various divisions never seem to be perfectly sync'd.

Point in fact: I got notice from DHL a few days ago that Intel is sending me something from Colorado. Curious, I asked Dan, my regular PR contact, what this 1lb package might be? He conferred w/ someone else in PR & in the end neither of them could tell me what it was. Christmas cookies? Chocolates? a new CPU? Turns out DHL has a description: "Earbud and heart sensor" Wearables, perhaps... Anyway, it's not clear which PR team sent this & the other PR tems don't know.

So maybe my sample was "corrected" later & I was never informed because no one knew to tell me. Seems pretty normal to be. :lol:

btw, is there any way to change the hysteresis of the fan speed? ie -- delay before speed changes? I haven't looked in the BIOs of a NUC for quite a while. Maybe this can be hacked?

hmsrolst
*Lifetime Patron*
Posts: 427
Joined: Sun Aug 11, 2002 3:26 pm
Location: Arlington, VA USA

Re: Haswell comes to NUC

Post by hmsrolst » Tue Dec 30, 2014 11:45 am

MikeC wrote:hmsrolst --

All your speculations could be right... but OTOH, Intel is a big company & in such, the various divisions never seem to be perfectly sync'd.

Point in fact: I got notice from DHL a few days ago that Intel is sending me something from Colorado. Curious, I asked Dan, my regular PR contact, what this 1lb package might be? He conferred w/ someone else in PR & in the end neither of them could tell me what it was. Christmas cookies? Chocolates? a new CPU? Turns out DHL has a description: "Earbud and heart sensor" Wearables, perhaps... Anyway, it's not clear which PR team sent this & the other PR tems don't know.

So maybe my sample was "corrected" later & I was never informed because no one knew to tell me. Seems pretty normal to be. :lol:

btw, is there any way to change the hysteresis of the fan speed? ie -- delay before speed changes? I haven't looked in the BIOs of a NUC for quite a while. Maybe this can be hacked?
Mike, I looked for an Intel option to change the delay and didn't see one. Howard

Post Reply