I don't know if Mike tested it, but I did. My VMWare server running on the DQ77KB and Akasa Euler is completely headless. No monitor. No keyboard. No Mouse.

Beautiful thing is I can still access it as if it did have all those things using vPRO AMT remote management! It sits tucked into a corner under my desk, yet I can access it using VNC Viewer Plus from my desktop - including power controls (on/off/reboot) and full access to the BIOS menus if I need it.

No, this was not done, and I can't verify it now because the board died mysteriously after it was removed from the Euler for a detailed review. Tried everything, and the thing refuses to even post... after 2 weeks of perfect operation. Crib death?

Anyway, I wonder why you'd use this as a file server? It only has room for one 2.5 drive and 1 mSATA. I equate server with high storage capacity, and the Euler doesn't have that. OTOH, there's really no need for fanless in a high storage server because the multiple HDDs will always make enough noise so that 1 or 2 very quiet slow spinning fans will never be noticed.

Just a minor point, but CPU location is part of the thin mini-ITX standard so that standardized cooling solutions can be used. I believe all the cooling solutions should be designed to handle 65W processors to officially meet the spec. Intel is really pushing this for AIO systems (e.g. iMac style), which is why most of these boards have an internal LVDS connection.

Some details in the catalog here: https://www-ssl.intel.com/content/dam/w ... atalog.pdf
Note that the CPU socket may be rotated 90 degrees, but the heatsink mounting points are the same since they're arranged in a square.

Additonal info: http://www.intel.com/content/www/us/en/ ... i-itx.html

Oh yeah, here is the standard thermal solution specified by Intel. Note that the case doesn't have to use this cooler, but the motherboard mounting holes must be in the right place to accomodate it and also not have any components in the way of the heat pipes.

http://cache-www.intel.com/cd/00/00/49/ ... 492064.pdf

Hardly at all. Similar to the Intel NUC, the power draw during 1080p playback w/just about any source/method is around 20W. I didn't bother providing details on this because even in Prime95, which is way more stressful, temperatures remained modest.

BTW, in case no one has noticed, the Euler is pretty easy to adapt to any 1155 Thin m-ITX board, even if it does not have exactly the same CPU location as the Intel boards. There's a question of whether the Thin m-ITX spec actually specifies CPU location; it appears to, but the reference in Intel's Thin Mini-ITX-Based All-in-One PC System Design Guide v1.1 is section 5.1 to 5.4... but that document ends at section 3.10 -- there is no section 5. So, if you choose to use another Thin-ITX board for whatever reason -- and there are quite a few, made mostly for industrial markets, for All-in-one systems etc, not retail -- and the board's CPU cooler mounting holes don't line up properly, you can simply MOVE and reposition the block. It would be slightly tricky but perfectly doable for someone with a bit of machine-shop experience.

That heat block is removable, as I showed in the pics. So remove it, mount it to the 1155 CPU cooler mounting holes of your motherboard, then goop up the other side with TIM or similar (to use as a marker), and install the board in the Euler. Then remove the board. You will now have a clear imprint of where the heatblock should go, and can use a drill press to tap holes for the 4 screws that mount the heatblock to the chassis.

How useful all that is depends on whether there are thin-itx boards which don't have the same CPU position as the Intels.

Spurred by the posts by drawz and Delta_42, I delved deeper into Intel's vision for Thin-ITX and ended up making extensive changes at the introduction of the article. There are now links to Intel pages and docs on Thin Mini-ITX, a discussion of the layout, etc.

MikeC - thanks for the additional info (especially the PDF of the spec) and also thanks much for the review. Clearly with the power we have in today's computing platforms and the easy availability of local networked storage and cloud storage, we are going to see more and more SFF and AIO systems. Thin mini-ITX is going to be key going forward.

Food for thought: AMD CPUs do not share the same cooler mounting mechanism, so they will inherently be incompatible with the spec put out by Intel. Unless manufacturers make an adaptation themselves or AMD changes the mount on future CPUs. It will be interesting to see how this plays out. Thin mini-ITX thermal solutions also do not account for discrete GPUs, either via PCIe or integrated on the motherboard. PCIe graphics would also be hampered by the use of x4 slots, which are needed to stay out of the way of the heat pipes (I think - all current mobos are x4 only).

As for the Atom boards, I think they don't need to use the standard thermal solution because they tend to use a small passive heatsink that is probably ok without heatsink & blower used for LGA1155 CPUs.

Thanks for confirming that

Did you need to alter any of the Bios setting such as "halt on Keyboard" that used to be in older bioses?

I can't do vPRO AMT remote management with a 2100t so i'll have to use TeamViewer until i get a Vpro CPU as i planned on doing this aswell



Sorry to hear that the Board is Borked , there are a few on the Intel forums that have had to return theres aswell so lets hope it's not due to a design fault

As for the fileserver that was a bad choice of words , what i'm building with the euler is a maintenance free low powered fanless download box i can leave on at night times downloading to a Western Digital 2.5 inch Scorpio Black 750GB 7200rpm so it's not intended for long term file storage , And apart from an occasional wipe down of the case i should never have to do anything but turn it on and off.

i think this case looks amazing, but 2 things i dont quite like. One is the laptop style power brick. I view that as somewhat cheating, though to be fair, you can hide that. 2nd is the limitations of this case as a whole. I think this has at most the power of a typical netbook.

Still, I can imagine these being used in university computer rooms, or offices with high frequencies of computers. Is this even achievable whilst making financial sense?

The power brick is part and parcel of Intel's Thin ITX spec -- every one of those boards runs on 19VDC.

Typical netbook? You must not have read the review -- the current Thin mini-ITX motherboard spec virtually mandates a socket 1155 CPU. Ivy/Sandy Bridge with Intel HD Graphics is the norm. They are as powerful as most desktop boards with the same components; the only serious difference is that Thin ITX generally does not support a discrete PCIe graphics card, or CPUs with higher than 65W TDP... but this is losing sight of the intended use. The case actually has no problem running a 55W Sandy Bridge part in normal use.

My Akasa Euler and DQ77KB arrived yesterday.............

The Euler is a lot bigger and heavier than i expected , But in a good way that it feels like you are getting a decent lump of metal for your money and it is very well made.

Shame they don't do a vertical stand.

On the other hand I'm not very impressed with the condition of my DQ77KB as the back of the board is still covered in sticky Flux that hasn't been removed and both the back of the board and all the silver parts are covered in sticky finger prints.All the original seals was intact so it's left the factory like it.

Registered here just so I could comment on this, I had a friend visiting england awhile back and got him to pick this up for me.

In the article you mentioned not being able to do 2560x1440, but the DQ77KB has displayport which is more than capable.

I used a i3 3220T and while I was installing everything it was hooked up to a dell u3011 and it did 2560x1600 just fine. Perhaps its your choice of a lower end cpu?

Also if you don't like the included adapter or they sell it in the states without (I noticed there are no UL marks on that one) any large business likely has an entire fleet of dell or hp laptops that use the exact same plug, see if they have spares or throwaways. A ton of ebay/amazon sellers have them super cheap too.

Where did you get it from? Mike -what are you doing with the SPCR copy? ... I am interested.

Ian

Woops, I made a mistake in my post. I meant ultrabook, not netbook :S. My issue is that the graphics would have to be inbuilt with the cpu itself. I guess my specific point is that, as you said, the case (and such Thin ITX motherboards) cannot support discrete gpus.
I think we can agree though, that one of it's intended purpose is for it to be a media pc. I was simply wondering whether the current cpus can play things such as blu-ray type media smoothly without lag issues when making this post, given the 55w tdp margin. After further finding out that for video playback, these cpus are good enough (looking at the performance of Intel HD 2500), I think it is safe to say I am wrong on this issue.

I will also yeild that I am somewhat ignorant about the power brick (a shame that they all use them). I still find them to be out of place and "cheating" though.

Finally, it turns out I was not complaining about the case at all, but rather the technological limitations it houses. Sigh, this is what you get for posting in the early hours of the morning.

FrozenCPU.com is an Akasa dealer and will special order it for USA based customers. I got mine through them. Send e-mail to support@frozencpu.com for a quote and a one-time order code for the website.

I'm in the UK and got it from dabs.com

@PigLover

With Intel Vpro KVM don't you find it annoying having to keep plugging in the monitor for getting the session key for remote login?

Not at all. I just disabled it in the Management Engine BIOS. KVM sessions are just accepted without acknowledgement.

Except for a couple of minor annoyances I find vPRO/AMT far superior to IPMI for remote management. It's faster, more responsive, better integrated with the video and trivially simple to use. The annoyances are minor:

- you can't specify a VLAN for the management engine IP interface
- Wth ESXi, if you do a 'pass through' of the USB3 controller you lose the ability to type on the remote keyboard. The AMT keyboard appears as a USB keyboard to the guest OS you passed the usb3 controller to (this is an uncommon config...I found it by accident)
- there is a weird IP address sharing scheme used if you get your vPRO IP address via dhcp and the Host OS also gets an address via DHCP on the same Ethernet nic. Some people have reported strange packet drops at high utilization that may be related to this.
- there is no access to CPU/mb environmentals - temps, fan speeds, voltages - like you get with IPMI.

These are minor nits...overall, I really like it,

This has been a sticky point for me ever ever since i've considered Vpro KVM that i can't get a answer for on the Intel Forums.

Doesn't disabling the Bios Key option mean anybody can start a remote session with the PC both internal to the Network and anybody scanning from the internet for a open session?

You still have to authenticate the session. At least a password is required. You can set up TLS certificate based authentication if you like.

Also, in my note above, I mentioned the lack of VPN support. The main reason you want VPN support is to move the management access onto a separate LAN with better security - including no access to/from the Internet. My servers IPMI connections - and now this one's vPRO port - are set up on such a LAN. So are the ESXi management interfaces. Without VLAN support its a bit more complicated, but not impossible.

I doubt it has to do with the CPU. The documentation for the board does say up to 2560x1600 is supported for Displayfort, so maybe I just didn't try hard enough. The handshaking between video card and monitor can sometimes take forever, which might have cause the auto input sensor on the Smasung monitor I was using to time out and switch to the next input. I'll go and change that -- if one person got WQHD, then it's obviously a function that is there.

Hi everyone,
I think I have found rather nasty flaw in this case. When trying to figure out why my system behaves so strangely I have noticed massive deformation of the moteherboard inside the case:



About 2mm flex at roughly 2/3 of the board width. When looking at it from the side the entire board is basically bent around the CPU socket and then back up by the rear panel.

I did not use any extreme force to tighten the CPU nuts (I used a pen kinfe as screw driver which makes it even harder to exert big torque) but there are no stops on the screws and the nuts themselves are open so you have no way of guessing if you have tightened enough or not.

I think I have killed my brand new DQ77KB.

Hmmmm.... So what you are saying is that the relative vertical position of the CPU heatblock vs the 4 mounting posts at the corners of the board are not correct, right? Ie, if the screw posts were a 1 or 2mm shorter, or the heatblock that much taller, this flex in the board would not happen? I wonder if this flexing caused our board to die? Time to put it back in and take a real good look.

FYI, a certain amount of flexing in the motherboard around the CPU has always been defined by Intel (and AMD) as part of the mechanical design for cooling heatsink mounting. Maybe this case forces the board beyond the safe flex range? I will check our board * base out and report back.

Nope, no such flexing, all the 8 points that anchor the board to the chassis are very close to identical height. I'd be really surprised if there was more than 1mm variation among them

LOL , This is so confusing like learning a foreign language for the first time.

Just to be sure we mean the same thing i'm talking abut the User Consent code that comes up in a black screen with a padlock that has to be given to remote computer user to log in.



If you disable this in the Bios can you then manually set a key as i thought this was the only thing stopping remote access?

Hi Mike,
I did not check if the posts are level. I think there are two problems:
a) I overtightened the CPU nuts. I tightened them enough to use the entire length of the thread in the nut but no more. Apparently that was way too much as the CPU socked formed very visible bulge on the MB. There is no mechanical stop on the CPU heatblock bolt to prevent you overtightening the screws (EDIT: This is false, apparently I am blind. See the posts below.) - the only indication that it is too much is the flex of the MB.
b) the IO shield cutout is way too high (or low) - if I align the IO ports the board is about 3mm off the standoffs at the back. This exacerbates the problem a) even more. Have you tried with the IO shield in place?

EDIT:
A picture worth thousands words:



I have just noticed when I remove the IO shield and place the MB in the case it seems to fit reasonably well. However puting on the CPU nuts and tightening them by hand revelas the bolts are either not of the same height or the MB does not sit evenly on the heatblock. This makes it even harder to judge how much you should tighten them. (EDIT: This seems to be result of my already bent MB and not the fault of the bolts.)

I think the slack caused by the IO shield lifting the board caused me to tighten the CPU too much in attempt to secure proper contact. I should have seen this the first time but I was too eager to get it running.

The SPCR review system was built without using the I/O shield. See the last photo and caption on page 3 of the review.

This is not true of the sample here. See this photo -- from page 2 of the review, 3rd from the bottom:

You can see that the screw threads run only 2~3mm at the tip.

I wonder if my sample is a later version, with minor adjustments/improvements from an earlier batch?

After considering the revelations here by Regis, it seems prudent for all current users of the Akasa Euler to power down and disconnect their system from AC, remove the bottom panel, and examine the flatness of the trace side of the board - by eye, with the aid of a straight edge of some kind. To try fitting the board again with the backplate, I, in the meanwhile, have to find the plastic bag into which all the DQ77KB parts were tossed and stored... somewhere in the chaos of this lab!

Actually you are entirely right Mike! I did not notice the the short threads - I admit I just assumed lookig on how bulged the CPU was.

I am starting to think the I/O shield is the sole culprit. Check this out:



I fitted the shield from the outside so it is reversed left-to-right but it is the correct side up. You can see on the USB ports how much it is off.

I did notice on my Euler that the rear-left MB screwpost was a bit too low (rear-left with the MB upside-down and the IO-shield area being called front - or in other words, the post closest to the corner of the CPU). It was low by 1-2mm and I compensated by dropping in two of the paper-insulating washers you used to get for MB installs.

Didn't think much of it at the time - just seemed like an expected manufacturing variance for such a small case. Maybe my expectations are too low or tolerance for quality issues is too loose?

As for the IO shield - mine fit and lined up perfectly the first time.