BGA Venting Thread ;)

@FredSRichardson The joy of having socket hardware and not BGA JUNK in Jokebooks!! Aka LGA http://valid.x86.fr/fs084d

 

@Papusan - that is certainainly sweet! =)

 

get a 5 ghz sub 1.4v would be amazing. i have seen some really good 4960x thats 5ghz sub 1.4v with 6 core so quad should get lower voltage.

 

haha thats pretty nice, but if its that big i'd just go for s9t portable http://www.ssiportable.com/products/portable-solutions/spark-s9t-2/. im still holding out buying it because waiting for skylake-X laptop and hope it might come true from what eurocom mentioned, if not im jumping on the S9T with an external UPS and AIO water for my CPU.

 

The three screens are nice!

I'd like to see a standard micro-ATX (or stx/itx) with MXM sockets so it could support a very flat form factor.

Sent from my SM-G900V using Tapatalk

 

issue with desktop is that its got no battery, which is one of my main reason i havent gone with S9T yet. because they use E-ATX or any desktop mobo in general it wont have a battery.

 

Tell me more, tell me more, do the fans purr like a car? Sorry, exhaustion and Grease popped into my head. Can you please give more, cite post, mention source?

Sent from my SM-G900P using Tapatalk

 

LOL!

Tell me more, tell me more...
Was it love at first sight?
Tell me more, tell me more...
Did she put up a fight?

 

Here's a "luggable" =D

http://www.modsrigs.com/detail.aspx?BuildID=36013

 

lol and its the 6700k too haha.

 

Yeah, I'm still looking to see if there is an mini-{a,i,s}tx form factor that has a MXM socket and can take a battery.

This would be pretty close to a generic laptop mobo.

 

desktop mobo is already standard they wont be able to carry a battery because too much work need to put into place. watt out, voltage out, voltage split, current etc similar to a psu. if u could find one pls let me know im dying for it too.

as for a mini atx that has MXM, prob none but could use PCIE extender + PCIE to MXM adapter card that would work so the PCIE card won't be perpendicular to mobo.

 

You do get some desktop boards that take a 19v input but they are the lower power soldered cpu boards. So even hooking a battery to that input would not get you very far.

The issue with a desktop is the lack of power management for not over drawing the battery and setting the thing on fire

 

I learned an interesting "feature" today about the Alienware 17 R4 when setting up a machine for a friend. The Thunderbolt 3 port is apparently wired to the Intel GPU, even if you have the GSync display and GTX 1080. So if you try to connect a display via the Thunderbolt port you end up with the Intel GPU being turned on and gaming performance is degraded on the external display. Now, there is a Displayport connector right next to the Thunderbolt connector on the back, but this essentially makes using a Thunderbolt docking solution a 3-cable setup (power + Thunderbolt + Display) vs. a 2-cable setup...

 

rofl was thinking in PC tiny UPS like battery for desktop look what i found hahahahah
http://www.logicsupply.com/explore/io-hub/think-inside-the-box-fail-proofing-with-in-pc-ups/

 

This thread is pretty interesting, some recent updates to it:

http://www.tomsguide.com/forum/65602-35-notorious-laptop

 

I saw a few references to folks who got battery power to "work", but now I wonder...

To prevent over drawing from the battery do you essentially set TDP limits?

Sent from my SM-G900V using Tapatalk

 

Exactly. The thing about desktops is that they suck up whatever power is available, with little to no regard for power management and downclocking, because desktops by definition are permanently plugged in—at least while in operation. Not so for notebooks, and it is why stuff like the P570WM, P750DM and P870DM are freaks of nature—the LGA socket and the CPU are free to draw so much power simply because they are not meant to be on laptops. Even Intel says so.

That's one way of putting it. Modern notebooks use either lithium-ion or lithium-polymer batteries, which are highly sensitive to discharge current. Granted, I = P/V, and the voltage discharged by a typical given lithium cell is nearly constant ( 3.8–4.3 V) hence we should be looking at power, not current draw. But it is current that kills lithium cells, so controlling that has to be the target anyway. Doesn't matter how it's done.

 

Is limiting power consumption in a desktop system really rocket science? It seems like all the tools to do it are thete. Perhaps the real issue is switching the "profile" when the system is plugged in.

Sent from my SM-G900V using Tapatalk

 

Measuring and managing power draw is a lot more difficult than people think

 

I have to admit this surprises me. I guess a laptop plugged in with a 300+ watt power supply is quite different than a laptop running on a 90+ WH battery...

Sent from my SM-G900V using Tapatalk

 

A 90 Wh battery gives you exactly that: 90 W of power, for one hour. Let's take the maximum FAA lithium battery capacity, 99 Wh. Most removable laptop batteries use packs of cylindrical 18650 Li-ion batteries. This sort of battery typically nominally outputs 4.2 V. The P870DM3-G has a battery voltage of 15.12 V, probably containing multiple 18650 batteries in parallel/series.

Let's assume one draws 330 W from this battery (single GTX 1080, i7-6700K system). The current through the battery is something on the order of 330/15.12 = 21 A. This is already fairly high, notwithstanding the additional thermal stress on the battery from the nearby, hot CPU and GPU. Lots of factors to consider, least of all that the battery is not just inflammable, but potentially explosive.

 

Yes, I get this. To run for 4 hours on a ~100 WH battery you essentially have to draw on average 25 watts.

Yes, at some point your just shorting the battery. I would like to think there is a current limiting system that prevents this, but the recent Galaxy Note 7 issues are sobering.

In a typical laptop, is power limiting implemented through through the BIOS and EC?

 

Generally, Li-ion battery packs already have built-in protection circuits.

Notwithstanding that, I think power limiting is important enough that notebooks have hardware protection, not merely software switches. That's why notebooks tend to hard reset when the CPU draws too much current (happened to me plenty of times), or the GPU is over-volted too much.

 

to solve the battery problem we could simply find a small UPS and put it into the custom case that would prob be enough. it'll be too difficult to have a specific battery build for a standard desktop mobo thats simply way too much work.

i mean with desktop mobo being around for that long, still can't find anything that comes with a battery this shows how difficult and rare is to get it done. mini UPS is the way to go.

 

That was an issue of expansion, the cells will expand under use and if restricted this could cause th cell to rupture releasing all the stored energy in one go.

Ever wonder why the P570WM had a physically larger battery than other models but was a similar capacity? The cells are further apart and given more room to expand/cool meaning you can safely free more from it.

 

Is there any argument that the junky BGA performance is tied to minimizing power requirements? That is would it be possible to design a socketed system that can run for 6+ hours on battery or does BGA score a point here? For this comparison, lets say we are using the same WH battery...

A small form factor PSU sized unit that behaved like a a UPS with an external power adapter (just like the bricks we lug around now for our laptops) seems very plausable.

Here's a 6700/1070 system that fits in the "node 202" case (~15 x 13 x 3.5 inches)

http://pcpartpicker.com/b/4xXH99

The Spark S3 case is pretty cool - this looks close to a laptop prototype box. I could even prop this on my laptop =P



Do you think the Spark-S3 could be outfit with a i7-6700K / GTX-1070 board using a mini-itx board like the GA-Z170N-Gaming 5?

http://www.ssiportable.com/products/portable-solutions/spark-s3/

EDIT: I noticed there's room for batteries in the box

 

Another interesting thing - I've seen a lot of server room combo keyboard - displays but very few portable versions. Something like this:

It would make a lot of sense to have a generic portable version of this to plug into whatever PC you want.

Semi-portable couch-top computing for all! =D

 

Tested Cinebench R11.5 with fully locked 6700hq aka BGA FILTH clock-speed. The fan could not bear to start in my machine. Massive 31C degrees max temp in the bench While the new thin modern flimsy JOKEBOOKS struggling with +90C in same benchmark test Passive cooling should be more than enough for this low powered filth.

 

tell that to @hmscott he thinks with just BGA mobile it means power saving LOLOL and reducing heat.

 

that looks hella ugly lol but i guess if im gonna use external keyboard why the hell not. issue is u gotta look upward while using that haha. i still prefer the S9T or S24T just cause triple screen and can fit EATX for LGA2066

 

I don't care too much about looks, I was more interested in a proof of concept performance DIY laptop. I can't tell much about the Spark S3 case other than that it takes a mini-itx board, the S3B version is supposed to have a 220W power supply and they left room for batteries in there but the specs don't give any details on what the intended purpose of the batteries is. The whole rig is less than 5" thick which is quite remarkable. That could still fit under the seat in an airplane which to me would be important.

So I wonder, could this Spark S3B case work with a 6700K / 1070? That would be pretty awesome and it really would be a DIY laptop. Especially if there is flexibility in the particular mini-itx board used and if the case has adequate cooling for a decent graphics card. But again, there are almost no details provided on the site and I couldn't find any other web references to the Spark S3(B) for example used in an actual system. So perhaps this is a new product? Or maybe an old product...

Yes, you have a very different set of goals with the S9 - and certainly not bad ones - but I understand you want a mobile desktop system which would be quite a bit thicker (9.5"). But three screens is really nice.

The Sparc S5 is another one that takes a micro-ATX board and is 7" thick and allows for more expansion. I think most mini-itx boards only have two ram slots.

The whole Sparc series of cases are very interesting. They look rugged and they all come with keyboards and screens which is quite nice. I can't tell if they are also meant to be water resistant.

 

Papusan, perhaps this gets at my question. Could one configure the 6700K (i.e. with throttlestop) so that a system using it would run for over 6 hours on a modest sized battery? I am just wondering if there is a case where one has to use the "low power" versions of this CPU or if it can be handled through a configuration of the 6700K.

 

Server rooms tend to have trollies with a screen and input devices on them for other systems that don't normally need a dedicated console.

 

Yes, I've used them before

This is very close to what I would like to see in a lap-based form factor. I've only come across one so far.

EDIT: maybe gluing a beanbag lapdesk to the bottom would do the trick...

Sent from my SM-G900V using Tapatalk

 

You want to carry around a rack mountable console? That's heavy and has no hardware inside?

You could go for a corsair lap desk and keep the mouse inside it. All you need to do is glue a display to it.

 

A company called "TRANS2000" made one of these as a custom order ( http://www.trans2000.com/oem-odm-services/). It looks like it will fall over, but maybe a heavy been bag tray would counter balance it:

 

They likely weighted the base to make it balance.

 

The Spark style enclosures on ssiportable.com I think are sometimes called lunchbox computers.

Here's one that can be configured with 6700K and a GPU:

http://www.ipctechnology.com/products/portable-computer/mpc-1500.html

It's 7" wide which is almost reasonable. 28 lbs is not light. I don't see this kind of thing going on anyone's lap, but it is a portable desktop system complete with keyboard and display in one case. You can separate the keyboard and put that part in your lap

I have seen a few references to battery powered lunchbox computers, but I don't know if any of these are in the high performance bracket we're talking about here.

EDIT: here's a thumb-nail sketch of how to run a desktop off of battery power with a caveat stating that you shouldn't do this for the reasons alluded to here - without proper fail safes, shorting a battery would not be a fun experience:
http://www.wikihow.com/Run-Your-Desktop-off-DC-Power

 

If laptops like the one I have do go extinct, that is exactly what I would move to. It would still be semi-portable, more so than a standard desktop anyway. And I love the all-in-one nature, yet still standardized components. I remember having one of those really tiny cases with special motherboard, and special power supply (250w?) which caused no end to my irritation when trying to upgrade GPU's. Plus carrying around the screen separately was a pain.

I built my own version a few years ago of what you linked to, it was a smallish but yet standard mainboard case, all standard components. I bolted a 21" screen that had speakers to the side panel that didn't open. All I had to do was grab keyboard, mouse and two AC cords and I was ready to go out the door. While it worked, was pretty cheap compared to a high end laptop, it was still limiting and I eventually moved back to portables.

I like my portability, and am willing to pay extra for it, as long as they keep making them how I want them.

 

So long as there is a decent sized market then the products wont go away.

 

Yes, the laptop I have is very convenient and I'm not eager to follow the DIY route, but I like having a DIY "proof-of-concept" out there mainly to keep hope alive. So I have been thinking in terms of what are the loose constraints for a portable PC / "laptop"? Something like:

Maximum thickness (5"? 7"?)
Has a built-in display and keyboard
High performance components (e.g. 6700K / 1070)
Can run on internal batteries for 4 hours

That last point is the tough one and I have no idea what is involved in it other than DC to DC converts etc... The system would have to auto-detec when it is on battery and switch the profile in throttle stop or some such thing.

 

I feel mostly the same way, but I would rank battery as the absolute last, and even optional part. I did use my beast this weekend on battery burning a DVD to watch and some other minor tasks. It was nice, but it would have been no big deal for me to plug into AC.

I have AC outlets all around the house, even outside where I sit in the summer, and even in my truck so I am always near AC if I need it. The few times I am not near it, like hiking/biking or whatever, I would not need to use a computer anyway.

 

The cool thing to me is that I think there is an overlap between the needs for "field" PCs - rugged desktops that can run anywhere - and high performance portable gaming systems. The differences between the needs are not huge. Just having a rugged portable all-in-one gaming system would be a big step.

I guess the reason this interests me is that I am a fan of open standards. Just having a viable open DIY portable gaming system that would please enough people in terms of portability, usability, performance and cost would be a big step. The proof-of-concept system with enough momentum can create viable niche markets for better cases, more specialized features in the mobos, more power options, better BIOS features, etc...

The ultimate goal? Keep DTR gaming laptop manufacturers on their toes. Large expensive "high performance" BGA systems start to look pretty pointless if you can make a viable DIY system with commodity desktop components that is "good enough". Definining "good enough" is a really interesting problem. Figuring what a large enough group are willing to accept for size, weight, usability (i.e. form factor) and cost to create a niche market. Success would yield many things - a better understanding of laptop design for very advanced users, a minimum standard for high performance portable gaming systems (e.g. laptops), a broader range of option for portable system builders, and with all that the evolution of better suited components to build these DYI systems with.

(Maybe better than battery would be a generic (unregulated) DC input for running off car battery or alternator.)

 

There is the problem that the top end of the market is quite small. Most of the gaming laptop market is at the mid range more portable, and affordable level. I suppose you would have to consider it not specifically gaming, but just high performance luggable box that could be for gaming, graphics work, or whatever. Then maybe it would be worth it.
As to the battery issue, that is certainly not a tough one. Laptops have always been able to know when they are on battery or mains. Giving a desktop the same ability is easily done, as is producing the voltages a desktop PSU produces using batteries.

 

You never want to handle something so critical at the OS level. The closer to the metal you are, the less chance there is of failure. That means a custom motherboard at the very least, or handling it all within the battery/PSU system before it plugs into the ATX connector.

 

The DC to DC converters have a lot of power/current limiting features. Windows already has a feature that detects when the system is running on battery - this must be based on some hardware signaling. This is where throttle stop / afterburner can be used step down the power consumption of the CPU / GPU.

 

And what happens when your software-level throttling program crashes right before you unplug the power? What if the user sets a process to realtime priority and spinlocks before the cable comes out?

 

Certainly we are talking about a small number of very demanding enthusiasts. But it is possible for some things to gain momentum (granted there are many instances where this fails as well). Success could mean more main-stream solutions for the even larger number of people who get suckered into buying big, expensive and poor performing laptops.

Yes, I am somewhat optimistic that there are high performance needs among those that use portable rugged desktop systems out in the field. There is clearly a market there.

Yes, I think you're right. There are ways to down step the CPU and GPU - at the very least with throttlestop and afterburner - that can be triggered by the system's power mode.

Is it just me - or is there some potential for a kickstarter project in there somewhere?

 

You would be surprised what type of mistakes general " non - enthusiast" sometimes makes. Not to mention bugs. And most of "signalling" isnt just hardware based, there are complex controllers that take care of a lot of lower level stuff.