Intel's Long Awaited BTX Form Factor
by Purav Sanghani on November 15, 2004 5:31 AM EST- Posted in
- Cases/Cooling/PSUs
Final Thoughts
Instead of an overnight switch akin to what we saw from Intel during the transition from 865 to 915, the move from ATX to BTX will be more of a slow gentle transition. We saw some exciting things in the AOpen/Intel combination today; smaller power supply, no dead heat spots (places in the case that air hangs around with no where to go) and smaller design. Although there are certainly advantages in BTX, no one would realistically expect Intel to force the new form factor down our throats. As Anand mentioned several months ago, "Switching CPU sockets is one thing, but force people to buy all new cases, power supplies and motherboards and you're bound to get some negative response." In the lab, some of us were fairly divided on BTX in general; Anand and I had spent the majority of dinners over the last year debating where the technology fit. When Intel sent us a microBTX board rather than a standard BTX design we were extremely hesitant on what we were ultimately proving here with this analysis.Years of planning, marketing and finally deployment have lead up to today's official release of Intel's first BTX product. Although the BTX solution we looked at today performed poorer than a full ATX desktop, we need to put things in perspective. Our microBTX case and motherboard ran a 3.4GHz Pentium 4 significantly cooler than a microATX case with the same volume, even though the microATX setup had an additional fan. The noise benchmarks further tip the scales in BTX's favor. How is Intel doing it? Aligning the core components into a strict path for airflow to follow while using a slower 120mm fan to "push" air. The internal wiring is reduced in favor of a single 24-pin adaptor, thing SATA cables are used instead of fat PATA ones; things inside the case are just less cluttered.
With the flurry of SFF computers over the last few years, we know smaller computing platforms are necessary. When we saw Shuttle, IWill and FIC designing their own small form factors to produce smaller and smaller desktop PCs, the results become obvious that an industry standard design was long over due. The original goals of BTX have always been to decrease the form factor size and progressively reduce the size of desktop computers in general. MicroBTX isn't the ground floor of the BTX standard, PicoBTX - even smaller than MicroBTX - was planned in the original design of the BTX spec. When we see BTX mature a little more, letting Shuttle or FIC run with the PicoBTX blueprint might really show us some revolutionary computing.
To succinctly put it, the BTX approach to a case/motherboard design is an approach in efficiency. As we saw in our uBTX measurements, there were virtually no dead spots in the case, and these results were obtained with fewer fans. The much larger ATX chassis with several additional fans (including a dedicated CPU heatsink fan) ran cooler, but we could see unused portions of the case. The ATX case used more space and ran louder. Don't forget that additional non-redundant fans increase the failure rate of a computer as well - more moving parts. Now back to the question as to why Intel sent us a microBTX design rather than a full ATX setup. Thermals and noise tests today showed us that a first generation microBTX setup ran slightly hotter - although not as hot as a standard microATX case - while reducing size, fans and noise. This is significant as we will undoubtedly see cases and motherboards moving to smaller designs. Many of us were skeptical for a long time, but after seeing the numbers and realizing Intel's long term plan to reduce the PC footprint, and not just on the OEM side, we can conclude that BTX is not hype. It's obvious why Intel waited for Monday morning to lift their BTX platform - they have a winner on their hands.
Facebook
Twitter
RSS
77 Comments
View All Comments
trikster2 - Tuesday, November 16, 2004 - link
I've got to say, this review is bad, even by anandtech standards.
Comparing an mBTX destop to a mATX TOWER to a Fullsize ATX desktop, litteraly comparing apples to oranges to pears, and then coming to sweeping conclusions like "any of us were skeptical for a long time, but after seeing the numbers ... we can conclude that BTX is not hype. " is simply ludicrouse.
Shame on you anand for letting such a review fall through your QC into the hands of the general public.
I'm not sure if you guys are incompetent or payed by intel to find a combo to pimp the BTX standard. I just did a refresh 5 times, no intel adds so it must be the former.
How about spending 49.95 on a decent desktop mATX case and compare apples to apples, instead of just grabbing whatever was laying around the lab (we've seen that whacked oppus case in other reviews haven't we?)?
I have the silverstonetek equivalent of the oppus matx tower. It is the worst mATX cases I have ever worked with. Give us a break and do a valid review, mATX single fan 1/2 height desktop to mBTX single fan 1/2 height desktop.
johnsonx - Tuesday, November 16, 2004 - link
I see nothing here in the BTX standard that couldn't have been better accomplished with a backwards and forwards compatible ATX II standard.ATX II compliant motherboards would have standardized locations for the CPU socket, two standard locatations for RAM (either or both locations could be used, as needed), standard locations for the power connectors, standard locations for the most common onboard ports (IDE, SATA, front USB, audio, etc.) and standard component heights (capacitors and etc.).
ATX II compliant cases would have standardized air intake and exhaust locations, fan mounts, and ducting. ATX II cases could be built to open from the right side, with the CPU on the bottom. The ducting would be removeable for use with a legacy ATX mainboard.
ATX II compliant heatsinks would be fanless, designed to be cooled by an ATX II chassis duct, but would have a fan mounting point for use in a legacy ATX chassis. Boxed coolers that come with retail CPU's would include the fan on a standard clip with instructions to remove it when used with an ATX II compliant chassis and mainboard.
ATX II compliant power supplies would basically be the same as the latest ATX supplies are now - 24pin power, 8-pin ATX 12v power, standard 4-pin drive power plugs, SATA power plugs, PCI-E graphics power plug. 24 to 20 pin converters and 8-pin to 4-pin converters would be standard with retail power supplies.
Any mix of ATX and ATX II components would work fine together, but a system built of all ATX II components would gain various cooling, noise level, assembly and operational advantages.
What does BTX offer that my hypothetical ATX II standard doesn't? Oh, yeah... ATX II couldn't give you a face of hot air...
xsilver - Monday, November 15, 2004 - link
To the guy b4 that was wondering where the air intakes are -- no amount of case fans are going to create a vacum inside your case, there's just not enough pressure -- unless you tape up all the gaps lol.... so the air will just "flow" in from any holes in your caseblackmetalegg - Monday, November 15, 2004 - link
Bah we're(work) using this motherboard in an application that shall remain anonymous(NOT oridinary PC) and the damn design make wiring a bitch to deal with. And the motherboard isn't all that stable, considering it's from Intel...justly - Monday, November 15, 2004 - link
BTW that 20.666715 "cfm" is assuming all of the area taken by the 2 fans has air flow, but we should all knoe that the motor and frame ocupy some of that space so the actual "cfm" would be lower.justly - Monday, November 15, 2004 - link
Why was "lfm" not converted to "cfm" for a direct comparison on page 11?If I am reading this correctly the BTX provides 40 "cfm" while the ATX requires 150 "lfm". This sounds like BTX is an overwhelming success, untill you do the conversion that is. So what exactly does this 150 "lfm mean in comparison, well if the ATX uses 2-80mm fans that would be 80mm x 80mm = 6400 square mm per fan, 6400 x 2 = 12800 square mm for both fans, convert that to square ft and you get 0.1377781 square ft, take that and multiply by the 150 "lfm" and we get (drum roll) 20.666715 "cfm". Thats right about half what the BTX has.
I don't know if this was intentional or not, but I will give the author the benifit of the doubt and assume it was not.
If I have overlooked something please explain my mistake, if not then I feel some injustice has been done to the readers that did not ask this question.
ShadowVlican - Monday, November 15, 2004 - link
like someone else said... the BTX form wasn't made for enthusiasts, since we already know how to cool our rigs properly..BTX is for the average consumer who doesn't know jack about computers, so BTX will help those people have quieter/cooler computers without them reading forums/articles
i'm just wondering how well the transition from ATX machines into BTX will go...
ZobarStyl - Monday, November 15, 2004 - link
Yes Five40, you can't directly upgrade from 754 to 939, but you can take your case and PSU with you...I'm using the same case I used in my AXP system now for an A64 with watercooling, but if Intel has their way, my next rig won't be able to use the same rig and I'll just have to spend more cash. Hence, we are rightfully pissed because the real question is "Why?" and when something lacks benefits but costs you mad cash: a whole new case and good PSU will set me back 120 bucks, for what? Nothing worth my time.MAValpha - Monday, November 15, 2004 - link
Anyone know how pBTX stacks up against FlexATX?five40 - Monday, November 15, 2004 - link
#50....umm I can't do a A64 to A64 upgrade so what's your point? 754 to 939 does me no good. I just see this as another socket update. A64 might not need it right now, but if all things where just made for what we need right now, we'll move forward at a very slow pace.