Please report all spam threads, posts and suspicious members. We receive spam notifications and will take immediate action!
Results 1 to 4 of 4

Thread: P67A-UD7 Most In-Depth Preview/Review!!

  1. #1
    Join Date
    Sep 2010
    Norther Virginia

    Default P67A-UD7 Most In-Depth Preview/Review!!

    P67A-UD7 Most In-Depth Preview/Review:
    With Intel’s release of Sandy Bridge processors around the corner we have a lot of new technology to look forward to; the P67 Cougar Point LGA1155 boards will be one of the top choices for Overclockers, Gamers, and System Builders alike. Sandy Bridge will fill in as the LGA 1156 replacement, and following in the footsteps of its predecessor with a Platform Controller Hub (PCH) instead of an IOH and ICH (Northbridge and Southbridge). The processor itself will carry an Internal Memory Controller (IMC) (the board lists 1066MB/s-2133MB/s), a brand new Internal Graphics Processor (IGP), 8 5GT/s(16 2.5GT/s) PCI-E lanes, and a wide variety of new features. Some to name are a Power Control Unit (PCU) inside the processor, an internal clock generator, and oups cache. All the new features should really raise the bar as far as processor engineering goes. Overclocking is supposed to change because the system bus will be tied to a variety of devices, the base clock that we change will probably have to stay close to stock, while the multipliers are raised. Turbo Mode 2.0 is going to be our friend instead of foe, and power delivery will become ever more important. SLI will require an add-on chip, because natively only one graphics card can be used with 16x lanes. GIGABYTE has a solution for everything, and today I will take a firsthand look at their flagship GA-P67A-UD7 rev 1.0.
    Some highlights include the VRM/PMW design (power delivery), and I go very in depth about the advantages and disadvantages of Analogue VS Digital VRM design. BIOS section including brand new BIOS settings including wattage and amperage to the implementation of EFI BIOS in the future. The USB 3.0 sub-system on this board is phenomenal tree type system along with SATA6G connectivity this board can power all of your devices to their fullest extent, and when you have all the SATA and USB ports filled you can enable USB 3.0 Turbo Mode to add speed and bandwidth. I take apart of the cooling on this board and put it to the test to find out how well it transfers and dissipates heat in the motherboard cooling section. I then take a look at both PCI-E busses in the NF200 and PCH P67 section, and analyze the benefits. Follow along with me as we take a look into the near future!
    The way I do reviews is in depth and on point. Right now I cannot post any benchmarks, Intel markings, or list any spec. What I can do is show you the board, go over its feature set, and most importantly explain in depth about how everything works. If you have read any of my previous reviews you know that I go in depth on every IC I can find and then I explain their uses, because that is what makes the board. I like to get up close with my Canon EF 100mm f/2.8 USM Macro Lens, so the pictures are just as in depth as my analysis.
    Just sit back relax and enjoy the ride!

    1. Introduction
    2. Unpacking Box & Accessories
    3. Board Layout and Overall Design
    4. Power Delivery at its Finest (The Nuclear Power Plant)
    5. The NF200 SLI
    6. The PCH and Other Connectivity
    7. Random IC’s + BIOS (including BIOS features)
    8. Motherboard Cooling (includes torture tests)
    9. Conclusions & Analysis

    Need printable PDF?? Here you go!:File name: P67AUD7Preview.pdf File size: 1.98 MB

    Unpacking Box & Accessories:
    When I received this package I was very surprised as GIGABYTE has changed a lot about their design as far as the box design goes. No glitter, just nice sleek black and gold design:
    Here are shots of the back of the box as well as the ISBN that lists general features and revision:
    Next we have a shot of inside the front cover, like other GIGABYTE flagship motherboards:
    This board has so many features. It is just incredible how much work went into the design. Motherboards aren’t just ICs and traces, now they are almost like pieces of art! Maybe I just enjoy them too much….
    Here are some shots of the many accessories included. We have 4 SATA cables (baby blue), as well as two eSATA to SATA cables (black), eSATA bracket, eSATA power cable, motherboard manual and drivers cd, 2x/3x SLI bridges, and the back panel. (please excuse the hot glue mess on the side, it’s my anti-static/work mat)
    Now there is something I want you guys to see, because I think this is a great idea. It’s the yellow warning paper below, showing the difference between 1156 and 1155 socket and CPUs!
    Here is an up-close and personal:
    It’s in every language, so there really is no excuse for taking a hammer to your old CPU and this elegant motherboard.
    Also let’s not forget this board has 2x copper PCB (8 layered, 2x more copper in power and ground layers), a feature of all Ultra Durable 3 Motherboards, ASUS just introduced this as well.

    Board Layout and Overall Design:
    This board design is very appealing because it’s almost all black; this is a case modder’s dream waiting to happen. Imagine a black case with black/hint of gold motherboard, black sleeved wires with gold molex connectors, black power supply, black video card, with gold ram, and water cooling with gold tubing.
    Black PCB design has really caught on in the past few years, now GIGABYTE is offering their design with smooth matte surface; it’s almost like an extra layer of PCB. I have a personal gripe with many black PCB motherboards because up close they look brown!!! It is because of the copper traces and PCB coming from underneath the black coating. On this board there is nothing on the sort. On the top PCB layer there is a little #1 in a box where it should say #8 for layers of the PCB, like it does on the backside, I believe that GIGABYTE didn’t just coat this board black, they used another layer of PCB to really make this a true black motherboard, no brown here!
    Let’s get up-close and personal:
    Processor area with all 24 phases, mosfet drivers underneath the coolers, as well as 24 chokes, go ahead and count them up. Make note that LGA1156 coolers will fit, if you have LGA 1366 cooler you need to buy a socket adapter, they are only a few bucks.
    Next let’s move on to the RAM slots, manual says support for up to 2133MB/s DDR3 4gb max per slot Dual Channel Configuration, that is blazing fast. RAM slots are black to match matte PCB:

    Here we have the connectivity area, the black matte PCB contrasts really well here:
    Now we move onto PCI-E with 3-Way SLI/CrossFireX powered by NVIDIA NF200 (NF200). There is also one PCI-E 1x port that got cut out of the picture; it’s above the top PCI-E 16x slot. I should mention that the NF200 blocks the PCI-E 1X port so that you have to use a low profile card. We have 4x PCI-E 16x/8x slots, as well as 2x PCI slots.
    I would like to point out a few things that are new, first are the two USB 3.0 Connectors both are pictured here in the center, they are wider and have many more pins than traditional USB 2.0 ports:
    Here we have the back panel, 10 USB ports!!!!(The 6 Blue ones are USB 3.0):

    Here is the 8 inscribed on the board, means it’s an 8 layer PCB. This is under the top side of the board, on the top side of the board in this exact mirrored spot, there is a 1 where there is an 8 here, I am guessing that this is a layer of black matte PCB, and that is why here it looks a bit brown on but on the surface on the top side of the motherboard there is no brown, just black.
    Quick Switches are also a great idea, and they have carried over to this board. The big button is obviously power, the small blue one is reset, and the small black one below is clear CMOS, in case your overclocking fails.

    Now on previous boards, there were heaps of LEDs, while many people enjoyed the Christmas tree effect, many did not. The truth is only some LEDs were automatic; others had to be turned on through the Dynamic Energy Saver program. This board only has 4 sets of LEDs and they serve a good purpose.
    To the left are the phase LEDs, 6 LEDs for 6 Channels, then we have Two memory LEDs showing normal voltage or high voltage memory operation.

    To the left we have ACPI LEDs, these can show you what power state your system is in (S0, S1, S3, S4/S5), and to the right are CPU VTT LEDs showing normal and overvolting.
    How to install CPU:

    Power Delivery at its Finest (The Nuclear Power Plant)
    Now I am going to switch to one of my favorite topics, Power Delivery. With Sandybridge being less about “how high your board will allow base clock to go?”, and more about “how high will my chip’s multiplier go?”, Sandy Bridge overclocking is heavily multiplier dependent. So naturally boards need to continue doing what they do best, power delivery.
    With Sandy Bridge comes new Voltage Regulator spec, Voltage Regulator Down 12(VRD 12) is Intel’s new spec for power delivery design. One major aspect is the addition of SVID.
    What is Serial VID?:
    With VRD12 comes SVID (Serial Voltage Identification), in the past Intel introduced DVID (Dynamic Voltage Identification), as well as plain old VID (voltage Identification, stock chip voltage). Let me explain these a little bit to you. DVID is dynamic voltage regulation protocol for the processor, when you change frequency of the chip voltage needs to be changed as well, DVID works by setting predetermined voltages depending on the situation, such as in C1E power saving state. DVID is a way to assign voltage values to frequencies and a way for the processor to know what voltage to operate at. In the BIOS you also have the ability to set DVID +/- voltage increments and under load it will bump the voltage up to what you set plus stock VID, so DVID of .250v + VID of 1.2V when your processor is under load Vcore will be 1.45v.
    With SVID you add a little kick to the situation; Serial VID is a simple but powerful way to link the processor’s power regulator(PWM) to the motherboard’s power regulator. SVID is a super high speed interface that features VID tables, fault response, and power states. SVID allows the PWM (ISL6366) to “talk” directly to the processor to find the best voltage for different frequencies. What you get is an ultra-smart system that can be fed suggestions, and then it will take the suggestions and make it happen in real time. For instance, multiplier overclocking on unlocked chips is a common occurrence and turbo boost is as well, what if you want to set a multiplier that the system isn’t ready for because the frequency is too high? You proceed to mess around with the right voltage trying to find what is best, and after hours of tinkering and overvolting you end up with a stable system or a fried chip. SVID is a way that you can give the processor info on what you want to do and it will determine the best way to proceed with voltage. It will take a matter of seconds to find the proper voltage for your overclock.
    The most savvy Overclockers will tell you that every processor has its sweet spot, and that if you overvolt performance will decrease and the same goes for undervolting. This is due to the fact that voltage isn’t the only thing that powers the chip, current(amperage) is also in play. Right now the user has no way to tell how many amperes are being fed to the chip, just voltage. With SVID this will change.

    The Analogue PWM (Pulse Width Modulator):
    The Interstil ISL 6366 is the new phase controller chip with 6 channels. The ISL 6366 is a Voltage Regulator Down (VRD) 12.0 complaint analogue PWM (pulse width modulator), meaning it has to conform and operate up to certain Intel standards to power the new Sandybridge Platform, most importantly it has to support SVID.
    This PWM is the brain of the whole power delivery system; it’s a smart little chip. What is great about Analogue PWMs is that they have ultra-fast response, no firmware required, no analogue to digital conversion, no errors to deal with, high switching frequency, and they are downright dependable. The ISL 6366 controls the driver MOSFETs.
    Six channel operation means that you take the whole phase array of 24 and divide it by 6 channels so you get 4 driver MOSFETS, 4 ferrite core chokes, and an array of capacitors per channel. This way every stage of the voltage is refined to provide the cleanest(less ripple) and most efficient (90%+) power to the processor. This technique improves overall efficiency while not reducing power output (which is huge).
    There is a reason GIGABYTE uses Interstil, and that is because they have a very good reputation for delivering some of the best Analog PWMs on the market.
    Let’s move on to the driver MOSFETs, the smart ICs that change 12v to 1v-2v and delivers more amperage than your power supply.
    The Vishay Siliconix SiC769CD is a common driver MOSFETs on GIGABYTE motherboards and for good reason. These bad boys and each put out 35-40 amps, 35 amps maintained load at 1 MHz switching frequency. As switching frequency increases from 300KHz-1+MHz, current (amperage) increases in a linear fashion along with it. Temperature also plays a large role in overall power output and efficiency, and these driver MOSFETSs operate best from 20C-50C. They have automatic thermal shutdown at 150C and when they drop down below 135C they start to work again. GIGABYTE has taken care of the temperature problem as you will see in the motherboard cooling section. The advantage of having such a large array of driver MOSFETs and chokes is enormous. Lower temperature, lower ripple, high voltage stability, higher amperage/voltage output, and the list goes on and on. This power array GIGABYTE has to offer hasn’t been reproduced yet. Other than GIGABYTE boards I have not seen 24 power phases on a board. What is unique about these driver MOSFETs are that they each have a high side MOSFET, a low side MOSFET, and driver chip in them; they are an all-in-one solution.
    If all phases are working at 30amps each (isn’t even the max) at 1.9v the whole 24 phase array can generate 1368 watts of stable, reliable, and clean power.
    Driver MOSFETs + Ferrite Core Chokes Improvements: GIGABYTE claims reduction in temperature of 16C on the driver MOSFETs and 27C on the Ferrite Core Chokes, from previous generations!

    Along with the analogue PWM we have a nifty little iTE IT2875e phase switching chip. This chip switches on an off phases in use and not in use. Some people asked me a few really great questions about automatic phase switching that I will address right now.
    Question #1: If this motherboard advertises 12 on and 12 off phases then are there only 12 phases on at one time, so in reality there are only 12 phases giving power to the CPU?
    Answer: Not exactly, this is how it works, you have to enable dual power switching in the Dynamic Energy Saver program, and then upon reboot if it is enabled phases will switch 12 on 12 off. If not enabled phase switching does not occur.
    Question #2: If this motherboard has “gears” where it turns off phases how do I know that all phases are working, or that phases are being switched?
    Answer: Phases will “gear” automatically at start up when at stock settings. Upon start up fewer phases will be used than on full load where all 24 phases will be used at stock situation. It isn’t that simple though, if the system detects and overclock then this “gearing” does not occur, and all phases will be activated at the same time automatically, as you can see from either the phase LEDs or the DES program.
    Last edited by sin0822; 12-16-2010 at 03:07 AM.

  2. #2
    Join Date
    Sep 2010
    Norther Virginia

    Default Re: P67A-UD7 Most In-Depth Preview/Review!!

    [Analogue or Digital Education for the Masses!!!:
    Analogue: In the past you used to have a high side MOSFET, low side MOSFET, and the driver which controlled the flow of electricity(all per one stage), now that has changed as many manufactures are using driver MOSFETs (instead of the MOSFETs you see on older and even newer motherboards). Response to voltage change is almost instantaneous and now with large arrays of MOSFET drivers you are able to control/regulate voltage as precisely as a digital PWM could. Other benefits of large analogue phase arrays include , higher efficiency, higher power output, no firmware, no processing errors, good ripple suppression, and high switching frequencies. Higher switching frequencies result in higher temperatures. In the past with all the different MOSFETS and drivers you wouldn’t be able to fit 72((2 MOSFET+1 driver) x 24 phases) MOSFETs + drivers on the motherboard just to power the processor. Another benefit of analogue VRMs is that you can increase power output by adding phases per channel because analogue PWM design allows for different arrangements of MOSTFET drivers and capacitors, whereas Digital PWM design is limited to a single design. This is why you see this Analogue controller with 4 MOSFET Drivers, 4 Ferrite Core chokes, and many capacitors per stage. This type design increases efficiency as well as power output.
    Digital: With digital VRMs it gets more complicated as you need an analogue to digital converter for everything, the main difference in Digital PWM design is the processor of how the pulse modulator works and receives a signal, other than that they are almost the same. Because of the digital aspect of the design, software and hardware are used instead of straight hardware, this has many benefits and downfalls. Digital PWMs do not need to operate at high switching frequencies and thus have some benefits such as much cooler temperatures. For Overclocking at high speeds you need to use a high frequency of 800KHz+. Digital solutions also offer good ripple suppression, good voltage regulation, the ability to use a single large capacitor, and the ability to change switching frequency on the fly/at users request. The problems with digital PWM/VRM are that efficiency is low and they respond slower because of the processing that needs to take place. You also have to deal with software and errors and bugs that come along with it, but when you OC you wouldn't know whether your system is unstable or you did something that caused a software error for your Digital PWM. Digital PWM/VRM can save you a lot of space on a board because only one large capacitor is needed for digital systems.
    Conclusion: If you compare digital VRM/PWM to low end, less than 12 phase analogue VRM/PWM the digital PWM would be the right option. If you are dealing with a large and high quality analogue VRM/PWM and high quality digital PWM both systems have a few things in common such as excellent ripple suppression and voltage regulation. Other than that here is a table for the best Analogue VRM/PWM VS the best Digital VRM/PWM right now:

    As you can see, both are just fine, and it comes down to how much the manufacturer spent on the voltage regulator and controller on your motherboard. GIGABYTE really dropped a pretty penny on this motherboard’s voltage regulation, and that means you will too, digital is the same thing, you really can’t have the best unless you spend the money. It is going to be hard to beat this array with any digital VRM solution.
    Another shot of the MOSFET driver:
    Moving on to NF200 and RAM power plants:
    The ISL 6322G is a 2 phase buck controller, this board has two of them. The first powers the NF200 chip, and even though it is only a two phase controller it does the same thing as what we saw with the processor; each phase is beefed up so in total there are 5 chokes and 4 Low RDS (on) MOSFETs that refine voltage and reduce ripple, by doing this you can use a lower voltage to power the NF200 and reduce heat and improve efficiency. Low RDS (on) MOSFETs and ferrite core Chokes for NF200:

    For the RAM, we have another ISL 6322G 2 phase buck controller. This time we have 6 low RDS (on) MOSFETs and 5 ferrite core chokes, thus again reducing ripple and making sure the DDR3 slots get stable power. You should also take into consideration that the processor feeds a large portion of the DDR3 power.

    NVIDIA NF200 the addition of SLI:
    A little history about this tiny piece of silicon:
    A time not too long ago as some of you may remember, NVIDIA made chipsets and one of the greatest features was SLI support, while young and very buggy it introduced a totally new concept to gamers and game designers. The NVIDIA NF200 was initially designed to provide SLI technology to chipsets NVIDIA did not produce. NVIDIA released this chip when it was still producing chipsets a few years ago. Many think NVIDIA did this because they knew that they were leaving the market, but still wanted to provide SLI support to their loyal customers, and thus keeping the demand for multiple GPUs. After Intel won its lawsuit against NVIDIA for making chipsets, SLI was still possible due to the NF200. Eventually Intel and NVIDIA made an agreement and SLI technology was made possible through the IOH/Northbridge. On the P55 chipset boards, the NF200 chip allowed full 16x SLI for both video cards. This same technology was used on many boards even before P55. With a PCH there is no IOH/Northbridge so SLI has to be integrated separately, just like on P55 platform. The NVIDIA NF200 chip supports 32 PCI-E 2.5GT/s lanes.
    Here is how it works, NF200 takes the native 8 lane 5GT/s bus and doubles the bandwidth to 16x5GT/s so you would have 32 lanes of 2.5GT/S PCI-E bus for 16x SLI. Do not be fooled it doesn’t add PCI-E lanes; instead it splits the ones already present and adds bandwidth. It is really great at consolidating and making sure the proper devices are fed full bandwidth. Although we have 32 lanes of PCI-E bandwidth GIGABYTE says the board only supports 3-way SLI. This is most likely because 4-way SLI would be 8x each card and performance would take a hit. This board also supports CrossFireX which is the ATI/AMD equivalent for SLI technology.
    On this board ALL the PCI-E lanes from the processor are fed directly into the NF200. On many X58 boards and P55 boards, only a portion of the PCI-E lanes are fed to the NF200, the reason is that USB3 and SATA6G work off the high speed PCI-E bus, but this board is one of the few exceptions. The new P67 PCH connects to all of these devices through a secondary PCI-E bus, that has its own clock generator, and its own set of 8 PCI-E switches. The PCH then communicates to the CPU through a high speed 20GB/s DMI interface. The NF200 connects to the 4 PCI-E slots through 2 switches.
    Many people have mixed views of the benefits of 16x/16x vs 16x/8x SLI, regardless the NF200 is needed for proper SLI on LGA 1155 motherboard. This motherboard will probably perform exceptionally at SLI with the NF200 because all the PCI-E lanes from the processor are used solely for the PCI-E slots.
    This board has two PCI-E clock generators, both are the same model: ISC9DB403DGLF. These clock gens’ spec says they put out a clock between 50-100 MHz. They are used specifically for PCI-E. I will cover the purpose of the second clock generator a little later in the PCH section. Along with clock generators the board has two PCI-E switches for SLI; if a 3rd card is inserted then one video card goes down from 16x to 8x. This board utilizes two ASMT PCI-E switches to switch a native NF200 16x,16x configurations to 8x on each PCI-E port for SLI.

    The Platform Controller Hub (PCH):
    With the introduction of LGA 1156 we had a new type of chipset on our hands. Instead of the traditional Northbridge and Southbridge, we now have the platform controller hub, which does everything from connectivity to PCI-E bus. I will focus on the PCH and what it can do for you. Since the LGA1155 CPU will control the primary PCI-E bus for the video cards, the secondary PCI-E bus is for peripherals, the PCH will control the flow of information and this flow continues to the CPU through the handy 20GB/s DMI bus.
    USB 3.0, SATA6G (Marvell), LAN, and the PCI ports all run off PCI-E bus, and all of these peripherals need switches. Let’s go over them in a list:
    USB3.0(NEC) x2
    Realtek (LAN) x2
    Marvell SATA6Gx2
    iTE PCI Ports (1x) x1
    PCI-E x1 (x1)
    Total PCI-E switches needed=8

    So we have 8 places where we need switches, so that if it is not in use the extra bandwidth can be used for other devices. GIGABYTE uses 9 ASMT ASM1440 and 2 P13L switches, 2 ASMT for the PCI-E slots which are connected to the NF200, the other 7 ASMT switch everything but the LAN to the PCI-E bus of the PCH. The ASMT switches are called multiplexers/de-multiplexers which is a fancy name for a PCI-E switch. The LAN use P13L PCI-E switches, such as on GIGABYTE X58 boards. There is 1 ASMT switch left over, it is used for the USB 3.0 Turbo Mode.
    This board also features 2x ICS 9DB403DGLF Clock generators, one for the actual PCI-E ports and the other for all the devices listed above. A PCI-E clock needs to be fed for all these peripherals, which are on a separate PCI-E bus. This is an excellent yet difficult design, but in the end your peripherals do not interfere with SLI performance and vice versa.
    (If you look closely the SMD resistor placement on the lower left-hand corner is different from that of the same clock gen in the picture in the NF200 section)
    The board features Legacy PCI slots, and two of them. Since LGA1155 Sandybridge is a mainstream product and not an enthusiast product, PCI ports are a must. Many people have PCI devices such as soundcards and PCI video cards that work perfectly fine and shouldn’t need to be replaced just because PCI-E is the new standard. GIGABYTE allows the use of two of these devices. The PCI ports are available because of their controller the iTE IT8892E PCI-E Bridge chip which is a PCI-E 1x to PCI bridge. The NF200 is also a PCI-E bridge chip, so I guess you could say this is the daughter PCI-E bridge chip.
    USB 3.0, Turbo, 3XPower, and ON/OFF Charge:
    Let’s take a look at something that is very impressive, GIGABYTE’s USB 3.0 implementation. Along with the new USB spec come some new challenges and GIGABYTE has tackled them pretty well. USB3 devices need extra power; GIGABYTE has come up with a very nice way to deliver it. GIGABYTE’s USB 3.0 system is designed like a tree, on top of the tree are two NEC (Renesas) D720200 USB controllers (each has its own BIOS), they provide USB 3.0. Each controller is attached to a hub.
    (The little 8 legged IC’s with a white dot are the BIOS chips for the USB controllers)
    The VIA VLI VL810 is the first USB 3.0 HUB allowing 4 USB 3.0 devices to be used on one controller without interfering with each other’s bandwidth. Each VLI810 has its own BIOS as well! This motherboard has two of these hubs. All the blue ports on the back panel are USB 3.0, there are also 2 internal USB 3.0 headers. That is 4 BIOS chips for just the USB!
    (Notice the BIOS chip for the hub on the left, the hub on the right has a BIOS chip as well, but it is unusually placed underneath the board.)
    Let’s move on to USB 3x power and ON/OFF charge. Each USB port is protected by low resistance fuses, these fuses are what allow the flow of more power to the USB ports. This implementation is very simple, replace your high resistance fuses with low resistance fuses, the problem is the price. For USB 2.0 the standard power output needs to be at least 500mv per port, GIGABYTE has increased that to 1,500mv per USB 2.0 Port. For USB 3.0 the standard is at least 900mv, GIGABYTE increased that to 2,700mv. GIGABYTE also has a technology called ON/OFF charge that allows you to charge portable electronics that charge through USB(iPod, iPad, iPhone) even while the computer is turned off or in power saving state.
    USB3.0 TURBO!!!!
    USB3.0 turbo mode allows you to add extra bandwidth to the NEC USB 3.0 controllers directly from the processor’s PCI-E bandwidth, so when enabled USB3.0 performance jumps 10% when all ports are being used on the PCH. There is a disclaimer, this USB Turbo mode will eat some of the PCI-E slots bandwidth, but you should be fine never the less.

    SATA 6GB/s & SATA3GB/s:
    (Two black ports on left are SATA3GB/s, the white ports are native SATA6GB/s, and the black ports on the right are Marvell SATA6GB/s)
    Let’s shift our attention over to SATA6GB/s or as many know it SATA III. I will call it SATA6G as to not confuse you with SATA3GB/s which is SATAII, can you see why it’s confusing? So the standard PCH (Platform Controller Hub) supplies us with 2x SATA6G ports(the white ones) natively. Intel always does a fantastic job in this section, so I am going to go out on a limb here and say that you should use this controller first for SATA6G because the Marvell isn’t on-die, it’s on-chip, and thus needs to go through channels before getting to the DMI. GIGABYTE does include more SATA6G connectivity, even eSATA ports that use the Marvell SE9128.
    One Marvell SE9128 is near the two black SATA6G ports that it supports connectivity too, along with its BIOS it sits there just in case you need it. Then in a very surprising place I found the second Marvell SE9128 along with BIOS, right behind the back panel, right next to the 24 phase power!! This was quite a surprise, I thought it was a PWM controller at first, but upon close examination it was none other than a good ‘ole 9128. Take count, 4USB+2SATA6G BIOS chips.

    IEEE 1394a Sub-System:
    Here we see the typical Texas Instruments supplied IEEE from the TSB43AB23 chip that is commonly found on motherboards, this chip can provide support for 3 IEEE 1394a devices.

    Audio Sub-System:
    Audio: The Realtek ALC889 is the high definition audio controller. Common for on-board audio, it has 6.1 channel capability as well as S/PDIF In/Out. This controller features Dolby Surround sound. I have had experience with this audio codec on past GIGABYTE boards; it’s not bad for onboard sound. Rated at 109dB. Support for flawless Blu-ray playback is part of the package.
    Last edited by sin0822; 12-14-2010 at 06:31 AM.

  3. #3
    Join Date
    Sep 2010
    Norther Virginia

    Default Re: P67A-UD7 Most In-Depth Preview/Review!!

    LAN Sub-System:
    Dual Ethernet: This board is equipped with Dual RJ45 Gigabit Ethernet ports, 2x Realtek RTL8111E chips power this feature, and provide excellent Ethernet capabilities. Most boards in this price range have this capability, but cheaper boards do not.
    Now you might ask, why do you need two Ethernet ports?
    The answer is teaming and bridging. The console for the Ethernet has the ability to team up the ports so that they can act as one port by combining two connections, so you can theoretically double your download and upload speeds on the computer’s side. Let’s say you have one port operating at 100mb/s, you can plug in another connection to your router and have it teamed to 200mb/s, and the connection shows up as a virtual 3rd port. Bridging is a bit different, let’s say
    you only have one port left on your router, yet 2 computers that need internet. You can bridge the ports so that one port can connect to the router and the other go to your second computer, sort of like a y-splitter for your Ethernet. Another thing you can do with bridging is control the flow of information over the outbound Ethernet cable, for instance if you are on an intricate intranet network and want to limit certain things.
    Also if one port goes out and dies, then you have another that automatically kicks in.

    I would like to take a second and mention the iTE IT8728F which is one heck of a Super I/O chip, I have personally had it on my other GIGABYTE boards and it is amazingly accurate at voltage reading. It communicates with the PCH P67 through an LPC bus. Voltages measured with a digital multi meter were within 1-3% of windows readings, it wasn’t even worth it to make voltage read points on this motherboard. Hopefully the same precision and accuracy will apply to this platform; there is no reason it shouldn’t. Other manufacturers use Windbond chips, ASUS has been using Windbond for over 8 years now, but this iTE chip is just fantastic, I have never had software read voltages so precisely. I think GIGABYTE wanted to carry that over from their previous boards. I need to also mention that this Super I/O Chip controllers the PS/2 Keyboard and Mouse port. GIGABYTE offered a single port on the back panel for a PS2 device, either a mouse or keyboard will work (its colored green/purple). I think that because this chip is so accurate, that GIGABYTE saw no need to include voltage read points. Nevertheless I will have voltage read points for you guys when I get a CPU .

    The BIOS!!!!
    Dual BIOS support is a great feature to have any board. GIGABYTE has been using this for a few years now, as have other manufacturers. When the main BIOS goes out the backup comes right on, and that is just fantastic for mishaps when flashing BIOS. Flashing BIOS should be done through USB drive and BIOS, not through windows. Just to make sure everything is nice and dandy, GIGABYTE made sure that users cannot access the backup BIOS, so you can never corrupt it! This makes it a great backup. There are two other things I need to mention here. #1 This BIOS has support for booting to 3+ TB drives that wasn’t possible in the past without partitioning. #2 GIGABYTE has a program called Smart 6; one very unique and spy-like feature is its ability to store passwords and dates/notifications in both the main and backup BIOS. To gain access you need the password. If your hard drive fails then you have a backup, same as if your main BIOS fails, you still have a backup. If someone goes snooping through your HDD/SSD they will never find your passwords/dates!!! What a cool feature that you only see in the movies.
    ALERT! I want to mention that each BIOS is a 32Mbit Flash Award bios, not 16Mbit like in the past, this means future support for EFI, as easy as flashing the BIOS!!! So ASUS has EFI and GIGABYTE has support for EFI. I like the fact that they give you the option, because right now we don’t know how stable EFI is for overclocking or in general!
    Preview of some BIOS settings:
    #1 Users will have the ability to set upper wattage limit on the CPU in turbo mode, to reduce heat and set a cap on how hot it runs.
    #2 Users will have the ability to set the limit on the number of amps pushed through the processor from the VRM, amperage is very important as is it directly related to power output and wattage. In theory you can raise amperage to a set limit and voltage to a set limit and calculate total wattage through P=IV Power (wattage)=Current(amperage) x Voltage.
    #3 Users now have the ability to control voltage in increments of 0.00625V.
    #4 Turbo USB 3.0 is an option through bios.
    #5 Control ACPI LEDs through bios.
    #6 Voltages you have control over include: 3 Levels of LLC (excellent carry over from previous generations), CPU Vcore, DVID, QPI/VTT, System Agent, PCH Core, CPU PLL, Dram, and Dram ref and termination.
    #7 Direct control over turbo multipliers on each core.
    One of the most useful tools a motherboard can have is a POST LED which goes through all the steps for booting. It will flash alpha numeric codes that show exactly what the POST cycle is going through at that exact time. When there is an error this LED will hang on a certain code, and in the back of the manual is a 3 page list of the POST error codes with descriptions of each. They are a standard set of codes for AWARD bios.

    The Brand New Cooling Solution Put to The Test!
    Motherboard Cooling:
    I have come too really like this new cooling design on this motherboard. Not only is it a break from the traditional heatsinks with fins, it is also very chic. Chic is an odd word to use for something like heatsinks, but I’m not just referring to the heat sinks, this new cooling apparatus complements the new motherboard design very well. The motherboard is very well dressed with 4 main heatsinks that are attached by no less than 3 black copper heat pipes. They are clamped down very well to the heatsinks. The bottom of the heatsinks is lapped, good for thermal contact.
    The whole system does it job as if it were the employee of the month. Not only did GIGABYTE use aluminum blocks, and have them precisely molded to resemble a radiator type design, but they also put gold accents on the heatsinks that basically sells the board. The seal between the heatsinks and the ICs is very tight, and should not be disturbed, so I took it apart to show you what it is made of. Please do not take the heatsinks off and apply your own thermal paste, there is no point because GIGABYTE put a lot of effort into the whole design, from thermal paste to thermal pads that do not need to be replaced.

    These gold accents have an eye-catching design, I don’t know who designed the cooling for this board but I would love to send them a thank you card. I would like to mention that this cooling solution makes the board very heavy, as the heatsinks alone carry a lot of weight, but that is a good thing because they can hold more heat, and with proper airflow disperse it.
    The Mosfet Drivers are cooled by two rows of heatsinks, one on the left carries the number 7 in gold, and the other on top is smaller and doesn’t carry any lettering.
    The PCH is cooled by its own low profile block (to not block long PCI-E GPUs) of aluminum and carries the GIGABYTE name in gold. The main heatsink on the NF200 sits where the Northbridge or IOH would be on traditional boards, it has Ultra Durable written on top, and has an angled cut on its underside to allow the Low RSD Mosfets and Ferrite Core Chokes to sit undisturbed and get some airflow.
    I would also like to mention that the heatsinks are held together by screws that make it easier to produce the same basic anodized silver/gray heatsinks and then add gold accents for the UD7 and blue for the UD5.
    Moving on to one of my gripes with many motherboards, their use of plastic push-pins to hold down heatsinks. A round of applause for GIGABYTE would be in order because they used 10 metal screws with springs and black rubber O-rings, a perfect design to compliment a perfect cooling apparatus. This is an improvement on the GA-X58A series which used two screws and 10 plastic push pins.
    When I first saw the cooling design, I thought to myself, is this really going to cool the components that well? I decided to put the cooling to the test. I had to devise a way to show performance without using the PCH, NF200, or mosfets to heat up the cooling apparatus, so I did the next best thing, I took apart my array of 10W 10omh resistors and used a few to heat up the heatsink. Artificial testing here we come!
    I know this picture is scary but no worries I am not modding the motherboard, I just felt like scaring you! The motherboard is probably perfect and doesn’t need any voltage modifications. I am using everything in the picture, including three digital temperature probes. They are pretty cheap but do the job well, my thermocouple was down, and so I used these thermal probes instead; they do the job very well. As you can see in the next picture both meters together show the same exact temperature which is ambient temp (it’s hot in here). Each meter has 3 probes attached to it and is wired to use each when switched to the certain probe. I will use only 3 probes for this experiment, as the other 3 are already integrated into my main pc.

    Experiment #1 Heat dispersion:
    This test shows me how well heat is dispersed from one heatsink to another as well as from the bottom of the heatsink to the top. One temperature probe for the resistor is placed between the resistor and the NF200 heatsink, another temperature probe is placed on top of NF200 heatsink, and then the last probe is placed on the bottom of the PCH heatsink to see how well the heat pipe transfers heat. Then I moved the probe from the top of the NF200 heatsink over to the top of the PCH Heatsink. Two 10 watt resistors are used, they get very hot to the touch and cannot be touched, they are powered by 12v with no limit on amperage, and they are held together by their soldering, then I add an extra bit of thermal past to the lapped surface of the NF200 heatsink, and tie them down with zip ties, the probe is in-between the resistors and HS.

    Test #1 results:

    As you can see the difference between the bottom of NF200 HS/top of resistor is a few degrees difference, about 3C in the beginning. When the heatsink starts to heat up that difference increases a small amount. The fact that it is within 10C is amazing, most high-end air coolers have a problem keeping within 10 of surface temp of the chip, but I believe because these blocks are one solid piece they are able to transfer heat much more effectively. When we move over to the PCH, which is heated through the heatpipe we can see heat exchange is very good, in the beginning there is a 3-6C difference. Once heat gets up to a point it seems that the vaporization that takes place inside the copper heatpipe transfers heat much more effectively when heated. We see the 6C difference decrease to only a few degrees, this indicated that the clamp and connection between aluminum block and copper heatpipe was done correctly. In the final moments of the test, I moved the probe from the top of the NF200 heatsink to the top of the PCH heatsink. While the heat is less on the PCH, the transfer of heat is about the same as from bottom of the NF200 HS to the top of the NF200 heatsink, this is most likely due to the fact that both are solid blocks and specific heat is the same for both blocks, so heat transfer would be about the same which we see is the case.

    Experiment #2: Heat Transfer. In this experiment I will add 3 more 10 watt resistors to the mix, and zip-tie them down to the mosfet heatsinks. I will use the stock thermal pad that is already there, because I want to test the system as is. In the past GIGABYTE used very thin heat pads, as is the case here. One thermometer is placed on the bottom of the NB block as usual, then one is placed on the PCH HS on top, and the third is placed on top of the mosfets HS on top. In the beginning I will let the system heat up to 50C (These resistors go up to 70C when on the original HS I had them on before testing), and then place a fan (silent 1500 rpm fan), to mimic case airflow. Please do understand that I am generating 50 watts, the NF200 is supposedly rated at 12W and the driver mosfets will be about 1 watt (at 5 amps a piece(24x5=120amps) which is more than enough) each under load conditions with all 24 activated, so that is 36 watts + PCH which I do not know. This 50 watt array of resistors is to mimic a heavily overclocked system, please take into consideration that not all 50 watts will be concentrated in the areas which they are in this test.
    Test #2 results:

    The system reacted as I thought it would, temperature quickly increases to 50C in under 90secs, the PCH’s heatsink really increased heat, but not as much as I expected, I was thinking that the increased heat from the mosfet resistors would add more heat to the resistor free PCH heatsink, but I was very surprised. The large amount of aluminum used for these heatsinks absorbs a good amount of heat, this then heats up the copper heatpipe and heat starts to transfer. In this case PCH temperature was only 2-3C higher than expected, but it is heat transfer none the less. When the fan as applied, it cooled down the heatsinks, but it took a few minutes, this is due to the heat capacity of the aluminum used for the heatsinks. The temperatures did drop a fair bit, which I did not expect because these resistors putout constant heat. The heat dissipated rather slowly but in a computer case this would be just fine because the constant airflow would be enough to cool the system.
    The results from both tests happily surprised me. The new heatsinks do their job flawlessly and provide added elegance to the board. Performance under heavy load and even sub-zero benchmarks should be excellent. This board’s cooling is really one of a kind, and it looks like GIGABYTE put a lot of money into R&D for this specific design.

    Conclusion & Analysis:
    The new LGA1155 platform has a lot to offer the overclocking and enthusiast community, from great implemenations of the old, and ever expanding features of the new, GIGABYTE products always pack that hard punch. From the great SLI design where peripherals do not eat up SLI bandwidth, to the huge tree of USB 3.0 devices, SLI and Connectivity have taken a turn for the better. No longer will your SATA6G SSD be slowed down because of your SLI, and vise versa.
    Beware that other motherboards might not implment these features such as this GIGABYTE board because there are many ways to route PCI-E bandwidth through the board, and many of those ways are much easier to implement than this. I wasn’t joking when I said that the engineers put a lot of effort into making the two PCI-E buses separate, that took a lot of effort and R&D. Power delivery is one of my favorite topics, after looking over some of the BIOS features such as amperage and wattage control, it is even more evident to me that power delivery is going to be playing a bigger role in overclocking the CPU than ever before.
    The reason I like this board so much is that its power delivery is perfect, responsiveness is super fast, and power output would in theory power 10 x i7 45nm 900 series processors (all at full load at stock), even though we are going to be dealing with single 32nm chips that run much cooler and proabably have less power leakage. Although this board doesn’t use digital PWMs, it is for good reason because Digital PWMs are in their infant stages, and do not offer high switching frequencies, fast reponsivness, and reliability of their Analogue counterpart.
    These new LGA1155 platforms have to be VRD12 compliant and that means implementation of Serial VID, which in turn means the VRM PWM and CPU PCU will be communicating with eachother at very high speeds continuesly. I do not know how well a digital PWN can keep up with this type of technology as it needs to decode and then re-encode to analogue signals which is time consuming. I can tell you this, GIGABYTE’s implmentation of the P67 PCH and NF200 along with power delivery is just fantastic. GIGABYTE has changed their PCB design and I really like it, and I don’t think I am alone. GIGABYTE delivered with a great black matte PCB along with a great cooling solution for the motherboard, they have improved on their older design without a doubt. I am looking forward to NDA lift so I can show you what this beast has to offer! GIGABYTE does deserve to have their name in gold on this board.
    I would like to thank GIGABYTE for making this preview/review of this motherboard happen! If you have any questions for comments please feel free to send me a private message!
    Last edited by sin0822; 12-13-2010 at 09:38 AM.

  4. #4
    Join Date
    Sep 2010
    Norther Virginia

    Default Re: P67A-UD7 Most In-Depth Preview/Review!!

    Printable PDF added by request: Need printable PDF?? Here you go!:File name: P67AUD7Preview.pdf File size: 1.98 MB

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts