MSI Z77 M-Power Big Bang - Testing with CL3P20

In this thread, I will be covering the highlights of the new M-power series Z77 Big Bang; covering the various traits of the board with various cooling methods and voltage tests.

For those of you not already familiar - Lets get acquainted!! {not your full run of the mill photos.. just some of the choice areas}

*unboxing - Certified pre-tested

Fan headers anyone..along with TPM and a special front-mount panel, its got bells and whistles; Check!

Naked Phases! OEM cooling removed for closer inspection..looking good MSI !

vDIMM mods arent needed.. but IC is here; to the upper left of the fan header.

uP1618A .. will do just fine for vcore 

Pretty standard SATA layout..good spacing between the bottom of the RAM slots and the first PCI-E !! Leaves room to have caps and mods on the back of your first GPU's PCB without interfering.

 Testing Platform:

Intel i7 3770k

2x4 Patriot D2's : 9-11-9-27 1T @ 2133mhz ; 1.65v

2x4 Team : 9-11-11-27 1T @ 2133mhz ; 1.5v

MSI TFII 5770's and TFIII 6950's

Antec TPQ 2.0 1200W

Cooling - Intel 980x stock cooler, OCN Marksman

 Initial impressions:

 Voltage components and layout on the PCB are very good. Premium inductors, mosfets and caps ensure that voltage delivery will be stout and clean [good start]. Droop options in the BIOS really tighten things up without being overbearing with unnecessary options. Using the Switching Phase options increases power consumption and temps a bit around the CPU area.. nothing major though. On an initial inspection/run through with the mobo, you can see the same voltage controller from many other MSI boards [the uP1618A]. Now for most Ivy Bridge enthusiasts.. +1.7vcore is simply not needed ; and only extreme users need worry about power delivery and characteristics of the circuitry when pushing more than 1.8vcore..  but we need a cold board and LN2 to properly test in that vcore range.

More to come.. Results on Air and DICE ahead.

shortcut to latest DICE results with 3770k and OCN Marksman

Testing on Air - stock cooler:

32mil is by far my favorite bench to tune for..and since I was unfamiliar with this set of Patriot Memory, I spent some time dialing in an average speed with timings as tight as the RAM was comfortable with.

vcore - 1.10v set = 1.113v real @ idle, 1.115v real @ load

vcci/o - 1.13v set = 1.16v real @ idle, 1.114v real @ load

DRAMv - 1.70v set = 1.71v real

PLLv - 1.81v set = 1.81v real

System Agent v - 1.02v set [no measurement location found yet]

PCHv - 1.06v set [no measurement location found yet]

CPU- 100*40

DRAM- 200*10

CL3P's_MicroXP 32m

Timings from above run:

Quick test of 3D'03: CPU stock Intel cooler / GPU's stock MSI TFII coolers

Catalyst 12.1 drivers

Win7 64 Ultimate

Quick change multi to 43x and vcore to 1.15

Pic of test rig on Air:

on to some cold today...despite ambient temp being a toasty 36c !! 

OCN Marksman Pot / Dry Ice

Dressed for cold weather: 10lbs of DICE and 6hrs of benching

Best runs from first session below:

bench run with 3770k @ 54x100, 1.52vcore set = 1.49vcore real [no other voltage adjustments]

RAM @ 200*10, 1.68v : 9-11-9-27 1T

CPU temp = -70.3c idle / 68.1c load with 4c/4t

32mil - 5.4ghz

3d '03

3d'05 - need to re-run with proper OS.. took 5th by mistake : http://hwbot.org/submission/2315694_

 ** Notes from -0 runs: 

 Mobo has a more significant droop after ~1.48vcore set [droop is >.3v from set to real voltage]. Although its still very stable with the supplied voltage..it would be nice if the Switching Phase Frequency was adjustable more to allow for compensation at these higher voltage ranges. With all power options set to "100%", the board does a great job keeping vdroop in the CPU circuit below .2v with the 3770k, all the way up to the plus side of 1.45vcore. At first I though it was amperage draw from the increased speed.. so I lowered CPU to stock multi and left vcore at 1.52v in BIOS.. I witnessed the same reading in windows after boot, of 1.49v. I do not think this 'loss' was heat related as PCB temp was well below 0c still too.

RAM and VTT circuits over-volt around .15v from BIOS settings.. just FYI 

 More to come!!

**May have to take you up on that some weekend 

 -Got into pushing RAM freq's a bit yesterday with this Patriot kit.. managed to boot 2400mhz 10-12-12-30 @ 1.69v .. was not even 1mil stable though regardless of volts or timings on CAS 10. Didnt bother to test CAS 11 .. think Id rather buy new RAM before I waste my time with CAS 11 testing .

Results on air for 4ghz efficiency test: 

vcore - 1.10v set = 1.113v real @ idle, 1.115v real @ load

vcci/o - 1.13v set = 1.16v real @ idle, 1.114v real @ load

DRAMv - 1.70v set = 1.71v real

PLLv - 1.81v set = 1.81v real

System Agent v - 1.02v set [no measurement location found yet]

PCHv - 1.066v set [no measurement location found yet]

CPU- 105.3*38

DRAM- 210.6*11

CL3P's_MicroXP 32m

Nothing amazing in a 1sec reduction in time for this test..moving tRCD and CAS killed much of what should have been a latency decrease for another 316mhz of RAM speed. Full timing set listed below:

 Definitely will be trying to bring these down some with cold next round.. ambients were again near 40c, didnt want to push too hard.

**I do still have the sinks removed.. so mofets are running with no sink .. just airflow. Even with it being hot out..the mosfets were doing just fine 'naked'. I dont have an IR gun.. but to the touch they were warm at best and couldn't burn a babies arse. 

More freezing .. soon.

Quote:

Originally Posted by Redwoodz 

Interesting location for the power/reset/ocgenie buttons. Nice start,anxiously awaiting further results!

I think its a '3x slot thing'... as it would cover anything on the bottom end of the mobo.. so on the side by the RAM is the only accessible area left. Pretty sure Asus moved all of theirs to this position after so many folks complained about the buttons being useless on the bottom with 3x/4x way GPU setups.

 **Will somebody build a movable reflector for the debug panel though? Would be nice just to twist the top of it like a parascope and have it point at me..almost have to stand with the board oriented like I do; to see the debug on boot. . Ok ..thats my one complaint so far. 

Vcore limit is a 'soft limit' gappo..  nothing on this mobo is going to prevent you from setting higher than ~1.8v.. you would be in dangerous territory if you did though. Spec for the IC is 1.8v max..I will do some air testing with better ambients to show the increase in droop as vcore starts to climb.

**Also - seems like sinks on the mobo keep quite a bit of heat "in" .. removed them from the board and Fet's are barely warm to the touch with 4.4ghz at full load, with PCB at room temp.

trying to work on some dewars I found for cheap.LN is definitely in order.. just unsure when right now.

**As for the vcore - Let me just say that you can set higher vcore than 1.8v.. but personally - this is where I have had boards with this IC controller fail. Efficiency starts going down as you approach 1.8vcore.. the droop in the circuit keeps getting marginally worse.

ie 1.6v set = 1.568v real  ... .4v droop

    1.7v set = 1.64v real  ... .5v droop

    1.8v set = 1.73v real ... .6v droop

**newer BIOS tightened up droop in vcore

.. That said, I would not want to push to hard as 1.8v approaches. The last board I blew with that controller took my CPU and PSU with it.. not very anxious to repeat if you know what I mean. This could be a fault of my PSU however.. so I am not discrediting the mobo for this issue yet.

**@ 636cc - I know what you mean.. P55 was the last series that I played with more than one manufacturer's boards. Couldnt hit +6ghz on the Trinergy with the vcore not cooperating over 1.82v set [mobo would not even boot].. Asus Gene took the same CPU to +6.2 with 1.96vcore...