X48 Rampage Formula Preview.

[EOD]

Dont know for certain, but it might be to do with your divider? And FSB Term Voltage, i set 1.40v, my bios/everest report 1.46-147v, i set 1.47v Vnb reports about the same.

I'm about to start considering that my board might be a female, I could swear it's that time of the month

I gotta stop pulling my comp apart because every time I do I end up spending 1/2-1 day fixing problems that I dont know how to fix by moving cables around and playing musical chairs with pci-e power and fan headers till eventually after a while it stops acting like a child that didn't get its toy it wanted and works normally again. Dont know if its my board or not. I would only imagine it is, because the problem it self appears as many fluctuating voltages in hardware monitor, by fluctuating i mean 0.05-0.1v, bouncing around like a pogo stick. Then during the exorcism I perform it does all kinds of evil things like corrupting and wiping my cmos!

Argh I can fix most things but this one has me baffled, it's been on going for 2 - 3 months and as many times as Ive fixed it up I still don't know how exactly I fixed it.

The only thing positive to come out of it is my CPU temps have plummeted from high 50s to high 40s and all I did was replace the thermal paste I replaced recently, and shimmed down my TRUE so it clamps better. Even Linpack temps are amazing now, 65c full load!

I wish I could run my ram at 1120, but 1:1 is all I can manage since they max out at 1088. I should have spent the extra $30 and got the 1100 set.

I've noticed that like CPUs, the X38/X48 NB chips OC differently, and no 2 OC the same. Anyways I am running a 465 FSB (hoping it is a lower chipset strap) with 1.59v NB with the Ai Clock Twister at Strong to see how it goes.

I notice my VTT voltages are not even close to what I set in bios. Both the bios hardware monitor and Everest report it as 1.17v, yet I set it to 1.26v for testing purposes.

And as for the Trans Booster, I cant even boot at 7...


But its great to here you have a nice temp drop with the reseat and shim. What TIM did you switch to?

[mikeyakame]

I wish I could run my ram at 1120, but 1:1 is all I can manage since they max out at 1088. I should have spent the extra $30 and got the 1100 set.

I've noticed that like CPUs, the X38/X48 NB chips OC differently, and no 2 OC the same. Anyways I am running a 465 FSB (hoping it is a lower chipset strap) with 1.59v NB with the Ai Clock Twister at Strong to see how it goes.

I notice my VTT voltages are not even close to what I set in bios. Both the bios hardware monitor and Everest report it as 1.17v, yet I set it to 1.26v for testing purposes.

And as for the Trans Booster, I cant even boot at 7...


But its great to here you have a nice temp drop with the reseat and shim. What TIM did you switch to?

still the same old AS5, redid my NB and SB too cause I hadn't done them in quite some time. I put a couple of nice long plier scratches across the circuit board to add to my collection cause of those god damn push pins! I have to use circlip pliers to remove some of them because normal long nose pliers just end up leaving scratches all over the place everytime they slip from the pin.

I scratched what appeared to be a bunch of traces though on the back of the board, so the whole amount of issues it's had I'm not sure if I just made them worse or what! It's possible, been stability and error testing for past few hours and letting prime95 run a while right now.

These are settings I use anyhow mate.

fsb 481
cpu multi 7
pcie freq 106
strap 333
ddr 1156
dram clk skew cha / chb adv 50 / adv 100
dram timings primary
5-5-5-16-3-65-7-4-2N
dram timings secondary
8-4-5-4-6-4-6
dram timings tertiary
14-5-1-6-6

ai clock twist moderate
dram static read enabled

performance level 7 or 8 (either works just using pl8 at the moment cause
not sure if pl7 is too tight for windows 7 and is corrupting my stuff, voltage I need for 7 or 8 at this fsb isn't much different, 1.47v for 8 vs 1.53v for 7)

no phases pulled in as less than PL7 on this divider is too inconsistent and generally eventually breaks some data

vcc 1.3875v
vcpu_pll 1.56v
vnb 1.47v
vtt 1.40v (1.47v real)
vsb 1.10v
vsb_pll 1.50v
vdimm 2.0v

cpu gtl 0.63
nb gtl 0.63
dram controller ref -10mv
dram ref cha ddr2_ref
dram ref chb ddr2_ref

loadline enabled


i think thats all the settings heh, thats just top of my head if I missed any important ones let me know!

my gskill 1100s the powerchip ics top out at around 1170mhz and 2.04v or so, anymore voltage or looser timings doesn't help. also I do remember trying to get 1:1 divider to POST at 490mhz fsb and I gave up at 1.61v Vnb so there is probably nothing wrong with yours needing so much for 1:1. Most I've tried was 1.65v on 5:4 divider to Pull in phases on PL6 to PL5. PL5 no post what so ever. 1.57 or 1.59v got me PL6 pha1/phb1 pulled in and 50.1ns latency or so at around 470ish fsb on some 4x1gb sticks I had that I managed to somehow get to clock up to 1190mhz with some crazy skews and loosening some of the precharge timings and burst ones with 2.40v

[Alien Grey]

I dont know how A-Grey and Mikeyakame are able to run such a low NB voltage for their FSB OCs.

If you think that 1.41V NB is low for FSB 465MHz. Wait till you read the following.

With BIOS 0410 and FSB 450MHz I only needed 1.25V in the BIOS for the NB to run IntelBurnTest stable and 1.27V in the BIOS for the NB to run Prime95+UT2004 stable.

If you don't believe me look here.

I think I know your problem. You're running 8 Giga memory and that puts a lot of strain on your NB. That's why I stay with 4 Giga. There is no program that I use that needs more than 4 Giga.

[mikeyakame]

no kidding 8gb + quad core + 1:1 + high fsb == blow up? heh well no but it's way too much strain on the NB I see why you use 1:1 even though it strains the NB, it's much less than the equivalent of using 12:10 would be. 1:1 only has a single phase to worry about, 12:10 has 5 mch and 6 dram phases to time.

[EOD]

With my previous setup I was running a Q6700 at 3.8GHz at 423 FSB x 9 and PL6 . I was running 1.45v NB and 1.8v vdimm. When I moved to 4x2Gb I only had to increase my NB to 1.49v and vdimm to 1.94v.

Here are some of my screen shots of settings taken from my thread at overclock (when I was running 2x2Gb):
THE Asus P5E/Maximus/Rampage Formula X38/X48 Thread

I've been considering getting rid of my 4x2Gb set of 1066 and just ordering the 1100 set and running the 6:5 divider.

[Zucker2k]

Start over... the key is to keep the NB cool.

Asus Rampage Formula x48, bios 0308
C2Q Q9550, C1 @ 8.5 x 471 (4Ghz)
4x 2GB G.Skill F2-7200CLD4, 8GB DDR2 942mhz Cas 4, PL7
WC, both chipsets watercooled.
2x 2-disk raid0
Windows Server 2008 SP2 beta (Regular OS)






8x450 fsb with nb and fsbt set to minimum 4GB ram @ DDR2 1080, PL7.

[mikeyakame]

Pretty impressiver Zuck! My settings broke last night after stripping the board down and redoing all thermal paste. Been fiddling about trying to figure out what the hell happened, but finally current Prime run is almost 2 hours, last 1 failed after 52mins. Seems whatever happened I had to lower Vtt and make some adjustments to get it stable at the same FSB.

[Zucker2k]

Pretty impressiver Zuck! My settings broke last night after stripping the board down and redoing all thermal paste. Been fiddling about trying to figure out what the hell happened, but finally current Prime run is almost 2 hours, last 1 failed after 52mins. Seems whatever happened I had to lower Vtt and make some adjustments to get it stable at the same FSB.

Yeah, stuff can get very sensitive; best way to get rid of the tantrums is to power off, better yet, pull the plugs for a minute, then power back back on. I really hope asus comes through with those bios enhancements; this board is a beast.

Edit: Anyone know what the voltage distribution is across the 4 slots?

[mikeyakame]

You know the funny bit though, my temps at same 3.38ghz clock were roughly 59-60c during prime runs and 75c during linpack before this, so it seems shimming the TRUE mount bracket with a couple of 1.2mm washers on each bolt and putting on new AS5 is making the world of difference. Cores are within 1c of each other as I sanded out the concave in the heat spreader while back but never really noticed it making as much difference as now, even the CPU temp is within 5-6c of actual reported core temps, its usually 15-20c even full load. Prime after nearly 3hrs now, cores are still 49-50c, and during a 10 min linpack 64bit run all cores never go past 65c, avg between 55-60.

I'd like to think that the TRUE mount not applying enough pressure and being loose was the problem, and not that I'm applying too much presure now and distorting the DTS values but I think If I was it wouldnt be stable as it seems to be. I guess I should be happy, always figured it was capable of more than I was seeing with a Delta 120x38 130cfm fan pulling and a Tuniq 120x25 90cfm pushing.

[ElAguila]

I consider myself very luck to get my 4x2GB of ram to run at 1116MHz with performance level 7 and be stable. I have backed my cpu down to 4020MHz for my 24/7 activity. In order to get stable at this fsb and with my ram at that speed my NB voltage is at 1.5v. Is that too much? I put a fan on the heatsink and replaced the thermal paste and my NB idle temp is 45C and the load temp is 51C. Are these temps okay for the NB?

[mikeyakame]

yeah mine used to be in the 30s but I guess long term voltage slowly takes a toll on it. Mines between 45-55c these days with 1.47-1.49v its no biggie really, just dont get above 60c normal op with such tight PL.

[Alien Grey]

I consider myself very luck to get my 4x2GB of ram to run at 1116MHz with performance level 7 and be stable. I have backed my cpu down to 4020MHz for my 24/7 activity. In order to get stable at this fsb and with my ram at that speed my NB voltage is at 1.5v. Is that too much? I put a fan on the heatsink and replaced the thermal paste and my NB idle temp is 45C and the load temp is 51C. Are these temps okay for the NB?

I don't think that 1.5V for the NB is bad with 4 X 2 GB at 1116MHz. I'm pretty impressed that you only need 1.5V to keep that stable.

[spoof]

Well, as i turns out - CPU can do more, MB can do more it's just a matter of time, patience and a lot of testing. So far, 460FSB owned! And voltages aren't that high as i expected at all. 1.37 vCPU, 1.58 PLL, 1.38 VTT, 1.45 vNB, PL8, 63x for CPU and 67x for NB GTLs. I hope, that with minor changes all be able to get stable on 465 too. Now i'll do full stability test just to be sure, thaht 460 is 24/7-ready and my next goal - 470 and more.
And yes, quad, high FSB and 4x2 RAM is really hard for the chipset, so high values for vNB is a must. After i got rid of lockups, memtest errors were fixed by increasing vNB.

mikeyakame, this fluctuation that you are talking about is way to much? I didn't understand what voltage do this? I think you have to send a message to ASUS, because 0.1 pikes can really do some serious damage. Try get rid of this LLC, for my it causes instability and since this board doesn't have any vdrop the only thing, what LLC do is to higher a bit you actuall CPU voltage. Maybe it have something to do with, don't know.

Question: What is this program, that Zucker2k use to monitor voltages and temps?

[Alien Grey]

LLC that's something i've never got a straight answer for form the Tech Support.
If I enable LLC in the BIOS it's going to end up with a system crash. The latest BIOS updates just made it worse.

My question to them was what it's supposed to do. I asked them if my CPU is stable at full load with CPU Voltage 1.24V with IntelBurnTest, should I be able to run it with LLC enabled with CPU Voltage 1.24V on load and idle?
This isn't possible with my system. I see no reason why I should use it when it can't do that.

The program that Zucker2k is using is ASUS PC Probe II that you can find on the ASUS support CD that came with your board.

[Alien Grey]

I notice my VTT voltages are not even close to what I set in bios. Both the bios hardware monitor and Everest report it as 1.17v, yet I set it to 1.26v for testing purposes.

Same thing here. I need 1.32V for vtt in the BIOS to have 1.22V real.

[Zucker2k]

You know the funny bit though, my temps at same 3.38ghz clock were roughly 59-60c during prime runs and 75c during linpack before this, so it seems shimming the TRUE mount bracket with a couple of 1.2mm washers on each bolt and putting on new AS5 is making the world of difference. Cores are within 1c of each other as I sanded out the concave in the heat spreader while back but never really noticed it making as much difference as now, even the CPU temp is within 5-6c of actual reported core temps, its usually 15-20c even full load. Prime after nearly 3hrs now, cores are still 49-50c, and during a 10 min linpack 64bit run all cores never go past 65c, avg between 55-60.

I'd like to think that the TRUE mount not applying enough pressure and being loose was the problem, and not that I'm applying too much presure now and distorting the DTS values but I think If I was it wouldnt be stable as it seems to be. I guess I should be happy, always figured it was capable of more than I was seeing with a Delta 120x38 130cfm fan pulling and a Tuniq 120x25 90cfm pushing.

Working my way up, but ram is limiting me a lot. Linpack packs the heat;

[greg.m]

Working my way up, but ram is limiting me a lot. Linpack packs the heat;

How is it possible to run your cpu in that freq with such a low volts?
I run my qx9650 prime stable @ 4050 (450x9) with at least 1.575bios voltage.
Here is my settings:

450x9
1080 my 2x2gb dominator 5-5-5-15 pl7
bios volts:
cpu 1.575
cpupll 1.5
nb 1.49
vtt 1.4
vram 2.08
my real cpuvolts under stress are 1.46-1.47 and idle 1.54
Whatever ive tried it didnt help. The only way to run stable @ 4050 is to
raise my cpuv.
Did i have such a bad luck with my f***in qx9650 or is there a problem
with my rampage?

[Zucker2k]

How is it possible to run your cpu in that freq with such a low volts?
I run my qx9650 prime stable @ 4050 (450x9) with at least 1.575bios voltage.
Here is my settings:

450x9
1080 my 2x2gb dominator 5-5-5-15 pl7
bios volts:
cpu 1.575
cpupll 1.5
nb 1.49
vtt 1.4
vram 2.08
my real cpuvolts under stress are 1.46-1.47 and idle 1.54
Whatever ive tried it didnt help. The only way to run stable @ 4050 is to
raise my cpuv.
Did i have such a bad luck with my f***in qx9650 or is there a problem
with my rampage?

Sell your qx for $350 and pick a Q9650 and you'll be a happy chap.

[EOD]

Start over... the key is to keep the NB cool.

Ok, I am taking your advice and starting over. I removed 2x2Gb of ram, cleared cmos, and just set some key voltages. Right now I am running 1.45v NB with a 445. The VTT is set to 1.26v (1.17v actual), PLL is 1.5v, vcore is 1.30v and all else is on auto. Before when I couldnt even boot load windows with less then 1.55v, it is now priming. But my PL is at 11

So I'm looking into a nice 2x2Gb 1120+ DDR2 ram kit. I was thinking of getting the G.Skill DDR2 1100 set.

[mikeyakame]

raise your vtt to 1.30-1.32v actual. you are using way too low of a vtt, pll should be about 1.58-1.60v real too or else it'll conflict with sb_pll voltage, manually set sb_pll to 1.5v so it reads 1.50-1.52v.

[EOD]

raise your vtt to 1.30-1.32v actual. you are using way too low of a vtt, pll should be about 1.58-1.60v real too or else it'll conflict with sb_pll voltage, manually set sb_pll to 1.5v so it reads 1.50-1.52v.

When I was running VTT at 1.3v I needed more NB voltage to keep it stable. But after dropping the VTT I was able to lower the NB voltage and maintain stability. With the PLL set to 1.5v, the actual is 1.52v, so I am right on the mark there. After this prime95 run, I will definitely try your recommendations. I am about 1.5hrs into Prime95 right now (seems stable) and have the following actual voltages:

[mikeyakame]

I've personally found when getting into higher fsb ranges keeping as many voltages apart from each other as you can will help with stability of the whole system. It just helps minimize the likelihood and severity of cross chatter across different clock domains (ie, pci clock, hd audio clock, usb clock, pcie clock, etc) during address strobes.

[EOD]

I've personally found when getting into higher fsb ranges keeping as many voltages apart from each other as you can will help with stability of the whole system. It just helps minimize the likelihood and severity of cross chatter across different clock domains (ie, pci clock, hd audio clock, usb clock, pcie clock, etc) during address strobes.

Thank you very much for the info. I will be pushing the FSB up to 467 and this will really help. BTW, I just ordered the 2x2Gb G.Skill 1100 kit to run 5:6 at 1120. Running 1:1 is killing my PL. I did some initial testing at 900 and I couldnt even run PL8, but once I set the ram to 1080 with 5:6, I was able to run PL7.

[mikeyakame]

yeah you'll find thats probably the case, 1:1 divider on the RF isn't its forte. 12:10 is its money divider for quad cores. It plays nicest and more consistently on that. My findings conclude that anyhow. I think you'll find most other guys here run 12:10 as well when going over 460-470mhz fsb.

My Vtt is jumping around in a 0.02v range, not a good sign. I think i'll be RMA'ing my board as soon as I can pick up Rampage II Extreme locally to drop in my system. Don't really want to go out and buy a crappy temp board when I can use the money for better purpose of upgrading. While the problem isn't critical it seems the VR is being temperamental and thats a good sign it's probably time to RMA.

[SoLoR]

Is there something wrong with "Read to Precharge" option? When I set it 1 in bios, it reads 4 with memset, when i set it to 6, it reads 8 in memset... Also is it normal that higher tREF values gives better results? If so, would be nice to have it in bios....

for example: tREF 2080T


tREF 16383T: