680i SLI is too expensive? Asus as the solution : P5N32-E SLI PLUS

[Jultsu]

lesterLL:

Bear in mind, though, that I am running my rig OUTSIDE on the balcony, where it's currently 1 degrees celcius. That 20 degree temperature difference makes a huge difference. But nevertheless, 450fsb is quite nice with this board.

[Radical_53]

That's one hell of a result for this board, really! Even if it's outside. But you should hurry, otherwise spring starts and you'll never get the chance to run it for 24/7 once

The last changes I made weren't all that successful. I did some new backplates for my mosfets, used 5mm aluminium bars this time. Now the board doesn't bend at all Even with some standard passive heatsinks on top, the temps do rise very slowly, vcore doesn't jump around. At least one success.
Then I ran into something really strange. I exchanged my 500W seasonic PSU with a 620W Corsair unit. Reported vcore is higher, all other voltages are lower. And my "normal" clocks aren't stable at all anymore.
I guess I'll need to swap again and see if the PSU really is the bugger here.

[lesterLL]

wow! hmm... but the NB's voltage are very high.. yikes.. i cant imagine the temps.. if running with passive coolers... hmmm.. i think i should buy watercooling just to cool down the NB and SB.. sigh

have you tried any other than 450x9 like anything with a x9 multiplier that boots??

overall though... you're the first one (i saw) to reach more than 3.0ghz with the board + quadl..l

[Snyxxx]

That's one hell of a result for this board, really! Even if it's outside. But you should hurry, otherwise spring starts and you'll never get the chance to run it for 24/7 once

The last changes I made weren't all that successful. I did some new backplates for my mosfets, used 5mm aluminium bars this time. Now the board doesn't bend at all Even with some standard passive heatsinks on top, the temps do rise very slowly, vcore doesn't jump around. At least one success.
Then I ran into something really strange. I exchanged my 500W seasonic PSU with a 620W Corsair unit. Reported vcore is higher, all other voltages are lower. And my "normal" clocks aren't stable at all anymore.
I guess I'll need to swap again and see if the PSU really is the bugger here.

Radical_53: What passive heat sinks did you use on the MOSFETS?

[Jultsu]

wow! hmm... but the NB's voltage are very high.. yikes.. i cant imagine the temps.. if running with passive coolers... hmmm.. i think i should buy watercooling just to cool down the NB and SB.. sigh

have you tried any other than 450x9 like anything with a x9 multiplier that boots??

overall though... you're the first one (i saw) to reach more than 3.0ghz with the board + quadl..l

400fsb works also (with 1.6v NB). I tried upping to 8x455 and 9x455 but it would just freeze. 460fsb gives the POST screen but hangs there.

Maybe I've just got a crappy quad? With a better one, maybe I could break 500fsb...

[lesterLL]

400fsb works also (with 1.6v NB). I tried upping to 8x455 and 9x455 but it would just freeze. 460fsb gives the POST screen but hangs there.

Maybe I've just got a crappy quad? With a better one, maybe I could break 500fsb...

im not sure if you hav a crappy quad... coz you already got it to 4.0ghz...

400x9 that would be 3.6ghz... on 1.6v NB... hmm.. ok... wow.. i am very impressed... really i am.. keep it up... im reading every inch of your progress

[Jultsu]

http://www.aijaa.com/img/b/00072/1406600.jpg

8x450 is rock stable.

[DarkCow]

http://www.aijaa.com/img/b/00072/1406600.jpg

8x450 is rock stable.

So im guessing 8x450 is stable where 9x400 with lower NB voltages wouldn't be as stable?


edit: so is this the pencil mod you did? http://www.xtremesystems.org/forums/...d.php?t=134100

Any information of proper technique on how to do the pencil mod will be appreciated, I bet I can do it fine, but the more info the better. (Just a bit tense because my other volt mod ended up in a broken mobo, but it was a harder mod)

[Radical_53]

@snyxxx: Well, I used some standard black aluminium heatsinks there. They fit the space nicely (3 of them on each "block" of mosfets), I had them here already so I tried them out.
Also, after making a backplate, I used a copper bottom (from a water block) as a base for the passive heatsinks. Gives me a nice contact and works really well so far

Looks somewhat like this:

[Snyxxx]

@snyxxx: Well, I used some standard black aluminium heatsinks there. They fit the space nicely (3 of them on each "block" of mosfets), I had them here already so I tried them out.
Also, after making a backplate, I used a copper bottom (from a water block) as a base for the passive heatsinks. Gives me a nice contact and works really well so far

Looks somewhat like this:

Thanks for th photo. I also have some of those Microcool PLL and other sizes of those heat sinks laying around as well from the old i478 days.

[Jultsu]

So im guessing 8x450 is stable where 9x400 with lower NB voltages wouldn't be as stable?


edit: so is this the pencil mod you did? http://www.xtremesystems.org/forums/...d.php?t=134100

Any information of proper technique on how to do the pencil mod will be appreciated, I bet I can do it fine, but the more info the better. (Just a bit tense because my other volt mod ended up in a broken mobo, but it was a harder mod)

Yep, that's the one. I just penciled the resistor on-the-fly Went back once in a while to check whether i had succeeded or not. Bit by bit, the voltages began to rose. To be honest, it's not completely 0 vdroop, but it's MUCH better than before.

I just used the (hard) pencil I use at school, and basically coloured it (like 100 sweeps). Kinda spooky doing it while running prime95 but everything seems to be working atm


And I'm guessing 400fsb is also rock stable (as is 450) but why settle there?

[lesterLL]

@jultsu
NB 1.744v SB 1.568v HT 1.664v VTT 1.616v --> dang... huhuhu... its a little bit high.. huhuhu... and i only hav the default passive heatsink... any suggestion on what to put on NB and SB??

[Jultsu]

I haven't fiddled with aircooling in a LONG while, so I have no idea what heatsinks are good nowadays. With the default heatpipe cooling, 1.6v is probably too much, unless you run your rig in 0 degrees ambient (like I do )

I can only suggest watercooling, if you really are persistent in overclocking with THIS board. 4 waterblocks will cost about the same as a new Intel p35 board, which probably will overclock to 450fsb easily with a quad.

It's all about choices.

I use MIPS waterblocks, http://mips-computer.de/ and the set of four blocks costs 155 euros (ASUS P5N / Striker Freezer Set)

[DarkCow]

great, thanks a bunch Jultsu. I'll probably OC as far as I can, if I'm still not satisfied ill do the pencil mod. I've done a bunch of practicing, and with several different pencils, and I found one that goes on nice enough and erases to leave it almost mark less. (Having a dud motherboard does still come in handy lol)


Now to finish modding my case, solder on my fancy new fan connector, and get my 175mm 220cfm fan to work . Now lets see how good that stock cooling really is lol.

Once I finally get my package from Ncix ill try out all your guys tips and I'll post back on what I've found. Hopefully I'll be able to get 3.4 ghz or more.

[Radical_53]

You can use the HR-05 for both chipsets. It works quite nicely for me so far, temps are pretty good. Smaller sized Zalman blocks do not work at all, at least not without a powerful fan in front of them.
I always like to keep as many waterblocks out of my loop as possible, at least all those that I don't need.

[DarkCow]

NOOOOOOOO!!!

well, the Video card came a day early (gotta love fedex). I put my computer back together, plug the power in, the green light turns on and everything, boot it, and I get one long beep pause, one long beep etc.

I'm guessing its a ram error because it requires 2.1 volts to run at spec but the motherboard is only sending 1.8v. Do you think thats what is going on?

EDIT: Nvm... fixed the problem.

Basicly what i was doing was putting the ram in the same colored slots and it wasn't booting...

When I put the ram right next to each other it booted fine.

But when it was one stick or the other stick, it refused to boot... weird... o well, at least it works

[Snyxxx]

NOOOOOOOO!!!

well, the Video card came a day early (gotta love fedex). I put my computer back together, plug the power in, the green light turns on and everything, boot it, and I get one long beep pause, one long beep etc.

I'm guessing its a ram error because it requires 2.1 volts to run at spec but the motherboard is only sending 1.8v. Do you think thats what is going on?

Did not see one long beep only, maybe it is one of these.

http://www.amptron.com/html/bios.beepcodes.html

[DarkCow]

Thanks snyxxx, I've been trying to look for those, in my manual its like "look below for the bios codes" and theres nothing there...

Well, I got it working, I got a g0 q6600 at 1.2875 vid.

quick question, how do you run the ram in Dual Channel? same color or right next to each other?

[Jultsu]

I'd go with the same colour.

BTW, does anyone have any idea what the GLT ref mod would help? Would it help make my quad go 9x450 stable, or is it about the FSB problems?

[Radical_53]

Same color. Use one slot leave one slot free, then use the next. That's DC.

PS: The what mod?

If there's anything that could help to stabilize the VTT voltage, I'd be so glad. Somehow it does have a serious impact on stability and according to OCCT, it's my "jumpiest" voltage. During the test I get differences of 3.88%.
Does anyone know from which PSU rail it's fed? 5V maybe, as this fluctuates the most out of the 3 PSU rails?

[Jultsu]

Hmmmm.... The board seems to require even greater NB voltage in order to drive 4 sticks. I'm now at a whopping 1.840v NB with 4 OCZ sticks @ 562MHz and 1h of memtest! I've also increased the 1.2v HT to maximum (2.016v).

Radical: are you giving it 1.55v from bios? Mine stays rock stable @ 1.616v according to Everest, I've never seen it fluctuate.

[Jultsu]

This board is just insane! (or is it just the NB?)

Here's some 4-stick-madness for ya:

http://www.aijaa.com/img/b/00023/1424142.jpg

Never ever would have thought that upping the NB voltage (again...) would stabilize it with 4 sticks. But I'm glad I tried

[Radical_53]

Looks quite nice But you could consider using 2 sticks. The newer 2GB sticks do run really nicely. I'm using the ReaperX from OCZ, roughly 2.1V gives me 557MHz 5-4-3, absolutely stable using OCCT or other tools alike. Less hassle

And no, I'm not using 1.55V on that one. VTT is set to 1.25V in Bios, which gives me something between 1.29-1.31V "normally". I saw that some lines where jumpy, so I monitored them one after the other with OCCT. And saw that the VTT was the one that jumped around the most. As only the 5V rail from my PSU was jumping too, I guessed they might be connected somehow. Also, the higher I raise the speed of the CPU, the more VTT jumps around...

PS: My board is quite funny regarding voltages. They have to be "just right", a little too high is just as bad as a little too low. Also it always wants the "right" combination of voltages for the according clocks. But well, got used to that by now

[Jultsu]

motherF------

My OCZ sticks are broken. Only one out of four will boot to windows. I ordered a cheap 2*2gb kit from a-data. I'll try breaking 500fsb after I get those babies

[trulez]

I just bought this mobo, along with some other new parts, in hopes of getting at least some OC done. Everything I've tried thus far fails.

I've gotten it to post 1280MHz QDR but anything above that is not posting. Everything but the stock 1066MHz fails to load Windows, it will just restart during loading.

If I set memory to 1066MHz manually it fails to load Windows. Upping voltage doesn't help and at 2.2V to mem it fails to post.

My settings were:

AI Tuning: Manual
Nvidia GPU Ex: Disabled
LinkBoost: Disabled

FSB - Memory Clock Mode: Unlinked
x FSB - Memory Ratio: 9
FSB (QDR), Mhz: 1280 (320FSB)
Actual FSB (QDR), Mhz: 1280
MEM (DDR), Mhz: 960
Actual MEM (DDR), Mhz: 960

Vcore Voltage:1.45
Memory Voltage:2.1
1.2v HT Voltage:1.25
NB Core Voltage:1.45
SB Core Voltage:1.55
CPU VTT Voltage:1.35
DDRII Controller Ref Voltage:AUTO
DDRII Channel A Ref Voltage:AUTO
DDRII Channel B Ref Voltage:AUTO

tCL (CAS Latency): 5
tRCD: 5
tRP: 5
tRAS: 18
Command Per Clock (CMD): 2 clock (2T)

Advance Memory Settings
tRRD: AUTO
tRC: AUTO
tWR: AUTO
tWTR: AUTO
tREF: AUTO
tRD: AUTO
tRFC: AUTO
Async Latency: AUTO

CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
MCP PCIE Spread Spectrum: Disabled
SATA Spread Spectrum: Disabled
LDT Spread Spectrum: Disabled

CPU Internal Thermal Control: Disabled
Limit CPUID MaxVal: Disabled
Enhanced C1 (C1E): Disabled
Execute Disable Bit: Disabled
Virtualization Technology: Disabled
Enhanced Intel Speedstep Tech: Disabled
LDT Frequency: 5x

PCIEX16_1 Frequency (Mhz): 100
PCIEX16_2 Frequency (Mhz): 100
PCIEX16_3 Frequency (Mhz): 100
SPP<->MCP Ref Clock, Mhz: 200

Should expect anything better than stock 1066MHz with stock NB and SB cooling?

My rig is:
C2Q6600 G0
2x 1GB A-Data PC-8500 Extreme Edition
2x nVIDIA 8800 GTS 512MB (G92) in SLI mode

Cooling:
4x 12cm case fans, two in takes and two outputs.
CPU IDLE 47C
GPU IDLE 54C