ASUS P5B-Deluxe; Problems & Fixes

[scharnhorst]

All right guys thank you for your help

[scharnhorst]

awesome!
I tried 9.5*422 =4000MHz (@1.3v in the bios)
And it works with all value in the bios at the lowest (except for the vcore of course)
Thank you

[Hobkob]

I have recently put a 4th hard drive in my computer and I am forced to plug it into one of the black sata ports on my P5B deluxe. Problem is, the computer doesn't boot when I plug it in and I don't recall installing any drivers for the black ports (jmicron I think). Is that my problem or might it be something else? And where can I get the drivers?

[rpsgc]

Hello,


I have a P5B Deluxe WiFi with an overclocked E8400 and now I'm thinking of upgrading to a Q9550. I've done some research and I found some success cases but I also found several posts recommending against it.
I was aiming for a 3,4GHz (400MHz FSB) overclock but now I don't think it'd be easily achieved with this motherboard and a 45nm quad-core.


Should I go for it nonetheless?

[scharnhorst]

Hello,


I have a P5B Deluxe WiFi with an overclocked E8400 and now I'm thinking of upgrading to a Q9550. I've done some research and I found some success cases but I also found several posts recommending against it.
I was aiming for a 3,4GHz (400MHz FSB) overclock but now I don't think it'd be easily achieved with this motherboard and a 45nm quad-core.


Should I go for it nonetheless?

I 've got a P5B Deluxe Wifi-AP (bios 1238)
- 2*1024mb Geil Ultra Pc6400 C4
- 2*1024Mb Corsair TwinX XMS2 PC6400 C4
- Intel Q9650

With this motherboard and cpu I obtained the following o/c very easily :

- 9*433MHz

[rpsgc]

I 've got a P5B Deluxe Wifi-AP (bios 1238)
- 2*1024mb Geil Ultra Pc6400 C4
- 2*1024Mb Corsair TwinX XMS2 PC6400 C4
- Intel Q9650

With this motherboard and cpu I obtained the following o/c very easily :

- 9*433MHz

Thanks for your reply. Did you have to increase the NB voltage a lot? How hot is it?

Many people keep saying that a P965 can't handle a quad-core properly and that it would be very hot, too hot.

[snnoopypw]

Many people keep saying that a P965 can't handle a quad-core properly and that it would be very hot, too hot.

why? I'm particularly interested in this as I plan to update my E6600 to Q9650 in the future with SCNJ-1000+120mm on it...

[scharnhorst]

Thanks for your reply. Did you have to increase the NB voltage a lot? How hot is it?

Many people keep saying that a P965 can't handle a quad-core properly and that it would be very hot, too hot.

All voltages were at the lowest value (vnb at 1.25 for example)
Except for the vcore (at 1.275)

[rpsgc]

why? I'm particularly interested in this as I plan to update my E6600 to Q9650 in the future with SCNJ-1000+120mm on it...

I don't know exactly why. Their justification was that a quad-core would stress a P965 too much. Even more with all RAM slots full. And therefore the NB would run too hot.
Could be all FUD though, or simple ignorance.

All voltages were at the lowest value (vnb at 1.25 for example)
Except for the vcore (at 1.275)

Once again thanks for your reply. I'll have to keep a close eye on mine though, for some reason as of late the MB temperature sensor is reporting 52º C at idle (as opposed to 42º / 44º C of before). Or maybe it's just because my CPU is running hotter.


I'd rather spend my hard-earned money on a spankin' new SSD than on another motherboard :P

[scharnhorst]

With around 25°C in the room I have no more than 42°C (in burn test) for the motherboard
But I have removed the small piece of "plastic" with ASUS writen on it and above the northbridge

[In-Fluence]

I ran a q6600 at 3.7 (410x9) for more than 6 months with 1.25nb and 1.2vtt and didn't get a single hang. One thing I did notice is the fact that the NB temps did seem to rise a lot, but I lapped the original IHS and bolted it onto the board which helped loads. Finally got a chipset block for it and 2 enzotech mosfet sinks, but couldn't go any higher on the quad. Running the same chip on my MFII now, and having trouble getting as stable as the P5B, a truly great board, I almost regret putting it on ebay!

I think from that point, I found that raising the NB volts any more caused it to run too hot/unstable, even with the waterblock.

[snnoopypw]

So basically if I were to install Q9650 it would be smart to install some proper NB cooler? (like Zalman...)

[rpsgc]

So basically if I were to install Q9650 it would be smart to install some proper NB cooler? (like Zalman...)

Bear in mind, there's no proof of that being true. Best thing is to try it for yourself and then decide.

[-n7-]

Biggest issue with quads on this board is the massive vdroop IMO...(yes, i am talking about the P5B-D, i have one as well as my P5Q-D)

[rpsgc]

Biggest issue with quads on this board is the massive vdroop IMO...(yes, i am talking about the P5B-D, i have one as well as my P5Q-D)

I'd say it's with any processor. My E8400 OC'ed:

BIOS: 1,275V
Windows idle: 1,25V
Windows load: 1,22V


So, eh. It's pretty bad already

[unclewebb]

After a simple pencil mod my P5B-Deluxe has the opposite of vdroop. When I start Prime95 Small FFTs, the voltage goes up either +0.008 or +0.016 with my E8400 C0. I didn't have much luck with a Q6600 G0 on this board but then it was a pretty crappy, high VID CPU so even on a good board with air cooling it probably wouldn't have gone too much higher. I bent the heat pipe a little and squared up the northbridge heatsink so it sits flat on that chip. I've been running 1.55 or 1.65 volts to it for at least a year without any issues or excessive heat. As long as the cooler is making good contact with that chip you won't have any issues. Replacing the crappy push pins with some nuts and bolts isn't such a bad idea either. The extra northbridge voltage comes in handy when playing with P965StrapTweaker.

[-n7-]

I'd say it's with any processor. My E8400 OC'ed:

BIOS: 1,275V
Windows idle: 1,25V
Windows load: 1,22V


So, eh. It's pretty bad already

Yeah.

Back when i had my Q6700 in it, i was getting 0.09v drop + droop under load.

Made things challenging.

I'd do 1.4875v BIOS for 1.4v load.

[unclewebb]

The 2x1GB PC2-9200 Reapers I tried worked well so I decided to upgrade to the, OCZ DDR2 PC2-9200 Reaper HPC 4GB Edition.



With my E8400, my 2x1GB Reapers could boot up at 7.5x533 but I couldn't get these ones to do that. I settled on 8x500 which booted fine and it has no problem going into or coming out of S3 Stand By with these timings.

Memory timings of 5-5-5-18 seem to work OK and 4-4-4-13 works too at this speed.

Here's a couple of screen shots using strap tweaker.





These sticks are rated at 5-6-6-18 using 2.20 volts. I'm using 2.15 volts in the bios and these sticks run very cool. I measured about 40C on the heat spreaders in an open case which is barely warm to the touch.

With this board I probably won't be able to get their rated DDR2-1150 speed out of them but those bandwidth numbers are competitive with newer DDR2 chipsets. The screen shot at 4-4-4-13 with the strap tweaker is not yet 100% stable but without the strap tweaker, you could probably run these at that speed 24/7.

The important thing is that these modules seem very compatible with the P5B Deluxe. No hangs at start up unless I tried some really tight timings and then it was easy enough to recover by cycling the power button for about 5 or 10 seconds on the back of the power supply. After that it would boot up on the next try and let me back into the bios where I could tweak some more.

$56.99 at Newegg is a great deal for 4GB. Highly recommended.

[Buckster]

interestingly I switch between 2 stable configs, one which is 9x373 odd with memory (4 banks worth) @ 1140 MHz CAS5

and another which is 9x392 with memory @ 785 CAS4

both have been stable. I have noticed occasionally in the past that the config 1 has been unstable - depending on hardware config - butI think I'vfe finally nailed it

the high freq RAM config was stable this week until I added a SSD drive - so instead of 2 HDDs - now 3 - the only thing I can think of is that when I get to a certain loading on the northbridge (ie numbers of HDDs etc) it gets to a point where it can't cope with all of those AND 4 banks of high frequency RAM. Oh well - doesn't really matter. Northbridge volts are set to 1.55V btw.

[HammerInTheNail]

I previously had this computer (P5B Deluxe Wifi with E6600, CPU watercooled, northbridge with stock cooling,OCZ 800MHz 2*1GB interleaved) running at 3.9GHz solid 24/7 under full load. However I have been running at stock for about the past year or two or sometimes 320 FSB since i have not needed the extra processing power. Now however it is impossible to get it past 320FSB (I am again using it for running simulations so need this). I can run 320 with almost stock voltages (1.35V CPU drooping to 1.32 something V etc. 1.3 FSB 1.9v mem) with the North Bridge up to 1.45V (I found I need this for stability although only when almost on 320). I find when I reboot I need to enter bios again otherwise it won't start, I believe this is a common problem with this board, but otherwise is completely stable. Increasing FSB to 321 or any other setting above 320 with any voltages results in the computer not booting. This is a common problem especially recently I believe however despite many many hours reading forums I have not found a solution.

Things that may have changed since it was running at 3.9 that I can think of:

Graphics card may have been changed to 9600 GT from previous X1950 (can't remember whether this was changed before or after)
Motherboard bios update

Things I have tried:

Increasing some or all of the voltages up to 1.5V CPU,1.65V Northbridge and max on FSB (1.5V?) and 2.15V memory
Different memory ratios 1:1 etc. (Making sure memory is 800Mhz or under)
Different BIOSes - I started with the most recent 1238, downgraded to 0604 however this would not allow me to boot using GRUB from an ide drive attached to the jmicron controller so I upgraded to 08 (also doesn't allow boot) and then 0910 which is the bios I am currently using (and allows boot using grub from ide drive). So I can say 0910 and 1238 have the same problem. I didn't test 0604 since I was panicing when my computer wouldn't boot and was busy trying to fix that by reinstalling grub etc
Various settings in bios e.g. changing static read control to disabled, disabling legacy USB
Different FSB speeds and multipliers e.g. 401 at 6X 401 at 8x 401 at 9x; 330 at 8x,9x; 360 at 8x, 400 at 8x etc. etc.

All the relevant stuff which should be disabled is disabled I think e.g. speedstep, C1E, thermal management, Vanderpool technology etc.

The only thing left to do that I can think of left to try is to downgrade to an even older bios version which I will do now or tomorrow, however I have the feeling this won't work and if it does it will be a huge pain to get linux to boot. (I did do this originally somehow by moving partitions around making a new Windows one with bootloader, creating a /boot partition in linux installing an ancient bootloader (lilo) to this partition instead of mbr, copying the first few sectors of this over to Windows and modifying the Windows bootloader to load this. In short a huge pain.)

Clearly something is changing at 321FSB to go from stable at almost stock voltages at 320 to no boot at any voltages at 321. I though it might be related to the strap but AFAIK this changes at 400 FSB not 320 FSB, 320 however is the number at which memory will run at 800Mhz under one of the memory ratios in bios so this may or may not be related. However when increasing the FSB I generally run at 1:1 (although I have tried other ratios)giving a memory rate well below 800 anyway so while it might be some strange memory controller problem I do not think it is a problem with the memory itself since this is well below what it is rated for. Since the 320 FSB "wall" seems to be a common problem I'm sure a real expert will know the solution but I've spent 3 days on this now (not including the many many days I spent originally when I made this system) and need help. Any ideas?

[Ancient1]

HammerInTheNail

Don't revert ( though you did) to 3 digit BIOSs , as ASUS stated that one can't at the time.

I suggest :
The first thing to do is CLEAN the mobo from dust.
Load BIOS Safe Defaults.
CLRCMOS
Try only 1 mem stick
Boot without Video Card
Downclock
Undervolt
disconnect PSU from the wall for the night.

* my Mobo developed strange Boot looping and eventually no Boot. dead.
http://www.youtube.com/watch?v=-_n5auumZ3g

[Falkentyne]

I remember something way back about a FSB wall at 320 FSB on the P5W DH motherboard, although I don't remember if that also happened on the P5B. This was AGES Ago.

I do remember that the answer was NOT to downgrade the BIOS (flash back to where you were, then unplug the PSU and clear cmos with the jumper!), but has to do with the memory timings or settings. I don't remember which combination of settings let you bypass the wall; whether it was auto/SPD or certain manual settings. It's definitely a memory related setting, though. I know the P5B had the "TRD" setting, which was important for higher FSB, and the P5W did NOT have this timing (that setting is too tight on 975X chipsets), but I honestly don't remember the exact solutions to passing that wall. See if you can find the P5W DH fixes thread and look at one of the first posts for 320 fsb wall, but I would have thought that was in the P5B thread too....

I don't even remember much about straps on the 975x boards...isn't the 333 FSB support unofficial, since unlike the P45/X38/P35 stuff, the 266 strap is "overclocked" to allow 333 fsb chips to work?

And make sure HyperPath is disabled.

[HammerInTheNail]

I just got in tonight, and I packed up for the night/morning yesterday after posting the last message. I will try your suggestions and let you know how it goes.

[HammerInTheNail]

Well I tried removing one memory stick. No difference (except it went to single channel mode and I have half the memory ofc). Still goes fine to 320 then fails above that. I tried a different dimm slot (furthest from CPU in the black slot, it was in yellow before) same. I have not removed the graphics card or tried different memory since I don't have another graphics card or DDR2 memory to try in it. I have tried downgrading 1238 -> 1004 and every inbetween bios version back to 0405. Strangely the bios version 0614 which has in it's notes that it now provides support for OCZ2P8002GK memory (my memory) was the only one that didn't work. I could get it to boot to the bios screen from a cold boot only, thereafter trying to save default settings or any other settings resulted in failure on boot. All bioses I tested both before and after this worked fairly similarly. The only difference I noticed apart from some different options in BIOS was that the 0405 (the earliest I test although there are previous version 0204 being the original I believe) would only go to 320FSB when ramping up in increments going directly there with the same settings would result in it not posting. With later BIOS versions e.g. current 1238 you can go directly from 266 (stock) but only if it's in "everything is ok" mode. If you try e.g. 325 (which fails) and then after cold boot (necessary to get it to POST) you put in 320 it won't work. You have to go back to stock and get it running in "everything is ok mode" before you can go directly to 320. On going from 266 to 320, in increments or directly, on saving settings (F10) the computer performs a "warm" reboot i.e. it restarts but the fans etc. don't power down. As soon as you go above 320 everything powers down before it tries and fails to restart. The only thread I read with any kind of solution goes something like this: "A: Oh great I've fixed it! B: What did you do? C: See [broken link] this is why it's happening A: Problem was I was WAITING for it to restart"
WTF? I mean what else are you supposed to do but wait - once you've applied the settings you can't do anything else right? However it does make sense in that only at 321 or above does it power down before restarting. I hate broken links lol.

With regard to Hyperpath there is no setting for this on the P5B deluxe (although there is on the P5W I have heard). I don't know whether this means it is always active or always inactive. If it is always active it could result in the memory controller being pushed too far however I doubt this is the problem on it's own unless memory controller latencies are being changed after 320FSB for some reason - otherwise a slight increase in voltage to the northbridge would stabilise it on going 320->321 I would expect.

With regard to TRFC the highest setting is 42 which is what it is on. This is lower than it should be for high FSB values I have read but it is the max you can set it too. Again I wouldn't expect this to be the problem on it's own since a slight increase in memory voltage should then stabilise going from 320->321. There is some step change on my system at that value which makes anything above unstable whatever the voltage (it is stable at 320 at almost stock voltages, but unstable at 321 with voltages of CPU 1.55v,Northbridge 1.65V,FSB 1.45V, memory 2.35V etc. i.e. the kind of voltages you would expect for around 420FSB not 321). I wouldn't even be comfortable running those voltages under load particularly the northbridge one since I have stock cooling on that but I tried them to test if it would post. I know this system is capable of 4.0GHz under full load previously although only if the room was cold. I used to run it at 3.9GHz 24/7 completely stable. I do remember it being a huge pain to get it above some low value probably 320FSB back when I first did it tho. It's like it's got a life of it's own. Maybe I've not kissed it's ar*e enough recently who knows.

By the way when it fails to POST it's not completely dead or anything: it starts up, keyboard lights go, dvdrom goes and you can press the button to make the tray go in and out, but then it stops with the red "hard drive" light on constantly though I think that light is also for other peripherals and the monitor on "power saving" mode i.e. it doesn't come on.

Any suggestions or should I call an exorcist?

[Falkentyne]

Ok that's interesting.

You're saying that when you try 320 (which works) or higher (which doesn't), the board powers down for 5 seconds, after exiting BIOS, and then powers up? (or 320 never powers down, but does post successfully, but 321 doesn't post (only powers on), is this correct?

If it's the latter, then the P5w DH used to do the same thing, but this was fixed in BIOS version 1407, if I recall. Everyone shouted about it because they would have their cooling loops and other devices shut down and back up. If I remember, the board still "technically" powers down, but the power supply is no longer shut down anymore.

Asus had replied with a technical explanation that when "going above 320 FSB, the board has to power down to reset the 975X chipset strap, and this was by design". I believe that originally, most 975X boards powered down like this, at least on early bioses. But with 1407+ on p5wdh, this was changed so the the board powers down without the power supply also powering down.

I would have THOUGHT that this PSU power down issue was fixed in the P5B series as well. There was a workstation board (P5ws64 pro, i think?) where the P5wdh fix was never applied (no idea if that still applies) and some users of that board also complained.

Anyway, that you are getting the no post, at the same FSB (321) that the power down occurs, shows that the chipset simply isn't accepting the strap reset or something. If RAM or chipset tweaking won't fix this, then there has to be some sort of BIOS bug with the current installed hardware.

Did you search the thread to see if that power down feature was never patched for the P5B? I know plenty of people have taken the P5B to higher FSB than the P5W (all due to being able to change TRFC, but both are FSB limited on quads due to no GTL ref tweaks ...

BTW if you do 320 FSB, what happens when you power down with the power switch and then power up? It posts successfully, right?

I just noticed you wrote this:

I find when I reboot I need to enter bios again otherwise it won't start, I believe this is a common problem with this board, but otherwise is completely stable. Increasing FSB to 321 or any other setting above 320 with any voltages results in the computer not booting. This is a common problem especially recently I believe however despite many many hours reading forums I have not found a solution.

This sounds a LOT like the exact same problem is happening at 319-320 FSB, with the only exception is that the board is not powering itself down, which lets you enter the BIOS again. But of course that's impossible on 321 FSB since the board shuts down and won't POST after it powers itself back on.

I need to ask you this:
Did you ever have this type of problem at 3.9 ghz?
Since you were obviously WAY above 320 fsb (400+?). Did you have to enter the BIOS and exit it, after saving settings, back when 3.9 ghz worked, to get it to POST? Did the board, back then, also used to power itself off and back on, after a reboot or BIOS exit at >321?