955BE on Biostar A770E3 overclocking issue.

[Tais]

CPU: Phenom II 955BE RB-C3
Motherboard: Biostar A770E3 v6.3
BIOS: 78XADB03BF (AMIBIOS 080014)
RAM: 2x 2GB PC3-10700 G.Skill DDR3
Powersupply: Antec 520W Gamer edition (single 40A 12v rail)
Vcard: ATI 5770

The problem I am running into while trying to multiplier overclock this 955BE is that anything over 16.5x multiplier results in a no post and the motherboard constantly beeping as if no CPU is socketed. 16.5x which is a half step over the stock 16x works flawlessly without any changes to any voltages.

17x, 17.5x fails with a no post and beeping requiring a reset of the CMOS no matter what CPU VID or NB-CPU VID I use. I've tried everything from 1.4-1.5 CPU VID and 1.1 to 1.3 NB-CPU VID. Temperatures at 3.3Ghz have yet to go beyond 45C at full load and since the problem is it won't even post I wouldn't think they are a factor.

So far I've had no luck finding anyone using this motherboard and CPU that could give me some insight into what the problem might me. I find it strange that a mere 1x multiplier increase is resulting in such an inescapable no post issue when a .5x works without issue.

The motherboard is speced for Phenom II 965,970,975 and up. http://www.biostar.com.tw/app/en/mb/cpu_support.php?S_ID=450

I wouldn't think that just changing the multiplier up on a 955 would throw it beyond what it is capable would it? Is it possible that perhaps I've just got a 955 that can't overclock much or is there advice any of you can give on something I am missing. Just getting 3.6Ghz would make me happy not looking to push the 3.8Ghz region.

 

[trents]

What's the TDP wattage rating of that motherboard compared to the CPU you plugged into it?

 

[Tais]

Biostar's specification site says TDP 125W. Phenom II 955 BE is a 125W one. I noticed the site is down atm, so you probably can't use my link atm.

 

[trents]

It would be helpful if you would post pics of CPU-z tabs: "CPU", "Memory" and "SPD". That will give us a snap shot of your hardware and a number of critical bios settings.

 

[trents]

Is their a newer bios version available for the board?

 

[xsuperbgx]

I had one of these boards for a while. I know that they were originally only rated for 95 watts, although it was probably an older revision. The one I had, had to be flashed to a newer bios to run the 955 that I had. I don't think it is really up to the task of overclocking much. It doesn't have a very strong circuitry for the power to the cpu.

If you are going to try to overclock it, I would make sure you have the latest bios first. Will it overclock much by fsb?

 

[Tais]

Adding the requested screenshots of the CPUZ, the only setting away from stock is it is running at 16.5x multiplier which is completely stable even with a fairly large voltage droop down to 1.344 on the CPU Vcore (stock setting is 1.4v)



@xsuperbgx

Yes the new revision of 6.3 is upped to 125W and comes with the latest BIOS for the A770E3. I verified this myself when I hit the problem thinking maybe this was it. I actually haven't tried to see what it can do with FSB overclocking. I was looking for the easier approach because I know my memory isn't that great for overclocking but I can see about trying it.

It may very well be it can't supply any more watts then the 4 extra watts of going to 3.3Ghz. Though as an electronic engineer myself, that seems an awefully narrow margin of tolerance error for a power circuit design.

Ah well I can't complain much if I can't get any more out of it. For $50 it is really proving to be very stable and the performance is much much better then my old Athlon 64 X2 5200+ overclocked to 3.1Ghz on the boards precursor the A770A2+.

 

[trents]

Yeah, all the CPU-z stuff looks appropriate. The board is apparently pretty marginal with regard to 125W TDP CPUs. It would be interesting to see how well it would overclock say, a Athlon II X2 Regor.

 

[trents]

Tais, I have an idea. If the problem is just that it won't post when you set the multiplier higher than 16.5, why not use bios to change all things but the multiplier and use AMDOverdrive to set the multiplier in Windows? Overdrive can be configured to automatically apply settings when Windows loads and it wouldn't do it until Windows loads. Get my drift?

 

[Tais]

I tried using the AMD overdrive software but for an unknown reason it will not show any multiplier higher then what the BIOS is set at. When it is at 16x and the sliders stop at 16x and there is nothing further to the right. When set to 16.5x the sliders stop at 16.5x.

 

[trents]

And unfortunately I don't think Biostar offers their own overclocking utility as do some of the other board manufacturers.

 

[Tais]

So far no luck at all getting it to post with a multiplier set at anything above 16.5. Too bad the motherboard has no features to try to under volt the CPU so that I can see if it really is the wattage that is triggering a fault protection.

So to anyone out there wanting to overclock a Phenom II 125W BE cpu make sure you get a 140W TDP motherboard! You really have to research too and find them, they aren't nearly as common as one might think it turns out and you will pay $ for them.

The real kicker is the day after I bought the 955. Newegg put up a promo code on 965s that put them at the same price as 955s... doh.

 

[Tais]

Further testing this morning has me beginning to doubt it is a TDP wattage issue with the board.

By my calculations:

955 BE stock voltage/clocks : 125W TDP (1.35v at 3.2Ghz) - POSTS

955 BE stock voltage/100Mhz overclock via multiplier: 129 W - POSTS

955 BE stock voltage/200Mhz overclock via multiplier: 133 W - NO POST

955 BE 1.42V/100Mhz overclock via multiplier: 147 W - POSTS

955 BE 1.45V/100Mhz overclock via multiplier: 153 W - POSTS

Um.. so it posts at higher watt draw via higher voltage... but not lower...? I wonder how helpful at all Biostar would be with this on their support cause.. it doesn't seem to be a power issue after all.

 

[trents]

It's beginning to look like bios issue with the multiplier itself. In my experience with biostar their bioses tend to be buggy and they don't like to acknowledge it. Don't be surprised if a future bios release for that board (if there will be any) fixes the problem.

 

[Tais]

Shot a letter off them to find out the answer. They list support for Phenom II z965, which is a C3 stepping Phenom II 965BE which is of course the one you can buy OEM and retail atm. So a 17x multiplier is required for that as is a 17.5x for the z970 they list and the 18x for the z975.

I have a hard time believing it is the CPU simply because I've never encountered a cpu that when it hit or was beyond it's clock limit just refused to even report to the motherboard and result in a constant rapid beep.

If I was FSB overclocking I'd look to suspect my memory as that is the most common reason for a neverending rapid beep and pre post failure as without memory the process can't even start.

Just multiplier overclocking shouldn't mess with the memory divider should it? I tried setting the memory to forced DDR3 1066 to see if maybe the board was switching memory divider and overclocking the memory but no joy either.

 

[trents]

. . . Just multiplier overclocking shouldn't mess with the memory divider should it? . . .

Shouldn't.

 

[trents]

Is weird that the board would support 965s and up but not 955s. But what you are describing is in fact what you commonly see when a non-supported CPU model is used in a board that supports other members of that CPU. It will work but in some basic mode but not be fully functional.

 

[Belial88]

I know it's a little late, but the solution to your problem is to up CPU Over Voltage to +.2v.

I had the same issue, it's weird. Specifically, I couldn't post at all with zero overclock if I enabled Custom-p states (it doesnt do anything, it's just an option for an option, as in, disabled does nothing, enabled gives you access to core vid/fid/did, cpu-nb vid/fid). I had to give +.1v just to enable custom p-states, and I had to give +.1v if I overclocked the CPU to 3.6ghz+

The weird thing is, it's not a vcore issue. Like I could drop the VID to very low, then give cpu over voltage from the motherboard to stock, and it'd post fine with custom p states enabled or even 3.6ghz (obviously stock volts on a c2 955 is not stable, but it'd post).

I'm not really sure why this is an issue. I'm guessing the motherboard just has to give enough juice to start up the CPU or else it won't post, hence why it doesnt matter if you raise VID to 1.55, if you dont have enough cpu over voltage, you won't post.