• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

m5a97 le vrm cooling

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Chuy

Registered
Joined
Dec 3, 2011
Location
Las Vegas
I already know that this mobo isn't the best for overclocking but since I have it I decided to see what it can do.With stock settings I took the temp of the vrm area with from the back of the motherboard with a laser temp gun and got a max temp reading of 65.5c. I mickey moused a fan over the vrm area which dropped the temps to 53.3c.

When trying to overclock to 4.3 I kept freezing up the computer and i would see the vcore displayed in cpuz would drop to 1.2something (forgot to write down the number). I took off both side panels and and even pointed a 6'' cool breeze fan directly at the vrm area on the backside of the motherboard. I should of taken a pic and posted it up because it looked hilarious, trying to cool my comp with a house fan lol. Once I had my fancy set up I was able to run prime95 for about 10 mins before a core would stop working, with a voltage of 1.425 in bios and a vcore reading of 1.344 in cpuz and ai suite II while running prime. The vrm temp reading was at 52.7c at the back of the mobo. I think the vcore voltage isn't stable so here come my question. If i add heat sinks and a second fan to lower the vrm/fets temps will I be able to squeeze out more voltage out of them? If not will I be able to get a more steady supply of what it is putting out? I'm overclocking the computer listed in my sig. Thanks in advance!
 
A bit more info really wouldn't hurt. Is that a FX8xxx? Any additional cooling will be beneficial but do you really want to invest much into that board. When I first OCd an 8350 I had an M5A 6 phase heatsinked board and was still limited to 4.6, that was wth additional cooling to VRM and socket area behind the mobo. Personally I don't think you would get much more out of it. The limit is still the 4 phase VRM that won't be able to supply enough " stable" power to feed an 8 core CPU.
 
Well he did not give actual motherboard model nor the cpu but saying he has no VRM sinks would indicate R1.0 of that M5A97. The R2.0 of that mobo has the VRM sinks. I think mandrake4565 had the other of the EVO or PRO whichever you did not have "johan" and he was limited to 4.5Ghz with his 8 core FX processor. Plus I remember caddi daddi saying his efforts on the M5A97 R2.0 and only a four core FX, was a disaster.

Put all of that together from known scenarios, I doubt that heatsinks and more fan would help that R1.0 motherboard. But hey it is somebody else's money, so they can try if they want to try. I just am guessing it would be more enjoyable to try overclocking on a board better suited to such fun and games.
RGone...
 
I filled out my sig but I guess I didnt save it? Just realized theres nothing there lol

Rig specs:
Mobo m5a97 le r2.0
Cpu fx 6100
Cooler sunbeam cr-cctf 120mm core contact freezer
Ram hyperx 4x2 (fill in speed later)
PSU corsair tx650
GPU sapphire 7870 ghz edition (one card)
OS windows 7 home premium
Samsung 64gig ssd & 1tb hdd
Case raidmax blade atx-298wbp with 1 - 120mm exaust fan at rear of case &
1 - 120mm 1 - 92mm intake fan at front of case

Im not planning on dumping a load of money, Im talking about some cheap heatsinks from ebay that come somewhere from China. Some zip ties from harbor freight or diy bracket from items I already have or find at work, I work in building maintenance and theres construction going so great opportunity to find some goodies. Considering its a fx6100 I dot really want to buy another mobo for it, correct me if im wrong but new mobo would mean new copy of os right? I having fun ocing not doing out of necessity, the games I do play are running fine on the stock settings. Squeezing the most out of this would be my little budget project.

I acually have a m5a97 r1.0 with the heat sink over the vrm that I thought was dead due to only getting a black screen even after reseating everything and pressing the mem ok button 100 times. It was months later messing with another computer did I learn to take out the mobo battery to reset bios. So I got a working mobo sitting around. I wanted to buy a fx6300 for this board but im starting to wonder if I should just use this for my fx6100 sell the m5a96 le and save up for a better mobo for my second future rig.

So would extra cooling get me more vcore voltage to feed the fx-6100?
 
If you are still considering trying to oc this board it would be helpful to everyone if you would post a screen shot of hwmonitor & cpuz (cpu, memory,& spd tabs) while running P95.
as for needing a new OP sys. when getting a new mobo the answer is no. windows might want to reactivate but that's not a problem.
 
Like I said earlier with a better power system on my M5A board I topped at 4.6 with an 8350 so 4.2 sounds to me like the resting place for that CPU/Mobo Combination. FX6100 will take a bit more voltage than it's piledriverr counterpart. So as I have said I wouldn't expect a great deal more. If they're that cheap just give it a shot it might get you to 4.3 .
 
Here are some screen shots after running prime95 for about 10 mins so I can't say this is a stable overclock. Both cover from the case a removed to help with cooling, once I get a stable oc I will put the covers on to see the temp differences. Voltage setting in bios is 1.4625 and according to cpuid and ai suite fluctuates between 1.368 and 1.380
 

Attachments

  • rig sc1.png
    rig sc1.png
    35.8 KB · Views: 1,943
well I walked away from comp while in the middle of posting this and one core dropped out at about the 20min mark. gonna up the voltage a bit.
 
Was messing around and was able to get to 4.5GHz with lots of vcore. Was able to run prime95 small fft, only tested for 30 mins and would fail the blend test in about 20 mins one core would drop out. I was googling around trying to find the danger temp of the mosfets. I found that they an operating junction and storage temperature range of -55 to 150c. Does this mean that in the case of running these in our computers temperature isn't what kills them but over working them is what takes them out? If so do more oc friendly mobos have beefier ones?

Well here some screen shots. If you got any tips on how to stabilize this OC please chime in. Can post pics of bios setting if it helps too. Posted both idle and load pics
 

Attachments

  • 4.5GHzidle.png
    4.5GHzidle.png
    53.2 KB · Views: 1,883
  • 4.5GHzload.png
    4.5GHzload.png
    52.3 KB · Views: 1,983
Chuy said:
. I found that they an operating junction and storage temperature range of -55 to 150c. 1.) I would imagine as relates to mosfets as seen on motherboards as VRM mosfets, that range is too wide. I too saw that range you mention but it was for a mosfet we do not see on motherboards. 2.) Without having to go drag out formulas you do not want to calculate to the answer...suffice it to say that generally a mosfet at 40c can handle nearly twice the current that it can at about 90c or maybe a shade greater temp. Here again we are talking about mosfets used in general on current motherboards and not industrial mosfets. At the very least you will have less power available to the cpu as mosfet temp increases. 3.) More phases comprised of heftier components and better cooling is a must for high current drawing cpus like the newer AMD FX processors if one plans to overclock them and wish the components to survive long-range. 4.) Over-heated and VRM circuits that are unable to keep up with the load "must" throttle the cpu or risk mosfet short and then possible cpu damage and other possible damage. 5.) To the best of my knowledge, even today, it is understood that a mosfet is far more trustworthy at 90c or less. Again at least within the computer motherboard arena.

Does this mean that in the case of running these in our computers temperature isn't what kills them but over working them is what takes them out? 1.) Well if taken literally; working our computers with a high current requirement which will also raise the temp, comes down to a balance between not overtaxing the VRM due to too much current draw or a too greatly elevated temp that will cause failure. 2.) So it is not a case of one or the other but that too much heat or too much demand on the VRM circuit will hasten failure. 3.) Demand brings heat and runaway heat can bring failure rather rapidly.

If so do more oc friendly mobos have beefier ones? This is the easiest to answer without much ado. Most all truly overclock friendly motherboards have beefier power circuits to support the CPU and this additional beef almost without exception includes hefty VRM heat sinks for lesser VRM temps.
 
RGone, thanks for the help and info. Little by little I'm learning what parts of the mobo do what.

Mandrake4565, when I get a chance I'll get a screenshot of the HWMonitor. I usually run stock due to it being good enough for web surfing, even the game I play don't make the cpu usage go to 100%
 
Yeah, we need to see the package temp under full load. It's the most important one. CPU temp is really CPU socket temp. Package temp gives a better idea of what the temp of the cores is inside the processor. Please adjust the frame of the HWMonitor interface accordingly.
 
Here are some screen shots, let me know what you think.
 

Attachments

  • 4.5GHZextendedHWM.png
    4.5GHZextendedHWM.png
    62.5 KB · Views: 1,819
I would say if your using 1.5v for 4.5 and it still crashes, you should drop back to you 4.3 OC and get it stable. Your motherboard is at it's limits I think, it just can't supply the power you need to go faster. So drop back and work from there to tune the CPU_NB and ram.
 
Just a heads up, the M5A97 r1.0 does have a vrm heatsink, but if the le r2.0 uefi is like the vanilla r1.0, LLC is either on or off. There is no in between settings on mine :-(

I know the M5A97 vanilla r1.0 isn't great with the cpu voltages. For example, I have my vcore set to 1.39v with llc on, and am currently seeing 1.416-1.464v. Asus definatelly skimped on the r1.0's LLC...
 
Noob question: cpu llc is the cpu load line calibration right? That is set to auto. Should I disable it or enable it?
The cpu/nb load line calibration is also set to auto.
The voltage is all over the place with this board lol, when it dips low is when it either freezes or core drops out of prime
 
so far with my vanilla r1.0 m5a97, I have found auto LLC and using voltage offsets seems to get a slightly smaller range of voltages. The LE does not have a VRM sink, and yours could be everywhere cause the increased vrm heat. :/
 
Back