• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

MP-05 CPU Block Mounting

C

CoW]8(0)

[H]ard|Gawd
Joined
Jul 25, 2005
Messages
1,427
Does anyone know where I can find a guide on how to mount the MP-05 blocks on the AMD64 socket? I'm assuming that the Pro, Extreme, SP, and SP LE use the same retention mechanism.

But to be specific (just in case), I'm using an MP-05 SP.....

The directions are extremely vague, they only provide information for the Intel socket....

If you could provide a picture of the mounting, that would be great :D

At the moment I've mounted it on an assumption of how I believe it should be mounted. And this assumption has give me 45C load temps :(
 
heres a picture of it mounted....not sure if this is similar to yours or not but it should at least show you SOMEWHAT of what you are looking for....

mp0520mounted2qs.jpg
 
Are those threaded rods or are the long screws which came with it going through from the back of the motherboard?
 
board-big.jpg


The black plastic square surrounding the CPU, could that be preventing good contact to the CPU?
 
Not as far as I can tell...I was able to slide a piece of paper with ease through the high points of the plastic square.

Perhaps my poor temps are because of flow rate? The MP-05 SP with NO3 nozzle plate is fairly restrictive. Could a Hydor L35 handle that?

Here's my current loop:
Swiftech MC-RES -> Hydor L35 -> 77' Bonneville heatercore -> MP-05 SP -> MCW-50 GPU Block -> Swiftech MC-RES

All this in 3/8" ID tubing...
 
Black plastic square - the heatsink mounting cage? For most waterblocks (including these), you only need the plastic blackplate (also known as an X-plate).

An L35 should be capable enough, though it's not an ideal pump for high density pin grid blocks like the MP-05/SP-05. It's possible your heatercore may be dropping flowrate enough to make a reasonable impact on the flow rate at the inlet of the MP-05, though it's quite unlikely. You may try re-routing your tubing so that the MP-05 lies immediately after the pump outlet. Your tubing size won't have any profound impact on flow rate, so long as the total run is not excessively long.

You may be the victim of a bad mount, or you may want to validate baseplate flatness.
 
phide said:
It's possible your heatercore may be dropping flowrate enough to make a reasonable impact on the flow rate at the inlet of the MP-05, though it's quite unlikely.
Click to expand...

Flow is the same throughout the entire loop (if it is a series loop)
 
The heatercore's barbs, tank and multiple tube bends cause a slight reduction in Flow and most certainly induces a Pressure Drop. Eliminating laminar flow in a pin grid block requires a fairly high pumping pressure.

For all intensive purposes, I'll use flowrate and pressure interchangably so as not to confuse.
 
phide said:
The heatercore's barbs, tank and multiple tube bends cause a slight reduction in Flow and most certainly induces a Pressure Drop. Eliminating laminar flow in a pin grid block requires a fairly high pumping pressure.

For all intensive purposes, I'll use flowrate and pressure interchangably so as not to confuse.
Click to expand...

flowrate is constant throughout the entire loop. The flowrate into the radiator is the same as the flowrate out of the radiator. You cant have it any different. There is a pressure drop across components in a watercooling loop. You can't use pressure drop and flowrate interchangeably, they are completely different.


PS. The grammar nazi in me wants to point out that there is no such phrase "intensive purposes". You meant to say "intents and purposes".
 
Intents and purposes, huh? That seems to make a bit more sense to me now - very well, then.

Fluid mechanics is something I'm currently concentrating on learning (a subject difficult to grasp without a higher education in mathematics). If the effective flowrate in a closed system is, as you say, constant, perhaps you'd like to educate me (via PM) a little further on the reasons why.

My suggestion to CoW still stands, of course.
 
phide said:
Black plastic square - the heatsink mounting cage? For most waterblocks (including these), you only need the plastic blackplate (also known as an X-plate).

An L35 should be capable enough, though it's not an ideal pump for high density pin grid blocks like the MP-05/SP-05. It's possible your heatercore may be dropping flowrate enough to make a reasonable impact on the flow rate at the inlet of the MP-05, though it's quite unlikely. You may try re-routing your tubing so that the MP-05 lies immediately after the pump outlet. Your tubing size won't have any profound impact on flow rate, so long as the total run is not excessively long.

You may be the victim of a bad mount, or you may want to validate baseplate flatness.
Click to expand...

So for the most part, we can agree that the Hydor L35 should provide enough current for my loop?

Here's my current loop:
Swiftech MC-RES -> Hydor L35 -> 77' Bonneville heatercore -> MP-05 SP -> MCW-50 GPU Block -> Swiftech MC-RES

All this in 3/8" ID tubing...


Just a reminder, my coolant is MCT5, could its viscocity be killing the flow?

Also I plan to do another remount very soon (it will be the third). Also to confirm my baseplate flatness, can I just use the graph paper test and see if there are any distortions in the reflection? Or is there a better test?
 
Well I just did a remount of the CPU block and here are the current results....

Idle:27C
Load:44C

I'm getting worried that it's a combination of a bad mount/not enough flow....

But just a note, when I had filled the system....The pump seemed to have no problem pushing the tiny air bubbles past the angled baffle:
mcresmicro4.jpg


For now I may just let the AS5 sink in and see what happens....

Just to let everyone know, I'm cooling a AMD64 3400+ (Socket 754 - Newcastle). This CPU is on a 130nm process which means it's temps are probably higher than those Venice cores...
 
AS5 will settle a bit more quickly by cycling it repeatedly. That is, periods of intense loading followed by periods of room temperature (or close to it). Boot up your machine, prime or burn (whatever pleases you) for an hour to two hours, then shut your machine down for an hour or two. Repeat this process about five times, and the AS5 will reach the point of being very near fully settled.

The graph paper test is okay if your baseplate is fairly reflective, yes. There are other methods, but they're largely dependent on a comparitive object being flatter than the baseplate. It's unlikely you have anything in your home that is more evenly flat than your baseplate. Odds are that if it has no glaring inconsistencies with regard to flatness, it's likely as flat as the manufacturer intended it.

If you want to find your overall flowrate (to a relatively fair degree of accuracy), you can always do the bucket test. Remove all components from your system and reassemble in the garage, outside, or what have you. Provide a source of water at one end (a bucket full of...water), all the components of your loop in the middle and have the water flow into another bucket. Allot 60 seconds for water to flow into the empty bucket, then find the total volume. There's your flowrate. I can't say what your flowrate will be with your pump and other components - you probably have a better idea than I do as to what is bad and what is not so bad. There's one gentlemen at ProCooling that has a simple Excel-based approximator, though I don't think he has data curves for the L35 or the MP-05.

MCT5 isn't vastly more viscous than water. So long as your pump it outputting enough flow to the block, the coolant's boundary layer is being broken by the turbulators (pins) in the block.

EDIT: It's hard to tell what a given pump is doing by looking at bubbles. My two D5s push an impressive amount of water, but I certainly can't tell how much by looking at what air pockets are doing.
 
Thanks for the comments and replies everyone...

Here are the current results:
Idle:27C
Load:42C

Ambient temperature is about the same (~1F variance)....

I'm hoping the AS5 sitll has more sinking in to do :D

If my temps aren't where they should be, I think I'll be doing a bucket test...Or I may first switch the accelerator nozzle mid-plate from the N03 to the N02. According to systemcooling's data , the N02 isn't so restrictive. The N02 probably won't create as much turbulent flow as the N03 and I'll probably lose some performance, but the increase in flow may offset this in my favor...
 
Here's what I've fixed so far on my cooling system.

Good news:
My pump was fairly noisy so I took it out of the system and found that the impeller was off it's axis. I reseated it to the impeller shaft and the noise is gone. :)

Bad news:
But unfortunately that solved nothing for my temperatures. I am really baffled, I've done over 4 remounts, changed the to a less restrictive accelerator plate, changed temporarily to 100% distilled coolant (for the least viscosity), and my temps are still maxing at 45C.
All my cooling parts feel cool to the touch, even the CPU block's copper base plate.


I'm starting to think that maybe my temperature standards are a bit strict. Maybe 45C for a 130nm Newcastle processor isn't that bad. They do consume a great deal more power than the Venice cores...
 
I think your pump just may not have the guts to use the full potential of the MP-05. Even with the No2 plate it is a restrictive block. That is from the micro pins, you cant really get around it.

My 165 is priming right now at 2.6 1.45V and it is loading at 39-40C on a TDX.
 
But my CPU block still feels cool to the touch when the CPU his running at 45C...Would it feel somewhat warm or hot if the water flow wasn't enough?

From several reviews and charts, the Hydor L30 performs almost exactly like the Eheim 1250. But I don't know how much this says because I don't know if an Eheim 1250 is considered a strong pump. But I hope I can assume that my Hydor L35 performs at least at a Hydor L30 level...
 
Eheim 1250 = about 320gph and i think 6.5ft max head.
Hydor L30 = about 320gph and 6.25ft max head.
Hydor L35 = about 420gph and 6.75ft max head.
 
domoMKIV said:
Eheim 1250 = about 320gph and i think 6.5ft max head.
Hydor L30 = about 320gph and 6.25ft max head.
Hydor L35 = about 420gph and 6.75ft max head.
Click to expand...
Hydor rates the L35 @ 450gph, but the fact that you could recall that from memory is impressive..

I know the head isn't very high compared to the 12V DC pumps, but as far as AC pumps go isn't the Hydor L35 pretty good?
 
CoW]8(0) said:
Hydor rates the L35 @ 450gph, but the fact that you could recall that from memory is impressive..

I know the head isn't very high compared to the 12V DC pumps, but as far as AC pumps go isn't the Hydor L35 pretty good?
Click to expand...
haha definitely wasnt memory, just did a quick google search. you're right, 450gph, my mistake.
a buddy of mine was running the L30 and it did just fine for a couple years, however he was running a diff block with it.
 
45c would make heatpipes luke-warm to warm. 50c would be warm. You simply can not rely on onboard temp measurements from the motherboard. I had experience too many bogus temps on motherboards and resorted to placing an micro 10k NTC Thermistor right under the cpu and between the pins. Now I know my temps are accurate and are typically 10-15 degrees celcius under what the mb reports. I'm getting an even better sensor that measures with accuracy of +/- 0.1C and will test it against a another known thermometer accurate to +/- 0.3C this weekend. I specifically had to resort to this because when I went to watercooling my temps sky-rocketed compared to air. Turned out that the thermistor on the motherboard is crack-pot placed not under the 939 area but closer to some hot mostfets and volt-regs to the left of the cpu. Free flowing air that used to be there with a fan was no longer there with water cooling thus resulting in bogus higher temps.
 
What kind of block was he running?

I hate to say this, but if the cause of these temps is my flow rate, I may have to part with my MP-05 SP :'(

It was the pride and joy of my new system, but my budget has been thoroughly spent and won't hold for a stronger pump (anything pump stronger than what I have now would cost me $70).

But I'm still not sure, I really do not want to give up this block, IMO, it's the been the Storm's underrated rival for some time. Does anyone have any suggestions of what I should do?

Before I even commit to putting this block on the market, I will need confirmation that my temps are bad. Keep in mind, I'm running a AMD64 3400+ 130nm Newcastle @ stock voltages/clocks.
 
scwam said:
45c would make heatpipes luke-warm to warm. 50c would be warm. You simply can not rely on onboard temp measurements from the motherboard. I had experience too many bogus temps on motherboards and resorted to placing an micro 10k NTC Thermistor right under the cpu and between the pins. Now I know my temps are accurate and are typically 10-15 degrees celcius under what the mb reports. I'm getting an even better sensor that measures with accuracy of +/- 0.1C and will test it against a another known thermometer accurate to +/- 0.3C this weekend. I specifically had to resort to this because when I went to watercooling my temps sky-rocketed compared to air. Turned out that the thermistor on the motherboard is crack-pot placed not under the 939 area but closer to some hot mostfets and volt-regs to the left of the cpu. Free flowing air that used to be there with a fan was no longer there with water cooling thus resulting in bogus higher temps.
Click to expand...

I agree, motherboard indicators are always suspect...But IIRC, the diode is on the CPU die for all AMD64's.

But one problem with the current 45C temp is that with previous WC system (Kingwin AWC-1), my max temp was about 52C. I really hate to think my new system is this close to such a low end system.
 
I just did some calibration tests with a very accurate digital thermometer against my thermistor and my themistor is dead on accurate. At 2750mhz on a opteron 170 using the stock heatsink and fan I was getting a full burn maximum temp of 54c. That was at 1.5 vcore. Heatpipes were very warm. This was with an open case and 22c ambient temps. This was also using some cheap white thermal paste too.
 
The MP-05 is a very restrictive block. More restrictive than that of a Swiftech Storm, but the L35 pump really shouldn't have any problems with your setup. My quess is that you have air trapped inside your block. With most blocks with a center inlet and one off-center outlet the outlet is supposed to be above the inlet to aleiviate any air from being trapped within the block. I would also try rocking your case back and forth an listen for sloshing sounds within your heatercore. Any sloshing will indicate that you have air trapped within the heatercore. Another thing to try is turing your fans around so they are sucking air through the radiator and exhausting it out the top.
 
According to Systemcooling.com's review, with the N03 accelerator it is more restrictive than the Storm, but with the N02 accelerator it's about the same...

Well I'm fairly certain I bled the loop properly because I've rotated the case on all sides by up to 90 degrees several times. My rotations were done with the pump on and off and at angles and patterns which should concentrate the air pockets where the pump would be able to flush them. I did this till I could hear no flushing sounds.

And yes, the fans are pulling air from the radiator to the top of the case :)
 
You must log in or register to reply here.
Back
Top