Replacement Water Pump Advise?

Just blew out my Danger Den CPX-PRO 1200lph pump.
Well, its not blown, the impeller has started to make its death rattle.
I've peeled her open and tried to do some rescue surgery, it has helped a little, but she's dead Jim...

I'm looking for a replacement for her.
She is currently running a Raystorm block with a Feser XChanger Radiator 360mm
Cooling down a 3930k.

http://www.amazon.com/Swiftech-MCP6...ie=UTF8&qid=1382757087&sr=1-2&keywords=MCP655

Looking at this guy for the replacement.
Unless I can get some other valid options from you guys.

Looking forward to any help you can provide.

Sorry to derail the thread, but the quote she's dead Jim is referring to starcraft?

That's 'just' a Laing D5. that about right price wise. Everyone Ive ever talked with on this subject says that the D5 is simply the best pump to use. I've used it myself and I definately can't complain. Plus there are tons of kits you can get to 'pimp out' the pump.

MCP655, aka Laing D5 aka Koolance PMP-450 aka Alphacool VPP655. One of the most reliable pumps out there in my experience. My pump manufactured in 2005 is still going strong.

Yup. Get a D5 variant.

Only other thing to closely consider might be an Eheim variant if you wanted an AC water pump. But just about no one does these days.

D5 is great. Even with future upgrades you can use it. There is been people running 5 radiators and 3 waterblocks with single D5 (Linus tech tips on youtube and even in such stupid loop it did good job). Well here is link skipp to like 35:30 to see bleeding action. http://www.youtube.com/watch?v=JtHDqdo5pR0 . BTW I am not related to him or think that that build has any sense what so ever.

Though tbh DCC Plus would do good job to. Even DCC Plus is over kill for single rad and single waterblock.

If you get it get MCP 655 and not MCP 655B. MCP 655B stands for basic and has no speed regulation. Speed regulation will be usefull to slow it down to lower a noise. D5 is relativley quiet if mounted proper.

I have both the MCP-655 and three MCP-355 pumps.

You might want to think about the MCP-355 and an XSPC acrylic pump top reservoir.....
that combination will almost rival the performance of a 655 at a bit less cost, take up less room, and give you the pump and revervoir all in a neat package.

My MCP-355 powers a loop of two dual 120 XSPC rads, a Heatkiller CPU block and two Heatkiller GPU blocks just fine.

magoo said:

I have both the MCP-655 and three MCP-355 pumps.

You might want to think about the MCP-355 and an XSPC acrylic pump top reservoir.....
that combination will almost rival the performance of a 655 at a bit less cost, take up less room, and give you the pump and revervoir all in a neat package.

My MCP-355 powers a loop of two dual 120 XSPC rads, a Heatkiller CPU block and two Heatkiller GPU blocks just fine.

Click to expand...

I'd take take a D5 over a 355 pump. D5 IME is more durable and less prone to heat problems. Also D5 runs a great deal quieter at any throttle speed.

+1 MCP655, personally I use it with the bitspower top replacement, AND, I have 2 of them in serial...which is...overkill heh.

I found this at Martins Liquid Lab......the MCP 355 and XSPC top is really a good combination. http://martinsliquidlab.i4memory.com/DDC32PumpTopTesting.html

I need to search out a comparison to the 655 though.
Here's a good read......MCP-35X vs. MCP-355 vs. MCP-655 http://martinsliquidlab.org/2011/02/25/swiftech-mcp-35x-reservoir/

Makes the MCP-35X look like a pretty good little pump!

magoo said:

I found this at Martins Liquid Lab......the MCP 355 and XSPC top is really a good combination. http://martinsliquidlab.i4memory.com/DDC32PumpTopTesting.html

I need to search out a comparison to the 655 though.
Here's a good read......MCP-35X vs. MCP-355 vs. MCP-655 http://martinsliquidlab.org/2011/02/25/swiftech-mcp-35x-reservoir/

Makes the MCP-35X look like a pretty good little pump!

Click to expand...

As an ex-35X owner...It might be good. But it is loud apples:apples to the D5...and further the MOSFETs inside the pump are known to reach ridiculous temps and unless sinked you can smoke the pump. Further I'm not the only person on [H] to have a 35X brain go stupid and render the pump unreliable. Mine lasted just 2 years (enough to be outside of warranty). The nice thing about the D5, no PWM brain that can go retard and inappropriately throttle the pump off regardless of temp.

It is a small pump with good performance hydraulically, but the D5 is a better piece of hardware in reliability and noise at every RPM level.

Swiftech actually makes a PWM version of the MCP655.

Skripka said:

As an ex-35X owner...It might be good. But it is loud apples:apples to the D5...and further the MOSFETs inside the pump are known to reach ridiculous temps and unless sinked you can smoke the pump. Further I'm not the only person on [H] to have a 35X brain go stupid and render the pump unreliable. Mine lasted just 2 years (enough to be outside of warranty). The nice thing about the D5, no PWM brain that can go retard and inappropriately throttle the pump off regardless of temp.

It is a small pump with good performance hydraulically, but the D5 is a better piece of hardware in reliability and noise at every RPM level.

Click to expand...

As Ive said, I have 655s and 355s. Never had problems with either.
The 355 does get warm.
I think both pumps make noise. Depends how you mount them.
Just throwing out ideas.

At this moment I do have MCP 655 and two DCC's. I do have XSPC Restop and double bay single and double pump reservoirs from XSPC and Alphacool plexi top. DCC stock top is just there to keep a impeller from falling out during the transsportation. If you want any performance out of it, you will get your self a aftermarket top. I use new rubber O-ring and alternative top always with a 35x. On other hand D5/MCP 655 will not need any top to work better and some top's might even slow it down. Only "drawback" of stock top on D5 is non removable hose barbs. So if you want compression fittings on it you will need other top. Be aware to check model you choose for performane figures.

If you mount DCC in bay res with side rubber mounts it will not make that much noise. If you mount it on the bottom of a case isolate it from case or it will be audible.

D5 on other hand asks for nice mounting. If you tighten screws to much and it can relay vibrations to a case it will induce headache over time. It will sound like some fan is on. Some noisy fan. At all speeds. But if you de-couple it and do good job it will be really quiet in almost all ranges of it's RPM's.

D5 will sound different with different loops and at different speeds. It changes character and it can almost always be adjusted for good performance with acceptable noise level.
Loops with high restriction will make it sound louder sooner in the RPM's (already from number 3). Low resistance loops will be less audible. Anyways from 4 and upwards it might be more audible. In loops I have build I have almost never seen any performance gain pass 3 or 4 setting. In restrictive loops 2 will raise temps for 3-5 dgree and minimum speed will not be enough even. But even in such restrictive loops there is 0 degree difference between 4 and 5. Those are my findings and some loops with not enough radiator surface might show difference but not in my loops. I tend to have slight radiator overkill. Although there are many theories I like thick low preasure radiators and would install Alphacool Monster everywhere

When I set up D5 I fire up some burn in program and regulate D5 speed until I can get lowert number I can without dropping any perfomrance. That is then reference point for me. This is with all fans and all system hardware on.

Then I fine tweak it to aviod any resonanse with a case. I disconnect mainboard and most of the fans and other hardware.
Do not keep a loop without fans to long though while pump does heat up water over long period of time. Once I find best position I confirm it with all fans and system on. This is just to avoid some resonanse with some other case vibrations made by other hardware.

In quiet systems, if D5 is not adjusted, I can hear D5 as irritant background noise. Gives me headache. Many people can not hear it in first few hours though. Later it comes out of the mix.

On other side DCC at full speed and ot good isolated is just noisy. Not that irritant.

D5 is heavy duty monster, kinda millspec device. DCC is elegant and durable product, more like civilian market highend part. Never had any of them failing.

BTW I still have Enheim 1250 somewhere and it is huge pump working on mains. Well D5 pumps same amount of water in same time and makes nice overhead preasure. It is 12V and way smaller. It is a overkill in most of the loops and it can run almost any loop with enough water speed.


Och yeah adjust D5 only after your system has been leak proofed and with zero bubbles and with proper coolant levels.

I find that DCC (600L/min plus model) with XSPC Restop makes best combo regarding a performance. Though nices for me was always bay ressolution. With Restop I find that it can handle two rads and two waterblocks max. Everithing over it will slow it down big time. I have tried 3 radiators and 3 waterblocks and water flow was way to slow. Same loop can be run with D5 at 3 to 4 setting. No need for 5 even.

Tbh I have not build a loop that was so restrictive that one D5 can not handle it. Some loops might need it but I have never build such a loop that asks for more then one D5. It is so strong.

BTW I almost never use 90 degree fittings and I do place my tubing so that water can move fast and I never mind on how it looks actually.

All those claims about DCC power and D5 Power I draw from my loops and with 1/2 hoses from Tygon (silver line). 3/8 is good enough but anything under it might have negative effect. 3/8 in small loop is just as good as 1/2 and routing of hoses will be so much easier. Hoses I use are just way to hard and thick and way to much work. Once in place they are good though.

I use lot of Bitspower parts mostly. Not because of the look but because they are high flow parts mostly. Way overpriced and can brake to but high flow parts. I use lot of normal fittings in true silver variant. Straight normal true silver fitting. Side wall is very thin and they are really not restricting water flow. Plus they might influence algee build up.

So my loops tend to look ruff but water flow is good. Same components with nice hose routing and lot of 90 degree fittings and U turns will ask for more pump power biut will look superb.

MD said:

I use lot of Bitspower parts mostly. Not because of the look but because they are high flow parts mostly. Way overpriced and can brake to but high flow parts. I use lot of normal fittings in true silver variant. Straight normal true silver fitting. Side wall is very thin and they are really not restricting water flow. Plus they might influence algee build up.

Click to expand...

I actually switched to more Bitspower because of the fitting options. I found that they generally worked better in addition to the flow advantages. I also started using quick disconnect fittings which were a bit more expensive, but made it far easier to swap things out when needed.

MD said:

I have build I have almost never seen any performance gain pass 3 or 4 setting. In restrictive loops 2 will raise temps for 3-5 dgree and minimum speed will not be enough even. But even in such restrictive loops there is 0 degree difference between 4 and 5. Those are my findings and some loops with not enough radiator surface might show difference but not in my loops. I tend to have slight radiator overkill. Although there are many theories I like thick low preasure radiators and would install Alphacool Monster everywhere

Click to expand...

thanks for the time to put post this - very informative.

What would classify as a restrictive loop as far as impacting temps when changing D5 from 1,2,3 and 4?

I am running 1 x D5 in a Koolance Bay res, 1 x120.3 rad, cooling just GPU and CPU - temps don't change from speeds 1 to 4, but I am in the process of adding another 120.1 rad and am wondering what effect this will have on restriction. I do use 90 degree and fittings and Koolance QDC's.

Currently, 1 is absolutely silent, but any higher speed starts to introduce vibration transfer which is inaudible when case side panels are off, but audible when on. Bay res is decoupled with silicone anti-vibration tape and side panels are decouple as much as possible using the same tape.
the noise levels at higher speeds are not loud, just not completely silent.

kickmic said:

thanks for the time to put post this - very informative.

What would classify as a restrictive loop as far as impacting temps when changing D5 from 1,2,3 and 4?

I am running 1 x D5 in a Koolance Bay res, 1 x120.3 rad, cooling just GPU and CPU - temps don't change from speeds 1 to 4, but I am in the process of adding another 120.1 rad and am wondering what effect this will have on restriction. I do use 90 degree and fittings and Koolance QDC's.

Currently, 1 is absolutely silent, but any higher speed starts to introduce vibration transfer which is inaudible when case side panels are off, but audible when on. Bay res is decoupled with silicone anti-vibration tape and side panels are decouple as much as possible using the same tape.
the noise levels at higher speeds are not loud, just not completely silent.

Click to expand...

Adding an additional 120 radiator will add a bit of restriction/noise and may not quite provide the performance you are looking for. You might want to think about a single 480 Radiator instead if you can fit it. XSPC makes a nice 480 radiator, I would go for the AX line (~$115). The best one out there is probably the Alphacool NexXxoS Monsta (~$145).

If you can't fit the 480, going with 2 radiators, you may end up wanting an additional pump. The Thermochill PA 120.3 are pretty low restriction, but radiators are still one of the highest points of restriction in your loop, especially when flowing in/out of the radiator. With just one CPU/GPU and a D5 with a performance top, you may be fine. But I know with 1 CPU and 2 GPUs and 2 120.3 radiators, I used 2 MCP655s with performance tops (mainly for barb fittings) and high flow fittings. That was just my preference.

Also I used push/pull on my radiators and hooked all the fans up to a fan controller and used dampeners (a little pia to install, but helped a bit) to help with the noise. I also bought gel packs to put under my pumps which worked like a charm.

Thanks Guys.

Ending up going this route.

http://www.sidewindercomputers.com/sicumclad5wi.html

Good choice. Sidewinders is a great place to buy from.
They do an excellent job and are very quick to help.

I have one of these on my D5 (mcp 655) http://www.sidewindercomputers.com/bid5modtop.html

On another front......I have a heatkiller cpu block, two heatkiller gpu blocks, a reservoir , a 120.3 and a 120.2 radiator all powered just fine by an mcp 655.....loop temp is a constant 40C at full load.

kickmic said:

thanks for the time to put post this - very informative.

What would classify as a restrictive loop as far as impacting temps when changing D5 from 1,2,3 and 4?

I am running 1 x D5 in a Koolance Bay res, 1 x120.3 rad, cooling just GPU and CPU - temps don't change from speeds 1 to 4, but I am in the process of adding another 120.1 rad and am wondering what effect this will have on restriction. I do use 90 degree and fittings and Koolance QDC's.

Currently, 1 is absolutely silent, but any higher speed starts to introduce vibration transfer which is inaudible when case side panels are off, but audible when on. Bay res is decoupled with silicone anti-vibration tape and side panels are decouple as much as possible using the same tape.
the noise levels at higher speeds are not loud, just not completely silent.

Click to expand...

You are welcome.

Your loop does not sound t that complicated and adding one more component should not be big issue for a D5.

Now disclaimer

I do not know what components are you using and I have no picture or video of your loop to be able to spot obvious errors so all I say I do say thinking you have done propper job. Well you are reading HardForum so you should know enough

Small radiators tend to be restrictive in some setups. Sometimes they do not help much either. At some point I was adding a radiator to every spot I could I have had my reasoning to Like not enough cooling potential (read not enough overkill). Or to lower a bit water temperature before it enters next water block etc. etc.
At this point I would always rather change main radiator for bigger or thicker version. Or even a new technology version of same size, one that simply works better.

Sure more parts in loop higher resistance pump has to overcome. Though loop can be ruined with single part. So in a way loop is as restrictive as most restrictive part of a loop. So introducing a bottle neck might slow down a loop to a point where you need to up D5 rpm's, Many times after a CPU block water speed drops the most there for some people tend to "time" their loop before and after CPU water block.

Jet another strange thing. D5 tends to sound different with different loops at same rpm's. Yes it does sound crazy. Same pump same case other components and at same pump speed and different sound signature. Seams like when car engine changes sounds when hitting a hill. With restriction comes more vibrations I assume. So I have had a loop where 4 sounded OK with x amount of components and x length of hoses and in same loop without two GPU blocks and different res. and way shorter hoses it sounds worse at 3 !???

Also major issue is to de-couple a pump from case. I have overtightened screws once and it sounded so bad. It is not enough to put some sponge under a pump if screws are not decoupled as well. Little rubber ring under a screw can help a lot. If pump is placed on the floor of the case you can dynamat a bottom of a case to. It helps.

Forgot to mention if noise of the pump increases when adding a side panels then pump is not decoupled good enough. Sure it always sounds bit louder with panels on because they are like a membrane of a speak. Though if pump is good de-coupled from a case it will not be that much louder with panels on. Decoupling a pump removes effect of a “voice coil”.

When pump is not decoupled well even sound dampening on side panels does not help that much.

Good luck.

Another thing to remember, a loop will come to an equilibrium over time.

You cannot make one segment "cooler".

The entire loop will eventually arrive at a constant temperature, derived from the function of the efficiency of the radiators and the heat impacted by the components.

I love threads like this with us all preaching to the choir.
You hear that young ones, LOTS of very good info up there ^.

magoo said:

Another thing to remember, a loop will come to an equilibrium over time.

You cannot make one segment "cooler".

The entire loop will eventually arrive at a constant temperature, derived from the function of the efficiency of the radiators and the heat impacted by the components.

Click to expand...

Main reason why order of the parts in the loop does not make much difference when tested for extended periods. If reservoir is at higher point then pump and it directly gravity feed a pump other components can be in any order and once system reaches it's balance it will be at same temperature as in any other order.

Still I like to have my order of cooling parts. Mainly I like to have res> pump> radiators> water blocks > res. This makes me sleeps better. Just my personal trip. This way I made my self believe that coolest water will enter heating elements, pick up the heat and bring it to radiators again. I found that radiators that I use restrict water flow way less then water blocks. Also seams like GPU water blocks I use are less restrictive then CPU water blocks. So there for I feel better when I can make loop so that most restrictive parts are closer to the end of a loop. Again it makes no sense because speed in one part of a loop will be just same as speed in other part of a loop. Coolant is not easy to compress. Still I believe with such order it is easier for pump to overcome loops resistance. Does not have to be so for all loops but worked good for me. See we all have our believes

Now if something works good for me is it a rule? I do not think so. Maybe if someone uses exactly same parts and loop configuration but as soon as one variable is changed it has no value for me. Only that counts then is to test it.

Water cooling components available now are so good that even bad loops perform at high level making people believe thing they do are way to do it. Most of the time people can not even reliable measurement at first place and tolerance differences bring them to wrong conclusions.

Some people make loops with big amount of coolant (long hoses, huge reservoirs etc etc), others make minimal loops. Some people use high pump speeds and some lowest speeds possible. When this works for them they think that is a rule. Well it might be but does not have to be so.

Every loop has maximum amount of heat energy that it can absorb and dissipate. Amount of heat they can absorb is directly influenced by amount of coolant in system as well as weight and materials water blocks and radiators are made of and OFC rate it is cooled, otherwise loop will reach sooner or latter it's maximum and overheat. Now amount of heat energy loop can dissipate depends on many things. Main influence is radiator surface and construction. Other elements are airflow and air temperature going over it's surface and water flow speed.

Some loops do not have enough radiator surface but have huge amount of coolant. This loops will perform good enough in short tests but once they are exposed to long term heating they will fail. Specially if water flow speed in those loops is to slow.

On other side some liquid thermal buffer is welcome. To little coolant and to high speed will not result in lower temperatures necessary. After optimal speed for that loop is reached, gain from higher flow speed is minimal.

But if loop is configured with some common sense it will regulate it self kinda. Like big loop with lot of components should have big radiators and will ask for higher hose length. Bigger radiators hold more coolant. Longer hose, more coolant they will contain. So big loops will have more thermal buffer in amount of coolant they move anyway. So only way to influence amount of coolant is to choose appropriate reservoir.

Sometimes people with thinner radiators use bigger reservoirs with good results. Well their radiators do not hold enough coolant to have decent/needed thermal buffer. This loops mostly ask for bit higher water flow speed as well. On other side people with think radiators might have good results with Swiftech microres.

There for I go with good fans and fan controller. This way I can dial in noise and airflow. If possible I go with pump that has option to regulate it's speed (D5) or with a pump with fixed speed and with pump controller. This way I can dial in enough water flow speed.

I like thicker radiators with medium to lower fin density combined with medium speed fans with high pressure numbers. Fans that are quiet when slowed down. For example if all I need is 1000 rpm fan I like to have 1300-1500 rpm version and regulate it to 1000 rpm using a fan controller. There for I like to have bit more radiator volume and surface then needed. I really do think 1000 rpm with a good 120mm fan is upper limit of what I want to have regarding noise.

I try to optimise water flow as much as I can and then I just regulate coolant speed. Then I make compromise on noise to performance and I live with it. Regarding a radiator overkill results are mostly good but over all costs are higher. Well over kill in radiator is good for future upgrades to. Only you have to keep it at some normal level. Like two 480 radiators are to much for single CPU water block

I am also using Dynamat or any other dampening product when ever possible.

I like good performance cooling with low noise if possible.

I use the opposite flow direction......

reservoir>pump>blocks>radiators

As I said, the loop will come to an equilibrium after just a few minutes.
From that point on you are merely circulating a constant temperature coolant on average.

My philosophy is push the lowest temperature coolant through the blocks, radiate all the heat and then return the coolest temperature coolant to the reservoir.

magoo said:

I use the opposite flow direction......

reservoir>pump>blocks>radiators

As I said, the loop will come to an equilibrium after just a few minutes.
From that point on you are merely circulating a constant temperature coolant on average.

My philosophy is push the lowest temperature coolant through the blocks, radiate all the heat and then return the coolest temperature coolant to the reservoir.

Click to expand...

Like I have said order does not matter for the msot part. Just what I prefer.

BTW pump dumps 18-24 watts heat in water before hitting your water blocks

There for res> pump> rad >/coolest water/> blocks >res is better LOL

It simply does not matter. Difference is mostly as big as margin of error anyways.

Pretty much its been summarized. It is important to note* Water temp equalizes in the ENTIRE loop, what does this mean for anyone? Setup your components where you want them, worry about routing last, the only real "convention" to follow is the res. being above the pump (in a vertical sense), otherwise its all preference

magoo said:

I use the opposite flow direction......

reservoir>pump>blocks>radiators

As I said, the loop will come to an equilibrium after just a few minutes.
From that point on you are merely circulating a constant temperature coolant on average.

My philosophy is push the lowest temperature coolant through the blocks, radiate all the heat and then return the coolest temperature coolant to the reservoir.

Click to expand...

I usually do Res > Pump > Rad > Blocks not because of the temp, but because typically the radiator is the biggest restriction on the loop, so I like having it closer to the pump in the loop where the flow will be strongest, I also tend to have my radiators toward the bottom of my case so I am not pushing water up and through the radiator.

I always put the system as res pump rad blocks.
Keeps the pump fed and the strongest push of flow into the rad to help bleed air from it asap.

NoOther said:

I usually do Res > Pump > Rad > Blocks not because of the temp, but because typically the radiator is the biggest restriction on the loop, so I like having it closer to the pump in the loop where the flow will be strongest, I also tend to have my radiators toward the bottom of my case so I am not pushing water up and through the radiator.

Click to expand...

It doesn't really matter which restriction comes first though, it all adds up together the same way.

only in a closed loop.

I almost went with the D5 on my last build, but decided on an Aquastream XT after some last minute deliberation, due to the reputation for silence, and easy monitoring and control via the desktop interface. Loop in sig.

At a low pump speed setting of 60Hz, the water/ambient temp delta is ~5C, with no more than 1-2C between idle and extended CPU/GPU load. In addition to what's in my sig, tubing is 10/8mm, and I'm running 4 AP15's on the rad at sub-1000 RPMs (not really audible from a few feet away). Maxing both the fans and pump speed makes only a negligible difference.

Conclusion: In my case, a non-powerhouse pump and quiet fans was more than adequate for good results.

Only thing I'd add is that even with adequate decoupling, the Aquastream is (subjectively) nowhere near silent, more of a quiet, low-pitched hum. Not bothersome to me, but maybe something to consider.

Got the D5 in, did a full loop clean after a year and a half. apparently my kill coil got overwelmed or failed and i had some growth in the loop. This pump is alot quieter than the CPX-PRO i was running. And my rigs 15c cooler than it was running to boot. She was soooo dirty.