Stren's D5 Top Testing (EK, Alphacool and Koolance so far)

[stren]

The organized version is now on my website: http://www.xtremerigs.net/reviews/water-cooling/2012-pump-roundup/single-d5-pump-tops-roundup/

Hello all I have a bunch of D5s and tops that I'm supposed to be testing. I've been busy with other stuff, but got time to ran the EK XTop CSQ Single D5 Pump Top (http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=59_439_770&products_id=34588) yesterday, pics of the top itself at the end:

Summary: Nice performance - consistently near the top and the price is right!

With a D5 Vario:

http://i.imgur.com/n2PU9ZV.png

With a D5 Strong:

http://i.imgur.com/47ZcB6M.png

How to read the plot:
- I tested 5 pump settings with a D5 vario, these are the horizontal-ish blue/red/green lines, the higher the line the higher the pump setting
- The orange lines are estimates of three "typical" loops, one basic cpu loop, one more average cpu + 2 gpus in parallel + 2 rads and some rotary fittings type loop, and then one "extreme" loop with one cpu, multiple series gpus, multiple rads, a pair of QDCs etc etc
- Choose an orange line based on your approximate setup, and then see where it crosses the RGB lines and this will show the difference in flow between the vanilla top and the EK top
- Alternative output refers to the optional output port on the top of the top. This has an internal 90 degree bend that should reduce flow

Conclusions:
- EK top increases flow slightly where it counts
- Gains are small (at best 10% but usually less and at worst not much at all)
- Alternative output port did not hurt as much as I thought it would

Pics:

http://i.imgur.com/uA6w3url.jpg

http://i.imgur.com/VFQW5Vll.jpg

http://i.imgur.com/6rwFiO5l.jpg

http://i.imgur.com/iVOGicRl.jpg

http://i.imgur.com/biPEFeil.jpg

http://i.imgur.com/IkHZOynl.jpg

http://i.imgur.com/U9MW7r0l.jpg

Under test:

http://i.imgur.com/haMxgO2l.jpg

http://i.imgur.com/9LgZnfxl.jpg

[stren]

Alphacool HF D5 (http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=59_439_771&products_id=31996)

Summary - A nice performing high flow top, not a fan of the o-ring clamping though. Higher flow, an extra input port and$7 less than the EK, but I did prefer the EK in terms of feel and assembly.

O-Ring drama - all the EK tops I've used have always felt very robust in terms of o-ring positioning and I've never had a leak. This one didn't and I had to really tighten down the screws more than I liked in order to fully seal the unit (first attempt had a slow leak). In fact it was so tight that the allen key slipped in one of the bolts and I nearly stripped the bolt. When I think about it seems like the clamping mechanism should be the same as EK's but I just didn't get the same feel from it :confused: EDIT - I took a look and the design of the block clamping uses the exact same idea, however the o-rings are different so I think that must be what I'm feeling. EK's feels a bit thicker and squishier.

Results:

Slightly higher in the region of interest (0.75gpm - 1.5gpm) than the EK. The alternative outputs' flow is a bit worse relatively because it has two elbows, one for the alternative input and one for the alternative output. Still impressive that it out performs the vanilla top even with two elbows!

Performance with a D5 Vario:

http://i.imgur.com/FWLSRxl.png

How to read the plot:
- I tested 5 pump settings with a D5 vario, these are the horizontal-ish blue/red/green lines, the higher the line the higher the pump setting
- The orange lines are estimates of three "typical" loops, one basic cpu loop, one more average cpu + 2 gpus in parallel + 2 rads and some rotary fittings type loop, and then one "extreme" loop with one cpu, multiple series gpus, multiple rads, a pair of QDCs etc etc
- Choose an orange line based on your approximate setup, and then see where it crosses the RGB lines and this will show the difference in flow between the vanilla top and the EK top
- Alternative output refers to the optional output and input ports. These have an internal 90 degree bend that reduces flow

D5 Strong testing was not completed as one of the clamp bolt heads stripped when tightening further to avoid a slow leak.

Pics:

http://i.imgur.com/ZGV9KhLl.jpg

http://i.imgur.com/JXCALwBl.jpg

http://i.imgur.com/5z0l0lPl.jpg

http://i.imgur.com/2ZBMsd0l.jpg

http://i.imgur.com/4Oykyj8l.jpg

http://i.imgur.com/OwJ7uiql.jpg

Under Test:

http://i.imgur.com/7krcospl.jpg

[stren]

Ok this time the Koolance:

Koolance COV-RP450 D5 Top (http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=59_439_772&products_id=26481) + 80mm (fat) x 120mm tall reservoir (http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=59_1165_368&products_id=29833) with base adapter (http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=59_1165_368&products_id=29836), and top with fill port (http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=59_1165_368&products_id=29829)

Summary: A very nice high flow top that also covers the pump body and can attach a reservoir directly to it.

Nice and high flow, with a nice meaty feel to the design. Metal plate used to clamp the pump in which means less screws than alphacool or EK. I do wish the screws were countersunk into the plate however. Covering the pump body is a nice touch as labels can be large and distracting aesthetically. Much more affordable than Bitspower's overpriced top + dress kit because this is $45 for both. (Additional $37 however for the reservoir parts)

Flow data:

How to read the plot:
- I tested 5 pump settings with a D5 vario, these are the horizontal-ish blue/red/green lines, the higher the line the higher the pump setting
- The orange lines are estimates of three "typical" loops, one basic cpu loop, one more average cpu + 2 gpus in parallel + 2 rads and some rotary fittings type loop, and then one "extreme" loop with one cpu, multiple series gpus, multiple rads, a pair of QDCs etc etc
- Choose an orange line based on your approximate setup, and then see where it crosses the RGB lines and this will show the difference in flow between the vanilla laing top and this top
- Alternative output refers to the optional output port on the top of the top. This has an internal 90 degree bend that should reduce flow

D5 Vario:

Just the top without any reservoir:

http://i.imgur.com/qQc2bvf.png

With Reservoir:

http://i.imgur.com/RHXNYX8.png

D5 Strong

Just the top without any reservoir:

http://i.imgur.com/CeQ8Gtv.png

With a reservoir:

http://i.imgur.com/a9n2VFs.png

Pics:

http://i.imgur.com/50mk9otl.jpg

http://i.imgur.com/TMLHpsVl.jpg

http://i.imgur.com/dzDsDuDl.jpg

http://i.imgur.com/IzhRtczl.jpg

http://i.imgur.com/TUHXJNwl.jpg

Reservoir - 3 parts - top, body and base:

http://i.imgur.com/HvLWuAul.jpg

Assembled:

http://i.imgur.com/CQoRN3vl.jpg

Adapter because this is the fat version:

http://i.imgur.com/vEav08Jl.jpg

Attached to the D5 top:

http://i.imgur.com/n1kascel.jpg

During Testing:

http://i.imgur.com/88fDeUrl.jpg

http://i.imgur.com/VnMh0Xtl.jpg

http://i.imgur.com/Rz94Ij0l.jpg

http://i.imgur.com/eiz9fdql.jpg

[stren]

Friday night update
- Added Koolance top with D5 vario data
- Added D5 strong data for the EK top.
- I tried to run the alphacool with the D5 strong, but again I had leaks, and again the screw head stripped. This time irreparably :( Either the o-ring is bad or I'm just assembling it wrong because that o-ring is not sealing it. For now I would avoid this top until there is confirmation that I did something wrong.
- Added more pics

[stren]

EK Older Style Rev 2 top:

Summary - Same performance as the Rev 3 CSQ style top meaning that Rev 3 was purely an aesthetic change

- I tested 5 pump settings with a D5 vario, these are the horizontal-ish blue/red/green lines, the higher the line the higher the pump rpm setting
- The orange lines are estimates of three "typical" loops, one basic cpu loop, one more average cpu + 2 gpus in parallel + 2 rads and some rotary fittings type loop, and then one "extreme" loop with one cpu, multiple series gpus, multiple rads, a pair of QDCs etc etc
- Choose an orange line based on your approximate setup, and then see where it crosses the RGB lines and this will show the difference in flow between the vanilla laing top and this top
- Alternative output refers to the optional output port on the top of the top. This has an internal 90 degree bend that should reduce flow

With a D5 Vario:

http://i.imgur.com/TzyRIMB.png

With a D5 strong at Min/Max on a 24V Koolance controller:

http://i.imgur.com/26s9paM.png

Rev 2 vs Rev 3 with a D5 Vario:

http://i.imgur.com/kZifis8.png

Bear in mind settings 2,3,4 are selected on an analog dial and manually set by me with a screwdriver. I try and match RPM from previous runs. It's not perfect data therefore as it's more influenced by me than settings 1 and 5 which are as far as I can turn the dial.

Pics:

Testing the alt output:

http://i.imgur.com/Iy2ZG50l.jpg

http://i.imgur.com/XRk9u3cl.jpg

http://i.imgur.com/MmvrXP9l.jpg

Testing done:

http://i.imgur.com/8jp1DPal.jpg

Rev 3 vs rev 2 clamp, holes line up, just a square vs rounded corners:

http://i.imgur.com/ueWIcUgl.jpg

[stren]

Got some more testing done with the D5 Strong:
- Added the Koolance D5 Strong data (above)
- Added the EK Rev 2 D5 Strong data (above)
- Added the XSPC Dual Bay Reservoir (below)
- Added a combo plot of all tops so far with the D5 strong (at the end of this post)

XSPC Single D5 Dual Bay Reservoir

Summary -

Nice looking update of the XSPC D5 dual bay reservoir. No more acrylic glued together is a good thing. However the performance is no better than the vanilla laing top which is a missed opportunity. Fill port on top is oversized but I'd rather see two individual G1/4 fill ports like the monsoon. On the other hand it's only $55 (http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=59_318_665&products_id=34300) which is pretty affordable.

Performance Data:

Performance Data with a D5 Vario Pump:

http://i.imgur.com/UzdQDhD.png

Performance with a D5 Strong Pump:

http://i.imgur.com/yWhmc51.png

Pics:

http://i.imgur.com/4MwtMaRl.jpg

http://i.imgur.com/RXWAcySl.jpg

This version comes with the pump, the $55 version at PPCS does not come with a pump.

http://i.imgur.com/3PCV77ml.jpg

http://i.imgur.com/bjB2hJql.jpg

http://i.imgur.com/Fu8fHN7l.jpg

Pics of volute:

http://i.imgur.com/vmXeBCfl.jpg

Pics of identical standard laing D5 top volute:

http://i.imgur.com/RnmmbDfl.jpg

[Church]

Imho we will see more performance gains over stock for DDC pumps with 90deg on stock top's inlet. No drama for D5 with not that bad stock one.
BTW, Stren: how about commenting on if not measuring about your impressions if noise felt/sounded louder or quieter, how easy it was to mount in case, how easy it was to fill/deaerate loop relative to others?

[stren]

Imho we will see more performance gains over stock for DDC pumps with 90deg on stock top's inlet. No drama for D5 with not that bad stock one.
BTW, Stren: how about commenting on if not measuring about your impressions if noise felt/sounded louder or quieter, how easy it was to mount in case, how easy it was to fill/deaerate loop relative to others?

Yeah this might be true though even the tops with a 90 degree bend at the input showed little loss in performance. The DDCs will have to wait til I'm done with all the D5's. Then the "other" pumps will follow after that :)

I've been taking noise measurements and it's hard to do it in a realistic way as I'm not mounting it to the case for these measurement runs. To avoid resonance of whatever the pump is sitting on (granite counter) I suspend it by the tube during the measurement. However then the pump is often so quiet that the top doesn't matter (Below Setting 4 on a vario, ambient noise becomes a huge issue and I have to wait for the house to get quiet). I've maybe seen 1dB amongst the tops, and usually the quieter ones are the ones with the attached reservoir presumably because of the extra mass of water. On the other hand if I don't isolate the pump from the surface it sits one then noise numbers will jump easily 6-10dB. So resonance is clearly a bigger factor than the top and at least as big as the RPM itself. I think the easiest way might be to detail what noise suppression if any is included in the mounting kit. Resonance really depends so much on a case by case basis though that what works for me may not work for you. Any suggestions on this would be appreciated! It has made me plan on trying to have my pumps unmounted and supported purely by tube for my next build iteration.

I feed the pump under test with a 400mm reservoir, so bleeding of a simple top is easy no matter what. When a reservoir is also involved although it makes it harder it's still simple. I think a fairer test would be to try them with a real loop without the extra reservoir. I think results will be similar though between all the regular tops when directly fed from a reservoir vs the bay reservoirs which will always be more of a hassle. For example the Monsoon has two fill ports so you can easily put water in one and let the air push out the other (less overflow). On the other hand the old Koolance bay reservoirs (rp452x2 v1.3) have the front mounted fill ports meaning you can get the pumps onto their back to make your life easier. That one I've tried in my own builds before and that was a major PITA despite that feature.

[Justintoxicated]

Looks like GPM isn't affected by the tops as much as pressure. Isn't GPM more important, or I guess that depends on what blocks your using?

[stren]

Looks like GPM isn't affected by the tops as much as pressure. Isn't GPM more important, or I guess that depends on what blocks your using?

Both are important lol, but yeah it depends on your loop. Your kind of setup will be similar to the middle orange line which moves from about 1.35gpm to about 1.45gpm on a D5 Vario on setting 5 across all those tops that I tested. So yeah the top only make a small difference in most cases. I did run the AquaComputer top with and without the aqualis top today (though I haven't put the data into excel yet) and that one looks surprisingly poor though I really need to plot it to see where it really lies. To me D5 tops are more about aesthetics, or about using a different tube size.

Tomorrow I'll take the data on the monsoon it's bleeding overnight. The inbuilt LEDs on the monsoon are super cool too.

[Justintoxicated]

Both are important lol, but yeah it depends on your loop. Your kind of setup will be similar to the middle orange line which moves from about 1.35gpm to about 1.45gpm on a D5 Vario on setting 5 across all those tops that I tested. So yeah the top only make a small difference in most cases. I did run the AquaComputer top with and without the aqualis top today (though I haven't put the data into excel yet) and that one looks surprisingly poor though I really need to plot it to see where it really lies. To me D5 tops are more about aesthetics, or about using a different tube size.

Tomorrow I'll take the data on the monsoon it's bleeding overnight. The inbuilt LEDs on the monsoon are super cool too.

And just when I was about decided on the aqwacomputer pump / reservoir!

[Avathar77]

Any plans on testing a Bitspower Mod Top V2?

[stren]

And just when I was about decided on the aqwacomputer pump / reservoir!

Lol IKR, I've taken the data on the monsoon, putting them both into excel now


Any plans on testing a Bitspower Mod Top V2?

Not right now, next time I talk to BP I'll ask them again but they weren't interested in participating last time. Sometimes seeing results from existing companies can change people's minds though

I do have the old BP dual top (http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=59_439_773&products_id=28420) though which I'll be running

[Avathar77]

Not right now, next time I talk to BP I'll ask them again but they weren't interested in participating last time. Sometimes seeing results from existing companies can change people's minds though

I do have the old BP dual top (http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=59_439_773&products_id=28420) though which I'll be running

Even better :)

[stren]

Aquacomputer D5 Top with Integrated Fill Level Sensor & Aqualis Reservoir with Fountain Effect & Nano Coating

Summary - Nice chunky top with a cool if debatably not very useful fill level feature. Performance however is surprisingly low. I was thinking nearly all the tops are equal only to have someone surprise me with a design worse than stock. The aqualis fountain effect is neat and has only a slight effect on performance. A glass body to the reservoir is nice and so this I can recommend if you want something cool to look at and the bubble guard works well too. I'd suggest if you want the reservoir though that you get the standalone version and just buy a different D5 top instead.

Performance:

- I tested 5 pump settings with a D5 vario, these are the horizontal-ish blue/red/green lines, the higher the line the higher the pump rpm setting
- The orange lines are estimates of three "typical" loops, one basic cpu loop, one more average cpu + 2 gpus in parallel + 2 rads and some rotary fittings type loop, and then one "extreme" loop with one cpu, multiple series gpus, multiple rads, a pair of QDCs etc etc
- Choose an orange line based on your approximate setup, and then see where it crosses the RGB lines and this will show the difference in flow between the vanilla laing top and this top
- Alternative output refers to the optional output port on the top of the top. This has an internal 90 degree bend that should reduce flow

D5 Vario:

http://i.imgur.com/s6LJbel.png

Same thing but with a D5 strong at Min/Max on a 24V Koolance controller:

http://i.imgur.com/M4QRSaG.png

Both "normal" and "alternative" outputs perform well below the stock laing top across the range of interest. Here I was thinking XSPC was a bit lazy for making an exact copy of the Laing top, but this is worse than that. I would theorize this is because the volute output does not go direcly to the output port but has a double 90 degree dogleg in the path.

Comparing the top using the "normal" ports with and without the Aqualis reservoir with the fountain effect shows a small performance hit because of the reservoir:

http://i.imgur.com/FpOk90f.png

This is a larger effect than we saw when introducing the Koolance reservoir and this can be most likely blamed on the fountain feature. It is therefore not free, but it is cool looking so you need to pick your poison.

Pics:

Pump Top:

Internals:

http://i.imgur.com/nmUjj2Yl.jpg

http://i.imgur.com/O0NDOFPl.jpg

http://i.imgur.com/y8foqdul.jpg

http://i.imgur.com/r3g3MYWl.jpg

Ports on top from left to right: Output, Input, Fill Level Sensor pressure sense point

http://i.imgur.com/5Y9udTQl.jpg

Aqualis Reservoir:

http://i.imgur.com/tcL88jnl.jpg

http://i.imgur.com/hHiwJnPl.jpg

http://i.imgur.com/FRw0jQAl.jpg

Together:

http://i.imgur.com/8YCJlYp.jpg

Left side Input/Output Ports, right side fill level sensor

http://i.imgur.com/qKbj3Lwl.jpg

Left side fill level sensor, right side mounting bracket screw holes

http://i.imgur.com/em70NSul.jpg

[stren]

- Took D5 Strong data on the Monsoon S2 Single D5 Dual Bay Reservoir
- Took D5 Strong data on the tops with their reservoirs so that bay reservoir data would be "fairer" i.e. apples to apples
- Waiting on my RPM meter to come back so I can test with the Vario

Monsoon S2 Single D5 Dual Bay Reservoir

Summary - A beautiful reservoir with some cool features. Lots of design work to isolate vibration from the case means that once mounted in a case this should be quieter than other bay reservoirs. While performance is necessarily compromised when comparing to a top and a cylinder reservoir it still performs well and the price is right for all the features you get.

Features
- Programmable lighting that remembers it's last setting. Lighting has two different level settings and seven colors.
- Mounting screws do not attach to the body but use a rubber compression shoe to isolate the case from pump vibrations
- Pump lockring covers the entire pump and has a removable foam covered backplate that further isolates noise
- Two sets of ports so you can run two parallel loops off the same pump though I really wouldn't bother with this.

Other remarks
- Bottom ports perform slightly better than the top ports
- If you run a D5 Strong at 24V (6K rpm) you may hear a noise increase by 4dB when operating around 1GPM. I would suggest sticking with a D5 Vario/Basic/PWM (Max RPM 4800rpm)
- Red O-ring performance difference is small
- Two fill ports means you can pour water in one and let the air escape out the other. This makes bleeding and filling much easier.

Store link as shown: http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=59_318_1235&products_id=37659

Data:

Performance:

- I tested 5 pump settings with a D5 vario and two settings with a D5 Strong on a Koolance 24V controller. These are the horizontal-ish blue/red/green lines, the higher the line the higher the pump rpm setting
- The orange lines are estimates of three "typical" loops, one basic cpu loop, one more average cpu + 2 gpus in parallel + 2 rads and some rotary fittings type loop, and then one "extreme" loop with one cpu, multiple series gpus, multiple rads, a pair of QDCs etc etc
- Choose an orange line based on your approximate setup, and then see where it crosses the RGB lines and this will show the difference in flow between the vanilla laing top and this setup

D5 Vario Data

http://i.imgur.com/UzdQDhD.png

D5 Strong:

All iterations of o-ring and ports:

http://i.imgur.com/56qcqAh.png

Here we see that for most setups the Monsoon outperforms the standard Laing pump top with a basic cylinder reservoir across the range of interest. The bottom ports are slightly better than the tops and the red o-ring is slightly better than the black. Because the difference is small I'd use whichever you are most comfortable with.

While the bay reservoir is necessarily compromised and can not compete at the same level of performance as a separate high flow top and reservoir it does compete well. This is ok because I wouldn't recommend *any* bay reservoir for more extreme loops.

Pics:

Box:

http://i.imgur.com/d177AfGh.jpg

http://i.imgur.com/DVY5B9th.jpg

Opened up:

http://i.imgur.com/38GI1Xuh.jpg

All the allen keys are supplied:

http://i.imgur.com/mCd2TpCh.jpg

Color coordination on the screws is a nice touch!

http://i.imgur.com/FtgShmhh.jpg

After looking at the instructions I found that this is actually upside down, I still need to replace this with a right side up photo!

http://i.imgur.com/1RDw7qGh.jpg

I.E. The fill ports preloaded with silver bullets should be at the top and the bank of LEDs should be a the bottom

http://i.imgur.com/4u8fpe6h.jpg

The reservoir is made up of five main pieces:

- 6mm Thick powder coated Aluminum plate
- Acrylic window that is cast into the main body
- Main body in cast glass fibre reinforced nylon
- Back piece in PVC to absorb vibrations that can be swapped for a multi port or a different type of pump mount
- Locking metal pump cover with molded silicone mounting hardware

The large pump cover is threaded and comes off quickly:

http://i.imgur.com/wAEi2GMh.jpg

The part of the top that sits close to the impeller is not angled in the same way that many D5 tops are

http://i.imgur.com/Y9JA90fh.jpg

The fill ports have a "fence" around them so that small spills are contained

Close up of the volute which is less tapered than the EK/Koolance/Alphacool:

http://i.imgur.com/jOVBZ6Vh.jpg

You can also see just how small that red o-ring is.

Body:

Lighting:

Pics of the lighting system. Left button cycles off -> low -> high. Right button cycles colors (Green -> light blue/teal -> blue -> blueypurple -> white -> red -> yellow -> lime green)

Firstly Low vs High - pics taken at 8am with a lightly overcast sky. White balance set to daytime, all settings constant through the photos:

http://i.imgur.com/ussgbkHm.jpg http://i.imgur.com/RFbZO6gm.jpg

Colors (the weird lines that looks like scratches are small bubbles as they spin about the reservoir):

Green:

http://i.imgur.com/RFbZO6gl.jpg

Light Blue/Aquamarine/Teal kind of color:

http://i.imgur.com/IBSBgKBl.jpg

Darker Blue:

http://i.imgur.com/LsgteEzl.jpg

Bluey/Purple (looks more red in reality, but still on the blue side of purple)

http://i.imgur.com/2Qz2anul.jpg

White (has a slight blue tint, but not as bad as the photo):

http://i.imgur.com/nzDP3jll.jpg

Red (is a bit dimmer than the others)

http://i.imgur.com/k1vcdPHl.jpg

Yellow:

http://i.imgur.com/SS3G59Hl.jpg

Lime Green:

http://i.imgur.com/PlRRs7El.jpg

Still to do:

- Noise numbers from a case
- Pic showing the top of the pump
- Pics showing the compression mount screws in more detail
- Pics showing the removal pump cover end cap being removed

[stren]

Updated all the posts with missing data plots now that all data is complete. Here are the comparison plots:

Performance Curves

Only Tops with D5 Vario:

http://i.imgur.com/sqSbDUK.png

Tops + Reservoirs and Bay Reservoirs with D5 Vario:

http://i.imgur.com/671XAXA.png

Only tops with D5 Strong:

http://i.imgur.com/UGk4tnq.png

Tops + Reservoirs and Bay Reservoirs with D5 Strong:

http://i.imgur.com/eROwPlH.png

"Flow in a Simulated Loop"

These are expected flow rates for each top vs pump setting for a simulated loop that's made from flow rate data from various parts. In this case the simulated loop is the same as the "Medium Restriction" orange lines on the pressure/flow plots. This just attempts to show the kind of differences you might see with the tops:

http://i.imgur.com/KcjuxS3.png

And because the differences are small, here's a zoomed in version:

http://i.imgur.com/6hQI1TM.png

Noise

This was an attempt to plot noise vs flow for the medium restriction loop. This is less useful here as all the tops are very similar but will be more useful when comparing to other pumps. Bear in mind this is noise measured when the pump is isolated from a chassis. Chassis vibration and resonance can cause large amounts of noise and increase these numbers by 6-10dB. It should be noted that the Monsoon is the only part supplied with a vibration reducing mount. This should help, but every case and build will resonate at slightly different frequencies and changing your pump speed either up or down may reduce noise.

Also bear in mind the ambient noisefloor varies between 32dB and 35dB. For the lower noise numbers I often have to wait a while in order to get a good reading. Even though 32dB is about the limit of my meter, if you have sensitive ears you can hear a bit quieter than that. I.E. you can hear a difference between setting 1 and 2 on a pump if you listen carefully and if the room is quiet. However it's extremely hard to hear setting 1 at all unless it's resonating with something else. For most users with a well isolated pump and modestly quiet ambient noise it will be hard to hear any of the pump settings below 4 on a vario.

http://i.imgur.com/ZLX2EdE.png

Conclusions

- For performance avoid bay reservoirs
- A good top will increase performance slightly over the Laing top
- The Koolance and EK tops are both good choices, they perform well and are priced fairly for their features
- I'm not comfortable with the torque required to avoid leaking with the Alphacool top
- Both the XSPC and Monsoon bay reservoirs perform decently. The Monsoon has more features, but a higher cost

[Justintoxicated]

Great review, and thanks for doing this, looks like I'll be grabbing an EK or Koolance top. Currently, I have my stock D5 Lang wrapped in closed cell foam (I had some leftover from car audio install) which has made it virtually silent (especially from outside the case) for the past 6-7 years. It's not much for aesthetics wrapped in foam, unless you like that laboratory experiment type look, but if you hide your pump then who cares as it greatly reduces all pump noise. :)

[stren]

Thanks - yeah D5s can be super quiet when you get them setup just right. I'm moving on to the dual d5 tops now. I also relocated my pump testing to an understairs closet as I can get much better noise measurements. Ambient noise is now around 30dBA vs 32-35dBA in the kitchen. Much less space, and I had to build a 10 ft molex extension from the PSU thats running further away. Still at least my wife will be happy that I'm not using up all the counter space again!

Hopefully tomorrow I'll get the last of the pics done on the single tops and write the review up properly on my site.

[stren]

I ran the BP dual D5 top today, initially I compared the data against the single Laing top multiplied by two. Because the Bitspower top was significantly lower I thought maybe the two pumps wouldn't scale 100%, so I ran the Laing tops on the two D5 strongs also and scaling/superposition does work as expected:

http://i.imgur.com/MDQ7Lfe.png

The D5s at 24V are pretty darn strong. For reference Martin measured the Iwaki RD30@24V to have a shutoff gain of 14.7psi:

http://martinsliquidlab.i4memory.com/img/IwakiRD30PQ-24v.png

I'm not sure yet as to why the BP top is performing so poorly. It could be the compromised nature of a dual top (we'll know once I test the EK dual top).

[stren]

I next ran the EK dual top (green line) to see if dual tops were inherently compromised. I compared the data also to two individual tops multiplied by two (red line):

http://i.imgur.com/cdudkiz.png

Conclusions:
- a dual top can be as efficient as two single tops
- EK > Laing > Bitspower
- It's also worth noting the EK and BP tops were both ~2dB quieter than the Laing top when running @24V (38dBA vs 40dBA)

[stren]

Put the top roundup on the website: http://www.xtremerigs.net/reviews/water-cooling/2012-pump-roundup/single-d5-pump-tops-roundup/

Next up the reservoir combos, and then back to testing dual d5 setups...

[SylarPowers]

Wow, I may need to rethink my Aqualis setup.