Pump shootout - noise

shadowscotland · 6 May 2009 at 17:26

Well by a series of untimely buys I’ve now own / or will own three ‘quiet’ pumps.

A Eheim Reserator2 pump (which I know is inaudible)
300l/h 0.5m head

A D-tek DB-1
500l/h 2m head

Laing DDC-1T Pro with res top
420/3.7 stock (enhanced by top but I don’t believe it double the flow)

Now I don’t know of anywhere that actually compared these pumps with a passive watercooling loop. So for my own research and that of the community I’m going to run some loudness tests. I don’t own a dB meter but do have ears so it will be a ranking not a value. But will some other standard pc parts in the list – hdd, fan, etc (happy to add other by request/if send to me)

I don’t have an Anechoic Chamber like SPCR but do have a fanless pc I know and use daily in a semi-rural area (medium-low ambient but day, very low ambient buy night) I also have a full set of low restriction block – CPU/NB/GPU/MosFet and HHD to play with.

In every review of these pumps they are teamed with an active rad – so fan noise could easily mask the noise signature of the pump?

Was going to run some quick loop tests for noise values but could do some temps as well – would just be simple loop as don’t have the time to do lots of variations with multiple mounts and re-mounts to average out TIM issues.

If anyone has a Spare XSPC RES/Pump combo’s or D5 vario that they can lend me - I’ll do a complete set of results with all 4 main pumps (plus the eheim).

Any comments/requests while I await the laing

shadowscotland · 6 May 2009 at 17:39

Don't have the stock top sorry (will ask the seller)
As to flow - baseline pump flow has been done buy other - But can rig up something basic with my full loop and count the rmp's

shadowscotland · 6 May 2009 at 19:29

Thanks Monkey Puzzle that would be great - will let you know tomorrow.
And will try the undervolt approach mike if the pumps need some thing more

shadowscotland · 10 May 2009 at 22:50

Ok some provisional results

First off my test loop.
Pump
Koolance VL4 quickconnectors (set)
Alphacool Silentstar
XSPC Edge
Alphacool Silentstar
MIPS MosFet block
DD Maze 4
MIPS MosFet block
Silverprop Cyclone Fusion HL
Koolance VL4 quickconnectors (set)
Pump

Loop laid flat on bed - both pumps visable, D-tek with QC's attached

On the subject of Koolance - there threads are sharp, note to self always use the compression top (even when tubing is thin wall)
Did make swaping the pumps easy

Ouch! one of three cuts

Had the D-tek DB-1 in the loop first - and during the day is was inauabable, but the vibration of the pump if it touched the case was significant.
Also great flow from this little pump - was really impressed. (sorry didn't take any images) But the one bubble that was in the loop
(QC's allow air when connecting) was smashed into millions of little bubbles that happly zipped round the loop (no res to remove them)

Next up was the Laing DDC-1T Pro with alphacool res top

Loop running on anti flooding apparatus

Res top making the removal of bubbles from the loop easy.

Now this did emit a sharp but quiet hum - but it's easily masked by fan noise.
The surprise was how little vibration it emited.
I'm sure the size and weight of the filled res help but still very impressed.
I didn't think it was massively more powerfull than the d-tek
as it needed some help (moving loop around) to clear the bubble that had formed on the outlet.
Once it had moved past the first few inches of tubing it pushed the bubble so fast it broke up into foam. I guess this is a primeing issue,
exaggerated by a 'flat' loop but can't be sure.

If I can stay up I'll do some ultra quiet ambient tests - my little boy was singing himself to sleep with the above results.

shadowscotland · 11 May 2009 at 12:07

After and hour or so last night a small leak had droped the waterlevel

(didn't bother with TPFE tape for the test and the seal on one plastic barb had failed)
The resulting vortex was clearly auadible over the hum of the pump as was the 'hiss' caused but thosands of tiny fast flowing bubbles.
Again if the fan was restarted you could only just hear the water with the res top off.
Worth noting that the bubbles were moving faster that with the previous d-tek.
I almost think the lack of restriction in the loop with such a powerfull pump is causing more issues - will try and slow it down.
or find something anticyclonic that isn't my finger

@ almighty - The Laing DDC Ultra is not silent.
It might be inaudible in your system

And that great for any actively cooled waterloop.
But inaudible when next to 4 or more fans is not the same as silent.

Even the submerged Eheim is not silent

@ Firestar - Yes I've hear very good things about the XSPC pumps - Just a shame I've not got one to do a comparision test with.
Submersion really helps with the 'air-born' sound level.
But it's actually the fact that the sound energy needs to pass from liquid to solid to gas that makes the differance.
Vibration can still be an issue but easily solved.

shadowscotland · 11 May 2009 at 13:08

Why not set the vario to 5 and push all the bubbles out - then drop it back down to 3 again?

Or is it a different type of bubble trouble issue.

shadowscotland · 16 Jun 2009 at 17:41

I've a MCP655 on route so will update this when I hear it

shadowscotland · 6 Jul 2009 at 00:26

Well I finally got round to adding another common pump (D5) to my round up
regarding sound values. So am also now in the position to rank them against each
other. Will also try and merge the martin liquid lab graphs so you can see some
flow rates.

But firstly the D5 Vario (MCP655) – Well it’s quite a chunky pump and not that
far off the size of my DDC pro with the Res top fitted. The molded ½ connects on
the pump are very similar to DD perfect seal barb ends (you can see a set on the
alphacool res top above) These are not my favorite barb profile – hot water is most
definitely needed – but even without clamps are very secure (cut the tube off)

What I find odd is the D5 is always pictured with the outlet pointed up? In fact I
nearly added the EK res to the wrong side when setting up the loop. Thankfully
the D5 has an arrow molded into the barb so show direction to avoid embarrassment
of potential loop builders that don’t read instructions.

The D5 Vario has 5 speed settings (ascending) and I had no problems changing
These on the fly. Vibration is similar to the D-tek pump and only gain slightly with
Increased speed. The DDC wins hands down regarding vibration noise, even without
a clever bracket. What is significant in the D5 is a high pitched electric motor wine
that is clearly audible at 4 and 5 (dentist drill like). This can just be heard at setting 3
with your ear near the pump but ramps up considerably at setting 4 and then is quite
noisy at 5. I’m no mathematician but would say the noise level rises exponentially
and above 3 it’s not even worth considering for a quiet loop.

I’d also like to point out, that at setting 5 my EK150 res was vortexing and dragging
air back into the loop. So an anti-cyclone insert is a must with any small reservoirs
and this pump on high. There is slight vortexing at 4 but it’s not a problem.

I would add that in the same way may folks 7v there fans or general use, can be
equated here – have the power to push out any air bubbles on first instillation of a
loop or for benchmarking is very handy, especially as the pump can been run
day-to-day at a much lower speed/volume.

Well here is the ranking and it wasn’t that easy to place them as each have a very
different sound. The D5 at 4 and 5 is by far the most intrusive, due to it’s pitch but
at setting 3 the hum of the DDC become ‘noisier’ than the D5.
The vibration noise of the d-tek pushes it down the ranking –
if you can isolate it, it would be almost as quiet as the D5 at 1.

D5 setting 1 – 0.62 GPM
D5 setting 2 – 0.92 GPM
DB-1 – 1.09 GPM
-------
D5 setting 3 – 1.26GPM
DDC pro – 1.53GPM (1.86 with xspc top – noise difference unknown)
-------
D5 setting 4 – 1.63GPM
CPX-pro - 1.85GPM (untested based on others opinion)
D5 setting 5 – 1.89GPM

The gallons per minute where taken from Martin liquid lab 3.1 flow calculator and
show a direct relationship between flow and noise (but the D5 and DB-1 do need
retesting). To be honest I’m quite surprised at this so I re-worked the excel with just
my xspc in the loop and they were no longer in flow rate order so it must be specific
to my test loop.

Recommendations

Ultra quiet / fanless setup – I never thought I say this but the D5 vario come out on top
as you only need 1 gpm for a good loop and the D5 almost achieves this at setting 2
with next to no noise, but some vibration. With a good mounting strategy you can’t
lose and have the extra power available to kick start bleeding a new loop.

SFF / HTPC - The DB-1 come a very close second – and due to it’s size and low cost is a perfect
small and quiet loop pump.

Everything else - DDC-pro with a XSPC res top it really blows the rest away in flow to noise ratio.
1.8gpm with 5 blocks and a rad (my silentstar are comparable with PA / RX rads) imho is more that enought for anyone.
And doesn’t have the noise of the ultra or D5 at 4/5

Edit: CPX-pro added as untested, Recomendations broken into groups

shadowscotland · 8 Jul 2009 at 09:10

Thanks for that Ice Tea
I must admit I nearly brought one to test as it's little brother is a great pump.
Would gladly test it if some could lend me one - please

If Martin's reply on the youtube link (below) is any thing to go by it's not a silent setup pump and 48dB ambient is quiet loud (relative to the ~35dB of my house at night)
It is a much lower pitch noise than the D5 with would be much easier to mask with fan noise / AcoustiPack

Yes, it's one of the more quiet pumps that I've played around with. I think at least as quiet, and likely more silent than the MCP655/D5 at setting 5.

Using the same blocksetup in Martin's flow rate estimator as the others above the CPX-pro gets 1.85gpm.
Based on the comment above and I'll insert it above the D5@5 with an untested note.
As with the DDC ultra / Xspc bayres it's not a ultra quiet pump

Edit: updated recomendations into use types

shadowscotland · 8 Jul 2009 at 11:30

The CPX-pro has a lower pitch noise that the D5 on setting 5 (think dentist drill in next room)

At setting 1 and 2 you only hear vibration noise if the pump touches anything.
If you hold it in your hand or have it on a folded towel it's below early sunday morning ambient aka near silent
It's a strong vibration and did travel up the tube to make my EK res vibrate too.
So I'd be carefull using it with a hard mounted bay res.

shadowscotland · 8 Jul 2009 at 13:20

I've ditch all the aluminum in my loop inc the Zalman with the internal Eheim.
But my tiny ATX case is not really big enought so may stick with the DB-1.

Really wasn't expecting the D5 to be the best... some decisions to make...

shadowscotland · 8 Jul 2009 at 14:53

Twin pump loops can work but you MUST use same pump.
The APOGEEDRIVE has a DDC 3.1 build into it - your better off buying a res top and converting it to a stand alone pump. Getting a top five block at the same time and your all set. (there a good how-too artical on the web somewhere)
But it is worth starting your own thread to try and keep this on OT (your get more help that way too

)

@Monkey Puzzle - don't have the DDC any more (only got it to test now sold on) so can't check anymore.
I did want to volt mod it but just ran out of time

Also didn't test it with the stock top for the same reason

But I don't see any reason why someone wouldn't put an aftermarket top on the pump...

shadowscotland · 9 Jul 2009 at 08:11

Lol – I’m still a Reserator fan boy – But experimenting with radiant rads.

The DiscreteHeat skirting rad I had in my last loop was aluminium but I’ve found a copper core
alternative made by Heatprofile – has copper connections too and no 90 degree bends

.

Don’t have issue with a mostly aluminium loop with a single copper element (that’s coated),
but a mostly copper loop with a aluminium res is just asking for trouble.

Oh and I have plans for the Res2

Feel free to ask question in my stagnate build thread as it’s just about to be reactivated
(might even get some done tonight

@wolvers69 - nice one, I really like your build - tidy / effecient and quiet.
And I'm guessing it inaudible due to light fan noise (Noctua plus the enermax psu)
I had that psu before my Fanless yesico - was the last fan I used in a finished build

shadowscotland · 9 Jul 2009 at 10:46

Welcome to the forum coolmiester - I do beleive that is the best first post I've ever read!
Now I have to go and look up cavitation*...

Edit: Now you don't what had happending - learn someting every day - thanks coolmiester now I know what it called
*wiki had nice pic but this link was better

shadowscotland · 13 Feb 2010 at 00:22

Did some more testing tonight with some other aftermarket tops
NOTE the massive noise drop of the D5 with the ek x-top.
and yes setting 5 is quieter that 4 (but that could be my old ears not hearing higher frequencey noise as well)
and the increased flowrates but reduced noise of the xspc tops over the older alphacool one.

Full review will be posted up elsewhere but wanted you guys to get first look at the provisional results.
A few may swap a place on second run through with different loop arrangement.

D5 setting 1 (EK X-top rev2) – 0.69GPM
D5 setting 2 (EK X-top rev2) - 1.00GPM
D5 setting 1 – 0.62 GPM
D5 setting 3 (EK X-top rev2) - 1.35GPM
D5 setting 2 – 0.92 GPM
DB-1 (Ek 2.2) – 1.09 GPM
-------
Ddc stock – whoosh like but quiet – 1.37 GPM
D5 setting 5 (EK X-top rev2) - 1.99GPM
D5 setting 3 – 1.26GPM
Ddc xspc top - 1.87GPM
Ddc xspc res top - 1.86GPM
D5 setting 4 (EK X-top rev2) - 1.72GPM
DDC pro (Alphacool res top) – 1.53GPM
-------
D5 setting 4 – 1.63GPM
CPX-pro (Ek 4.0) - 1.85GPM (untested based on others opinion)
D5 setting 5 – 1.89GPM