water pump comparison: eheim1250 vs eheim 1048

[sunama]

(for the long story, read on, if u wanna go straight to the results, scroll down to the "Test (cooling) Results" section).

eheim pump specs
model . .Output ................................. Dimensions (L x B x H mm)
1046 .... 300 lit/h (79 US gal/h) ............ 145x75x103
1048 .... 600 lit/h (158 US gal/h) ............147x75x118
1250 .... 1200 lit/h (317 US gal/h) ..........180x95x120
1060 .... 2280 lit/h (602 US gal/h) ......... 218x116x120

History: air cooling and why i switched to water
so ive been wanting a super quiet rig for quite sometime, and it seemed that watercooling was one of the best ways to achieve this. last year, i was giving serious consideration to going the watercooling route, however the cost was just too prohibitive. however, during the last 2 months or so i made the commitment to put together a great watercooling setup regardless of the cost.

i wanted 2 things:

noise reduction (over air cooling); and
superior overclockability;

in that order.

so i put together my watercooling setup and initially i was pleased with the setup (eheim1250, maze2, DD chipset block, koolance hard disk block, heatercore, surgical tubing, 1/2" throughout, inline, not submerged). when using air cooling, i was able to run my amd 1ghz cpu @ [email protected] using a thermoengine and a thermal take 60mm fan running at around 2000rpm. this setup was much much quieter compared to the delta38@7v which allowed me to run my cpu at [email protected]. However, although i was getting quieter, i was no where near silent and as Summer approached there was no way that i would be able to cool a [email protected] using a 2000rpm fan. the answer: watercooling.

after spending £250 i had the watercooling system in place. this allowed me to run my cpu @ 1.54ghz @ 1.85v. however, due to the water heating up, i had to lower the voltage and hence drop my cpu speed down to 1.48ghz. now after spending this sort of money, i demanded that not only should i be able to overclock to higher speeds than air, but i should be able to do this with absolutely no noise at all (ie. with silence). the first thing i did was to put the 2 fans on my radiator on a rheostat and run these down to around 6v. this made it so that i couldnt hear when these fans were on or off (ie. silent). however, i could still hear something. i could definetly hear the seagate baracudas but also i could hear that damned pump. i run the pump 24/7, which means when i switch off my pc, i expect total quiet. unfortunately i couldnt achieve this with the eheim1250.

the definition of silence
definition of silence: if im testing some computer peripheral, i would define it to be silent if: i can switch it on, and not be able to tell whether it is on or not by noise alone. i normally close my windows and concentrate hard on listening to the noise produced when i switch on a peripheral (making sure that i hold my breath, so i can hear any noise). if i can get it to a stage where i switch it on, yet cannot hear a change in the noise level then, i conclude that it is "silent". this definition is what i shall use for the remainder of this comparison.

eheim 1250
i tried to make this pump silent. i ran this pump without my pc on so that any noise being created around my pc was coming only from the pump. the double glazed windows were closed so that i could really get some quiet and listen to the pump.

now the problem with this pump is that it vibrates. it vibrates alot. and if placed on a hard surface it will vibrate so much that u can hear it outside the room. so to make it quiet, u gotta place it on some vibration absorbing material. now if done correctly, this really does make the pump quiet. however, i was unable to make it silent. i used bubble wrap, sponge, clothing, but for the life of me i was unable to make it silent. at one stage i put some clothing on top of the pump and this worked, but discovered that the pump was in fact getting very warm/hot, so i had to remove the clothing. this made the pump audible again .

while this is all happening, i was reading posts that people have got their 1250s silent. i then began to think that their definition of silent wasnt as strict as mine (some people seemed to think that panaflow L1A 120mm fans @ 12v were silent (30db). as far as im concerned, 30db is NO WHERE NEAR SILENT. i would guess that my fans at about 20db are silent...30db is what i would call LOUD and im also beginning to think that it is these same people who are also suggesting that the eheim1250 is silent).

i asked for some advice from different people and some suggested that i would be able to easily get away with using a eheim1048. also on procooling.com, the guys who post on the forums are really into watercooling and while most seemed to suggest the use of the eheim 1250, one guy (web medic) who tested many pumps and rads was really adamant that any pump that whose flow rate was over 90 US gal/h, was just a waste and that in all his test the best flow rate was around 60-90US gal/h. [note: eheim1250 flowrate=317US g/h, while the 1048 flowrate=158 US g/h)]. no one argued with this guy cuz he was shutting all the "high flow rate pump lovers" up and he seemed pretty sure of his finding. moreover, no one else seemed to have tested as many pumps/rads as this guy. also, i was told that the 1048 was quieter than the 1250. from what i had read, i decided to downgrade my pump to the 1048. this would give me lower flow rate but lower noise, a tradeoff that i was willing to take. if the lesser pump gave me an increase of anything upto 1-2C in cpu temps with quiter operation, i would be pleased. anything more than this and i would probably make do with the 1250.

so i ordered the eheim1048 and it arrived today (well, yesterday). i ran some tests with the 1250, swapped it for the 1048 and then repeated the tests.

Test (cooling) Results
computer system: see sig
cooling system:
eheim1250 & 1048
heatercore matrix baught from my local motor store
maze2
DD chipset waterblock (cooling the chipset)
koolance hard disk waterblock (cooling 2 hard disks)
surgical grade tubing (1/2"
350ml plastic reservoir
2 L1A 120mm panaflow fans (@6v) in a pull-push arrangement on the rad, shrouded
water wetter
deionised water

rad and pump both outside the case, inline.

test procedure:
windows were first closed. then the system was left idle for 30mins. idle temps were then recorded.
then i ran prime95 for 60mins. full load temps recorded using a the thermal probe on my motherboard.

when i changed pumps, i did a straight swamp: stopped the pump, then pulled out the tubes and plugged them into the next pump, trying to keep water leakage to a minimum (i think i lost about 3 tablespoons of water during the pump change).

as summarised in the table below. all temps taken in degrees C.

ambient temp throughout was 21.5C.

pump..........idle(C)...........prime95(C)
1250.............37.................41
1048.............37.................40

so the 1048 seemed to give me lower temps (1C at full load) than the 1250, used in an idetical setup.

Noise
this was the main reason for shifting to the 1048. how did it compare to the 1250?

eheim1250
at its quietest i managed to get the 1250 so that i couldnt hear it more than 3 ft away, however, when i would put my ear 6" from the floor (anywhere in the room), i could hear a very low pitched rumble. when i would goto sleep cuz the pump would run 24/7, i could hear a feint low pitched rumble-quiet but audible! keep in mind that if u r going from air cooling (say a delta38) and moving to watercooling, u will probably think this thing is very quiet, almost silent. note that i am being very very strict with the noise requirements and most of u wont be as strict as me, so having a 1250 rumbling in your case may not bother u so much, as it would me.

eheim1048
although this pump did vibrate, it vibrated with less intensity than the 1250. at its quietest this setup is inaudible while lying in my bed. ie. when i switch it on, i cannot tell if it has been switched on or not. it only becomes audible from about 1ft away from the ear. when sitting at my pc, i cannot tell that the pump is on by noise alone.

the 1048 is significantly quieter than the 1250...not by a massive amount, but i was able to tame the noise (ie. keep changing the position of the pump until silence was achieved) of the 1048 in less than 15 mins whereas the noise of the 1250 could not be tamed at all, even after 3hrs or so.

the 1048 wins here!

power usage
the 1250 uses 28W
the 1048 uses 10W

although your electricity bills arent going to be reduced by probably more than £5/month, the 1048 does indeed use less power and will save u some cash, compared to if the 1250 was used. the 1048 beats the 1250 again here.


size
1048 measures 147x75x118 (mm)
1250 measures 180x95x120 (mm)

the 1048 is most definetly smaller, i would estimate in volume, its about 50% the size of the 1250. its lighter too. to me this makes little difference, however, for people who have all their stuff in the case, the smaller the pump the better, especially for people who dont have particularly large cases.

i believe the 1048 wins again here.


Flow Rate
at last this is something the 1250 can shine in.

the 1250 has a flow rate of 1200l/h
the 1048 has a flow rate of 600l/h

the 1250 can pump out double the amount of water that the 1048 can pump out.

1250 has double the performance of the 1048. 'nuff said.


Conclusion
so all in all i would say that i did the right thing in buying the 1048. since i didnt take any performance hit by going from the 1250 -> 1048, i would love to try out the 1046. we can summarize the comparison thus:

noise: 1048 is quieter (vibrates with less force) than the 1250
flow rate: 1250 has double the flow rate of the 1048
cpu temps: 1048 gave me a lower full load cpu temp (by 1C) than the 1250
size: the 1048 is about 50% of the size in volume of the 1250
power usage: the 1048 uses less than half the power of the 1250
price: the 1048 is cheaper by about £9.

i would suggest that for anyone deciding on which pump to buy, unless u r going to be running very long distances of hoses and are going to have alot of 90 degree bends go with the 1048. even with 3 waterblocks (soon to be 4, gpu waterblock will be added later this week), a heatercore and reservoir, the eheim1048 had no problems at all matching and surpassing the cooling performance of the 1250. i cant believe that people will be using more waterblocks than what i have tested these pumps with. also note that i was using 1/2" (12.5mm) tubing. if using narrower tubing, perhaps a higher output pump would be better due to the greater resistance that narrower tubing would create for the water. in my system i have no 90 degree bends and have tried to keep the hose pipes with as little bend as possible, and where there are bends i have tried to make the angle as obtuse as possible. also keep in mind that these pump findings are limited to my watercooling system alone. different people may well get different results (i think one guy said that he got lower cpu temps when he went from 1048 -> 1250).


now i gotta sell this eheim1250 pump, although i mightve shot myself in the foot here by recommending the 1048

[next on the agenda is my hard disks. the seagates are indeed quiet, but by not means silent...i know exactly what im gonna do to silence them, i will do a write up on this if u want me to, should work for even the noisy hard disks...thats later on this week, along with the new geforce 4 waterblock, and new mobo...must get everything finished by this Friday]

any questions, feel free to ask and dont be scared to ask stupid questions either. this is a forum questions and answers is what this place is for...isnt it?

 

[Iron Hawk]

Ever talk to anyone about having this put on the front page?

 

[sunama]

how do i do that?

 

[pHaestus]

Dunno if it is wise to extrapolate one person's results to water cooling design as a whole. I take issue with a couple of the comments. First of all, not sure the 1048 is going to produce half the flow rate of the 1250 in a system with plenty of restrictions (and pressure drop). Do you have an idea of how different the PSI produced by these two pumps might be?

Regarding lower flow rates being better, I personally see temperature decreases as flow rate is increased with my Eheim 1250 (using a ball valve to adjust flow from ~ 0.5 to 1.5 GPM). I assume you had to remove and reassemble the cooling loop between runs, and thermal paste variation is also a ***** (if your temp monitoring is good enough to notice it). That would bring us to the final point. Take care in drawing conclusions about performance differences of 1C observed with temperature monitoring that has a +/-3C accuracy.

On a positive note, the vibrational differences is welcome info to see posted. I had my Eheim 1250 elevated off my wooden desk with a mouse pad on the desk, a roll of duct tape holding it above, and then a pot holder directly under it and I could hear it if I put my ear really close. That isn't exactly what you see online normally: "Eheims are all silent" is more typical. Also no mention of size? The 1048 fits easily into my Lian Li PC-50; the 1250 would have been a big challenge.

Keep in mind that there are always two sides to every coin. If you notice temperatures increasing with a bigger pump, you could just as easily use a larger radiator and the same large pump (for even LOWER temperatures) or you can go with the same radiator and a smaller pump. Despite your and webmedic's observations, this doesn't mean that high flowrate and big pumps are useless. Waterblocks ALWAYS improve in performance as flow rates are increased. What your observations mean is that for your particular cooling system they aren't required, as your radiator isn't handling the heat load especially efficiently at higher flow rates and so your block can't reap the benefits of higher flow.

But please don't try to overinterpret that result that flow rates don't matter or that "big pump guys" don't know what they are doing. As it stands, blanket statements against the benefits of higher flow rates are an indefensible position. You are in effect trying to argue against thermodynamics (might be fun for a while, but then someone always digs out some pesky Laws).

 

[sunama]

Dunno if it is wise to extrapolate one person's results to water cooling design as a whole. I take issue with a couple of the comments. First of all, not sure the 1048 is going to produce half the flow rate of the 1250 in a system with plenty of restrictions (and pressure drop). Do you have an idea of how different the PSI produced by these two pumps might be?

indeed when the water is pumped right through the system of hozes and waterblocks, the waterflow of the 1250 wont necessarily be double that of the 1048, however, these specs are states without any resistance, much like we rate fans. we cant say what the air flow will be after u place a fan ontop a densly packed heatsink, u can only rate a fan when it is standing freely on its own. the pumps are rated in a similar way. although i did read that one guy measured the real flow rate, and it didnt match up to that quoted by eheim.

Regarding lower flow rates being better, I personally see temperature decreases as flow rate is increased with my Eheim 1250 (using a ball valve to adjust flow from ~ 0.5 to 1.5 GPM).

when using pumps in different watercooling setups u would expect different pumps to behave in different ways. some setups would work best with a flow rate of say 1.2gpm, other setups would perform better at a flow rate of say 2gpm, it depends on each system, and i do make that clear at the end of the review. however, what im trying to do is give people an idea that the slower pumps will not necessarily yield higher temps (webmedic on procooling found exactly the same thing).
.

I assume you had to remove and reassemble the cooling loop between runs, and thermal paste variation is also a ***** (if your temp monitoring is good enough to notice it). That would bring us to the final point. Take care in drawing conclusions about performance differences of 1C observed with temperature monitoring that has a +/-3C accuracy.

nope. i didnt even open up my case. i just detached the 1250, then immediately slotted in the same tube into the inlet/outlet of the eheim1048, it was a very speedy process and i had a towel underneath while doing this to prevent a massive mess. like i said, in total i lost about 3 tablespoons of coolant.

On a positive note, the vibrational differences is welcome info to see posted. I had my Eheim 1250 elevated off my wooden desk with a mouse pad on the desk, a roll of duct tape holding it above, and then a pot holder directly under it and I could hear it if I put my ear really close. That isn't exactly what you see online normally: "Eheims are all silent" is more typical. Also no mention of size? The 1048 fits easily into my Lian Li PC-50; the 1250 would have been a big challenge.

i did mention the size difference. i stated that in volume, the 1048 was about 50% the size of the 1250.

Waterblocks ALWAYS improve in performance as flow rates are increased. What your observations mean is that for your particular cooling system they aren't required, as your radiator isn't handling the heat load especially efficiently at higher flow rates and so your block can't reap the benefits of higher flow.

agreed.

But please don't try to overinterpret that result that flow rates don't matter or that "big pump guys" don't know what they are doing. As it stands, blanket statements against the benefits of higher flow rates are an indefensible position. You are in effect trying to argue against thermodynamics (might be fun for a while, but then someone always digs out some pesky Laws).

i didnt want to infer this. the point i wanted to get across is that high flow rates caused by larger, louder pumps DONT ALWAYS get u lower temps. in fact, in some instances (eg mine and web medic's) a lower flow rate can yield lower or identical temps. its ok to go for a smaller pump without fearing higher temps. right now, it seemed to me that web medic was the only guy on the net who was preaching this. everyone else was going and doing what the "big pump" users were telling them to do. indeed i myself baught the 1250 because of what most people had suggested (ie. the bigger the pump the better). now i can say, that that is not strictly true (although like u said, it maybe true in some cases).

 

[neo86]

The guy you were speaking of who said that any thing above 90 gph is uesless didn't know what he was talking about. If you look at head of the pumps (the highest point they can keep water flowing) it would've been a different story.

I also believe that your temperature differences were not accurate. I think it's because your radiator was unable to dump off the heat from the pump, and the processor. That's why you had a lower temperature using the 1048.

I don't own a 1.3ghz proc, but I still think those temps are a bit high for watercooling. Of course I'm not sure about that so if I'm wrong I'm wrong.

However, I'm pretty sure abou the rest of the stuff.

 

[sunama]

u must understand that im using at tbird. at 1.48ghz these cpus were notoriously high in temps. alot of 1.4ghz tbird owners report idle temps in the 40C area. also consider that im using a 1ghz cpu (not a 1.4 tb).

another thing to consider is that the temp probe im using on my mobo has some thermal compound on it, to aid heat transfer from the back of the cpu to the thermal probe. this increases temps.

another thing is that im using slow fans, not the loud fast fans that most other people use. the point of watercooling for me is to reduce noise (first and foremost) then to overclock. it would be a bit stupid for me to spend time and money to reduce noise of the pump then have loud fans spinning. so yes my rad couldnt dump the heat fast enough.

not enough people seem to have used different pumps and ran a side by side comparison and the result is that everyone "assumes" that the faster the better...which may not necessarily be the case...as is what i have found (and webmedic).

believe me, if the 1048 meant higher cpu temps i wouldve made do with the 1250, but as it happened, the 1250 didnt help my temps at all.

i am only reporting my findings, make of them what u will. and the guys over at procooling.com seem to know their stuff, and NOONE was arguing with web medic...not a single guy in about 40 post long thread. (that included posts made by the moderator of the forum). if he didnt know what he was doing, then im sure someone wouldve dared say that...but they didnt.

here is a link to that thread...read for yourself....high flow vs low flow...its a good read.

http://forums.procooling.com/vbb/showthread.php?s=&threadid=2978

 

[Angelslayer]

Great piece of info, especially for someone like me who is looking to w/c my Athlon 1Ghz.
Where did you mange to find the pumps in UK I keep looking but can't seem to find them. PM me if you can't post link etc.

You guys are wealth of information, I all spare my time at work on the forums, I get loads of people ask "what the hell are you reading?" So I tell them and they just stare blankly at me and go away shortly after!


Thanks

A.S

 

[sunama]

go here

www.aquatics-online.co.uk

they have the 1250 for £42
and 1048 for £34 but www.theoverclockingstore.co.uk are doing them for £32.90 (i found this out after i baught the 1048).

rather than spend the money on watercooling have u thought about spending it on a new cpu. u can run it at stock speed and use a very slow fan, so u get the advantage of a quiet pc. although if u wanna watercool, just for the heck of it, then u will probably get 1.4ghz at least out of your cpu, depending what stepping it is.

 

[ColdMiser]

I have a 1048 that I have managed to completely silence, so maybe this will work for your 1250.

If the pump touches anything hard it will resonate. Even if the tubing close to the pump is touching something hard it may resonate. The surgical tubing you have usually takes care of that though. I placed my pump on a two-layer sandwich of the following materials:

Top layer: A ¼" thick piece of neoprene from Dangerden.
Bottom Layer: A ½" thick piece of open-cell foam that was the packing material in my mobo box.

The neoprene alone wont work because it is too stiff to absorb all of the vibrations. The foam alone wont work because it is too compressible, so the pump flattens it and the vibrations still penetrate. The key to thier success is that when used together, the neoprene is firm enough to spread the weight of the pump over the whole piece of neoprene, which spreads the weight over the whole piece of foam. The foam does not get compressed and can therefore absorb all of the vibration. I kid you not, I have to check an indicator light to see if my pump is running or not.

Hope that helps.

 

[sunama]

hey thanx man. but ive already agreed to sell the 1250. i will be keeping the 1048 as the temps seem to be at least as good with the slower pump as with the bigger 1250.

however, your tips might come in use for someone else who is reading this thread and who also wants to quieten down their pump, and doesnt want to buy another one.

edit: i just noticed that your pump is the 1048 (the same that im now using). it took me about 15 mins to make my 1048 totally silent (by my earlier definition). i placed it on some bubble wrap and some old rags. i cant hear it anywhere now over 1 ft away from it. now the 1250: i mustve spent at least 3 hours on trying to quieten it down with no success...it just vibrates way too much. it pumps out a phenomenal amount of water though, so if flow rate is your thing, i cant argue with the 1250. keep in mind though that i have a stricter definition of "silence" than most.

 

[pHaestus]

How much do you want for the Eheim 1250? I might be interested

 

[sunama]

sorry dude its already gone

i live in london though so the postage to the usa wouldve been prohibitive.