Pumps and Cavitation......

[jtroutma]

I have had enough of this stupid Sedra pump that I paid too much for!

The damn pump is the LOUDEST part of my whole computer by a LONG shot and I am tired of listening to it. (I cant even hear my 7200 RPM HDDs AND MY Sunon 120mm fan over the noise!)

I tried converting my whole system over to 1/2" equivalent and it is still just as loud.

/rant mode off

How well does the Eheim 1060 handle backpressure? I am thinking about getting the 1060 over the 1250 simply because I am not sure that the 1250 will provide enough flow. BUT I dont want to get the 1060 if it will just cavitate and give off the same amount of noise that this Sedra is already doing. I am stuck!

:shrug: (This has NOT been the best month for me at all!)

[Skulemate]

The Eheim 1060 is really a poor choice for a water-cooling set-up. Sure, it has plenty of flow when running unobstructed, but it doesn't have much greater head capacity then the 1250, and that's where it counts. I'd stay far away from the 1060 if I were you.

[MadDogMe]

1060 = no, no...

I had bad problems with my 1250 making 'cavitation' noises, it turns out it was drawing air in from around the inlet~housing,(rubber O~Ring), I only found out cause when I changed from a reservoir(not airtight) to a closed loop it started leaking water(very slightly), I guess it did'nt need to with the res?, air'd get drawn in and leave through the res?, but with the loop closed it had to loose some water to replace the space the air was taking?. I don't know that's my guess though. this was a new pump, it was probably just a bad seating...

You could try gluing the impellor if it's loose on the shaft/magnet~rotor...

[Volenti]

An Iwaki, with the japanese motor option, their quiet, and the RLT models shrug at back pressure.

MD20RLT or MD30RLT would be good options, be warned their pricey though, quality comes at a cost...

[MadDogMe]

Yep!, don't buy the WMD model, it uses a US motor that runs hotter and dies quicker :shrug: ...

The MD20RZT has the best head/pressure, it only has a max flowrate of 3.5~4GPM but it'll maintane it well. Cathars block is considered a high resistance block I think. very well named block IMO ...


...

[Cathar]

Originally posted by MadDogMe
Cathars block is considered a high resistance block I think.

85% of the volumetric flow rate through a complete system as the Maze 3 (one of the most freely flowing blocks I've tested).

I'd actually classify it as "medium resistance".

[Turbokeu]

Originally posted by jtroutma
I have had enough of this stupid Sedra pump that I paid too much for!

The damn pump is the LOUDEST part of my whole computer by a LONG shot and I am tired of listening to it. (I cant even hear my 7200 RPM HDDs AND MY Sunon 120mm fan over the noise!)

I tried converting my whole system over to 1/2" equivalent and it is still just as loud.

/rant mode off

How well does the Eheim 1060 handle backpressure? I am thinking about getting the 1060 over the 1250 simply because I am not sure that the 1250 will provide enough flow. BUT I dont want to get the 1060 if it will just cavitate and give off the same amount of noise that this Sedra is already doing. I am stuck!

:shrug: (This has NOT been the best month for me at all!)

jtroutma, I suppose you mean the Sicce Idra? (http://www.sicce.co.uk).
I had one of them too, leaked a little and rattled as hell
I use it submerged now in one of my fishtanks...
Went for a 1250 now.

Be aware that Swiftech's new MCP300 pump is nothing else but a Sicce Idra (with modified Alu inlet & outlet)...

CD

[BillA]

you guys might want to wade through this thread

a tad involved, but approaching a 'conclusion' (of sorts)
(I need to add today's 'lesson')

most (I would say almost all) of WCers' pump problems derive from too much restriction in the suction (pump inlet) line
- that's why the impeller 'lifts' and rattles, they suck air, etc

the pump is being blamed for what are in many systems simply hose configuration errors
- sort out the pump sypply before buying another

-> feed the pump with oversize hose, as big as you can fit

[Brians256]

Turbokeu, long time no see! Good to see that you aren't dead or gone from the computer cooling.

As for noise, I am very happy with my Eheim 1250. Yes, it isn't the best that money can buy but it does keep things relatively cool in my system. No rattling, cavitation, or sucking at the intake seal for me.

[airspirit]

BillA is right on the money. If there is restriction, your pump will get pissed. The less work it has to do immediately at the inlet, the better it'll behave.

[Brians256]

As another ditto to BillA and airspirit, these magdrive pumps degrade very quickly as flow restriction is placed upon them. This is the main reason I have not done any NB or GPU watercooling yet with my single pump system. Those blocks add so restrictiction that my CPU cooling would be seriously compromised. If I ever watercool my GPU, PSU and NB, I'd use a separate pump because trying to cool all those things with a single Eheim 1250 would be like trying to drive a large car at 5mph when in 5th gear. You might get some movement, but it would make a lot of noise trying to do it.

Listen to BillA and airspirit. They know what they are talking about.

Has anyone used a different type of pump like a peristaltic or positive displacment pump? I don't know much about their flow rate, MTBF, efficiency or noise level, to be honest.

[bigben2k]

Cieprus came closest to a positive displacement pump (but it wasn't). Flexible rubber impeller, noisy as hell.

I've seen many peristaltic pump specs, but none were in the flow range that we use. They're mostly for dosing applications (i.e. deliver a specific volume, then stop).

I agree with you Brian: I wouldn't bother cooling anything else than a CPU with one pump; I'd get a second one.

[Brians256]

Well, if those peristaltic pumps were too slow, then just overclock them! Hook up the 220VAC and watch them go! LOL. I've seen the peristaltic pumps on IV delivery systems and didn't know if there were higher volume versions that we could use. I know that we wouldn't want those specific models, as anything designed to put stuff into the human body likely doesn't deliver 350GPH (I hope not!) unless it is an artificial heart or a kidney dialysis machine.

[bigben2k]

They should have a high pressure rating, but I have yet to see one that's high flow.

The problem also, is that the tube needs to be replaced regularly, since this pumps works by squeezing a fluid through that tube, which probably wears out very quickly, especially at high speed.

[jtroutma]

WOW!

How the heck did we go from "does the 1060 cavitate" to "WW not our firend" to "Overclocking peristaltic pumps" ?!?!?!?!?!


Anyways, Well if you guys are interested here is what I have setup in my cooling loop right now;

Res. = 1 1/2" PVC pipe ~1' long at higest point in system
1/2" tubing out the bottom of res. in to 90degree elbow to pump
SEDRA 3500A 1/2" intake to NOW 1/2" exhaust
1/2" "T" going to (2) 3/8" lines

WaterBlock 3/8" line TC-4 (no turbulators)
Gemini GPU Block 3/8" line

1/2" "T" taking 3/8" lines back to 1/2" line to radiator
DTEK Custom radiator 1/2" lines
3/8" line from radiator to 3/8" intake on res.
(Yes I know that 3/8" ine between radiator and res is NOT good; will convert to all 1/2" as soon as I can cut some more holes in case)

I think the intake on the pump is rated for 1" thread and the exhaust is rated for 1/2" thread.

Found link to details on pump (thanks BigBen)


Sedra 3500A

Now do you think that I could be starving the pump of water from the intake? Would going to 5/8" lines from res. to pump would "cure" this pumps noise problem?

[jtroutma]

Two more strange details that I should mention:

(Yes I do beleive that these deserve a new posting)

1) Once I put in the new custom radiator from DTEK, I started getting greenish film covering everything in my water loop. I was using purified water at the time. Fearing for the life of my cooling loop, I added WW and now it has subsided. My system is all copper/brass/plastic/tygon. Where the hell could I be getting this corosion?!?!?

2) The whole setup was done so that I could use a 1/2" return side and a 3/8" pressure side of the system (AKA 1/2" from res. to pump & 3/8" from pump -> blocks -> res.) Now I am almost all 1/2" and it is still noisier than hell.

I am desperate to quiet this thing down but I dont have a huge budget. NEEED IDEAS!!!!! PLEASE!!

[BillA]

"Now do you think that I could be starving the pump of water from the intake?"

yes, that is just what was said

1" thread ? -> then try 1" hose
NOTE that a powerful pump may be ok when submerged w/o any inlet plumbing
but that such added tubing and fittings at even its 'threaded' size may inhibit the pump

(that's what that big thread was all about, the WW was an extra bonus)

are you sure its corrosion ?
use an algaecide

[bigben2k]

First of all: DON'T PANIC!

then, "you're welcome!"

Now... We need to know what kind of flow rate you're getting.

Since your pump has a max head of 9 feet, and a max flow rate of 350 gph, the pump will be at its best somewhere in the middle, say 150 gph with maybe 5 or 6 feet of head, most probably.

If you want to experiment, you can put/fit/cram a big a** 1 inch tube at the pump inlet, but I doubt that would have any effect.

[BillA]

"If you want to experiment, you can put/fit/cram a big a** 1 inch tube at the pump inlet, but I doubt that would have any effect."

Ben

is that an opinion ?
or do you have some actual experience, and measurements, upon which you are basing the above ?
(did you take the effort to read the linked to thread ?)

jtroutma is asking for assistance
when he tells you his flow rate, what are you going to do with that info ?
give him some BS about line velocity ?

-> the problem is at the pump inlet
do you know what NPSHA is about ? did you follow g_f's link to pump info ?
(it is apparent you did not)

jtroutma
close-couple your res to the pump inlet with as big a pipe as you can (and I mean big), eliminate that ell if at all possible

[bigben2k]

Actually, I did read it, and thought it was most interesting.

But given a (relatively) low pressure (if that is the case), I don't believe that this would be the problem. Do you?

Maybe this pump needs an impeller fix, like the 1048.

I happen to know that this setup is fairly new, so I would quickly rule out algae, but a number of factors might make it algae, hard to say. Has it been exposed to daylight? (i.e. direct sunlight). What do you mean by "purified water"?

If it is corrosion, it would have to be between the copper and brass, since there's no other metal, right? Water Wetter may make it look like things are stabilizing, simply because the red/pink color will reduce the "visible" green-age.

On the other hand, it might simply be a residue left in the heatercore. Did you try to clean it out, before installing it?

[myv65]

OK, I didn't read the other thread, yet. What I would say is that many people don't seem to understand cavitation and what pump manufacturers call NPSH. My apologies in advance for covering "old territory". Consider this an explanation of BillA's reference to NPSH above.

Cavitation is simply a condition where the pressure on a liquid is insufficient to maintain it in the liquid state. Some of the liquid converts to vapor. When the vapor hits the impeller and the pressure rises, the vapor bubbles collapse. Both the collisions and the collapsing are the sounds of cavitation.

NPSH is net positive suction head. For every pump a curve may be generated for NPSH vs flow rate. NPSH is the minimum permissible pressure that the pump requires to avoid cavitation. The curve for NPSH tends to look much like the curve for P/Q, but going in the other direction. At zero flow, you essentially require zero net suction head. As flow grows, so does NPSH, slowly at first and rapidly as you approach the max flow of the pump.

So how do you calculate NPSH? It is nothing more than the static pressure that exists at the pump suction minus velocity head and the fluid's vapor pressure. Static pressure is basically one atmosphere +/- depending on your elevation and your particular system. System configuration doesn't matter much as one atmosphere is roughly 33' (10 m) of head. Velocity head is V^2/2g, where "V" is the average velocity (flow rate / area) and "g" is the gravitational constant (in whatever units you use).

So what does this mean? It means that as flow rate goes up, you're more likely to have cavitation. If you have cavitation and throttle the outlet, the cavitation will generally diminish or cease. Here's the kicker. For a given flow rate, a higher volume, higher pressure pump will not cavitate as easily as a smaller volume, lower pressure pump.

Many times when people think they have cavitation it is nothing more than air getting into the pump inlet (not truly cavitation) or the rattling of a cheap impeller under load (also not cavitation).

For anyone seeking a technical explanation from a pump supplier, feel free to look here.

[Brians256]

That link was a goldmine of information. Thanks BillA. Thanks for a quick summary, myv65. That thread is dense with info! I'm still reading it.

[jtroutma]

BillA:

Not quite sure what you want me to do.... find a large res. and attach it directly to the pump intake?

Also I am sure that is it not alge. It has a greenish blue color to it and it is a very fine powder. Yes, the water that I was using is not necessarily the best (VONS "purified" water; they claim that it was deionized), I thought it was the knock-off equivalent to distilled water.


BigBen:

I did not flush out the radiator before I installed it; my mistake But I didnt think that I needed to clean it out before using it. Oh well. I just went back to using WW.

I thought that brass and copper doesnt react the same way AL and CU do in a closed loop. Not to say that I didnt expect a very small bit of corosion but to see this powder film over all my tubes and building up in all the corners of my blocks?!

What is this pump fix? I have heard about it before but figured it was only for the MAG pumps. I may have to try it.


One question I do have for you BillA. If I have a 1/2" line comming into the pump and only have a 3/8" exhaust, then how could I still be starving the pump for water when it should NOT be able to get rid of it that fast? Not challenging you, just curiosity.

BTW thanks for all your input on the matter.

[BillA]

I know that myv65 will always do better than I at answering the 'why' of things, but . . .

as the quantity of liquid is the same at the inlet and outlet, it will be only the pressure and velocity that differs
reducing the outlet dia will create more resistance hence also higher pressure
reducing the inlet dia will create more resistance but the pressure relative to the impeller is negative, and if 'excessive' will induce one or more of several types of cavitation (see g_f's link)

what you need to attempt to do is to reduce this inlet flow resistance as much as possible

the 'best' solution is to submerge the pump in the res with no inlet piping at all
- failing that put the res as close to the pump inlet as possible and connect it with the largest short piece of pipe possible, w/o any fittings or reductions in the ID

no idea on the particulate corrosion 'product', sure sounds like copper
'something' is in the system, that is not normal
and you want distilled water, DI water is more reactive

[freeloadingbum]

If I understand myv65 correctly, It's doubtful that you have cavitation at all. It should be easy to verify for certain by pinching the output hose a little (throttling) and see if the noise diminishes. If that doesn't make any difference then you know it's problably the other two reasons that myv65 suggested, a trapped air pocket or a bad prop.