• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Redundant Array of Inexpensive Pumps

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Volenti said:


ahh, that looks pretty good in this almost final step, you did test the pumps in original condition, for comparason right?


A little burn-in test but unfortunately no performance tests...

Volenti I think I have a little problem here. Due to the pumps being so close together the hall-effect sensors inside seem to get interference from the other pump. The 2nd pump fails to start immediately around 50% of the time. It jerks back and forth for a few seconds before spinning normally. This NEVER happens when the 1st pump is not running.

Have you experienced this with your pumps? Maybe I should've added magnetic shielding between them?
 
Last edited:
sounds like safemode needs to take a chill pill and take a walk around the block
 
Some informal tests:

raip5.jpg


raip6.jpg


raip7.jpg



Cheers.
 
RonnieG said:



A little burn-in test but unfortunately no performance tests...

Volenti I think I have a little problem here. Due to the pumps being so close together the hall-effect sensors inside seem to get interference from the other pump. The 2nd pump fails to start immediately around 50% of the time. It jerks back and forth for a few seconds before spinning normally. This NEVER happens when the 1st pump is not running.

Have you experienced this with your pumps? Maybe I should've added magnetic shielding between them?

Don't worry about that, it will only happen when the pump/s don't have any restriction on the inlets or outlets, when you get it "in system" they will start normally, I noticed this effect in some previous pump mods of mine.
 
since you are getting an on head from just one of the pumps running you should be fine for a while if one of the pumps fails, you should have enought cool to keep everything ok, for a little while anyways.
 
Volenti said:


Don't worry about that, it will only happen when the pump/s don't have any restriction on the inlets or outlets, when you get it "in system" they will start normally, I noticed this effect in some previous pump mods of mine.

Hooked up:

raip8.jpg


...but the problem still persists :( albeit with lesser frequency...

I'm currently investigating the problem. I was wrong about the hall-effect interference thing. It seems that this particular pump (Resun SP-600) has startup problems when water is already flowing through it. This makes it a very poor choice for a serial setup :mad:

Well time to get a pair of Resun SP-850s :D

I'll probably be dismantling this and turn it into a parallel setup though I wonder how the acrylic will come off with all the strong adhesives...

The saga continues....... :cool:
 
RonnieG,

how bout a shot at dd cooling???

I've done direct-die before.... well actually direct-slug - there weren't exposed dice before, until when I removed the K6 cap... DD didn't do well then, even with the low-power K6 (just 30 watts I think), and won't do well now with our 100W+ progressively smaller dice... Well maybe it'd work if you had a high-pressure/volume pump (think pressure washers) then you'd have to worry about eroding the die...

...

So, I've transformed it parallel (valve-less):

praip2.jpg


praip1.jpg


...slapped on some steel plate "magnetic shielding" and a coat of spray paint.

It's done.

Conclusions:

1) Pumps in serial add up the head

2) Pumps in parallel develop lousy pressure. With both pumps on it could only do 13 inches, contrast with the 26" of the serial setup.

3) When 1 pump dies or is turned off in a valve-less parallel setup, some water will inevitably short-circuit through the other pump BUT water will still circulate through the system and save it from catastrophe. With my low-drag components my temp increased only ~1*C when one pump is turned off.

4) Some pumps have startup problems when in serial (e.g. Resun SP-600). It seems that an existing flow through the pump, like what would happen in a serial setup when the other pump starts first, prevents the pump from deciding which direction to turn. This seems to happen when the pump is of a fixed-impeller design (ie the impeller is rigidly fixed to the rest of the rotor). Some pumps have worked around this problem by allowing the impeller some degree of movement independent of the rotor, so that the rotor can start spinning freely WITHOUT the burden of the drag caused by the impeller blades. So when attempting a serial pump mod choose pumps that are NOT fixed-impeller.

So if you see a pump mod on the internet that tries to reduce noise by fixing the impeller to the rotor. Don't. The impeller is "free" for a reason...

5) "Redundant" doesn't imply "unused" :D Serial or parallel pumps ARE redundant. When one pumps stops spinning due to a stuck rotor, faulty electronics, etc, the system lives.

6) If you take care not to put in sewer primodial soup into your system you will NEVER have to worry about clogging your pump (unless you do something else stupid). As such you really don't need to protect against such incident.

7) If you put in distilled water you will NEVER have calcium buildup in your pumps.

8) If you have an all-copper system you will NOT have to worry about galvanic corrosion clogging up your pump...

9) Someone needs to take a vacation... :D :D :D


Cheers. Be safe
 
RonnieG said:


I've done direct-die before.... well actually direct-slug - there weren't exposed dice before, until when I removed the K6 cap... DD didn't do well then, even with the low-power K6 (just 30 watts I think), and won't do well now with our 100W+ progressively smaller dice... Well maybe it'd work if you had a high-pressure/volume pump (think pressure washers) then you'd have to worry about eroding the die...


You can successfully direct die cool modern cpu's, I know I've done it.
 
I understand why the impeller is "free" for a reason, however, having a quiet pump sure is nicer than one that vibrates. also, for those of us who run our pumps 24/7, we dont really have to worry about the pump having trouble starting. Good points though.

Keep up the good work.
 
Back