Redundant Array of Inexpensive Pumps

[safemode]

giving an excuse like a trickle of water is still enough to save your system makes having two pumps in the first place "over engineering" the problem. If you can survive on no flow then why have the second pump at all? You're already over-working the problem then. The whole point of having two pumps is so that if one fails you're still well within being safe because you have still a decent amount of water flow. If the water is being mostly to completely blocked, you rely mostly on water convection to save you, which makes the fact that you have a second pump pointless. You can do that with just one. I've already proven that works. And your post digerr simply doesn't work here.

AS for ****ing on the thread, Nobody can deny that a parallel setup makes more sense and is easier to build with the pieces he already has than a series setup. Saying that the project is contradicting it's name is telling the truth. Just because you run dual pump doesn't mean you're in a redundent setup. They have to do the function of being redundent. It's not like a pump clogging is a marginal case, it's quite the common one during pump failure. If your "redundent" system cant even handle the common cases what's so redundent about it?

as for single directional valves (backflow valves), they're just circular doors that only open one way from pressure. There is nothing complicated about that. two is all you would need to protect the flow path in a parallel setup completely, while also protecting the secondary pump from possibly getting clogged by whatever was in the water that did in the first or getting eaten away or whatever.


If the water was good then there would be no need for redundency in the first place. calcium buildups come mostly from tap water and poorly treated distilled/deionized water. you can get algea buildup to clog a pump. You can have your impellor shaft eaten from acidic water due to battery effects and have that mess clog a pump. I've seen quite a few pictures of people dumping out nastiness from parts due to the flow being cut off and them not noticing until it's almost too late. Once the flow cuts down a lot things (if they still can) find it much easier to grow and continue to cut flow rate. You cant give me an example saying that a properly made system would not have the pump clog from buildup, because a properly made system would not need a redundent pump setup. Contamination may be uncontrollable or perhaps you dont care to take that much effort into it, whatever the case it's been deemed that redundency is going to be used and it makes no sense to only look at the "best cases" of pump failure in order to determine if a setup is viable.

for someone who is apparently ****ing on the thread, i seem to be the only one acknowledging the huge gaping flaw in the setup that makes it all but pointless to do. you dont build safety systems for the "best cases" and my examples of where it utterly fails are hardly uncommon among system failures.

 

[[EG]~NaTz~]

hey man i like it looked at the first pic and under stood. prolly cause i saw that other setup with 3 already though and the 2 layer thing. good luck i hope it does u goo dim currently working on a multi loop system with 3 pumps all differnt gph... hence the multi loop... ...kinda hard to explain so just wait till i get all the supllys and take pics...

 

[Diggrr]

giving an excuse like a trickle of water is still enough to save your system makes having two pumps in the first place "over engineering" the problem. If you can survive on no flow then why have the second pump at all? You're already over-working the problem then. The whole point of having two pumps is so that if one fails you're still well within being safe because you have still a decent amount of water flow. If the water is being mostly to completely blocked, you rely mostly on water convection to save you, which makes the fact that you have a second pump pointless. You can do that with just one. I've already proven that works. And your post digerr simply doesn't work here.

I didn't say that it'd survive on no flow, now did I?
I said that (with more words used so you can follow) if one pump failed, and the restrictions involved with having the pumps in serial meant that your flow was much lower, the cpu would survive. It wouldn't be running at a nice cool 30C, but it would still be better than none at all.
When a watercooling pump fails, as has been posted about here coupled with my own experience, they can actually last quite well, but the weakness is that the tubing often becomes mis-shapen and leaks water into your system. This can happen with plastic topped blocks as well.
We also know that it's the moving water that's the key to cooling, so cooling would continue, though not as efficiently. Adding a second pump into the system gives you that redundency without having to consult an electronics guru on circuit building, nor an engineer on flow valves and other junk you don't need....$25 bucks for a second pump, some homebrew engineering, and your done.

And one more thing. Most watercooling pump failures are operator error, bearing wear, and power failure. We're not talking an industrial environment, we're talking about the insides of a guys computer case. If he's spent anytime here, he knows not to use tapwater and such. But I believe I pointed that out already.

He's doing a fine job, and it's a good looking setup, and it does have redundency....one pump fails, the system lives! And that's besides the point that seeing the impellers do their thing just plain looks cool.

Have a groovy day!

 

[dagamore]

I knew you're going to get confused...

Here's a new pic:

I'm going to seal the area marked RED. Does it make more sense now?

what if you just leave the red area out, and then have a 'Y' connection between the two outlets of the pumps, feeding in to the system. this would give you better redundancy and more flow, and i think flow in watercooling is king. but any way you do it, it looks tight right now

 

[redPEPPER]

If you use a simple Y splitter to run the pumps in parallel there is almost no way that any flow is going to go backwards through the dead pump to create a circular circuit.

The force of the incoming water should be enough to prevent anything like that from happening if your piping is designed with the proper flow direction in mind.

Water is going to go through the path of least resistance. 90° turns in Y splitters would hurt the flow indeed, but would it hurt it more than going through the waterblocks, radiators, and all the tubing? The odds are that it wouldn't. Without a valve to prevent backflow, a good part of the water would loop through the dead pump.

 

[redPEPPER]

for someone who is apparently ****ing on the thread, i seem to be the only one acknowledging the huge gaping flaw in the setup that makes it all but pointless to do. you dont build safety systems for the "best cases" and my examples of where it utterly fails are hardly uncommon among system failures.

I understand your points, specifically that, despites the fact that a serial setup is easier to build (no valve needed), it doesn't solve a clogging problem.

I don't understand why you need to be so arrogant about it though. I think people would accept your points a little easier if you got off your high horses.

I don't understand either why you insist that labelling a system with dual pump as reduntant is an error. It seems to me that if you're pumping the same water through the same circuit with several pumps where any one of them would suffice, you have redundancy. You're free to have other standards but that doesn't mean everybody else is wrong about it.

I just read this project for what it is. The goal here is to solve pump failure. It's hopeless against severe clogging, yes. But it's not the purpose. On the plus side, a serial setup might deliver more pressure than a parallel one. It's not like this setup is an out-of-this-world design. You can find commercial kits with dual pumps. Serial. Typically, one pump pushes out of the res, the other pump pulls into the res.

You have a point with clogging. That doesn't mean serial doesn't make sense. And instead of suggesting nicely a possible improvement you made an a** of yourself, which alienated everybody else in the thread against you, regardless of how valid your point might be.

Ah well...

 

[I.M.O.G.]

ahh yes, its always a pleasure... some of you know what i mean.

 

[Since87]

ahh yes, its always a pleasure... some of you know what i mean.

What?

To watch a Mr. Negativity make a fool of himself nearly every time he posts. (While he seems to thinks he's making a fool of others.)

re·dun·dant
Exceeding what is necessary or natural; superfluous.
Needlessly wordy or repetitive in expression: a student paper filled with redundant phrases.
Of or relating to linguistic redundancy.
Chiefly British. Dismissed or laid off from work, as for being no longer needed.
Electronics. Of or involving redundancy in electronic equipment.
Of or involving redundancy in the transmission of messages.

I don't see any mention that redundancy implies unused.

Some redundant systems:

RAID - Redundant Array of Inexpensive Drives. Guess what, except in fault conditions all the REDUNDANT drives run all the time.

N+1 power supply systems - Power supplies which contain multiple "small" power supplies which loadshare, and when one fails, it can be hotswapped out and replaced. Guess what, except in fault conditons one more "small" power supply than is actually needed is running.

 

[dagamore]

Water is going to go through the path of least resistance. 90° turns in Y splitters would hurt the flow indeed, but would it hurt it more than going through the waterblocks, radiators, and all the tubing? The odds are that it wouldn't. Without a valve to prevent backflow, a good part of the water would loop through the dead pump.

t connections have 90 degree bends, y's have a bend more like 135 degree bends. and you could easly put in a one way valve to stop back flow if one of the pumps failed.

 

[safemode]

I understand your points, specifically that, despites the fact that a serial setup is easier to build (no valve needed), it doesn't solve a clogging problem.

I don't understand why you need to be so arrogant about it though. I think people would accept your points a little easier if you got off your high horses.

I don't understand either why you insist that labelling a system with dual pump as reduntant is an error. It seems to me that if you're pumping the same water through the same circuit with several pumps where any one of them would suffice, you have redundancy. You're free to have other standards but that doesn't mean everybody else is wrong about it.

I just read this project for what it is. The goal here is to solve pump failure. It's hopeless against severe clogging, yes. But it's not the purpose. On the plus side, a serial setup might deliver more pressure than a parallel one. It's not like this setup is an out-of-this-world design. You can find commercial kits with dual pumps. Serial. Typically, one pump pushes out of the res, the other pump pulls into the res.

You have a point with clogging. That doesn't mean serial doesn't make sense. And instead of suggesting nicely a possible improvement you made an a** of yourself, which alienated everybody else in the thread against you, regardless of how valid your point might be.

Ah well...

Just so you know red, I did politely give an alternative improvement to the design at first as i was asking questions relevant to the project that were not given. Just re-read the beginning of the thread. Then Ronnie decided to be sarcastic and uncooperative. So if things are the way they are it's not because I didn't take the effort to be polite. But people are selective in memory so I dont get caught up in stuff like that.

It is dictionary redundant but practically the redundancy is a facade. You cant fully depend on such a system set up in serial because one can directly effect the other. To me if you're going to do something why spend your time working on something that only solves part of the purpose of why you're doing it when there are easy ways to get all of it done? How can you say this is solving pump failure when you get the same performance under a common condition as if you had only the one failed pump? What makes this project any different than your dual pump kits if it doesn't aim to primarily solve pump failure? In those dual pump kits they aren't there for redundancy, they're used because the people who made those kits and systems found it to be more efficient to use two small pumps rather than a larger single pump. In many cases this works out to be true. It's not there as a saftey harness but under a number of conditions that may be a pleasant side-effect. I'm not seeing the difference between that and this project except this aims to use a piece of plexi in place of tubes. And the tone of my response is a reaction to what i've read as the thread. It is a hyped up dual pump enclosure. No more no less. The only differentiating feature is plexi. The whole beginning post when i read it had you thinking there was going to be some kind of ingenuity and cleverness in design to solve this problem when it turns out, as you've pointed out as well, that it's just like any kit using two pumps in series. .. there is no addition to the design in favor of solving problems of pump failure, simply a plexi pump enclosure. If the title of the thread had read that i wouldn't have given it any interest and if i did it would have been positive because it's right on track.

It's got nothing about superiority or arrogance, it's more about being blunt and realistic. My "wooos and wows" aren't free like other peoples. It's one thing to have a thread about your project regardless of difficulty or function ...but it's another to hype it up with a name designed to attract people to the thread. I find it unecessary to be polite to people who play that game. You see it on sites in reviews and articles and companies, it's like an old evil marketing trick.

Already though, this thread has gotten much more negative attention than it would have if the creator had just been true to what he was actually doing. This thread is well beyond overkill by now as it's devolved into a discussion about tone and attitude so i'll let it die off.

 

[safemode]

What?

To watch a Mr. Negativity make a fool of himself nearly every time he posts. (While he seems to thinks he's making a fool of others.)

re·dun·dant
Exceeding what is necessary or natural; superfluous.
Needlessly wordy or repetitive in expression: a student paper filled with redundant phrases.
Of or relating to linguistic redundancy.
Chiefly British. Dismissed or laid off from work, as for being no longer needed.
Electronics. Of or involving redundancy in electronic equipment.
Of or involving redundancy in the transmission of messages.

I don't see any mention that redundancy implies unused.

Some redundant systems:

RAID - Redundant Array of Inexpensive Drives. Guess what, except in fault conditions all the REDUNDANT drives run all the time.

N+1 power supply systems - Power supplies which contain multiple "small" power supplies which loadshare, and when one fails, it can be hotswapped out and replaced. Guess what, except in fault conditons one more "small" power supply than is actually needed is running.

If you read the very first post i made you'll see that I said that it would be a better redundent system if one was unused. I said you could use them both at the same time, but it wasn't ideal for handling possible pump failures (like in the case of wear and tear damage causing failure). And for someone who is constantly annoying people about semantics i'm certainly trying to be trounced. Yes having two of something that is doing the same thing is redundent, but is it useful in the context of one taking over (or continuing) the function of the other when they both dont work when one doesn't work to call them redundent? I stated that under a failsafe context in which the thread was focused (or the name implied to be focused on) independence was implied by being redundent. All of this has already been said and can be seen cleary if you want to. Being on all the time was always just a graduation of redundency with pumps, it was never a condition of redundency.

The only thing that makes me look foolish is that i dont simply ignore these kinds of posts sometimes. This is a forum of mostly technical discussions and questions, and I'm not the one one who brings it off that rail, and I didn't here either. As you can see from the progression of posts, I was only asking questions about a thread and suggesting mostly uncontested better ways to go about things. Eventually things stunk of hiding behind a flashy name for getting a pretty simple mod to a common thing a lot of attention and that's where things are now.

 

[Shroomer]

...Some people don't take this cooling thing quite so seriously. So what if he wants to name his work something that is maybe 85% correct rater than 100?

Some people are into it for the fun... And others are into it to suck all of the fun out!

It will always be this way, though. ...Sigh...

 

[Volenti]

Aww, looks like I missed the party

RonnieG, nice work, glad to see someone having a go.

 

[Al666]

So what if he wants to name his work something that is maybe 85% correct rater than 100?

Shroomer - Well, neither RonnieG's idea nor the RAIP name that he's giving to his work<?> is unique. I've used the exact same term RAIP - Redundant Array of Independent Pumps long time ago, Oct 2001 to be exact.
Check this thread over at HardOCP forums:

RAIP over at Hard/Forum - Circa Oct. 2001

and this one over here again in 2001:

RAIP - overclockers.com forum - 2001

So RonnieG, there's absolutely no reason to be arrogant! After all neither the name, nor the idea of YOURS <?> is unique. People are running pumps in series for a long time...

 

[RonnieG]

Whoa! What happened here? This thread's a mess

Anyways it's starting to shape up...:

The acrylic isn't bonded to the pump yet. I'll be doing that in a few moments, with this Devcon Plastic Welding adhesive.. probably the best adhesive for plastic-to-plastic. Unfortunately it isn't compatible with copper so I had to use weaker epoxy for that pipe...

I'll post updated pix tomorrow...

 

[RonnieG]

Aww, looks like I missed the party

RonnieG, nice work, glad to see someone having a go.

Volenti, I saw your triple pump work at the other forum. Wow it was awesome!

 

[RonnieG]

So RonnieG, there's absolutely no reason to be arrogant! After all neither the name, nor the idea of YOURS <?> is unique. People are running pumps in series for a long time...

Arrogant? Me?

 

[Shroomer]

Ho hum... I am truely sorry. I seriously didn't see that one comming.

I guess I'm refering to Safemode getting weird about the whole thing.

To me this stuff is fun. I don't take it too seriously at all, but I see people like this all the time. Man, who cares if it's original (maybe he didn't know about other's efforts), who cares if it's 100% right. As long as the guy had fun doing it and accomplishes what he set out to do, three cheers for him.

PS... I still don't understand how that's in series, but you're the expert.

Good luck with it.

 

[Volenti]

Whoa! What happened here? This thread's a mess

Anyways it's starting to shape up...:

The acrylic isn't bonded to the pump yet. I'll be doing that in a few moments, with this Devcon Plastic Welding adhesive.. probably the best adhesive for plastic-to-plastic. Unfortunately it isn't compatible with copper so I had to use weaker epoxy for that pipe...

I'll post updated pix tomorrow...

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

 

[RonnieG]

Bonded:

Plastic Welder for the mating surfaces and some steel-filled epoxy caulking for an added layer of protection.

No that vinyl hose isn't permanent It's only there to keep the epoxy from plugging the outlet...

It's looking rough right now but I'll have it sanded as soon as the epoxy cures. Good thing I have a "cure accelerator" (epoxy cures faster at higher temps); a styrofoam chest with a hair dryer inside. Turn it on high for a few seconds and the inside will be a nice 60-70C...

Now where to place the LEDs??? Do you think the dye-lite will keep the impellers from being visible??? Maybe I should aim the LEDs from the inside...