Flow and PSI and ?

[lead]

Well if I understand what Smart people have said the Presure is the same anywhere in a closed circuit (water) and I belive it but I design and have built and tested the system below(Attachment below) I could not --and no way- have the 1048 loop join on the pressure side of the Aqua Via. Common sense tells me it won't flow. I had run out of tubing to test but I got my Tygon today and started my first cut and a light went on (a dim light) about the fact " pressure is the same" so I put my knife down and decided to post and maybe somebody will look at it and see something I don't. I would like to have reduancy if one pump go out and that means joining the 1048 loop before the block or at the block with T or Y--( I have no 5/8 Y but several 1/2). A "Y" would be close on clearance so I will say a T.
If pressure is same, I should get the same flow through the 1048 loop if it enters a T right on the block as I do with it at the suction side of the Via. Common sense tells me "not" . The 1048 loop has nothing on it now but the Rad but when I decide to add a video block or MB or Hard drive cooler or a chiller, I will have the loop. I plan on keeping the 3/8 OD tube on it for now as that gives me a lot more options for Acess. blocks.

Be fore everybody says it Over kill--I know it now. I had the swiftech sys. and it flowed about 39 gal a hour and a danger Den block (maze) and BI w/1048 and it flowed (I forget but about 35Gph) and I wanted the Magic 1 gal Per min. So I bought the core from D-tek and the A V 1300 from Be coolin and found out with just the core, the 1048 would give me the Magic 60 GPH when i swap tops on the 462 swiftech (1/2 X 1/2). But now I got them and I gonna use them. And actually the Via 1300 has 12 Gph on the secondary loop with the 1048 off and that means it got to flow through it--that should be at least as much resistance as a block on it. Any thoughts or comments welcome--But it has been running like this for several days (leak testing) and My new MB arrives tommorow and I will probably do something bad or wrong before Thursday like not leak test it after I put new tubing on it. Should of posted last week but only when I put the blade on the tygon did the LiL light come on. Thanks if you read all this.

[RedTalon19]

yea.









thats a really confusing pic, i just hope you realize that. ok.








well you could simplify the whole thing and have 2 seperate loops for everything. i'm quite tired and going on 27 hours without sleep... and i have another 6 hours to go min before i can get some sleep... so i dont feel like thinking too hard right now. maybe later tonight i'll come back... maybe

[gmat]

OMG ! I'll (try to) answer to that tonight when I'm back home and relaxed.

[Volenti]

that's....the most confusing diagram I've ever seen, i can't imagine how you can have 2 semi-seperate loops and still have redundancy, you could run a serial setup with both pumps, the different gph of the pumps won't matter, you'll still get more flow than with either one alone (I've done flowrate tests with my 2 (different) pumps) and you'll retain some flow if one fails.

[bigben2k]

Errr... I looked at that graph for 10 mins, and I still couldn't visualize it.

[bigben2k]

Got it.

The problem is that the Via pump is recirculating the coolant without going through the air trap.

As for the redundancy aspect, there's 2 choices:
1-use both pumps in-line, giving a better overall flow, as the second pump acts as a booster pump. If a pump dies, then the flow is greatly reduced, as the dead pump becomes a flow limiter.

2-use both pumps in parallel, with flapper valves, to prevent backflow through the pump that dies. The result is a slightly better flow, but not as much as #1.

[lead]

looking at it this Am--it is confusing-sorry--Last nite it was clear to me Lol-
The main thing I was asking or thinking to get reduandancy ,I would have to move the botton T to the block. Then the 2 pumps would be "butting heads" and I just think there wouldn't be any flow through the top loop. The way it is now-If I shut off the A Via 1300 the water from the 1048 loop goes up the T and across the block, back to pump. Only thing to do is try it. That will be a test for another day. Last week when setting this up and testing, I decided to see how loud the fans were gonna be at 12v and 7v. So i set a 2x4 across top of case to hold a bowl of water to act as a reservior. Then set up my enermax 465 w/a jumper and usually a old cd player to give me load--that day just picked up my new maxtor ata133 to use and started checking noise with the 3 120mm fans and spliceing fan connectors to make 7v. I looked up and seen water was getting low in bowl, so got my gal jug and added some water to bowl and hit the 2x4 and launched the bowl of water on to the PSU and hard drive. It made severl LOUD pops before could turn power srip off. Scratch 1 hard drive-PSU lost 1 of its fan and only puts out 4.65 v on its 5v line.I got it to post when MB is at FAILSAFE settings but it won't run windows. I swore off testing for at least a week. It sounds like something from the "3 stooges" but it scared the XXXX out of me at the time.

I think I am gonna leave it set up like it is. Thanks

[lead]

bigben2k


I was posting and didn't see your post. But I have decided to leave as is and test it myself later on. Thanks for the reply.

[bigben2k]

Well, ok but...

you are basically running two independant loops, that are joined between the CPU waterblock and the Via pump/1st rad (i thought that was your HD/GPU WBs)

You would fare MUCH better like this:

[bigben2k]

Or more accurately:

[ECUPirate]

*spelling police

Heay, lead, yuo misspeled "Louisiana", en yuor bio...

might want to look into that.

[lead]

ECUPirate

Thanks- Officer

[Skulemate]

Here's my $0.02:

Before I get started, let me just say that pressure is not the same everywhere in a closed loop (Bernoulli will tell you that... even if the water is not moving the pressure is not the same).

Ok then. I like both of Bigben's suggestions concerning the pump setup. I don't think that one can be recommended over the other however, since the two pumps are different (and as such, have different Q-H curves). As far as I know, there are two ways to determine which is better in this system. First, you could simply try both, and see what happens. Otherwise, you can calculate a resistance curve for your setup for various flow rates, then create a combined pump curve for either series or parallel operation, and find out the operating point for both setups (not that I think there's enough information available to do this... when I asked DD what the pressure drop through their Maze 3 was all I got was ignored ). The only suggestion I have is to try the rads in parallel rather than in series, since I bet they're more efficient that way.

[ECUPirate]

[gmat]

Yep.
Run everything in series apart from the rads which should be in parallel.
Put the 2 pumps one after each other, before the rads.
The go to CPU, GPU, NB and back to pumps.
Put the refill/airtrap part where it will fit best (ie less tubing from/to it)

[bigben2k]

Like this (option 1):

[lead]

Thanks for all the answers--and the pics (bigben2k) .

I like the option 2 on the drawings better than option 1
It just seems to me that pump to pump, that you would limit your flow to the lowest rate one,although with a lot more head capablities (sp.) I hope that spelling cop ain't here
I think I ask for reduancy and yall gave me the answers. But what I really wanted was reduancy with no flow loss. I don't think I can get both. (my fault). Right now getting 120 GPH with the AV1300 and don't want to go down. When I check it with the 1048 running it only barely improves (29--27 sec. to fill gal jug). a big problem is I don't want to change out the 3/8 OD hose on top loop as have come to like the Quick connect fittings and the 3/8 hose is almost invisible compated to 1/2 ID sillicon. Am gonna move the 5/8 T at the bottom and put in next to block and see if I get flow through that loop. If it does, will set it up that way.Will be pump-- rad-- T--block-- T--back to pump twice. If not will set it up like I got it with aprx. .25--.30 water going through 2 rads and the rest just 1 rad. The ONLY way is for me turn this puter off and go do it.
Thanks all for the responses and I kinda feel bad about asking for advice and deciding not to try it. It will be a little while before I can get back as last nite I wired the 2 pumps together so would only have one plug and I will test them separately and together(just because I want to know) . Will have to undo that and set them up. But thank you all and will let yall know the flow rates. But it probably won't be today.
P.S. last nite was the first time I had used MS paint

[bigben2k]

No prob.

Option one: pumps in series, it'll work, and the second pump will act, like I said, as a booster pump. In terms of pressure, the booster pump increases the pressure above the input pressure, as it would normally if it was running as a single pump. The limitation is that the resulting pressure on the second pump's housing is more than normal, so you have to watch out for leaks. For these 2 pumps though, it should be ok.

In case of failure however, the dead pump will restrict flow a lot, so it's not so good.

Option 2: pumps in parallel, you'll be limited to a MAXIMUM flow rate that is somewhere between the MAXIMUM of the two pumps, but you'll have a better head, which translates in better flow. Not as much as option #1 though.

In case of failure, the dead pump does nothing. In the worst of cases, the pressure from the first pump will leak in the dead pump, resulting in a lower flow than what the working pump can provide. This is fixed with a flapper valve. (don't ask me where to get one!)

Just as a thought, your idea is good, but I don't think that you should worry about pump failure. These things are fairly reliable.

You do need however, a plan to handle failure. The first thing to do is install MBM, unless your mobo has OCP.

If you're worried about pump failure, then I have to assume that there are a number of other items that you also worry about. If you think about it though, it all comes down to the CPU overheating, which MBM or OCP can cover (unless your waterblock falls off, in which case only OCP can cover ya).

[airspirit]

That Via 1300 is a POS. I bought one and the pressure was terrible on a test loop with less than half the complexity of yours. If you use them both, put that first, because you might lose as much pressure to pump resistance as you gain in circulation force. I ended up scrapping that Via without putting it in my machine at all. I feel that putting them in series (I can't explain how I came to this decision ... I can picture it, though) with the Via feeding the Eheim might (I dare to say WILL) give you the best results. Putting them in parallel might cause minor backfeeding from the Eheim loop to the Via lowering your net system flow and your Via, besides lower flow rate through the entire system will serve as nothing but a heat battery.

[bigben2k]

Originally posted by airspirit
That Via 1300 is a POS. I bought one and the pressure was terrible on a test loop with less than half the complexity of yours. If you use them both, put that first, because you might lose as much pressure to pump resistance as you gain in circulation force. I ended up scrapping that Via without putting it in my machine at all. I feel that putting them in series (I can't explain how I came to this decision ... I can picture it, though) with the Via feeding the Eheim might (I dare to say WILL) give you the best results. Putting them in parallel might cause minor backfeeding from the Eheim loop to the Via lowering your net system flow and your Via, besides lower flow rate through the entire system will serve as nothing but a heat battery.

Normally I'd agree with you, but the Eheim 1048 here, is only rated for 158 GPH, so if anything, the Via (400 gph) would run through the Eheim.

[airspirit]

I think that the numbers Via advertises aren't worth wiping a dogs bum with. They say 400 GPH and 6' of head, but I found, in a low resistance system, that I was looking at about a quart per minute at 2ft lift. I can't see that thing getting 400 GPH if you were pouring water through it over a cliff, and even in solid hose I couldn't get more than 3.5ft of lift out of it. Maybe mine was defective, but it was just absolutely terrible. At first I thought my system was plugged until I put the other pump on it and I had flow like crazy going through it. Then I tried the lift test and chucked it into my closet junk bin.

[bigben2k]

Something's up here... That Via, at (?) 400 GPH (max 4 ft) should even outperform an Eheim 1250 (317 gph, max head about 6 ft.)

I know that it's possible to take the Eheim apart. Can you do the same with a Via?

[gmat]

Originally posted by lead
It just seems to me that pump to pump, that you would limit your flow to the lowest rate one,although with a lot more head capablities (sp.)

Wrong. Our pumps (centrifugal ones) are *not* volumetric, i.e. they dont work by moving a given volume of water at each engine turn (volumetric pumps are quite bigger, noisy and way more expensive).
They work by applying a pressure differential between outlet and inlet. The flow is only a product of this pressure differential minus the 'backpressure' of the circuit.
What that means is putting 2 different pumps in series will increase total head pressure, thus flow. Beware if your pumps are, say, 300 and 150gph the total flow will *NOT* be 450gph but i guess you've figured that already.
The 'lower' pump will only add its head pressure to the system.

Now as of putting them in series or parallel... i'd say series but tests need to be done on this (someone has experience on that matter ?)

Originally posted by lead
P.S. last nite was the first time I had used MS paint

I was about to make a nasty comment But since it's MS paint you're excused.

[airspirit]

I do have experience with pump setups from my stint in the Navy (blech). Centrifugal pumps build backpressure easily since there is no real "force" being put to the fluid. Unless you are shrinking the size of the tubing for the parallel section and running two pumps with identical flow rates you never want to run parallel with centrifugal pumps. It would be a whole different story if these were volumetric; then you'd want/NEED them to be parallel. In a case such as this, series is the only way to go.

[bigben2k]

Originally posted by airspirit
I do have experience with pump setups from my stint in the Navy (blech). Centrifugal pumps build backpressure easily since there is no real "force" being put to the fluid. Unless you are shrinking the size of the tubing for the parallel section and running two pumps with identical flow rates you never want to run parallel with centrifugal pumps. It would be a whole different story if these were volumetric; then you'd want/NEED them to be parallel. In a case such as this, series is the only way to go.

I agree!

There's nothing like military experience to prove a point!