Pump mystery

Think pressure causing flow, and open allowing flow.
- and my first guess is a switching T designed to empty one tank before the second empties (used for automatically switching into reserve tanks), a collapsing line (wrong or under-designed hose), or less likely, the hansen effect
If I have the facts correct:
1) the hose between the tanks is the same diameter,
2) the pump intake hose has a smaller diameter than than the hose connecting the two tanks
3) The connection between the tanks is - stinger pipe, hose, T, hose, stinger pipe.
4) The tanks siphon properly, and the fluid seeks the same level when the pump is not operating.
5) The tanks do not siphon or seek the same level when the pump is running.
6) When the pump is running, the tanks are drained one at a time.
For flow to happen in a T, there must be pressure differences at each end of the T, and open ports.
But here there is no flow on one port, and then there is - and what changes the flow condition is the flow from another port. So pressure changes, and possibly ports are blocked/restricted.
I would first look at the outside of the T and check that it is a simple T and not an unbalanced T.
If it is indeed a simple T (longitudinal axes to tanks, transverse port to pump, length of tank port legs of the same length), then:
Second, I would then look at the lines as the pump is running, especially at bends. If there is anywhere the line is not perfectly round DURING pumping, the hose is collapsing and pinching off flow. (Especially "transparent" line, which to my knowledge is not rated for any suction to speak of. Pressure only.)
If the lines are all perfectly round during operation and no fluid is leaving the tank connected by the lines:
Third, I would look inside the T and see if it is actually a shuttle T (ball/vane inside) rather than a simple T. - if it has a ball or a vane, it is doing its job in draining one tank first. - if it is a simple casting, then:
Fourth, a hose can be blocked by the inner wall of the hose going inside the fitting path when the hose is connected to a fitting. When fluid flows into the fitting at that port, the separated inner wall is moved back out of the way of the flow. When other fluids enter the fitting form other ports and/or the flow is reversed out that port, the split-off inner wall is pushed into the path and blocks the flow. Acts like a hidden check valve. Since the hose inner wall being inside the fitting, it reduces the diameter of that port relative to other ports, creating an inherent pressure differential in the multi-port fitting.
In your case, theory says that since the line is larger from the unblocked side, the mass from the flow will force the inner wall into its port opening, creating a pressure difference and blocking the flow from that port until the mass no longer is present and the pressure difference shifts.
(I call it the hansen effect because when I first had it as a problem, I was using 3000-10000 psi hansen outer/inner sleeved fittings.)
or something like that.
(And if this problem wasn't metric but rather more innately human-friendly , the solution would be more easily spotted. Wrap your mind around 18 and wrap your mind around 1. Esoteric process former, instinctual process latter.)
This is not some academic

Dumb question, If we say the letter "T" as shown in this sentence is your fitting, do the flows from both tanks enter the fitting from the right and left with the exit flow out the bottom? If not, then you'd get a lot more resistance from the stream that had to turn 90 deg than the one that went straight through.

"Aldo" wrote in news: snipped-for-privacy@adelphia.com:
Let's take the easy case first. Thru flow in a tee is when the flow is just flowing along the long leg (of the tee not the letter) with the branch entry valved off. Branch flow is when the flow enters the short leg and turns 90 degrees to exit the tee with one of the thru entries valved off. In this case, there will be 3 or 4 times more head loss with branch flow than thru flow. There is more going on, however, than just having to turn the corner. A standard 90 degree elbow will have only 1.5 to 2 times the head loss compared to the thru flow tee, much less than the branch flow tee. The thru flow tee will have 5 to 10 times the loss of a similar length of pipe.
The additional flow loss in the branch tee over the elbow and the thru tee over the pipe is caused by the more complex flow allowed by the more complex shape of the fittings. When you add to that the more complex flow due to the interaction of two distinct entering flows when all entries are open in the tee, things get harder to predict.
It is true that with all the tee ports open, you would expect branch input and thru output to be less even than thru input and branch output. Even meaning here, equal incoming flow rates. Relatively small perturbations in the incoming flows, however, can dramatically change the situation by choking off and restricting flow from one incoming branch. Whether you could actually get to the point where branch output actually had more head loss than thru output is an open question. I have seen set ups that had pretty simple problems that nobody really noticed that dramatically altered the flow from what people expected.

Just to clarify: The T-fitting is positioned so that it is symmetrical with respect to the tanks, i.e., the tank hoses enter horizontally from opposing ports. Flow is thence upward to the pump inlet. can make a lot of difference. However, it's still puzzling to me that the tank located at the greatest distance empties first. Paul

It still sounds like there is simply more head pressure on the longer run. (more mass/ higher energy potential) So therefore the higher flow rate will be from such higher head pressure first since it can easily take over from the lower head pressure.
-- James M Driscoll Jr Spaceman

Yes, but how did it get that higher pressure relative to the other?
More pressure one one or more resistance on the other...

The length of the pipe differences.

And Paul - you are sure that no part of the hose on the short leg is partly closing (flattening or the inner wall collapsing) when the pump is first running?

symmetrical If I understand what you are saying, that pressure difference of which you refer is the difference in pressure due to an equal flow in different lengths: same flow in a long pipe gives as greater pressure drop than it would in a short pipe. - but no flow in the short leg initially means that there is no "pressure from flow".
Here, there is the same pressure: from atmosphere at the two fluid surfaces relative to the common pump intake.
So the delta P from fluid to pump is the same, and any difference in flow is caused by differences in resistance in the path.
But for what happens here, the shorter pipe clearly has more resistance initially, even though there is no flow in the shorter pipe.

Obviously you have not told us about something important. As someone else observed, If the 2 tanks are connected by a hose where both ends are always under liquid and no air is in the lines then the 2 tanks should always equalize in level when the pump is not running due to siphoning. Since you are filling vehicles I assume the pump is not running a lot of the time. Therefore something in the piping from one tank to the other is preventing siphoning - what is that something?
-jim

The same pressure but 2 different volumes in each pipe length. the larger volume in the longer pipe would simply contain more potential energy.
Could be but like I am trying to say, one pipe definitely has a higher potential of energy.
Yes, I would love to see the total setup to really find out more "stuff" about it all. :) Fun physics project for a school if the pipes have no difference in restriction. :)

If it is important to you for more than curiosity. Why not swap the hoses on the T then swap the hoses on the tank etc etc. As you have 2 of almost everything, the old fault finding technique of SAWP TILL YOU DROP should sort it out for you.

Dear hob:
I wonder about the entrance into the hoses. How close to the bottom are they, how are they cut, stuff like that.
David A. Smith

Here is a more detailed description of the plumbing. All threads and bore diameters are NPT (tapered US std pipe threads):
Stinger: 36" length 3/4" pipe Elbow/nipple: PVC 3/4" female right angle to 1" hose barb Hose: rigid 1" ID PVC Nipple: straight 1" hose barb to 1" NPT male T: cast iron, 1" x 1" x 1" riser: 1" PVC 18" length pump inlet: 1" female pump: vane type, 0.5kw DC motor, rated 15 gpm, outlet 3/4" meter: impeller type, inlet/outlet 3/4" hose: 3/4", 10 ft long nozzle: standard auto-shutoff fuel dispensing type all metal parts Ni-plated for compatibility with neat biodiesel all exposed gaskets/seals replaced with biodiesel-compatible materials
As far as I can tell, there is virtually no flow in Tank A until Tank B is quite low. So, theories about flow resistance cannot be correct.
Paul

You sure there isn't a check valve in the piping? In particular, a check valve in tank A? If the check valve took a certain pressure differential to pull it off its seat, it might not open until tank B got low. It would also prevent siphoning. Otherwise, if something isn't completely obstructing the flow from tank A to B then siphoning should equalize the 2 tank levels over night.
-jim

Dear Paul Mathews:
Well, they *can*, but...
Do you have a picture of the setup? If you don't have any webspace, you can email picture(s) to me, and I'll post back with a link to it/them so everyone can see. My email address is:
snipped-for-privacy@D.T, with M = dlzc, D = aol, and T = com
David A. Smith

What he said. But the original poster is already convinced of what cannot be causing the problem. That's always a bad sign.....
Brian Whatcott Altus OK

I think I've been open-minded about any explanations, and I don't have any hypotheses of my own, so I don't understand the comment. I appreciate troubleshooting suggestions, and we'd like to have these tanks empty in unison, so I'll eventually get to a thorough diagnosis. However, I'm procrastinating for these reasons:
a) PVC hoses must be inserted into boiling water to soften them enough to get them over hose barbs. This fueling station is solar-powered, remote from mains power, and getting boiling water on site is a hassle. b) biodiesel is quite messy to work with, so I try to avoid such tasks
The plumbing really is as I've described in detail. I'll work on getting some photos. Paul Mathews

Other posters have pointed out that, once the pump is shut off, you have a siphon between tanks A and B, hence the liquid levels should equalize.
Does this in fact happen?