Another fuel tank plumbing option

[nkemp]

Maybe it is time for a "process check".

The original purpose of this thread was to create a better fuel pump, fuel pick-up & return installation. I, as the instigator, hate the current method. The preferred method keeps the cost reasonable (we all know that we can do anything with sufficient funds).

There has been good discussion on how the GM sumps work. The question discussed is if the GM unit's fluid level can be above the tank level. This discussion is pertinent to the original purpose since it relates to internal components and operation.

I think the consensus is that the GM unit can have a higher fuel level than the tank. If so, the discussion can go back to how we can secure it to the top of the tank and what components to use. A couple alternatives for securing to the top have been discussed (notably casting a thread and lock ring and re-purposing the existing boot).

If there is uncertainty relative to the GM sump level being above the tank fuel level, try this mind game...

The unit is plumbed so that 100% of the fuel returns to the sump. Will the sump be higher than the tank? A: Yes ( I presume we all agree here)

As above except we divert some percentage of fluid flow from going into the sump. Will the level in the sump be higher than the tank? A: Depends and is a function of inflow rates through the backflow valves and the percentage returned.

So let the discussion go on ........................

Nick

[DMCVegas]

The unit is plumbed so that 100% of the fuel returns to the sump. Will the sump be higher than the tank? A: Yes ( I presume we all agree here)

Nick

No, I disagree. The fluid level in the Sump/Baffle can only be as high as the reserve level in the tank. You might have a "check valve" system that keeps fluid escaping, acting as a one-way baffle, but with out a secondary pump whose sole purpose is to keep the Sump/Baffle filled with fluid, it will never rise above the fluid level in the tank.

As such it's hard, if not impossible to to use a replacement baffle that is separate from what we have now if you're going to maintain the same volume of fluid for the pump to be submerged in. We cannot defy gravity and raise fluid levels without a secondary pump.

Even if we were to calculate in the amount of residual fluid left in the return line from the Primary Pressure Regulator all the way back to the return line and used a Sump with a Check valve to prevent leakage back into the tank it won't happen. Let's say you start the car up, the pump begins extracting fluid, the level drops, and then fluid in the gas tank flows in to replace whats missing. Fine. You'll then have an equally balanced fluid level between the inside and outside of the Sump. Even if we figure in the volume of fluid from the return line that RAISES the fluid level inside of the Sump, that gain is only temporary. With the engine consuming fuel, we have a deficit of fuel of what it removes versus what it returns. At the same time the fluid pressure INSIDE of the Sump is greater than the fluid pressure inside. Sure we've got check valves, to keep fluid from flowing out, but the higher hydraulic pressure inside will prevent fresh fluid from flowing in.

Thus while at start-up there may be an initial surplus of fluid inside of the sump, it's only temporary. The pump will simply extract this fluid and it will not be replaced until there is room for additional fluid from the outside to flow in. And it will automatically stop ones the fluid levels equalize.

So no, the Sump level will not be higher than the tank because that is impossible. They'll always be equal.

[Ron]

FWIW
Before reading his post, I was thinking "a sump doesn't have check valves or holes here" and thought about asking about it but it didn't seem to be worth mentioning at the time oops-I should have...Sometimes ya step on one toe trying to miss another...
Afterwords, I only mentioned it as a suggestion for why he and I weren't on the same page (guessing he missed the last design upgrade/pic or ???)...Subtle indeeed!

Actually, I take credit for mudding it up...I should of said, "...but the later discussion was about a [proposed device] with sump/baffle* with check valves (not free flowing holes) and a return into the baffle[], after pointing out it was a baffle [since it sits on rather than below the bottom of the tank, in this case]".

No, I disagree. The fluid level in the Sump/Baffle can only be as high as the reserve level in the tank. You might have a "check valve" system that keeps fluid escaping, acting as a one-way baffle, but with out a secondary pump whose sole purpose is to keep the Sump/Baffle filled with fluid, it will never rise above the fluid level in the tank.

NOW we're on the same page!

(Add "except with KOEO and I'm right with you totally)

The question discussed is if the GM unit's fluid level can be above the tank level.

OK my mistake there...I thought you dave and dave were discussing the unit you were proposing. But I still think the equation you gave and the expansion on it indicates the level in the baffle will never raise above the tank level unless the pump is running and the engine is not.

[nkemp]

Enough talk ... time to go to the lab (garage)

I ran a GM pump unit in a bucket of water to see what happens. Here are the results

• When I put the GM unit in the water, the level inside and outside leveled
• When I removed the partially full sump from the water it does not leak out (works very well!)
• When I ran the pump with all water returning to the bucket, the level remained the same inside & outside (as best I could tell)
• When I ran 100% of the water into the sump, the sump raised to the sump overflow level
• With the sump full, I directed the water to the bucket (none to the sump). The sump level DID NOT DECREASE (This was not expected) It stayed as full as it can get.
• I let the sump sit "off" for a while. The level stayed as full as it can get
  
• I removed the gm unit from the water (With water in the sump) simulating a empty tank. With the pump running, the liquid gets pumped down to empty.

The sump level not decreasing when the return flow was directed to the bucket and not the sump was a surprise. I thought there would be a slight decrease but not returning to the tank's level. Balancing somewhere in-between.

And YES the sump level will get higher than the tank level... and stay there until the tank is empty.

Here is a crude diagram of the GM sump I used (there are other variations). It has a pickup screen that sits on the bottom of the tank, outside the sump (preferred).
Sump-n-pump.jpg

Nick
897

[Bitsyncmaster]

Then that GM pump is drawing fuel from outside the baffle? Does it also release some of that outside draw into the baffle? It must be working like the two pumps DMCvegas suggested.

The ultimate set up would have four pumps. One drawing from each corner of our triangle shaped tank feeding into a baffle for our stock pump.

[nkemp]

Then that GM pump is drawing fuel from outside the baffle? .

Outside the sump :-) ...Yes until the outside is empty then it pumps from inside the sump which has a fuel level equal to the overflow.

Does it also release some of that outside draw into the baffle?

I don't think so. The fuel level inside the sump does not change once it is full. There was no sign of fuel overflowing when pumping from the outside and no return to the sump.

I checked a few things out. Pumping water through everything makes it a little better when using your mouth to check flow direction and backflow valves.
- The valve in the bottom pickup screen prevents backflow from the sump to the tank.
- The pump has two inlets on the bottom, no internal backflow valves between the two. It appears the inlets are part of one chamber inside the pump(based on airflow tests (mouth) & air noise when testing)
- The 2nd inlet screen inside the sump does not have any backflow valves.

It is not clear why it pumps preferentially from the 1st/bottom inlet screen until the main tank is empty.

Nick

[nkemp]

Did anyone look closely at the unit displayed at the DMCH open house to see how they were going to attach the top of the pump unit to the tank? Pictures would be great.

[DMCVegas]

Enough talk ... time to go to the lab (garage)

<SNIP>

I ran a GM pump unit in a bucket of water to see what happens. Here are the results

• When I ran the pump with all water returning to the bucket, the level remained the same inside & outside (as best I could tell)
• When I ran 100% of the water into the sump, the sump raised to the sump overflow level
• With the sump full, I directed the water to the bucket (none to the sump). The sump level DID NOT DECREASE (This was not expected) It stayed as full as it can get.
• [COLOR=red][COLOR=black]I let the sump sit "off" for a while. The level stayed as full as it can get

And YES the sump level will get higher than the tank level... and stay there until the tank is empty.

Nick
897

OK, we already have a problem with this test. You can't recycle 100% of the fluid back into either the Sump nor the tank itself. You need to start reducing the amount that goes back into the Sump to truly simulate a real world environment since the engine consumes a certain amount of the fuel that is collected by the pump. So *technicly* you need to split return into two hoses, with a valve on one end simulating consumed fuel (just run that into another bucket).

Now, that's just as I would see a Sump with a check valve. To understand more, does anyone have a cross-section diagram of the GM Pump internals? Perhaps since I cannot see that, that's why I'm having a difficult time understanding your results.

[nkemp]

Here is the thinking relative to the test...

• If 0% of the fuel goes back to the sump and the sump does not loose level, then anything were some of the fuel goes back would only add to the level (until it reaches the overflow). 0% return is the worst case scenario or is an EXTREME load.
• 100% return is unrealistic during operation... but remember that when you fill the tank it will be filled higher than the sump and fills the sump to the max.
• Based on the test, the only time it draws from the sump is when the tank is empty.
• It is not relevant if the tank looses gas as it is consumed. Would it make any difference how the sump works if you left the car run at while filling with gas? (Not recommended)

I've never seen a cross section on the web (Other than the poor example I made). Also, I've never seen anything on the web that explains the units operation. And there are multiple designs some more confusing than others. But if someone has info I'd love to see it.

I encourage anyone interested to ask for a failed one from a GM repair center(or any service center). Tell them to break off the ports if they are concerned that you'll try to use it. From a junk yard they will likely charge you for one, especially if they have to remove it. Removing one is a challenge..on an Astro and other trucks you have to drop the tank at a minimum.

The test may not be perfect but it does generate usable data. I believe that it shows that the sump will operate at a higher fuel level than the tank fuel level.

[nkemp]

Here is one method of securing the top:
Custom_securing_ring_from_top.jpg

Note that a ring with a matching flange is needed that either mounts on the fuel tank opening lip or goes to the horizontal surface of the tank. If on the lip, something as simple as a hose clamp may be enough to secure the top to the tank.

Just an idea...