Fram Endurance Flashlight Test in canister

[4WD]

The flow field entering a constriction and exiting a constriction changes. For flow entering the constriction, the fluid accelerates. For flow exiting the constriction, the flow decelerates. This induces a pressure drop in addition to the pressure drop of flowing through the orifice itself. This additional pressure drop reduces the flow that will go through the constriction at any given total pressure drop available. That is my point. The calculation used above assumed a plain rectangular passageway for the rectangular slot, but used an orifice calculation for the round hole, which included the effects of the constriction and expansion. That is apples and oranges. You need to use the same methodology for both of them.

Given the same flow area and flow, which was the assumption of the above calculation, the rectangular slot with a very large aspect ratio will have a higher pressure drop than a round hole at the same flow rate. The velocity through both of them will be the same.

Of course if the “leak” geometry remains the same - but the dP slowly increases as the filter media loads - a higher percentage of flow will divert towards the “leak” …

 

[Chumango]

Of course if the “leak” geometry remains the same - but the dP slowly increases as the filter media loads - a higher percentage of flow will divert towards the “leak” …

Of course the flow through the leak path increases as the filter media loads, but this entire discussion has been about a calculated flow at a specific pressure drop. I pointed out that using an equivalent area round hole gives a higher flow than will actually occur in a rectangular slot with a high aspect ratio. This holds true regardless of the pressure drop available - the round hole flows more than the slot with the same flow area.

The original calculation, using the round hole, gave a high number for flow through the slot and presented it as fact in support of an argument about the impact of the gap on filter efficiency, when in fact it was overstated. I'm not saying there is no impact from the gap, I'm just saying that the numbers presented were overstated. In reality, it would be very challenging to calculate in any precise manner the actual impact of that gap. Even computational fluid dynamics would be challenging to do, coming up with the correct input conditions. Testing is warranted.

 

[garageman402]

The flow field entering a constriction and exiting a constriction changes. For flow entering the constriction, the fluid accelerates. For flow exiting the constriction, the flow decelerates. This induces a pressure drop in addition to the pressure drop of flowing through the orifice itself. This additional pressure drop reduces the flow that will go through the constriction at any given total pressure drop available. That is my point. The calculation used above assumed a plain rectangular passageway for the rectangular slot, but used an orifice calculation for the round hole, which included the effects of the constriction and expansion. That is apples and oranges. You need to use the same methodology for both of them.

Given the same flow area and flow, which was the assumption of the above calculation, the rectangular slot with a very large aspect ratio will have a higher pressure drop than a round hole at the same flow rate. The velocity through both of them will be the same.

I think you're gonna get the "positive displacement pump" explanation from @ZeeOSix. When cross section area decreases, the velocity increases, then when the area increases, velocity decreases. Volume is constant. The positive displacement pump will send a constant flow (volume): if it decreases through the filter, it will increase through the leak or bypass valve. The oil is going somewhere, that value is constant.

 

[Chumango]

I think you're gonna get the "positive displacement pump" explanation from @ZeeOSix. When cross section area decreases, the velocity increases, then when the area increases, velocity decreases. Volume is constant. The positive displacement pump will send a constant flow (volume): if it decreases through the filter, it will increase through the leak or bypass valve. The oil is going somewhere, that value is constant.

This is correct. However, this is not the point being discussed. See my post above.

 

[twX]

My point exactly. Better yet, use a round pipe for the round hole rather than an orifice with a slight constriction. The orifice calculation probably uses a discharge coefficient of around 0.62, which probably won't amount to a lot if you only have a small constriction, but it will still be there. Using straight pipe, that will not be a factor.

The width to height ratio for the case being considered here is 65.

Here's a calculation for a 20:1 ratio rectangular orifice, which I think is more realistic based on the photos. For a 20:1 rectangle, pressure drop is ~3.5 times higher than it is for a circular opening. To account for this using the circular orifice model, we'll use 0.8 psi / 3.5 = 0.23 psi as the target for the calculation, which should be equivalent to 0.8 psi with a rectangular orifice. I used the same parameters as ZeeOSix.

Bypass flow is 0.32 GPM. That's still 10% of total oil flow if we assume a media flow rate of 3 GPM. Restriction increasing by 3.5 times only cuts the flow in half due to the non-linear flow-dP relationship. This means that the flow is not all that sensitive to the shape of the gap. These calculations should be in the right ballpark so long as the estimated area is correct.

 

[Ronn]

This is correct. However, this is not the point being discussed. See my post above.

Chumago hit the nail on the head :
“Even computational fluid dynamics would be challenging to do, coming up with the correct input conditions. Testing is warranted.”
There are simply way too many variables and interactions to use a formula in this case. So…. the old adage…”Garbage In, Garbage Out” applies.
We need a test and autopsy, or we will find ourselves mired in endless speculation.
As for my personal situation, NO FRIGGIN’ way I’m yanking my Endurance. I’ll rely on the test to date…many of which are less than a year old, lack of leak testing notwithstanding.

 

[4WD]

Of course the flow through the leak path increases as the filter media loads, but this entire discussion has been about a calculated flow at a specific pressure drop. I pointed out that using an equivalent area round hole gives a higher flow than will actually occur in a rectangular slot with a high aspect ratio. This holds true regardless of the pressure drop available - the round hole flows more than the slot with the same flow area.

The original calculation, using the round hole, gave a high number for flow through the slot and presented it as fact in support of an argument about the impact of the gap on filter efficiency, when in fact it was overstated. I'm not saying there is no impact from the gap, I'm just saying that the numbers presented were overstated. In reality, it would be very challenging to calculate in any precise manner the actual impact of that gap. Even computational fluid dynamics would be challenging to do, coming up with the correct input conditions. Testing is warranted.

I understand hydraulics - but all you have to do is look at post #230 to know CL and/or Fram has made a mistake not having an elastomer seal with metal end caps …
If that gap is over 20-30 micron - Done.

 

[Ronn]

Bypass flow is 0.32 GPM. That's still 10% of total oil flow if we assume a media flow rate of 3 GPM. Restriction increasing by 3.5 times only cuts the flow in half due to the non-linear flow-dP relationship. This means that the flow is not all that sensitive to the shape of the gap. These calculations should be in the right ballpark so long as the estimated area is correct.

"Bypass flow is 0.32 GPM. That's still 10% of total oil flow if we assume a media flow rate of 3 GPM"
Even if one assumes this calculation, which is a leap of faith, you then have 10% leaking past the filter....but then getting recirculated back through the filter continuously! So, the only way really to quantify the ultimate result is to conduct a thorough filter test.

 

[ZeeOSix]

The rectangular calculation is treating it as a section of pipe, not an orifice, and does not account for the constriction entering the narrow slot and the expansion leaving the narrow slot (i.e. what an orifice actually experiences). If the online calculator cannot do a rectangular orifice, for a correct comparison, do the round orifice as a single piece of pipe, not an orifice, and see what that says.

After digging some more, I agree that the on-line flow/dP calculator for the square pipe isn't accounting for entrance and exit effects like with an orifice. It's just calculating the flow/dP of the square pipe flow path only without any orifice coefficient.

However, I did find an in depth study that investigated the orifice coefficients of equal flow areas but with different shaped orifices. Based on that info (below), a square orifice will have a flow coefficient that's around and average of 5.2% less than a round hole orifice. Therefore, one can use the round orifice on-line calculator and simply decrease that flow result by 5.2%.

Using the round orifice with 0.03 sq-in of area (2 x 0.020 x 0.75 inch as reported by @Glenda W.) gives a flow of 0.665 GPM. The flow reduction factor due to being a rectangle orifice coefficient would be 0.665 x 0.948 = 0.630 GPM.

If it's modeled more accurately as two separate orifices each being 0.020 x 0.75 inches, the flow through each is 0.30 GPM x 0.948 = 0.284 GPM. Two leaks at 0.248 GPM = 0.57 GPM total leakage. So that's 3.0 (media)+ 0.57 (leaks) of total flow, which is 0.57/3.57 = 16% of the flow leaking past the media.

 

[ZeeOSix]

Those two models aren't directly comparable, since the orifice model contains a large area reduction whereas the straight pipe model doesn't. If you set D1 to be just slightly higher than D2 in the orifice model, you'll get a more comparable estimate, and the result should be almost the same as if you were using the circular straight pipe model.

Comparing rectangular vs circular straight pipe models, a square pipe with a 10:1 ratio of width to height seems to have around 2 to 2.5 times the restriction of a circular pipe. For a 40:1 ratio, the rectangular pipe is around 10 times more restrictive.

See my last post where a coefficient correction factor for a rectangle orifice as used.

 

[ZeeOSix]

The original calculation, using the round hole, gave a high number for flow through the slot and presented it as fact in support of an argument about the impact of the gap on filter efficiency, when in fact it was overstated. I'm not saying there is no impact from the gap, I'm just saying that the numbers presented were overstated.

After the last model calculation with the orifice coefficient correction factor, the difference between a round orifice and a rectangular orifice is only around 5% less flow through the rectangular orifice under the same flow conditions. Just using the round orifice calculation in the first place therefore wasn't very "overstated" IMO.

In reality, it would be very challenging to calculate in any precise manner the actual impact of that gap. Even computational fluid dynamics would be challenging to do, coming up with the correct input conditions. Testing is warranted.

I think the calculation is close enough to get an idea of just how much a small leak path area will flow with very little dP.

The real test would be to spend a ton of money on actual ISO 4548-12 tests with the goal to equate leak volume past the media in terms of impact on the efficiency of the filter assembly.

 

[Ronn]

So in spite of all this speculation and numerous calculations, we have an independent Fram test conducted last Nov which showed superior filtration of the Endurance compared to other models in their line up. The “autopsy”was done but doesn’t include a flashlight test. However, this test falls well within the time period when the bypass defect was known to be present. Now I'm waiting for the chorus to chime in..."BEATING THAT DEAD HORSE" refrain..
If you want to see HOW they tested the video is here

 

[ZeeOSix]

My point exactly. Better yet, use a round pipe for the round hole rather than an orifice with a slight constriction. The orifice calculation probably uses a discharge coefficient of around 0.62, which probably won't amount to a lot if you only have a small constriction, but it will still be there. Using straight pipe, that will not be a factor.

The width to height ratio for the case being considered here is 65.

If you apply the the orifice coefficient for a rectangular orifice as shown in the table in post 269 above (avg around Cf = 0.50 for a rectangular orifice) to the flow found in post 249 using a just the rectangular pipe, then the flow at 0.8 PSI dP comes out basically the same as using the round orifice with the delta correction factor of 5.2% as pointed out earlier ... specifically: 1.14 GPM (square pipe model) x 0.50 coefficient for a square orifice = 0.57 GPM. The flow found in post 269 with the coefficient correction factor of 5.2% less flow for a square vs round orifice was also 0.57 GPM. All at the same viscosity and dP conditions with the same flow area. Just corrected for the shape of the orifice.

 

[Chumango]

See my last post where a coefficient correction factor for a rectangle orifice as used.

The rectangle is not defined, and given that it is not much different from a square, the aspect ratio is not very large. I maintain that a slot with a very large aspect ratio would have a very significantly different coefficient.

All of this is fairly moot when it comes to determining the effect of the gap. That is something left to testing. Or to CFD analysis.

 

[ZeeOSix]

Here's a calculation for a 20:1 ratio rectangular orifice, which I think is more realistic based on the photos. For a 20:1 rectangle, pressure drop is ~3.5 times higher than it is for a circular opening. To account for this using the circular orifice model, we'll use 0.8 psi / 3.5 = 0.23 psi as the target for the calculation, which should be equivalent to 0.8 psi with a rectangular orifice. I used the same parameters as ZeeOSix.

Interesting to see another modeling angle, but I'd think using the orifice correction factor would be a more direct method. Your assumption on the target dP of 0.23 PSI to achieve the equivalent answer may be off some.

 

[Pablo]

So in spite of all this speculation and numerous calculations, we have an independent Fram test conducted last Nov which showed superior filtration of the Endurance compared to other models in their line up. No “autopsy” was done, but this test falls well within the time period when the bypass defect was known to be present. Now I'm waiting for the chorus to chime in..."BEATING THAT DEAD HORSE" refrain..

View attachment 239468

Those last two on the right must have been sealed pretty well. Maybe they were older (?) or ???

 

[ZeeOSix]

The rectangle is not defined, and given that it is not much different from a square, the aspect ratio is not very large. I maintain that a slot with a very large aspect ratio would have a very significantly different coefficient.

All of this is fairly moot when it comes to determining the effect of the gap. That is something left to testing. Or to CFD analysis.

Then you tell us what the coefficient impact would be for different aspect ratio rectangles. The study didn't say anything about how the aspect ratio of a rectangular orifice effects the discharge coefficient. Maybe it really doesn't matter that much - can you show proof otherwise? Or they used one that wasn't crazy in aspect ratio ... they didn't say what the dimensions were.

 

[twX]

The real test would be to spend a ton of money on actual ISO 4548-12 tests with the goal to equate leak volume past the media in terms of impact on the efficiency of the filter assembly.

There's another ISO test that specifically tests bypass valve performance, ISO 4548-2. It includes a test for bypass valve leakage. This test is a lot simpler than the efficiency test. It requires the filter element to be blocked off, a pressure applied to the filter inlet, and a method of measuring the leakage flow.

A garage scientist could probably do a test like this with a container of oil, gravity-fed to a filter head, elevated to produce ~1 psi, and with a bucket & stopwatch used as a flow meter.

For a canister type filter, you'd have to cut open the canister, glue a sheet of material over the filter element to block it off, then reassemble the filter well enough for it to handle the spring pressure and ~1 psi of oil pressure. Some JB Weld should do the trick.

 

[Ronn]

Those last two on the right must have been sealed pretty well. Maybe they were older (?) or ???

You can see they took out bypass valves, but no up-close pictures. The ones shown look identical to each other and similar to bypass with defects shown here and elsewhere.

 

[Chumango]

Then you tell us what the coefficient impact would be for different aspect ratio rectangles.

A function of the hydraulic diameter (4 x flow area / wetted perimeter), but probably not a simple relationship. Probably experimentally determined.