The same model in OG Ultra construction would probably have a bit lower dP vs flow curve than the new Ultra. In the BR flow test the new Ultra came out pretty decent, same with the Endurance. As you know, the total media area went up quite a bit on the new Ultra to help retain the flow and holding capacity performance of the OG version. If Fram was smart, they should have made the new Ultra with at least a nylon backed media to keep the pleats more under control, even though it's no longer full synthetic.
The Myth of "Flow Over Efficiency"
- Thread starter ZeeOSix
- Start date
The pleats tore due to wide pleat spacing allowing them to bend sideways from the high dP across the media when the oil is that cold. Brittle weak media is also a factor. Millions of vehicles get cold start-ups well below 10C too. Of course all oil filters will have more dP as they load up, but if they are desiged right they should be able to take the expected loading over the recommened OCI to help keep the filter out of bypass. And the bypass valve should also be designed well enough and set to protect the filter and engine if the dP gets to the point where the bypass valve starts to open.
With cold winter start-ups, it's always a good idea to keep the engine revs down until the oil warms up pretty good ... that's the best way to control filter dP in cold start-up conditons. Obviously, using the correct W grade for the cold start-up temperatures is important. The oil pump is also going to be in pressure relief much sooner and cutting back oil flow to the engine in very cold starts, so that resluts in a reduction of oil flow into the enigne. Another reason to keep the engine revs and load low until the oil warms up.
The dP vs flow data on the new filters shown here all flow close enough to each other at a high flow rate with hot oil that the dP level doesn't really matter. No engine (expect for those over pumped Suarus, lol) is so sensitive that a 2-3 PSI of dP difference at 9-10 GPM of flow is going to cause a problem - if an engine is that senisitve to oil volume, then it's a bad engine design. Once the filters starts loading up then the dP vs flow curve is going to change dependant on the holding capacity and level of loading. Larger filters with more media area help keep the dP lower with the same loading. Also, lower efficiency filters will shed more debris as they load up compared to a higher efficiency filters.
It helps reduce dP and lower efficiency. A filter with a big tear in the media will have lower dP vs flow. That might actually qualify as "flow over efficiency".How would leaky bypass valves affect these measurements
Last edited:
The dP vs flow data is on one graph to easier show how they compare at 5 GPM and 9 GPM.
X-axis is the filter number from the Filter Key table.
Y-axis is the dP in PSI.
Test conditions are on the graph.
Note for the PGI filter fans, the STP Ext Life (Filter No. 18) shows to have a pretty decent dP level (middle of the range), and they are rated pretty high in efficiency too (99% @ 20u).
X-axis is the filter number from the Filter Key table.
Y-axis is the dP in PSI.
Test conditions are on the graph.
Note for the PGI filter fans, the STP Ext Life (Filter No. 18) shows to have a pretty decent dP level (middle of the range), and they are rated pretty high in efficiency too (99% @ 20u).
Last edited:
I've seen this happen as well but never understood why. Shouldn't the bypass open before the media collapses or tears?
If the media can't take dP up to the bypass valve setting then Huston has a problem. Wide pleat spacing is the root cause of media tearing.
Last edited:
I think many people are missing a key point that ZeeOSix is making: that “flow vs efficiency” is a false dichotomy. In a traditional oil pump setup with a fixed displacement and a recirculating bypass, there’s is no “flow vs efficiency.”
Rather, it’s flow 6gpm with X efficiency in the filter, or flow 6gpm with Y efficiency.
The higher efficiency filter does not reduce flow if the bypass is closed. Rather, the penalty for higher efficiency is reduced oil pressure after the filter. Which meant the pump will go into bypass sooner as flow or viscosity (or both) rise.
Let’s say an oil pump goes into bypass as 50psig. This would be the cracking pressure of the bypass— at its location. Once the bypass cracks, the pump can’t increase flow (or raise oil pressure) except by saturating the bypass.
So if you have a pump that bypasses at 50psig and your higher efficiency filter drops 3psi more, then you’d have 3psi less pressure feeding the main oil supply. This is a non-issue.
Rather, it’s flow 6gpm with X efficiency in the filter, or flow 6gpm with Y efficiency.
The higher efficiency filter does not reduce flow if the bypass is closed. Rather, the penalty for higher efficiency is reduced oil pressure after the filter. Which meant the pump will go into bypass sooner as flow or viscosity (or both) rise.
Let’s say an oil pump goes into bypass as 50psig. This would be the cracking pressure of the bypass— at its location. Once the bypass cracks, the pump can’t increase flow (or raise oil pressure) except by saturating the bypass.
So if you have a pump that bypasses at 50psig and your higher efficiency filter drops 3psi more, then you’d have 3psi less pressure feeding the main oil supply. This is a non-issue.
What are the fluid dynamics of pasta makers?
Here's a side project I did for my own curiosity (and parallels your comment) that shows the effect of 15 PSI more flow resistance (ie, added back pressure) to the pump's pressure relief valve - which is a good amount of added dP from any oil filter for whatever reason. This example is using a Melling oil pump with the same swept volume per rev (0.96 in^3/rev) that the OEM pump also has. The graph shows the pump output flow performance curve when running on a GM LS engine with 5w30 oil at 200 deg F.
Note that the GM LS has the filter bypass valve built into the filter mount, but I'm using this example to get a general point across about how added dP in the oiling system impacts the pump output volume when it gets in to pressure relief. The oil pump in the LS is a PD pump with a spring loaded pressure relief valve. Also, stock LS engines redline around 6500 RPM. A modified one can rev higher with the right mods.
The purple line represents the oiling system flow with a normally flowing new oil filter, let's say one that has 8 PSI of dP at 9 GPM with hot 5w30 grade oil. The green line would represent the pump output flow drop due to adding 15 PSI more dP on top of the initial 8 PSI dP from the filter. These curves are without the filter going into bypass ... say the filter's bypass setting was over 23 PSI (8 from the filter +15 added dP = 23 PSI). If the filter had a lower bypass valve setting, then the green line would not drop as far down from the purple line - ie, if the bypass valve was set to 16 PSI (8 PSI from the filter + 8 PSI added dP), then the green line would only drop about 1/2 GPM below the purple line. If the filter only loaded a slight amount over the OCI, then the green line would not be far below the purple line.
I ran 5 different brands of oil filters on my Z06 and never saw any oil pressure difference beyond maybe 1-2 PSI max (digital gauge with 1 PSI resolution) at the same oil temperature (also a digital gauge with 1 deg F resolution) and all the way to 6000 RPM. So the difference in dP vs flow of all those filters was small enough to barely see any oil pressure difference if I held oil temp and RPM at the same operating points. And some of those were high efficiency filters, like the yellow PureOne which was advertised as 99.9% @ 20u back then. Also an ACDelco Gold when they were wire backed full synthetic media back then, were also high efficiency.
The loss of flow with 15 PSI of added dP in the system would cause a 1 GPM loss in flow from about 4000 RPM to redline for an LS engine. The oiling system should be designed well enough to operate adequately and keep the engine safe with that level of pump flow reduction - that's a main goal when matching an oil pump to the oiling system. If that isn't done properly, then a bad pump to engine mismatch could cause issues (rarely seen). The pump shouldn't be specified to run on the ragged edge of destruction if an oil filter has 2-5 more PSI of dP at 9-10 GPM when the filter is new. The oil filter designer should be designing the filter with a proper also, with good dP vs flow, adequate holding capacity, and a bypass valve setting that flows well enough to keep the engine happy with oil flow if the filter dP gets too high.
The engines that specify an OEM filter with a high bypass setting like 23+ PSI (regardless of filter efficiency) are basically saying that they are confident the engine will still be getting enough oil flow if the filter does hit bypass for whatever reason, which means the oil pump in those engines would also be cutting back some flow to the engine similar to this example. An oil pump with a higher pressure relief setting also helps keep oil flow volume up as the system dP increases for whatever reason - ie, filter clogging, thick oil at start up, etc.
Last edited:
The engines that specify an OEM filter with a high bypass setting like 23+ PSI (regardless of filter efficiency) . . .
Like Subaru.
If I am understanding this, one takeaway is that when considering which oil filter to buy, don't worry about "flow" as all filters (or at least the ones tested here) provide more than adequate flow (all other things being equal), and instead, focus on particle filtration/micron rating.
Is that correct?
Is that correct?
Excellent post.
Note that the regulation points for both curves are at RPM ranges outside of typical cruise and street operation. This illustrates another key point Zee was making and I was echoing-- below the regulation point, the curves are the same. There is ZERO difference in flow between filters because the pump is positive displacement. The curves for the two filters are identical until they hit their respective regulation points.
Many vehicles now have variable output oil pumps--still positive displacement, but variably positive. This deserves a separate posting.
A traditional fixed positive displacement (PD) oil pump--whether gerotor, gear, vane, or other is sized in such a way that compromises the extremes of its capacity. If it is sized for cranking performance and low oil delay, it will be significantly over-sized for the engine's higher RPM demand and go into regulation quite early. Cummins pumps are typically larger like this owing to the low RPM at which they can be loaded. The points @ZeeOSix identified above a the regulation points are well under 1500rpm on my Cummins diesel (L6 and up at least). This is because not only are the engines low-rpm engine, they have oil squirters, turbochargers, and other accessories that need rapid oil pressure rise when cold.
The LS data above shows the GM is using a proportionally much smaller oil pump that doesn't regulate until above 2500rpm. (which is not to say undersized or incorrect-- just that the reg point is higher).
Many vehicles now-- including my own Honda with the K20C4-- have variable oil pumps. The variable pump allows for a larger capacity pump at cranking without the inefficiency of gobs of recirculation at higher RPM. These pumps have a bypass that almost never opens because it's only for emergency overpressure relief; rather, the oil pressure is regulated by reducing pump flow. They are typically vane pumps that vary pump displacement by shifting the center axis of the housing relative to the center axis of the pump head/rotor.
The image has been shown here on BITOG before:
Having established that oil filter restriction has no effect on oil pump output flow below the regulation point, and having had @ZeeOSix highlight how the flow changes above regulation with LS exemplar, we should not ask how oil filter restriction affects a variable pump design.
This depends on the pressure reference. If the oil pump is taking pressure reference from a main oil supply rifle (as one might expect-- AFTER the oil filter), then the variable pump essentially mimics the fixed-geometry pump in terms of being insensitive to filter restriction.
In the unlikely event that the pump's pressure signal is taken immediately at the pump's output (before the filter), then the filter restriction would add to the oil system's restriction to define the pressure signal to the pump's regulation mechanism.
The LS data above shows the GM is using a proportionally much smaller oil pump that doesn't regulate until above 2500rpm. (which is not to say undersized or incorrect-- just that the reg point is higher).
Many vehicles now-- including my own Honda with the K20C4-- have variable oil pumps. The variable pump allows for a larger capacity pump at cranking without the inefficiency of gobs of recirculation at higher RPM. These pumps have a bypass that almost never opens because it's only for emergency overpressure relief; rather, the oil pressure is regulated by reducing pump flow. They are typically vane pumps that vary pump displacement by shifting the center axis of the housing relative to the center axis of the pump head/rotor.
The image has been shown here on BITOG before:
Having established that oil filter restriction has no effect on oil pump output flow below the regulation point, and having had @ZeeOSix highlight how the flow changes above regulation with LS exemplar, we should not ask how oil filter restriction affects a variable pump design.
This depends on the pressure reference. If the oil pump is taking pressure reference from a main oil supply rifle (as one might expect-- AFTER the oil filter), then the variable pump essentially mimics the fixed-geometry pump in terms of being insensitive to filter restriction.
In the unlikely event that the pump's pressure signal is taken immediately at the pump's output (before the filter), then the filter restriction would add to the oil system's restriction to define the pressure signal to the pump's regulation mechanism.
*NOW ask. Sorry, caught too late to edit.
Yes, that was an example to show that even on the high volume pumps used on Subarus, and with the OEM filter that has such a high bypass setting, the oil pump is going to be cutting back oil flow to the engine if the pump is in relief and the filter is producing dP that high for whatever reason. I would really like to see the dP vs flow measurement on an OEM Subaru oil filter, it can't be any lower than say the Toytoa OEM filter which is only a couple of PSI lower in dP at 9-10 GPM compared to some good high efficiency oil filters. So if filter efficiency is a goal, don't let the myth that high efficiency filters are "too restrictive" get in the way of using them.
That's basically the bottom line for 99.99% of the engines running around on the streets. Obviously if someone is racing their car in track use it would be wise to use a filter that's known to have a lower dP vs flow for peace of mind. Also, don't go to the track with a possibly loaded up dirty oil filter. The bypass valve in the filter is still there to help ensure no lack of lubrication if the dP across the filter does get to bypass level, but it would still be good to try to avoid excessive filter bypassing if possible. A relatively flow restrictive oil filter with a high bypass setting could actually result in less oil to the engine at high RPM if the filter for some reason has experiences high enough dP to make the filter bypass valve open. Using an over-sized filter with more media can also help.
Last edited:
It seems that it's not only OEM filters that have that high bypass setting, as it seems at least some aftermarket filters do as well; I assume they are designed to meet the OEM spec or close to it. For instance, the CarQuest Premium filter (thanks to your advice) I use on my Subaru shows a Relief Valve PSI of 20-30, which is much higher than the one I use on my Camry which is 14-15 PSI, and is similar to the Toyota OEM filter.
^^^ Yes, most aftermarket filters with high bypass valve settings are mostly trying to reflect OEM bypass valve settings, regardless of how well they flow or not. Part of the higher bypass setting may also be because the filter media is a bit more flow restrictive, so the bypass is set to take that in to account and not open as often. Or they set the bypass higher with an expectation that the filter loading will add a good chunk of dP to the filter over the OCI.
A higher bypass setting keeps the filter out of bypass better, but if for some reason the dP does get high enough to hit bypass then the oil volume will be cut back a bit more when the pump is in pressure relief at high RPM compared to a filter with a lower bypass setting. These car manufactures that spec oil filters with a high bypass valve setting must know that cutting some flow volume output from the pump at higher RPM is still going to be enough to keep the engine from damage.
A higher bypass setting keeps the filter out of bypass better, but if for some reason the dP does get high enough to hit bypass then the oil volume will be cut back a bit more when the pump is in pressure relief at high RPM compared to a filter with a lower bypass setting. These car manufactures that spec oil filters with a high bypass valve setting must know that cutting some flow volume output from the pump at higher RPM is still going to be enough to keep the engine from damage.
I would assume the tearing forces and mechanism are dynamic at cold start up with extreme delta P (partially emptied filter and mains drills), otherwise the pressure will be equal on the exterior side(s) of the pleats regardless of the formed pleat angle. I am envisioning the forces to be higher mid-fold and near the endcap adhesive, as exterior forces should be equal at the peak of the pleat. I have worked closely with plastic (LCP) mold design engineers. Our startup had a "semi-supercomputer" to run flow vector analysis and mold fill for hours overnight. I am sure same is done with filter design. Too bad we closed shop a decade ago, we could model this with my bud Larry
99.9% of this stuff I don't even comprehend. I'm nothing more than you're average " Joe" self taught mechanic. But, as a kid, 12-13 yrs old, I was always told " if it's cold outside, let the engine idle for awhile so the oil can warm up". It can get pretty cold in NY ..10-15 degrees in the winter. I always let the car warm up to get fluids up to near operating temps. Taught my wife to do the same. No blown radiator hoses, tranny leaks, or engine problems. Yeah, you waste some fuel, but big whoop.
There have been many photos of filters posted here over the years with torn media showing that the pleats are abnormally bent to one side from the oil flow dP force. The excessive bending is what tears the pleat where it meets the end cap because that's were the pleat bending force is concentrated. If the media is brittle and weak, and tears pretty easily, then tearing at both ends of the bent pleat is most likely going to happen - that has also been shown in this forum where both ends of a bend over pleat were torn.
The media is definitely more likely to tear on a cold start combined with too much engine revving before the oil warms up, and that cam cause the dP level to hit the bypass valve setting. The wider the pleat spacing and the weaker the media combined with high dP makes it more likely for torn media. The wide pleat spacing is a manufacturing defect IMO, as the same oil filter brands & models with correctly spaced pleats don't ever show torn media, so they can take the dP up to the bypass valve opening and not cause media damage like an well designed and manufactured oil filter should.
Last edited:
