Filter Flow Performance - Any Good Ref Data ?

[edhackett]

My Bonneville would be a good example of where a more restrictive filter might cause problems. It has twin oil pumps; one for normal high pressure lubrication flow and a high volume, low pressure pump to insure adequate flow at low rpm for cooling. I can see where a restrictive filter might reduce the output of the cooling pump.

Here's the beta ratio for the WIX specified for the Triumph: 2/20=13/52 with an 8lb by-pass. That's just a boulder catcher compared to a PureOne, and I'm sure has a lower flow resistance.

Ed

 

[SuperBusa]

Originally Posted By: edhackett
My Bonneville would be a good example of where a more restrictive filter might cause problems. It has twin oil pumps; one for normal high pressure lubrication flow and a high volume, low pressure pump to insure adequate flow at low rpm for cooling. I can see where a restrictive filter might reduce the output of the cooling pump.


Never heard of an oil system setup like that. What doesn't make sense is the term "high volume, low pressure pump" ... because in order to get high volume through the same fixed flow resistance you would need higher pressure. Maybe in the low RPM mode, the pump sends the high volume oil to a different circuit in the engine which has much lower flow resistance? ... that's my assumption of how your system works.

Originally Posted By: edhackett
Here's the beta ratio for the WIX specified for the Triumph: 2/20=13/52 with an 8lb by-pass. That's just a boulder catcher compared to a PureOne, and I'm sure has a lower flow resistance.


"Boulder catcher" is a good analogy. Yes, I believe any filter with a low by-pass setting is inherently less restrictive to flow if the designer set the by-pass accordingly. IMO, I think the PureOne by-pass is set higher to compensate for its increased flow resistance.

 

[edhackett]

Your assumption is a good one. I'm curious as to how the twin pump system works myself. I haven't been able to find anything online. I can think of several ways this could be done. I'll be just a couple of blocks from the dealer tomorrow afternoon. I'll stop in and find out exactly how the dual pump system is configured.

Ed

 

[Gary Allan]

Quote:
BTW - they are also not only "points of fluid acceleration" but points of "flow restriction" ... which is exactly the context of the discussion we are having with regard to "holes"

But they don't translate to flow loss. At least if there is no relief event at the pump.

Quote:
Tell ya what ... go close up 7 of the 8 holes in the base of your oil filter and then tell me how much the delta P across the filter assembly changed ... it will be significant.


With a few conditions understood, not much, imo. It will vary with a few applications.

Quote:
You can't tell me that the pressure drop with the same flow being forced through it will not increase significantly and that only the flow "accelerated" without a corresponding pressure drop.


Without a relief event at the pump, there is no such thing as a pressure drop. It's an elevation of pressure upstream.

..but let me show you these lame schematics and see if it changes your conceptual view.

 

[SuperBusa]

Originally Posted By: Gary Allan
Quote:
BTW - they are also not only "points of fluid acceleration" but points of "flow restriction" ... which is exactly the context of the discussion we are having with regard to "holes"

But they don't translate to flow loss. At least if there is no relief event at the pump.


It's certain that a very small hole will cause a flow loss compared to a great big hole if the same delta P is across it. Likewise, it would take way more delta P to make the same flowrate go through a very small hole compared to a very large hole. This is basic fluid dymanics.

Originally Posted By: Gary Allan
Quote:
Tell ya what ... go close up 7 of the 8 holes in the base of your oil filter and then tell me how much the delta P across the filter assembly changed ... it will be significant.


With a few conditions understood, not much, imo. It will vary with a few applications.


If there is any significant oil flow trying to go through a filter that had 7 of its 8 base holes closed up, there would be a very significant pressure drop between the inlet and outlet of that oil filter.

Originally Posted By: Gary Allan
Quote:
You can't tell me that the pressure drop with the same flow being forced through it will not increase significantly and that only the flow "accelerated" without a corresponding pressure drop.


Without a relief event at the pump, there is no such thing as a pressure drop. It's an elevation of pressure upstream.

..but let me show you these lame schematics and see if it changes your conceptual view.

Those aren't that lame ... gives a basic concept of what's going on. In the 1st schematics above, the oil filter is basically the only "resistor" at the time. But, it's still true that the delta P across the filter is still solely dependant on the viscosity and flowrate of the oil going through it. In the 2nd schematic above, the engine is now also added to the flow circuit, and conceptually you are indicating that it is MUCH more restrictive than the filter.

If that's the case, the engine "resistance" has now choked the flowrate to a mere trickle compared to the capability of the filter. But is this really accurate? Without actual numbers, can you really say the engine's resistance is magnitudes higher than the filter in the series circuit? I would like to see real numbers if possible ... what's a typical LS6 engine oil pump put out for flowrate just before the relief valve kicks in?

And also, if an engine's oil circuit is really that restrictive, then it could very well be that an oil filter never goes into by-pass mode because the flowrate through it is never high enough to create a delta P great enough to cause it to by-pass. Without real numbers of an engine's oil pump output vs. RPM, the discussion is all "theoretical" IMO.

 

[Gary Allan]

Quote:
It's certain that a very small hole will cause a flow loss compared to a great big hole if the same delta P is across it. Likewise, it would take way more delta P to make the same flowrate go through a very small hole compared to a very large hole. This is basic fluid dymanics.


In wrapping your head around this, you have to concede that there is no flow loss without an oil pump relief event. Where else can it go? If that relief event doesn't occur ..then 100% of the sensible flow MUST travel through any and all intermediate restrictions. It has no choice. The fluid will either accelerate or decelerate as required.

Quote:
If that's the case, the engine "resistance" has now choked the flowrate to a mere trickle compared to the capability of the filter. But is this really accurate? Without actual numbers, can you really say the engine's resistance is magnitudes higher than the filter in the series circuit?


Yes. Many times higher 30/40/50:1 ..it depends on a few factors ..but generally speaking it WAY HUGE.

Quote:
If that's the case, the engine "resistance" has now choked the flowrate to a mere trickle compared to the capability of the filter.


Not exactly. The engine is developing high pressure due to the identical ..uninterrupted ..unaltered ...flow having to acelerate to exit the various points of egress that are collectively much smaller than the filter in comparison.

How much oil is passing through a 2" pipe @ 5gpm?
How much oil is passing through a .000000000024" pipe @ 5gpm?

A: 5gpm. The difference will be the velocity and the pressure developed. The 2" pipe may develop 5psi ..while the .0000x" pipe may develop 2500PSI.

Naturally, one would need an incredibly HIGH oil pump relief setting to achieve this ..but it should demonstrate the concept.

Quote:
Without real numbers of an engine's oil pump output vs. RPM, the discussion is all "theoretical" IMO.


Yes. I've constructed this from observed events. I can't quote volumes ..merely pressures and their inter-relationship(s).

When you're in relief, you'll see the peak pressure above the filter and the reduced pressure below the filter. The upstream pressure remains the same (usually) ..the the downstream pressure rises to match it. This is due to flow reduction due to relief action. As the oil gets more fluid ..or overcomes its static inertia, it will close the gap between pump output flow ..and realized engine flow. Once the relief is closed, there's virtually no pressure elevation from the filter. While in relief, it IS a pressure drop. A drop that is limited by the bypass valve setting.

 

[SuperBusa]

Originally Posted By: Gary Allan
Quote:
It's certain that a very small hole will cause a flow loss compared to a great big hole if the same delta P is across it. Likewise, it would take way more delta P to make the same flowrate go through a very small hole compared to a very large hole. This is basic fluid dymanics.


In wrapping your head around this, you have to concede that there is no flow loss without an oil pump relief event. Where else can it go? If that relief event doesn't occur ..then 100% of the sensible flow MUST travel through any and all intermediate restrictions. It has no choice. The fluid will either accelerate or decelerate as required.


We are kind of going in circles now. Yes, I agree that if there is no pump relief going on then ALL the pump's output must go through the fiter/engine circuit. BUT, if you suddenly created a huge resistance in the filter/engine ciruit, the oil pump would go into relief mode becasue the pressure would increase significanly when trying to stuff the same flowrate through a smaller hole (ie, higher reistance). This is exactly the concept I was describing above - my quote: "Likewise, it would take way more delta P to make the same flowrate go through a very small hole compared to a very large hole." That "way more delta P" probably just caused the oil pump to go into bypass relief mode.

Originally Posted By: Gary Allan
Quote:
If that's the case, the engine "resistance" has now choked the flowrate to a mere trickle compared to the capability of the filter. But is this really accurate? Without actual numbers, can you really say the engine's resistance is magnitudes higher than the filter in the series circuit?


Yes. Many times higher 30/40/50:1 ..it depends on a few factors ..but generally speaking it WAY HUGE.


I have never heard of any hard numbers of the ratio ... but it makes sense. Based on your pressure gauge photos in the other thread, the ratio was more like 8 psi delta P at 80 psi, or sounds like maybe closer to 10:1.

Originally Posted By: Gary Allan
Quote:
If that's the case, the engine "resistance" has now choked the flowrate to a mere trickle compared to the capability of the filter.


Not exactly. The engine is developing high pressure due to the identical ..uninterrupted ..unaltered ...flow having to acelerate to exit the various points of egress that are collectively much smaller than the filter in comparison.

How much oil is passing through a 2" pipe @ 5gpm?
How much oil is passing through a .000000000024" pipe @ 5gpm?

A: 5gpm. The difference will be the velocity and the pressure developed. The 2" pipe may develop 5psi ..while the .0000x" pipe may develop 2500PSI.

Naturally, one would need an incredibly HIGH oil pump relief setting to achieve this ..but it should demonstrate the concept.


The context of my quote above was with respect to the resistance ratio of the filter vs. engine oil circuit. I agree that if the engine's oil circuit is so restrictive that it caused the pump's pressure to increase to point of relief pressure, then you would naturally see a decrease in oil flow through the filter/engine circuit - the difference in flow being the amount bypassed back to the sump.

 

[Gary Allan]

Quote:
We are kind of going in circles now.


Well, it's easy to get lost in here. This challenges most conceptual views of flow through restrictions. When adding the oddity that you apply different physics in one state ..and then abandon them in another adds to the struggle.

Most of these events, in most cases, are transitional. I imagine if we could get into it tight enough, we'd see that, even at cold visc, there is not a static way to view the mass that you're moving. The friction of the laminar shearing on the exterior of the passages ..friction modifiers ..(whatever) probably make the first few moments like "giving birth" type thing where the visc may not be as much of a player as would be statically viewed.

That is, I think it's got a few more dimensions to it than what we're discussing here.

Quote:
Based on your pressure gauge photos in the other thread, the ratio was more like 8 psi delta P at 80 psi, or sounds like maybe closer to 10:1.


In my use of those gauges, the PSID (non-relief) was about 2 max on down. The only time that spread was present was in the relief event. You would watch the two needles sweep up ..then the downstream slow and stop ..and the pump side continue up to the peak pressure. Then, again, in "transition" ..the downstream needle would come up to about 2psi below the supply. Then ..a short time later, they would both retreat off of that.


I also installed a true pressure differential gauge. I thought that it was broken with a new filter (swapped at the time of install). I then installed a 9k used filter with really heavy oil on a relatively cold overnight shut down. It finally gave me a reaction ..but it didn't barely touched the rated bypass level ..and only stayed there for a very brief time.

 

[SuperBusa]

Originally Posted By: Gary Allan
Well, it's easy to get lost in here. This challenges most conceptual views of flow through restrictions. When adding the oddity that you apply different physics in one state ..and then abandon them in another adds to the struggle.

Most of these events, in most cases, are transitional. I imagine if we could get into it tight enough, we'd see that, even at cold visc, there is not a static way to view the mass that you're moving. The friction of the laminar shearing on the exterior of the passages ..friction modifiers ..(whatever) probably make the first few moments like "giving birth" type thing where the visc may not be as much of a player as would be statically viewed.

That is, I think it's got a few more dimensions to it than what we're discussing here.


Yes agreed, but IMO those initial flow conditions at startup are very quick transitions (2~3 seconds max), and once the system stabilizes it's much easier to understand what kind of flow dynamics are going on. Once the system is pressurized is what I concentrate on. And besides, if the oil filter stays full due to the anti-drain back valve, then the system is almost instantly pressurized. When someone does an oil change and starts the car with an empty oil filter, then there might be some different flow dynamics going on worth talking about.

Originally Posted By: Gary Allan
In my use of those gauges, the PSID (non-relief) was about 2 max on down. The only time that spread was present was in the relief event. You would watch the two needles sweep up ..then the downstream slow and stop ..and the pump side continue up to the peak pressure. Then, again, in "transition" ..the downstream needle would come up to about 2psi below the supply. Then ..a short time later, they would both retreat off of that.


Humm ... so if the max spread was only 2 PSID and in relief mode, then it sounds like that filter's bypass was set to 2 psi? That's what it sounds like. Maybe the gauges weren't very accurate. Sounds like the true delta P gauge you used below is an accurate assessment.

Originally Posted By: Gary Allan
I also installed a true pressure differential gauge. I thought that it was broken with a new filter (swapped at the time of install). I then installed a 9k used filter with really heavy oil on a relatively cold overnight shut down. It finally gave me a reaction ..but it didn't barely touched the rated bypass level ..and only stayed there for a very brief time.


That's good data ... shows the filters (even the dirty one) flowed good enough to basicallynever go into bypass mode, or at least very briefly on cold start. What kind of filters were they, and was the engine revved up significantly to produce some good flowrate through the system?

 

[LoneRanger]

I do understand the concept Gary Allan described with the back pressure created and sustained from the oil being forced into the bearings and galleries after exiting the filter, and considering that the fluid involved in this circuit is a non-compressible it makes perfect sense that the flow characteristics of the filter become secondary. I grasp that much.

 

[Gary Allan]

Quote:
Humm ... so if the max spread was only 2 PSID and in relief mode, then it sounds like that filter's bypass was set to 2 psi? That's what it sounds like. Maybe the gauges weren't very accurate. Sounds like the true delta P gauge you used below is an accurate assessment.

No. The normalized "out of relief" PSID was about 2PSI or less. It's the relief event that caused all appreciable PSID to exist to begin with.

It was the same (mostly) with a used filter with 9k on it. The difference being that when in a relief event, you moved more toward the filter's bypass setting, and could "surge" the flow to produce a "flare" of PSID until the oil warmed more. Since you were effectively using a smaller filter (aged) ..the "absolute" properties of the flow at visc were detectable.

Quote:
That's good data ... shows the filters (even the dirty one) flowed good enough to basicallynever go into bypass mode, or at least very briefly on cold start. What kind of filters were they, and was the engine revved up significantly to produce some good flowrate through the system?


The new one was a tea cup sized PureOne (same as PH3614). The 9k was a PH16 sized PureOne.

I'll try and find the link to my post on it. The images are gone due to Sony's image hosting going bye-bye. Unfortunately, I relied on the graphs to speak my thousand words with a few open ended statements .."I saw THIS" type stuff.

Basically I swapped out the dual gauges for the differential gauge. I think you could detect a minor twitch upon start up, but it was so slight, if at all, you could attribute it to your imagination. I really was sorta disappointed that I saw nothing. The two gauges, at least, provided something to look at.

When I took the images, I put 15w40 oil ..and had an overnight temp in the subfreezing level. The starting temp was 27F. My digital temp gauge didn't start reading until 70F ..so

(I upgraded the analog to a digital).

I can only describe the transitions as sorta a "phase shift" (and correction) of sorts. If you understand anything about inductive circuits .. the voltage leads the current for a brief time. In a DC circuit, the inductor itself creates its own impedance to the flow (it does this in AC, as well - but ..) ..so it appears like an open circuit for a brief time. After 5 time constants, it's a straight piece of wire.

In the same conceptual view, the max pressure (voltage) is being applied to the circuit ..but the flow (current) takes a bit to catch up.

You're reading your meter across the filter in place of the inductor.

This is an abstract view that doesn't directly apply. It would require some massaging and some allowances. That is, you could nitpick it if you wanted to ..but look at it as where I'm trying to take you ..and not in offering you something to dismantle.

 

[Gary Allan]

Originally Posted By: LoneRanger
I do understand the concept Gary Allan described with the back pressure created and sustained from the oil being forced into the bearings and galleries after exiting the filter, and considering that the fluid involved in this circuit is a non-compressible it makes perfect sense that the flow characteristics of the filter become secondary. I grasp that much.


Great

It really is counter intuitive and a very hard thing to grasp given that all of our existence is telling us otherwise (water faucets, etc.)

 

[SuperBusa]

Originally Posted By: LoneRanger
I do understand the concept Gary Allan described with the back pressure created and sustained from the oil being forced into the bearings and galleries after exiting the filter, and considering that the fluid involved in this circuit is a non-compressible it makes perfect sense that the flow characteristics of the filter become secondary. I grasp that much.


My conclusions from all these discussions are:

1) Probably all major oil filters out there will flow well enough for 99.9% of the conditions they are used under on street cars.

2) An oil filter is most likely to go into bypass mode if the engine is started in cold weather (high oil viscosity) and then revved excessively before the oil is warmed up. As the oil warms up and the viscosity goes down, the chances of the filter bypassing disappear unless the element becomes excessively clogged due to loading.

3) Using a filter with a higher bypass setting can't hurt as long as the filter flows well. Stay away from filters that have a combination of high flow resistance and a low bypass setting. Of course, nobody knows which filters these are (if there are any) because the manufactures won't tell you the flow parameters.

 

[FZ1]

So,then, are we saying that flow rate thru the filter is not a real concern and we should use the filters that filter best with only secondary consideration to flow rate?

 

[SuperBusa]

Originally Posted By: FZ1
So,then, are we saying that flow rate thru the filter is not a real concern and we should use the filters that filter best with only secondary consideration to flow rate?


That's basically the conclusion I've come to. Go with one that has very good filtering performance ... and of course one that has good build quality. Looks like I’ll be choosing the PureOne, as it has very good filtering performance, and I’m sure it’s flow rate is well above what the demand will be on practically any motor.

 

[Gary Allan]

The only time the static flow capability of a filter should make a difference is if you have some worn oil pump or a scored and leaking relief valve. This is where switching to a different filter may make a difference. These are obscure and isolated instance and the vast galactic mass of engines and drivers can pay it no mind. Buy to your level of utility. If you go long, but large ..if you go short, buy anything you want.

 

[FZ1]

Thanks,guys.

 

[LoneRanger]

This has been an informative thread. I learned to step out of the mistaken conception about flow that there is no resistance on the exit side of the filter, and that the back pressure actually existing on the exit side during normal engine operation means I can stop being so concerned with the filter flow rate having any meaningful effect on oil pressure.

My Subaru uses oil pressure to move the i-AVLS parts. intelligent-Active Valve Lift System is Subaru's variable valve timing system, similar to Honda's VTEC system.

 

[SuperBusa]

Originally Posted By: LoneRanger
This has been an informative thread. I learned to step out of the mistaken conception about flow that there is no resistance on the exit side of the filter, and that the back pressure actually existing on the exit side during normal engine operation means I can stop being so concerned with the filter flow rate having any meaningful effect on oil pressure.


So what filter are you going with after this revelation?

 

[LoneRanger]

I might just stay with Wix. That or Purolator.