Oil Filter......Flow vs. Filtration?

[FetchFar]

Good discussion. A bit long winded, still good. 1 micron to 40 micron HARD particles do the damage inside an engine. This has never been studied enough. Common sense tells you that the hardness (cragginess) of those free-roaming particles have the ability to scrape out internal engine metal, not the relatively soft carbon, hydrocarbons, and carbohydrates in there.

To get the hard iron particles out, use a magnetic drain plug, then all you can do is hope your air filter gets most silica out. After all that, let the oil filter just clean up whatever it can. Enough said.

 

[FetchFar]

The nearest thing I've seen to "proof" in a real engine about the above statements I just made, there was a poster elsewhere on bitog who wrote a bunch on how the tiny hard aluminum oxide grit in ScotchBrite scrubbing pads can ruin a rebuilt engine if the insides were scrubbed using it. The aluminim oxide particles were in the 1 to 50 micron range, oil filters don't clean it all up, and they quickly abrade bearings. If those were carbon particles instead of alum oxide, you would have beneficial ARCOgraphite oil!!!!!!

 

[dnewton3]

Originally Posted By: FetchFar
Good discussion. A bit long winded, still good. 1 micron to 40 micron HARD particles do the damage inside an engine. This has never been studied enough. Common sense tells you that the hardness (cragginess) of those free-roaming particles have the ability to scrape out internal engine metal, not the relatively soft carbon, hydrocarbons, and carbohydrates in there.

To get the hard iron particles out, use a magnetic drain plug, then all you can do is hope your air filter gets most silica out. After all that, let the oil filter just clean up whatever it can. Enough said.


I would disagree with you here. You essentially have this completely backwards.

I agree that silica is ingested; not much you can do but try to filter it out at the intake.

But soot is actually genereated in the engine; it's not ingested. Soot (incompletely burned hydrocarbons) is one of the most abrasive things in an engine; moreso than the small metal particles of Al, Fe, Pb and Cu. Cr is fairly tough, but not like carbon, and Cr is rare; often only in 1 or 2ppm increments total for a UOA. Soot starts out sub-micron in size; really, really small. Only when it would agglomerate with other carbon/soot particles, would it become large enough to do damage. Once it would reach 5um or larger, it does become an issue. And this is where the add-pack comes into play. The anti-agglomerates (or if you prefer, dispersants), keep the particles separated and don't let them cojoin. So as long as your OCI is reasonable, then the add-pack will be healthy enough to keep the soot in suspension and apart. As long as that happens, the wear will be really low. Again, look over those studies and my "normalcy" article. The proof is right there for all to see.

Metal particles are hardly ever (pratically never) bigger than 5um, and even if they were, they would not show up in a UOA because 5um is the upper end of the effective range for that method of viewing the construct of the particles being observed. Metal particles typically start small, and stay small, because they do not possess an affinity to cojoin like soot does.

If you believe otherwise, then please show the data backing up your claim.

Here is what I see from my perspective:
http://machinerylubrication.com/Read/51/soot-oil-engine
http://machinerylubrication.com/Read/712/diesel-engine-oil-particle
Check out that site; lots of reading. I will note that I do disagree with some of Jim's assertions, but generally I think his conclusions are sound. I disagree with some statments about linearity of metals; he casts a broad net whereas I'd be more definitive.

 

[Wilhelm_D]

Originally Posted By: dnewton3
Soot (incompletely burned hydrocarbons) is one of the most abrasive things in an engine; moreso than the small particles of Al, Fe, Pb and Cu. Cr is fairly tough, but not like carbon, and Cr is rare; often only in 1 or 2ppm increments total for a UOA. Soot starts out sub-micron in size; really, really small. Only when it would agglomerate with other carbon/soot particles, would it become large enough to do damage. Once it would reach 5um or larger, it does become an issue.

To soot add silica.

I have never been able to understand the popularity of low-efficiency oil-soaked air filters. We stopped using them because silica through the intake is not good for engines.

 

[FetchFar]

Originally Posted By: dnewton3
If you believe otherwise, then please show the data backing up your claim.

Here is what I see from my perspective:
http://machinerylubrication.com/Read/51/soot-oil-engine
http://machinerylubrication.com/Read/712/diesel-engine-oil-particle
Check out that site; lots of reading. I will note that I do disagree with some of Jim's assertions, but generally I think his conclusions are sound. I disagree with some statments about linearity of metals; he casts a broad net whereas I'd be more definitive.



Funny, you have no "data" ever, yet you constantly demand negative-proofs. ... And, your articlesyou cite actually contradict most of what you said. "The researchers found in a study of roller follower wear on engines with increasing soot levels that the rate of wear is sporadic and not systematically related to the soot level." That article supports my points made earlier that its the hardness of the particles that plays a huge role, and soot isn't a hard particle. Silica and metallic particles, other minerals, do most of the damage. Seems simple.

 

[Wilhelm_D]

Originally Posted By: FetchFar
Funny, you have no "data" ever, yet you constantly demand negative-proofs. ... And, your articlesyou cite actually contradict most of what you said. "The researchers found in a study of roller follower wear on engines with increasing soot levels that the rate of wear is sporadic and not systematically related to the soot level."


You're reading it out of context.

Lubrizol's research didn't contradict the fact that soot causes wear. What they found was that it was not soot load per se that was directly related to component wear but a complex relationship between soot load and dispersant level that affects the rate at which wear is generated.

In other words, you can mitigate the soot load with dispersants.

 

[FetchFar]

Originally Posted By: Wilhelm_D
To soot add silica.

I have never been able to understand


According to dnewton3, you have it all backwards if you think silica is an issue with wear.

 

[ZeeOSix]

dnewton - I appreciate your long and detailed write-up above. So, let's just focus on some details of a UOA test with respect to the suspended particle count data.

What is the particle size range that a typical UOA test measures?

 

[Wilhelm_D]

Originally Posted By: FetchFar
According to dnewton3, you have it all backwards if you think silica is an issue with wear.

He didn't dismiss silica. He didn't mention it. He noted that soot is a bigger problem than some previously mentioned suggestions such as metal particles.

 

[dnewton3]

I most certainly did mention both soot AND silica; read the post above, and allow me to quote myself ...

Originally Posted By: dnewton3

I agree that silica is ingested; not much you can do but try to filter it out at the intake.

And here's why I said that ...
Together, hard soot and silica are likely to do more damage than the metal particles every will.
- we try our best to keep silica out by filtering the air, because engines don't create silica, it's ingested
- we try to control soot as best we can by adding anti-agglomerates to the lube, because incomplete combustion creates soot; it's not ingested

My point was that silica is taken into the engine, not created in the engine. Silica does not agglomerate (cojoin), whereas soot does. Meaning that silica stays at it's general size. Soot will grow if the add-pack becomes overwhelmed; the dispersants keep it in check until they become degraded to a point where they cannot keep up with the soot loading. Additionally, the dispersants don't do as good a job on silica as they do soot; don't ask me why - I'm not a chemist. I accept facts as they are presented; I don't pretend to know everything there is to know about chemistry.

I stand by my statements. Soot is a VERY hard particle, whereas sludge is a softer issue. Soot is contained within sludge, but the sludge also has other byproducts in it such as oxidation insolubles, etc. Don't confuse soot with sludge. Sludge is a mixture of many combustion byproducts. Soot is the hard hyrdocarbon byproduct of incomplete combustion; it is VERY abrasive. Sludge is "gooey"; soot is "gritty".

The general condemnation limits for most metals is available in a thread I posted a few years ago, from many diesel OEMs. There generally is not a hard rule for gasser engines, but the limits may not be far off of diesel limits. For metals, it's reasonable to allow up to 100-150ppm of Fe, but you'll never see that in a UOA here, because folks are too anal about their OCIs.

There are some reasonable limits for soot in a UOA; often you'll see 3% or 3.5% as a condemnation limit. But that does not speak to the SIZE of the particles, only the percent by volume. As long as the add-pack is controlling soot, it's percentage can be allowed up to these levels. However, if the add-pack were weak, then those levels would result in much wear.

You cannot look at the perceived "hardness" of a metal, and think that is going to automatically translate into an assurance of damage to the engine. Some metals are reasonably soft; Cu, Pb, Al. Some are moderate, such as Fe. Others are hard like Cr. Ti is a very hard metal, and yet it's used as an ADDITIVE to REDUCE wear in some products like Conoco's lubes. So, if having metal in the oil is such a horrid thing, then please explain why Ti and some other metals are actually ADDED to lubes for wear reduction?



As to the particle sizing in a UOA, depending upon equipment and source you use, you can presume that the spectro-analysis will see up to about 5um, perhaps 7m and no larger. I don't think there is a finite lower end, but it's easily well below sub-micron, and may be able to be registered in nanometers.


This is a topic about POTENTIAL for damage to occur. While I agree that large concentrations of large, hard metals may result in wear, those conditions are very, very rare. Typically we don't get lots of large metals. What we get is small concentrations of small metals particles; those are pretty much harmless. We also get an influx of both soot and silica; they are controlled differently, but result in the same destructive effect.

Never confuse the size of a particle with the percent of population (ppm). 10 particles of .5um in size will give the same UOA result as one particle that is 5um in size on a "ppm" basis in spectral-analysis. Here's the difference between a UOA and a PC.
- The UOA allows us to see the elemental make up of the lube; we can get a sense of what's in the oil on a "part per million" basis, but it tells us ZIPPO about the SIZE of the particles.
- PCs allow us to quantify the SIZE of stuff in the lube, but we have ZERO idea of what each particle represents in terms of elements.
There is no direct relationship between the two whatsoever.


A UOA will tell us what's in the oil as a percentage of concentration, but not size.
A PC will tell us about particle size and quantity for any delineation, and can inform us about the efficiency of a filter, but cannot tell us what makes up the particulate.
UOA macro and micro analysis for "normal" field operations shows that once a reasonable level of filtration is in place, a tighter efficiency does not statistically yield less wear.
Lab testing will show that more efficient filters do produce a cleaner resultant lube, but lab tests don't always manifest into reality on the road because filters are only one leg in the tripod of wear reduction; OCI duration and tribochemical barriers play heavily into this.


I hope that clarifies my statements and position.

 

[ZeeOSix]

Originally Posted By: dnewton3

As to the particle sizing in a UOA, depending upon equipment and source you use, you can presume that the spectro-analysis will see up to about 5um, perhaps 7m and no larger. I don't think there is a finite lower end, but it's easily well below sub-micron, and may be able to be registered in nanometers.


That's why I was saying you could never accurately associate an oil filter's ISO tested efficiency/beta directly to a UOA particulate analysis in the link below. The measured particle size range is just way down in the mud compared to what a normal full-flow oil filter is designed to filter out effectively.
UOA Particle Count Size too Small for Correlation

Or in other words, there is not a good correlation between a UOA (aka, a "field test") to the ISO beta spec because the measured particulate size range is so far down in the mud as compared to what a filter can effectively filter out, that even the most efficient filter may not show much change (if any) in the measured particulates in a UAO.

If UOAs could accurately measure the suspended particle size in the range of 60 microns and below, then we would have good data to correlate a filter's measured ISO tested efficiency with "real world" performance.

Originally Posted By: dnewton3

A PC will tell us about particle size and quantity for any delineation, and can inform us about the efficiency of a filter, but cannot tell us what makes up the particulate.


The best you could hope for is that every oil filter will still catch some of these sub 7 micron particles. I would think that the more efficient an oil filter is rated at 20 microns, then the more efficient (relatively speaking) it will also be at

 

[Wilhelm_D]

Originally Posted By: ZeeOSix
Or in other words, there is not a good correlation between a UOA (aka, a "field test") to the ISO beta spec because the measured particulate size range is so far down in the mud as compared to what a filter can effectively filter out, that even the most efficient filter may not show much change (if any) in the measured particulates in a UAO.

The primary measurement in the UAO is not particulates but elements ppm. From that we infer engine wear. If engine wear is lower with Filter A than with Filter B, but Filter B shows a superior beta score, we can conclude that the beta score is poorly correlated to the engine wear.

In turn we correlate UAO inferences by actual engine teardowns, which the average car owner doesn't do, but fleets can track engine repairs and rebuilds and match them to their records on UAOs and oil and filters used.

Quote:
If UOAs could accurately measure the suspended particle size in the range of 60 microns and below, then we would have good data to correlate a filter's measured ISO tested efficiency with "real world" performance.

There's no need to measure the suspended particle size. We have the beta on a given filter based on standard lab tests. We have the results of the element analysis in a UAO. Since our goal is low wear, not small particular size per se, we have everything we need to assess whether the beta correlates with low engine wear.

 

[ZeeOSix]

Originally Posted By: Wilhelm_D
Originally Posted By: ZeeOSix
Or in other words, there is not a good correlation between a UOA (aka, a "field test") to the ISO beta spec because the measured particulate size range is so far down in the mud as compared to what a filter can effectively filter out, that even the most efficient filter may not show much change (if any) in the measured particulates in a UAO.


The primary measurement in the UAO is not particulates but elements ppm. From that we infer engine wear. If engine wear is lower with Filter A than with Filter B, but Filter B shows a superior beta score, we can conclude that the beta score is poorly correlated to the engine wear.


Originally Posted By: Wilhelm_D
Quote:
If UOAs could accurately measure the suspended particle size in the range of 60 microns and below, then we would have good data to correlate a filter's measured ISO tested efficiency with "real world" performance.


There's no need to measure the suspended particle size. We have the beta on a given filter based on standard lab tests. We have the results of the element analysis in a UAO. Since our goal is low wear, not small particular size per se, we have everything we need to assess whether the beta correlates with low engine wear.


Engine wear can be contributed to more than just particulate wear. Things like coolant leakage, fuel dilution, etc that effects the lubrication properties and protection of the motor oil and the engine wearing surfaces can come into play.

When talking only about the amount and size of the suspended particulate in the oil, the less of it the better. And as mentioned already, the particle sizes that a UOA measures are well below what a normal full-flow filter can catch effectively. I'm willing to bet that if Filter A has a better rated/ISO tested efficiency than Filter B, then you are at least getting some minor benefit out of using a more efficient oil filter to help reduce some engine wear. Is it significant or measurable ... maybe, maybe not without doing some very controlled test and engine teardowns and measurements to correlate data.

 

[Wilhelm_D]

Originally Posted By: ZeeOSix
Engine wear can be contributed to more than just particulate wear. Things like coolant leakage, fuel dilution, etc that effects the lubrication properties and protection of the motor oil and the engine wearing surfaces can come into play.

True but not relevant to this discussion.

Quote:
When talking only about the amount and size of the suspended particulate in the oil, the less of it the better. And as mentioned already, the particle sizes that a UOA measures are well below what a normal full-flow filter can catch effectively.

The particle size the UOA measures is irrelevant. We don't use that to compare them, we use the elemental analysis and infer the wear.

The only time we'd measure particle size directly in use would be if were in a design phase on some new filter design, where we are trying some new material, approach, or design.

Quote:
I'm willing to bet that if Filter A has a better rated/ISO tested efficiency than Filter B, then you are at least getting some minor benefit out of using a more efficient oil filter to help reduce some engine wear.

If we tear down the engine, or infer the wear from a UOA, and Filter B's wear results are better than Filter A's, it doesn't really matter that A has a better beta measurement. We are after lowering engine wear, not picking beta gnits.

 

[ZeeOSix]

Originally Posted By: Wilhelm_D
Originally Posted By: ZeeOSix
I'm willing to bet that if Filter A has a better rated/ISO tested efficiency than Filter B, then you are at least getting some minor benefit out of using a more efficient oil filter to help reduce some engine wear.

If we tear down the engine, or infer the wear from a UOA, and Filter B's wear results are better than Filter A's, it doesn't really matter that A has a better beta measurement. We are after lowering engine wear, not picking beta gnits.


Has anyone actually done a correlation exercise to determine if actual engine wear rate does correlate with Filter A and Filter B filtering efficiency specs? Filter A being much more efficient than Filter B - like using a PureOne for Filter A (99.9% @ 20 microns) and a Toyota OEM for Filter B (50% @ 20 microns).

I doubt that level of correlation has ever been done in a totally controlled way, and if so the data probably isn't public anyway.

If there was any correlation at all, I'd certainly think that the wear rate with Filter A vs. Filter B (with all other variables held perfectly constant, ie: oil, OCIs, driving conditions & environment, starting/heat cycles, filter ADBV performance, filter bypass frequency, other engine maintenance, etc, etc - MANY factors involved) would show some measurable level of decreased engine wear - even if it was ever so slight.

I would be very surprised if less wear would be seen with the much less efficient Filter B, and is so then there would obviously be some other factor(s) involved for the higher wear rate to show up while using a much less efficient oil filter.

The other scenario is that there would be no measurable engine wear different seen between Filter A & B if all other factors were held perfectly constant.

I'd like to see the results of a valid study like that. If anyone has a link to one, post it up.

 

[Wilhelm_D]

Originally Posted By: ZeeOSix
Has anyone actually done a correlation exercise to determine if actual engine wear rate does correlate with Filter A and Filter B filtering efficiency specs?

That's exactly what we've been talking about.

This love of specs is characteristic of BITOG dilettantes rather than people who do this for a living. People who do it for a living are focused on results.

 

[kschachn]

I will guess that the manufacturers have test data. I further wager that given the relatively low efficiencies of the Toyota and Honda filters, those manufacturers have determined that it does not matter, at least for their applications.

Lower efficiency specs don't mean a filter won't capture a particular particle, it just determines when.

 

[dnewton3]

Actually, there has a been a very well controlled "garage" test that has been done. Read the normacly article, and focus on the "micro analysis" data from the Ford Vulcan.

In that example, the person (was not me, but an individual I'm acquainted with) ran about 175k miles with Mobil 1 and Pure Ones, then switched to MC5K with a Puro Classic for another 110k miles. As you can see from the UOA data, the stacked-deck potential for advantage of both syn and premium filter did not appreciably alter the wear patterns at all, compared to the "normal" products.

The data in that example showed less than one sigma deviation for all metals, regardless of the oil and filter used. After the change in products, Al and Cr actually ticked up a tiny bit, whereas Fe, Cu and Pb were down. But they were all "normal", and well within a single standard deviation. In short, there was no change to the wear pattern even though he went from premium products to normal products. I will note here, that the shift was in tenths of ppm for all metals; were talking tiny variance here. This is a pittance and not even worth any BITOGers time when it comes to engine wear, but it was such a well controlled test that it's hard to ignore. He never pushed the OCIs out to where a disparity may have evolved. But under those conditions, the Mobil 1 and Pure One did absolutely nothing for the extra money spent.

Clearly the Pure One filter is "better" than a Classic, and yet even with the use of a syn, there was no statistically significant variance with it's use.

Being a "micro analysis" example, this was just one vehicle with one operator over several years of use. About as "controlled" as one can get without being in a lab. And just about as good a real-world example as I can hope to see.

 

[Wilhelm_D]

Originally Posted By: kschachn
I will guess that the manufacturers have test data. I further wager that given the relatively low efficiencies of the Toyota and Honda filters, those manufacturers have determined that it does not matter, at least for their applications.

It is not all about "efficiency" tests

Filter performance in the real world is a lot more complicated than beta testing.

 

[ZeeOSix]

Originally Posted By: Wilhelm_D
Originally Posted By: ZeeOSix
Has anyone actually done a correlation exercise to determine if actual engine wear rate does correlate with Filter A and Filter B filtering efficiency specs?

That's exactly what we've been talking about.

This love of specs is characteristic of BITOG dilettantes rather than people who do this for a living. People who do it for a living are focused on results.


Yeah, I know - I'm in this conversation. So what's the answer to my question? Does anyone have or know a source of information where someone has actually collected data to correlate filters with efficiency of widely different levels with particle count measurements and engine wear rates?