Using Oil Analysis to Determine Oil Filter Effectiveness

Status
Not open for further replies.
Joined
Feb 19, 2019
Messages
7
Location
HAZLETON
Has there been any comparison oil analysis tests done that would indicate the effectiveness on one brand or oil filter compared to another? All tests would need to be done under the same conditions and in the same vehicle for close to the same number of miles.

Any ideas on whether or not this would be feasible or a waste of time and expense? Would results prove that one brand's cheapest filter was indeed considerably less effective than using a more expensive premium brand filter? You see all sorts of comparisons on You Tube videos that compare how filters are constructed but who cares. Bottom line is... Does one actually perform better than another?

Your thoughts...
 
The proper method of measuring a full-flow oil filter efficiency is:

ISO FIltration Test

A particle count as part of a UOA may be interesting but the test above is the correct method.

You cannot test anything twice "under the same conditions" in the real world.
 
There is proof and truth. Good luck with either. Science is well science. All results in a lab.....In a perfect world!
 
I doubt that when Toyota & Honda torture test their vehicles they swap out their oem filters for a Fram Ultra.

I think that rationally speaking, the most important filter feature is construction and bypass reliability.

After that, one might want to look at capacity for example for extended ocis or when the engine is dirty and more frequent oil changes will remove sludge (I've never understood the advice to use a cheap filter for cleaning because one has no idea when a cheap low capacity filter will be loaded).

Very high efficiency is a nice to have. By definition, it keeps the oil cleaner which I *believe* means the oil lasts longer.

In my case, I used a Fram Ultra because I want to run multiple ocis on a single filter. However, in European cars, I use OEM filters, with far less efficiency compared to an Ultra cartridge because reliability is the most important aspect and Ultras have been known to crush.
 
UOAs can discern major changes in wear rates fairly easily. But that's not the whole story.

If you have two different oil filters that exhibit vastly different efficiencies, that would presumably induce a major difference in particulate loading. The inference is that more particulate of larger size would induce more wear. This is proven in some SAE studies, not the least of which is the GM filter study (881825). However, that same study also tells a different story when reality sets in. The differences of slight filter efficiency variation (as seen in the choices of today's typical filters) does not reveal any discernible shift in wear rates. This is because modern engines with short-to-moderate OCIs do not have enough contamination in them to statistically differentiated relative to filter selection. The same GM filter study that "proves" finer filtration is better, is also the very same filter study that acknowledges that "normal" OCIs will never exhibit evidence from the lab study. From the study:
"Used oil analysis from engines in the field will not typically show such a clear correlation since wear metals generated between oil changes will be at much lower concentrations."
In other words ... there's not enough wear in a normal UOA to distinguish any difference between filters. And that study was with filters that had a wide variation in efficiency. Today's typical filters are much more closely matched; brand to brand and beta to beta.

Essentially, there's not enough wear material in a normal OCI to tell the difference in filter efficiency performance. This is because the contaminant PM loading is so low that filter differences simply cannot have a major impact on wear rates.
 
Last edited:
I think one can simply say that car manufacturers' service intervals are very conservative and are designed to avoid even a modest likelihood of a problem. In fact dnewton3 you've said just that quite a few times as I recall, though perhaps in slightly different words. :)

You push well outside of that manufacturer schedule based on micro data and you may see differences due to filter selection. Or not. :)
 
I tried using Blackstone's particle counts about 10 years ago to compare filters. I tried very hard to hold as many variables constant as I could, considering the vehicle was a daily driver.

As hard as I tried, I could not get consistent results enough to reach any conclusions. The only thing interesting I found was that Amsoil filters came in with the worst particle counts, but I question those results, as it only happened on two filters. I actually use Amsoil filters today, as a matter of fact.

I think the only way you can use particle analysis to test oil filters is if you are in a laboratory setting, where you can hold more variables constant.
 
Originally Posted by btanchors
I tried using Blackstone's particle counts about 10 years ago to compare filters. I tried very hard to hold as many variables constant as I could, considering the vehicle was a daily driver.

As hard as I tried, I could not get consistent results enough to reach any conclusions. The only thing interesting I found was that Amsoil filters came in with the worst particle counts, but I question those results, as it only happened on two filters. I actually use Amsoil filters today, as a matter of fact.

I think the only way you can use particle analysis to test oil filters is if you are in a laboratory setting, where you can hold more variables constant.

Well, all you've done is realize that life is much larger than a lab.

This is exactly why I tell folks not to put too much stock into SAE studies they see where HALT methodology is used. HALTs (highly accelerated life test) use inputs and influences that illicit a specific result, as "proof" of concept. They often tell us exactly what we want to see. They are not worthless, but they are not to be considered completely transferrable to the real world.

Filter studies often show that finer filtration matters a lot. And yet when we do UOAs from our garage, we don't see the same effects as the HALTs. Why not? Because HALTs hold a lot of variable in a consistent manner; that's good. But they also manipulate things into an unrealistic situation to embolden the results; that's bad.

Great example ... the infamous GM filter study (SAE 881825). They "proved" that finer filtration resulted in less wear, on a factor of 8x "better". But there were several key things that most folks miss by not reading the study, as well as asking some fundamental questions. Did you know, for example, that the GM filter study that just about every bypass filter maker uses in their marketing hype is based on these conditions:
- the "base" filter was 40um nominal
- the "best" filter was 7um nominal
- they dumped in fine dust to simulate contamination; 50 grams each hour for 8 hours (this is the equivalent of 570,000 miles of "normal" silica ingestion)
- they never once changed oil, in that simulated 570k miles
- they only changed filters when the dP across the media went to 20psi
- they noted that all filters eventually clogged up to a condition where 10um was essentially the blinding point

And the coup de gras? They admitted in the summary that normal field results will never show the disparity in filter performance, because UOA data is never this significant in normal use. Why? BECAUSE NO SANE PERSON RUNS AN OCI FOR FIVE-HUNDRED-SEVENTY-THOUSAND MILES !!! Because all filter essentially stopped being effective at 10um, that means to ALL particulate matter below 10um stayed in circulation for the entire 8 hours. What happened is that the filters would eventuailly blind off to around 10um; it just took the more poreous filters longer (obviously), and so more wear happens when when more stuff above 10um is in play.

But in your garage, do you only OCI once every 570k miles????? I didn't think so. Generally, if you have a decent air filter, and a decent oil filter (no matter what brand), the wear rates are not going to change much; they'll do what they do. Soot is WAY too small in moderate OCIs (out to 15k miles). Soot starts out around 4nm in size (that's nano-meters). It has to grow via amalgamation/agglomeration about 100x larger just to get to 4um !!! As long as your oil additive package is good, the anti-agglomerates are going to keep this from happening. And if your air filter is doing it's job, not much Si is getting in. Hence, no matter how good or bad your filter is, when there's not a lot of PM (particulate matter) present, it's a moot point to talk about oil filter efficiency.

Congratulations! You're one of the few BITOGers that actually understand and experienced the reality of why filter conversations are misunderstood and overblown. Once filtration is "good enough", making it "better" has no effect in the real world, because filtration is not the only factor controlling wear.
- OCI duration
- TCB formation
- additive package effect
- filtration
These all affect wear. And when you cannot even control PM via your filter selection, that pretty much means it's a non-topic.

The reason you were not getting consistent results in your trials years ago is because normal life variation of your engine's duty cycles was actually larger than your input influence. In short, the normal daily wear is greater than what the filter has effective control over, for your chosen OCI duration.
 
Isn't BSLabs one of the outfits whose particle counts are done on only one particle size, and then the rest of the report is assumed based on a normal distribution?
 
Originally Posted by bulwnkl
Isn't BSLabs one of the outfits whose particle counts are done on only one particle size, and then the rest of the report is assumed based on a normal distribution?


10 years ago it might have been a pore counter, I believe that is the way they work (although I was thinking it was a couple different sizes).

They now use an optical counter. It should probably be noted that the drawback to that is that it can't do oil that is too dark.


Originally Posted by btanchors
I tried using Blackstone's particle counts about 10 years ago to compare filters. I tried very hard to hold as many variables constant as I could, considering the vehicle was a daily driver.

...


Did a similar exercise, though on a small scale.

I had a particle count run on an Ultra and on a Hamp Shorty on my Honda back to back (in fact I think I ran two on each).

I picked these two filters because the Fram should be among the highest efficiency and the Hamp the lowest. The Hamp is a A01 that is about half size, it has no end caps and Honda is not known for placing a high value on filter efficiency. If it is over 60% efficient at 20 microns I'd be shocked.

I was careful to try to minimize variation in my sample methods and usage, I pulled sample with a clean tube of the same length, on hot oil and as soon as I could get the sample after the run so that particles had little time to fall out of suspension.

The two filters were very similar in particle counts, too similar, and it resulted in much consternation and attacks upon my data, particularly since the Ultra was fairly unimpressive when viewed in a vacuum even.

I learned a few things from that test, that particle counts are a reliable indicator of filter quality/efficiency was not one of those things.
 
Last edited:
Originally Posted by DuckRyder
Originally Posted by bulwnkl
Isn't BSLabs one of the outfits whose particle counts are done on only one particle size, and then the rest of the report is assumed based on a normal distribution?


10 years ago it might have been a pore counter, I believe that is the way they work (although I was thinking it was a couple different sizes).

They now use an optical counter. It should probably be noted that the drawback to that is that it can't do oil that is too dark.


Perhaps you're right. It's odd, though, that they still report in the 5, 10, 15, 25 etc. micron sizes that go along with a pore counter.
 
Originally Posted by bulwnkl
Originally Posted by DuckRyder
Originally Posted by bulwnkl
Isn't BSLabs one of the outfits whose particle counts are done on only one particle size, and then the rest of the report is assumed based on a normal distribution?


10 years ago it might have been a pore counter, I believe that is the way they work (although I was thinking it was a couple different sizes).

They now use an optical counter. It should probably be noted that the drawback to that is that it can't do oil that is too dark.


Perhaps you're right. It's odd, though, that they still report in the 5, 10, 15, 25 etc. micron sizes that go along with a pore counter.


Unless they are outright lying to me (they aren't) I'm right. I even know the model number were I inclined to go back and look.
grin.gif
 
Status
Not open for further replies.
Back
Top