Toughguard vs wix

[ZeeOSix]

Originally Posted By: Nyogtha
Originally Posted By: ZeeOSix

Even if the manufacturer only gives one data point (ie, 65% @ 20 microns from my curve example above), then anyone knowing how filters work would realize that the filter is even higher efficiency at particles above 20 microns, just as the curves show.

Therefore, you really don't need to know the "multiple points of data" unless you were splitting hairs and focusing directly on one or two specific particle sizes for some reason.


I disagree. You focused on the Filters A&E curve in the graph you posted. That graph shows Filters B&H behave similarly to Filters A&E but at a higher efficiency than filters A&E. Filter F is also similar. Then look at the curves for all the other filters on that graph (filters C, D,& G). Completely different curves.

Extrapolating or interpolating along a small step size from a given data point or points (say from 25 microns to 20 microns with some reasonable directional guesstimation on increase or decrease)is probably OK, but 37.5% of the filters on the graph you posted do not follow the capture efficiency curves of those who showed linear behavior over a wide particle size range.


I was going to mention that in my previous posts too, that when there is a huge difference in efficiency the curves are harder to compare. Yes, there is a "family of curves" that would all reside someplace between the most efficient (Filter D) and the least efficient (Filters A & E) curves.

But keep in mind that if an oil filter is rated very high in efficiency at 99% @ 20 micorons (or at >20 microns), that it is also going to be much more efficient for particles below 20 microns than the filter that come in at say 85% @ 20 microns. Just look at the family of curves and you an instantly see that.

Once you know how oil filter efficiency vs particle size curves basically look (as I've shown with that graph), it's easy to see that when comparing two filters at the same particle size. it will instantly tell you which one is better at every particle size. To me, that's all that matters when looking for a filter that meets my efficiency requirement. Don't have to make it rocket science, because it's really not.

Therefore, I stand by my comment of:
Originally Posted By: ZeeOSix
Therefore, you really don't need to know the "multiple points of data" unless you were splitting hairs and focusing directly on one or two specific particle sizes for some reason.


Originally Posted By: Nyogtha
Personally, I'd prefer it if they all reported a standardized test efficiency from ISO 4548-12 at 10 microns. If the general consensus is particles in the 10 to 20 micron size range cause the most damage, why not report efficiency at the small end of that range instead of the large end of that range?

My reasonable guess - it's so a larger number can be reported for efficiency across a wider spectrum of media types.

I agree ... and I also agree that they use 20 or 25 microns (or even larger) so that the efficiency numbers looks good. And some of these manufacturers will simply say "99% efficient" without any particles size. "Oh, we forgot ... that was at 100 microns". Yeah, nice.

 

[Nate1979]

The only reason there is an efficiency number on the box is marketing. They want it to be easily understood by the average consume. You really think they are going to put an efficiency data curve on the box? Maybe on their website but that is very sensitive proprietary info that only the BITOG folks care about.

 

[ZeeOSix]

Originally Posted By: Nate1979
The only reason there is an efficiency number on the box is marketing. They want it to be easily understood by the average consume. You really think they are going to put an efficiency data curve on the box? Maybe on their website but that is very sensitive proprietary info that only the BITOG folks care about.


Can't even get some (ie, WIX) to tell you what efficiency test standard they use. Trying to get Flow vs. Delta-P or Efficiency vs Particle size curves would take a miracle.

 

[ExMachina]

Business 101: Never tell any more about your product you're selling than you absolutely have to. It might invite unwanted comparisons, and besides, shiny boxes and bright colors are all most look for in an oil filter. Sad.

 

[ZeeOSix]

Originally Posted By: ExMachina
Business 101: Never tell any more about your product you're selling than you absolutely have to. It might invite unwanted comparisons, and besides, shiny boxes and bright colors are all most look for in an oil filter. Sad.


Don't forget bikinis ... those would help in advertising too.

 

[Nyogtha]

ZeeOSix, I still respectfully disagree. Filter G, a blended media rated as 98% efficient at 25 microns, shows a markedly superior and differently shaped efficiency curve than Filters B and H with glass media which in the table shows a 98% efficiency at 15 microns.

Something that's immediately apparent to me is markedly different from what you "instantly see". And I agree, this is far from rocket science. It's not even jet science. It's automotive science in this case.

But the correlation that collection efficiency is only increasing as particle size increases, yes that's always been immediately apparent to me.

 

[ZeeOSix]

Originally Posted By: Nyogtha
ZeeOSix, I still respectfully disagree. Filter G, a blended media rated as 98% efficient at 25 microns, shows a markedly superior and differently shaped efficiency curve than Filters B and H with glass media which in the table shows a 98% efficiency at 15 microns.


I agree ... but now that is yet another side discussion topic.

I tried to simply sum it up with all my examples that:

1) Oil filters that are "xx% efficient @20 microns" is the basically saying the same thing as "xx% for particles >20 microns". The Efficiency vs Particle Size graphs show that clearly.

2) If comparing many filters to each other that are in the same media family (ie, either cellulose, cellulose/synthetic blend or full synthetic), and compared via ISO 4548-12 at the same particle size (or close to the same, ie, 20 or 25 microns), then if Filter A is more efficient @20 (or >20) microns, then Filter A is also more efficient over the whole particle spectrum, as the curves show.

3) Always have to wonder what's being hidden or manipulated/skewed when a filter company won't even tell you what test specification their efficiencies are based on.

 

[dnewton3]

I do not disagree with your general premise, but you, sir, are going to be seen as a heretic here on BITOG ... I should know; I am often viewed the same way

See my comments inserted in your quote ...


Originally Posted By: Nyogtha
If you can't find ISO 4548-12 test data for a brand of filter from that filter manufacturer, the best place to look is comparison charts from a competitor such as the previously linked data from Champion Labs or as another example the Amsoil Ea filters data sheet.

http://www.amsoil.com/lit/databulletins/g2192.pdf

Any "leaps" (extrapolation, interpolation, extreme hypotheticals, etc.) are at the discretion of the jumper.

What I personally choose as reasonable "leaps" for me are:

1. For a given media composition / construction, filtration efficiency reported using ISO 4548-12 at a given particle size will be virtually the same regardless of size of the filter element.
This is very logical in my opinion. I say this because UOA data shows us that minor manipulation of this variable (size) does not have any bearing whatsoever on wear rates. So very many folks here use the theory that larger is better, but the data just does not show this to be true. Even if true, the effect is so infinitesimally small that normal variation far overshadows the effect they seek.

1A. IF one has tabular data from ISO 4548-12 testing across the ISO Medium Test Dust range, one can construct a graph of that particular filter's (or media's) efficiency vs. particle size, and see what the efficiency is at any given particle size along the range of particle sizes used & measured (as ZeeOSix has shown). For example, if we had such data for all the filters shown in the previously linked Champion Labs testing, we could reasonably what the efficiency was at 20 microns instead of 25 microns for all the filters data was reported for. On the flip side, reporting at 25 microns may have been done by plotting data at measurement points not equal to 25 microns, then finding the mathematical calculated efficiency was at 25 microns through graphs or data regression analysis / curve fitting. Without the actual data along the range of particle sizes extrapolation or interpolation should be limited to very small steps from the reported data point (or points) as the test dust used does not have a linear particle size distribution and it is unlikely the filter / media will have a perfectly linear efficiency across the spectrum of particle size distribution. Extrapolation from any single data point is always unwise. This can all be tempered to varying degrees by first asking "would it be reasonable . . ."
I agree; we can essentially extrapolate and find the reasonable estimates for the data points, or we can test for them. But to test at every single point to a degree that the variance would be known and true range established would be STUPID expensive and time consuming. I accept that some data evolves out of math rather than results.


2. Dirt holding capacity reported using ISO 4548-12 is a function of both filter media and bypass valve opening differential pressure. So dirt holding capacity is not necessarily increased or decreased as the physical size of a filter element is increased or decreased, nor necessarily when counting media pleats or media surface area between filters with different bypass valve opening differential pressures. For filter applications with no bypass valve (bypass is in the engine) it will depend on what that engine's bypass opening differential pressure is (and how stable it remains over the life of the engine).
I don't really have a strong opinion here, but you are swimming against the current of most of the Filter Faithful here at BITOG. Most here believe that a larger filter will most certainly hold more. Much of this depends upon surface loading versus depth loading, which is a trait of the media design and not size. The real question most ignore is how much of the available capacity is used in the first place. Having "more" capacity, well past any sensible amount one uses is not a benefit, but rather a waste. Those who use a FU rather than the EG, simply because they was "more capacity" for a 5k mile OCI, are foolishly misunderstanding the concept of the need for capacity. The question of how much a filter holds is only secondary; the REAL QUESTION that should FIRST be answered is this:
How much capacity do you need?

It will be interesting to see what these results will look like when Fram starts reporting the average of an increased number of filter models / sizes in their offerings as Jay has indicated. I may find my personal "leaps" remain reasonable or I may need to revise them.

 

[Nyogtha]

OK, well lets have a side discussion.

Filters B & H are rated 98% efficient at 15 microns in the table you posted. However the graph you posted shows they actually are about 65% efficient at 15 microns (purple line).

Originally Posted By: ZeeOSix




Filter G rated as 98% efficient at 25 microns in the table you posted shows it's still close to 98% efficient at 15 microns as its curve is very flat in that section.

So only knowing the points in the table you posted, one could not reasonably predict / extrapolate what efficiencies are at other sizes not listed in the table for these 3 filters without the curves in the graph you posted (and I think whoever manufactured Filters B & H wasn't being very up-front). Personally, I wouldn't purchase Filters B & H after seeing the graph. But if I only saw the table you posted with one point for each filter, how would I know differently?

Even with only 2 points, you can see using some straight line approximation between any 2 points for Filter G increases the inaccuracy of the approximation in proportion to the distance between the 2 points. For eample, if we had the rated 98% efficiency at 25 microns, and additionally only had another point at a much smaller size, say 6 microns where the graph shows about 83%, and drew a straingt line between those 2 points, we would significantly underestimate Filter G's efficiency anywhere in the 12 to 15 micron range.

Think of the Fram Ultra being rated as 80% efficient at 5 microns as has been posted on BITOG in addition to its published advertised rating point off 99+% efficiency at 20 microns. Since we don't know the magnitude of "+", we'll just make the approximation 99% efficient at 20 microns.

If I draw a straight line between those 2 points I get a straight line. It's all I really can do with only 2 points. I can then make some sort of prediction of efficiency for particle sizes between 5 and 20 microns using this line as an approximation. But if for sake of example the data between those points isn't really a straight line , say something like Filter G, the estimate will be signidifancyly lower than what the filter actually accomplishes. Small step sizes = small errors in estimation. Larger step sizes increases the risk of greater errors in approximation. However if the shape of the curve is similar to Filters B & H, a linear approximation introduces much smaller errors.

So by only knowing one point, extrapolation on anything other than small steps sizes risk high error. Even knowing two points, extrapolation beyond either of those two points as well as interpolation between those two points has risk of significant error.

Originally Posted By: ZeeOSix
Therefore, you really don't need to know the "multiple points of data" unless you were splitting hairs and focusing directly on one or two specific particle sizes for some reason.


I am focused on a specific particle size range - 10 to 20 microns - as this appears to be the general consensus of the size range that causes the most damage.

This is why I disagree. But I can agree that we disagree. I've simply explained why I see what I see.

 

[Nyogtha]

dnewton3, I've been branded a heretic quite often through my life and career, saved my employers millions & millions of $$$ by being one, and earned a comfortable living doing so enabling retirement when my health precluded working further. I've never been one to "just go along with everyone else just to make things easier".

As far as testing at multiple particle sizes, do you (or does anyone other than maybe Motorking know) how many data points are collected using the test apparatus in Fram's video? It would seem other data can naturally and quite easily be found as some filters such as WIX XP, Napa Platinum and Bosch D+ all report a single value at 40 microns. Purolator Synthetic reports a single data point at 25 microns. How did they choose these points without having a number of data points to compare to choose from?

How about just 2 points - 10 microns and 20 microns - for efficiency reporting if that's the size range that causes the most wear? To me personally, collection efficiencies outside this range are of much lower interest based on that widely published premise. I know efficiency below 10 microns will be less than at 10 microns but how much should I care what that value is (for example, why should I care if Fram Ultra is 80% efficient at 5 microns if that size is not a significant concern for wear)? To me that info simply lets me know collection efficiency of the Fram Ultra at 10 microns is somewhere between 80% and 99+%. Similarly, I know collection efficiency above 20 microns will be greater than that at 20 microns, but why should I care exactly what that value is?

 

[Nyogtha]

If those 2 points are too much to ask for I go back to this.

Originally Posted By: Nyogtha
Personally, I'd prefer it if they all reported a standardized test efficiency from ISO 4548-12 at 10 microns. If the general consensus is particles in the 10 to 20 micron size range cause the most damage, why not report efficiency at the small end of that range instead of the large end of that range?

 

[dnewton3]

I would agree that it would be nice to consolidate the data down to a few points. 10um and 20um would be good to know.

However, and this is just from an engineering POV, if you want to know something with reasonable accuracy, you might do well to sample a bit below and above, as well in the middle.

If you had an air gage for filling your tires, and the typical fill was 32 PSI, you don't want a gage that stops at 35psi. You want to go AT LEAST 20% above (or below) your target so the inaccuracy of the edge of the envelope does not interfere with your desired range of measurement.

For filters, I would like to know the info at perhaps 8um, 15um and 25um. That would give me three points that "surround" the concerning range from 10-20um; one below, one in the middle, and one above. Three points also can indicate a curve, if prevalent. Two points can only ever give you a straight slope. If you understood the variance of your data, and the gage R&R, you can reasonably find any point along the "curve" with just three points. Which is why you want to grab one below, middle and above your interest range.

But it's not like the industry is going to stop what they do and pick up my idea; it's just wishful thinking.

 

[ZeeOSix]

Originally Posted By: dnewton3
I do not disagree with your general premise, but you, sir, are going to be seen as a heretic here on BITOG ... I should know; I am often viewed the same way

See my comments inserted in your quote ...

2. Dirt holding capacity reported using ISO 4548-12 is a function of both filter media and bypass valve opening differential pressure. So dirt holding capacity is not necessarily increased or decreased as the physical size of a filter element is increased or decreased, nor necessarily when counting media pleats or media surface area between filters with different bypass valve opening differential pressures. For filter applications with no bypass valve (bypass is in the engine) it will depend on what that engine's bypass opening differential pressure is (and how stable it remains over the life of the engine).
I don't really have a strong opinion here, but you are swimming against the current of most of the Filter Faithful here at BITOG. Most here believe that a larger filter will most certainly hold more. Much of this depends upon surface loading versus depth loading, which is a trait of the media design and not size. The real question most ignore is how much of the available capacity is used in the first place. Having "more" capacity, well past any sensible amount one uses is not a benefit, but rather a waste. Those who use a FU rather than the EG, simply because they was "more capacity" for a 5k mile OCI, are foolishly misunderstanding the concept of the need for capacity. The question of how much a filter holds is only secondary; the REAL QUESTION that should FIRST be answered is this:
How much capacity do you need?



I'm gonna have to disagree with the statement:
"Dirt holding capacity reported using ISO 4548-12 is a function of both filter media and bypass valve opening differential pressure. So dirt holding capacity is not necessarily increased or decreased as the physical size of a filter element is increased or decreased ... "

Dirt holding capacity is a very strong function of the media surface area when two identical medias are compared to each other. That is one reason people like to run an over-size of the same filter. Going over-size also will help keep the delta-p down which gives more insurance that the bypass valve doesn't open as soon.

 

[dnewton3]

Zee -

I agree in theory, but I would like to see a lot of data before I committed to the idea. Which is why I said I don't really have a strong opinion about this yet. I agree conceptually; I have yet to see proof.

 

[ZeeOSix]

Originally Posted By: Nyogtha
OK, well lets have a side discussion.

Filters B & H are rated 98% efficient at 15 microns in the table you posted. However the graph you posted shows they actually are about 65% efficient at 15 microns (purple line).

Filter G rated as 98% efficient at 25 microns in the table you posted shows it's still close to 98% efficient at 15 microns as its curve is very flat in that section.


Not my table or graph ... something I found on the 'net to use as an example of the topic we were first discussing when I posted it. Yes, some of the graphs don't quite match the table, and that's why I used the line for Filters A & E as my example in the first place. I used the line for Filters A & E to show what a typical Efficiency vs Particle Size curve might look like for an oil filter, and to use that curve to try and convince some here that "@20 microns" and "for particles >20 microns" basically meant the same thing. When shown graphically, it's hard to deny that is the case.

Originally Posted By: Nyogtha
So only knowing the points in the table you posted, one could not reasonably predict / extrapolate what efficiencies are at other sizes not listed in the table for these 3 filters without the curves in the graph you posted (and I think whoever manufactured Filters B & H wasn't being very up-front). Personally, I wouldn't purchase Filters B & H after seeing the graph. But if I only saw the table you posted with one point for each filter, how would I know differently?


I can look at just those single data points in the table and I can easily rate all of those filters from best to worse in efficiency. The filter(s) that are best at 98% efficiency are also best at any other micron level than the others. That's the beauty of comparing efficiencies ... it CAN be done with one data point IF the same data point is used. Look at the family of curves on that graph. You can draw a straight line up from any microns size and see which filter is best and which is worse ... just like the single data point in the table shows.

Originally Posted By: Nyogtha
So by only knowing one point, extrapolation on anything other than small steps sizes risk high error. Even knowing two points, extrapolation beyond either of those two points as well as interpolation between those two points has risk of significant error.


I think you're using the graph wrong. Draw a straight line vertical from any particle size and you'll see which filter is better than the next.

Originally Posted By: Nyogtha
Originally Posted By: ZeeOSix
Therefore, you really don't need to know the "multiple points of data" unless you were splitting hairs and focusing directly on one or two specific particle sizes for some reason.


I am focused on a specific particle size range - 10 to 20 microns - as this appears to be the general consensus of the size range that causes the most damage.

This is why I disagree. But I can agree that we disagree. I've simply explained why I see what I see.


Even if you draw two vertical lines on the graph capturing the 10 to 20 microns range, it's easy to see which filters are better.

 

[ZeeOSix]

Originally Posted By: dnewton3
Zee -

I agree in theory, but I would like to see a lot of data before I committed to the idea. Which is why I said I don't really have a strong opinion about this yet. I agree conceptually; I have yet to see proof.


Agreed ... some theories are hard to prove. But I'm sure you can agree using engineering logic that a filter with the same media composition but twice the surface area will theoretically hold twice as much debris before the delta-p across it is the same as the media that is half the area.

And engineering logic would also say that the same media with twice the surface area will have about half the delta-p (delta-p vs flow is slightly curvature, but 1/2 would be a close estimate).

The reason filter manufactures keep down sizing their filters is because they know that with today's cleaner engines and better motor oils that the filter will still hold enough debris if it's changed out on a regular basis per the vehicle's maintenance schedule. And they down size to save money on materials which raises their profit bottom line.

 

[FutureDoc]

Zee, I think you missed the point about media size. It was not solely that media size is not a factor, but rather that size is less relevant when design and media type is not controlled. So for your example, media size is important when all things are equal, but I think Nyogtha was point to when things are not equal (which is more common than not). So if we are keeping to a specific brand and media type, I would agree with you but not across brands.

Also, what about filter B/H and F. They "exchange" positions. At about 10 microns BH looks better but at 20, F looks better in the graph (the table is a different story). I just have a hard time thinking that filter media across the industry is linear with equal retention rates depending for a given point.

 

[ZeeOSix]

Originally Posted By: FutureDoc
Zee, I think you missed the point about media size. It was not solely that media size is not a factor, but rather that size is less relevant when design and media type is not controlled.

So for your example, media size is important when all things are equal, but I think Nyogtha was point to when things are not equal (which is more common than not). So if we are keeping to a specific brand and media type, I would agree with you but not across brands.


I qualified my comments - didn't miss anything. I never said you can compare totally different media types by surface area. I said IF the medias are the same composition, then you can look at such things as surface area, etc.

That's why a Fram Ultra (or any full synthetic media filter) can hold more debris with LESS surface area because full synthetic media is a "3- dimensional depth filtering" design instead of just 2-dimensional filtering like plain old cellulose media.

Originally Posted By: FutureDoc
Also, what about filter B/H and F. They "exchange" positions. At about 10 microns BH looks better but at 20, F looks better in the graph (the table is a different story). I just have a hard time thinking that filter media across the industry is linear with equal retention rates depending for a given point.


Yes, I saw the cross-over. Pretty negligible IMO. Note they are slightly different media compositions. All this graph was meant to be uses for is to show what a "typical" Efficiency vs Particle Size distribution curve would look like. Keep in mind that this table and graph shows comparisons of different types of media also, and that also gives one an idea of what the curve might look like for different types of media. Different media types yield different shaped curves.

But my statements still stand that you can simply look at that table of graph and easily see which filters are better amongst the group - even if they are different media compositions. The graph pretty much backs that up too. Except for Filter G and B&H, which seems to have a mis-match between table and graph. I think the table has some mis-prints, so just look at the graph.

Per the graph, the best to worse filters would be:
D, C, G, F, B&H, A&E.

 

[Nyogtha]

ZeeOSix,

FWIW I never said those were your table & graph - check what you quoted from my post to verify.

Any explanation on why table says B & H are more efficient than F?

Without the graph, there's no way to see F is better than B & H.

I'm not sure you understand the meaning of extrapolation and interpolation . Extrapolation is trying to predict something outside a data set. Interpolation is trying to predict something at a point within the bounds of the data set, but not directly measured in the data set.

So using just the data from the table, which is a single data point typical of what the consumer gets from a filter brand or manufacturer, can you reasonably predict say the efficiency of the various filters at say 20 microns, a data point not given in the table? Remember, not using the graph. Which filters are likely to have the most error trying to do this? Those with their single data point in the table further away from 20 microns, or those with their single data point closer to 20 microns?

Then let's give an example of interpolation. You can see a straight line drawn between the 2 data points you selected on the graph, greater error will result for Filter F due to it's curve shape, but if you only have those two data points (no graph) there's no way to know that.

Notice extrapolation and interpolation for the data we are given without the graph is very different? Clearly the graph was generated with more data than the points in the table. But we have no idea what those data points actually were, other than maybe the points given in the table (and not all those are consistent). It is quite reasonable to think the graph curves themselves have considerable interpolation built in as it would be impossible to test all the data points along a given line. We have no idea whether any extrapolation is involved on the graph but since the curves all have the same bounds, we can reasonably think those were the bounds of the data set used to generate the curves.

If we as consumers got such graphs on the side of the filter box, we wouldn't have to deal with trying to extrapolate from a single data point being givenor outside a set of two data points if the luxury of 2 data points are given - nor interpolating between only two data points if the luxury of 2 data points are provided.

dnewton3 - I agre the industry is unlikely to change what they're doing. But clearly more data is generated testing than is published otherwise how do manufacturers or brands choose the point or points they publish? So I don't think there would be an incremental cost to provide a tested value a my preferred value 10 microns.

If one filter is higher efficiency at both 10 and 20 microns than another filter that's all the info I'd need to make a more informed decision It's hard to imagine any media efficiency curve would be less than linear between those 2 points. It may be better perhaps in an upward curve shape vs. a straight line, but if that turns out to be the case, it would only be a bonus for me, and perhaps that manufacturer would share that data to differentiate that design.

I agree I doubt it will change in the real world.

 

[dnewton3]

Regarding the slope of linearity in the efficiency curve, I would like to see a lot more data.

As much as I abhor the infamous GM filter study, one thing it did show is that most filters will saturate (if run long enough, i.e. way too long ....) and end up setting a dP large enough to go into bypass. But the study also showed with clarity that regardless of the initial pore size (40u, 25um, 15um), the eventual performance capture limit in terms of size was around 10um. Stuff smaller than that just kept on passing through, which delayed the dP tripping the BP. The study clearly states that there was a convergence of the particulate size at 10um, despite the huge variance in initial pore sizes. More porous filters will take longer to get to that 10um point, but they all did get there eventually. So what it means is that for any given filter media design, along with the applied characteristic such as the BP setting, etc., that the response "curve" is not truly linear at ALL particulate sizes. That is why I want to see three points; so that I understand the curve and can predict it's salient saturation size.

No - I don't really care about how efficient a typical filter is at 5um, but between 8-25um, I want to see at least three points.