Fram Endurance / Walmart End of Life ?

[ZeeOSix]

All but one of the FRAM Ultras represented in that table are the older OG Ultra with the full synthetic media. The one other particle count I've seen for a newer synthetic blend Ultra came in ISO at 13 (>14 um), 2 ISO codes lower than this leaky filter. That PC was from 2023, so maybe before the leaky end caps became common. Still, there's not enough data to say for certain if this was due to the leaky end cap or just random variance.

It could be any filter with a metal-to-metal "sealing" interface could have some level of leakage, and obviously if the leaf spring stamping is bad with ruffles and warpage, then it could vary all over the place depending on the manufacturing quality the day they were stamped as seen by the different levels of that seen in C&Ps. The only way a metal-to-meal leaf spring interface it going to seal the best is if the leaf spring is nice and flat and very smooth where it seals on the raised ring around the center tube on the end cap.

Here's an updated table. I included another table with particle counts for filter models that I only have one particle count for.

View attachment 290528

What would this data look like if the lower ISO 4406 codes were compared?

 

[twX]

What would this data look like if the lower ISO 4406 codes were compared?

As I mentioned in a previous thread, it's difficult to compare the results from older and newer PCs from Blackstone, at least for 4 and 6 um particle sizes.

The ISO 4406 standard was updated in 2017 (from the 1999 version), and by 2018 most Blackstone PCs started reporting much higher ISO codes for the 4 and 6 um particle sizes (at least 4 ISO codes higher). The results for >14 um particles don't seem to have been affected much, if at all.

The other issue is that with the older method, the 4 and 6 um PCs seem to have been completely extrapolated instead of measured, which makes them pretty much meaningless. With the newer tests, there are some significant differences at 6 um, which suggests that they're now actually measuring the smaller particles.

Here's a table with results for 4 and 6 um, but only for filters tested with the more recent method, and only for filters I have multiple data points for. The filters rank about the same at 6 um as they did at 14 um with the larger data set, but with the BOSS edging ahead of the OG Ultra. Most of the differences between filters in this table aren't statistically significant due to the small sample sizes.

 

[ZeeOSix]

As I mentioned in a previous thread, it's difficult to compare the results from older and newer PCs from Blackstone, at least for 4 and 6 um particle sizes.

The ISO 4406 standard was updated in 2017 (from the 1999 version), and by 2018 most Blackstone PCs started reporting much higher ISO codes for the 4 and 6 um particle sizes (at least 4 ISO codes higher). The results for >14 um particles don't seem to have been affected much, if at all.

What changed in ISO 4406 that made the reporting output change? Was is calling out ISO 11171 instead of ISO 4406 for the particle counter calibration method? That's what would change the PC output. But ISO 11171 came out in 1999, so if Blackstone didn't adopt it until 2017 that seems strange. BTW, typically when a PC is done using a particle counter calibrated to ISO 11171, they use the subscript (c) after the PC data. I looked on Blackstone's website and couldn't find any direct info saying exactly what particle counter calibration method they use, or if they changed it at some point. The Blackstone PC reports don't reference the subscript (c) in their reports.

The other issue is that with the older method, the 4 and 6 um PCs seem to have been completely extrapolated instead of measured, which makes them pretty much meaningless. With the newer tests, there are some significant differences at 6 um, which suggests that they're now actually measuring the smaller particles.

Here's a table with results for 4 and 6 um, but only for filters tested with the more recent method, and only for filters I have multiple data points for. The filters rank about the same at 6 um as they did at 14 um with the larger data set, but with the BOSS edging ahead of the OG Ultra. Most of the differences between filters in this table aren't statistically significant due to the small sample sizes.

This again doesn't make sense when the Boss looks better than the OG Ultra in this data set. We know from Andrew's official ISO 4548-12 efficiency testing that the OG Ultra smoked the Boss in efficiency. The ISO 4548-12 claims from the manufacturers (OG Ultra 99+% @ 20u vs Boss 99% >46u) also correspond with Andrew's ISO 4548-12 testing, but not when compared to this Blackstone PC data.

There is a lot PC range in each ISO code tier, so for me looking at the actual particle counts instead of the ISO code number gives more insight, and plotting them out against each other like I showed in post 83 makes it easier to see how they compare.

 

[ZeeOSix]

Particle counting method ISO 4406 reference info.

https://ayalytical.com/iso-4406/

 

[twX]

What changed in ISO 4406 that made the reporting output change? Was is calling out ISO 11171 instead of ISO 4406 for the particle counter calibration method? That's what would change the PC output. But ISO 11171 came out in 1999, so if Blackstone didn't adopt it until 2017 that seems strange.

I don't know what changes were made to ISO 4406 in 2017, and I don't know if the changes to that standard are even responsible for the different particle counts, but both changes happened at around the same time.

ISO 11171 has been updated four times since 1999, including in 2016, so maybe that could be related as well.

Thinking about this some more, I really doubt that the ISO standards would change in such a way that measurements would start coming in 5 ISO codes higher than with older standards. My guess is that Blackstone wasn't doing their tests strictly to the standards in the first place. I don't have copies of the standards, but the extrapolation they were doing doesn't seem like something that would be allowed by the standard.

I'm guessing that they stopped using extrapolation for the counts of the smaller particles and started actually measuring them at some point. They were extrapolating using a standard particle size distribution curve which was probably more appropriate for hydraulic or bypass filtration than for typical automotive applications, and the actual measured values ended up being a lot higher than the previously extrapolated values.

 

[twX]

Particle counting method ISO 4406 reference info.

https://ayalytical.com/iso-4406/

"Interpolation is acceptable. Extrapolation is not permissible."

Well that answers that question. I'm more convinced now that Blackstone just wasn't testing properly prior to ~2017 since they were clearly using extrapolation.

 

[ZeeOSix]

I don't know what changes were made to ISO 4406 in 2017, and I don't know if the changes to that standard are even responsible for the different particle counts, but both changes happened at around the same time.

ISO 11171 has been updated four times since 1999, including in 2016, so maybe that could be related as well.

Thinking about this some more, I really doubt that the ISO standards would change in such a way that measurements would start coming in 5 ISO codes higher than with older standards. My guess is that Blackstone wasn't doing their tests strictly to the standards in the first place. I don't have copies of the standards, but the extrapolation they were doing doesn't seem like something that would be allowed by the standard.

I'm guessing that they stopped using extrapolation for the counts of the smaller particles and started actually measuring them at some point. They were extrapolating using a standard particle size distribution curve which was probably more appropriate for hydraulic or bypass filtration than for typical automotive applications, and the actual measured values ended up being a lot higher than the previously extrapolated values.

Hard to say about Blackstone. Maybe they weren't using ISO 11171 to calibrate their particle counter until around 2017. The old calibration method (ISO 4402) used for ISO 4406 particle count testing would skew the lower ISO 4406 particle sizes - 6u and lower, and more so as the particles go from 6u to 1u. Agree, it wouldn't skew it 4-5 ISO codes, but maybe that along with extrapolation really threw it off.

https://www.beckman.com/resources/industry-standards/iso-11171/calibration/reporting

Graphical comparison from a different source:

Also ... FYI.

 

[twX]

Hard to say about Blackstone. Maybe they weren't using ISO 11171 to calibrate their particle counter until around 2017. The old calibration method (ISO 4402) used for ISO 4406 particle count testing would skew the lower ISO 4406 particle sizes - 6u and lower, and more so as the particles go from 6u to 1u. Agree, it wouldn't skew it 4-5 ISO codes, but maybe that along with extrapolation really threw it off.

If that's the case, the PCs for >14um particles could read higher by up to one ISO code with the new method. It isn't obvious from the data if this is the case.

This again doesn't make sense when the Boss looks better than the OG Ultra in this data set. We know from Andrew's official ISO 4548-12 efficiency testing that the OG Ultra smoked the Boss in efficiency. The ISO 4548-12 claims from the manufacturers (OG Ultra 99+% @ 20u vs Boss 99% >46u) also correspond with Andrew's ISO 4548-12 testing, but not when compared to this Blackstone PC data.

One thing to consider is that Andrew only tested a single Ultra. There's going to be some variance in ISO test results, even for the same model of filter coming off the same production line. Add in other variables, like different model numbers, factories, batches of media, QC, production years, etc, there's probably a lot of variance. The Ultra Andrew tested might have been one of the best examples.

If you look at the standard deviations in the PC data, you'll notice that the synthetic media filters (OG Ultra and Amsoil/Royal Purple) have a lot more variance in their PC results than the cellulose or synthetic blend filters. The OG Ultra produced some of the best PCs as well as some of the worst. Maybe that's an indication of more variability in the quality of the synthetic media used in these filters.

There might also be differences in how a filter performs in an ISO test where it's filtering test dust in controlled conditions, vs in an engine where it's mostly filtering organic material, and where it's exposed to heat, higher flow velocities, pressure and temperature cycles, etc.

With 13 particle counts, I think it's safe to say that the OG Ultra doesn't tend to perform as well in the field as Andrew's ISO test suggests that it should.

There is a lot PC range in each ISO code tier, so for me looking at the actual particle counts instead of the ISO code number gives more insight, and plotting them out against each other like I showed in post 83 makes it easier to see how they compare.

I haven't been averaging the actual particle counts since there are a lot of PCs that I only have the ISO code for. I might make a formula to convert particle counts to fractional ISO codes for the filters I have actual particle counts for.

 

[ZeeOSix]

One thing to consider is that Andrew only tested a single Ultra. There's going to be some variance in ISO test results, even for the same model of filter coming off the same production line. Add in other variables, like different model numbers, factories, batches of media, QC, production years, etc, there's probably a lot of variance. The Ultra Andrew tested might have been one of the best examples.

Regardless, what Andrew's ISO 4548-12 showed is that those filters pretty much ranked in efficiency as the manufacturers' ISO efficiency claimed at the time. I'm not going to believe that the OG Ultra Andrew happened to test was a unicorn. We can say it was not an internal leaker since it has the fiber sealing ring for the leaf spring on the end cap. I'm betting if he tested 10 each of all those filters in the test, they would have all pretty much repeated the same ranking outcome. This is why when people latch-on to BR's "efficiency testing" by comparing particle counts, I have a hard time believing that their ranking is accurate.

If you look at the standard deviations in the PC data, you'll notice that the synthetic media filters (OG Ultra and Amsoil/Royal Purple) have a lot more variance in their PC results than the cellulose or synthetic blend filters. The OG Ultra produced some of the best PCs as well as some of the worst. Maybe that's an indication of more variability in the quality of the synthetic media used in these filters.

Probably more likely related to the fact that the UOAs were pretty much from many different engines and taken by many different people. All the filters with a stamped leaf spring may have more variance due to the level of possible internal leakage due to messed up leaf spring sealing on the end cap.

And who really knows how accurate and repeatable Blackstone's UOA particle count testing really is. Only way to have some accuracy of in field UOA to look at the filter efficiency would be to use different oil filters on the same car with the same oil over the same length OCI, and have the same person take the oil samples the same way every time. If the OCI is ran longer, there will be a higher level of particles that can't be fully filtered by the oil filter that will show up in the oil UOA PC.

There might also be differences in how a filter performs in an ISO test where it's filtering test dust in controlled conditions, vs in an engine where it's mostly filtering organic material, and where it's exposed to heat, higher flow velocities, pressure and temperature cycles, etc.

A filter should catch particulate, regardless if it's organic material or not.

With 13 particle counts, I think it's safe to say that the OG Ultra doesn't tend to perform as well in the field as Andrew's ISO test suggests that it should.

I don't recall if some of those 13 are not OG Ultras. If so, some could have been internal leakers, thereby reducing their efficiency. I will always use ISO 4548-12 efficiency over a PC to rate filter efficiency, unless someone can show that they have officially validated their home made garage testing apparatus results with correlation to the official ISO 4548-12 efficiency tests on the same filters.