Summary of non-destructive Fram filter testing (and a Carquest Premium added)

[Hatsuwr]

I tested three FS3600, three FE3600, four XG3600, and because of the disappointing results, added a Carquest Premium 85516.

I went into this testing expecting Fram to perform very well, fully planning to use them in my vehicles. Based on the results of these tests and Fram's failure to respond to my messages for weeks now, I will not be using Fram products anymore. I will run the 3rd XG3600 for a brief time for an engine flush and then cut it open for more testing, but the others are going to be returned and I'll be using the Carquest until something better comes up.



First test was to attach a vacuum to the outlets of the filters with a debris catching filter in between (I tried both cotton and cellulose). I wasn't expecting as much as there was from the Fram's, otherwise I would have used a better filter (maybe media from an air filter). I did not find anything from the Carquest. The following pictures are all from the Fram filters.

My takeaway from this, (besides maybe not getting Fram and maybe looking more into alternatives to fiberglass/microglass media), is to take a few seconds and vacuum out any new filter before installing. Very little effort, and there is clearly benefit at least in the case of these filters.

The next test was pressure differential, mostly to identify leaks. This was again done with a vacuum. I believe there are two things that can be gained, if we assume that leakage paths are a greater cause for differences in pressure differential than construction variance within a model:

• An oil filter that is less restrictive to air will likely be less restrictive to oil. In the absence of particle size and efficiency information (which takes priority in my opinion), and if we assume no leaks (at least, no leaks for the lower pressure differential filter), a filter with a lower pressure differential is likely better than a filter of a DIFFERENT MODEL with a higher pressure differential in the context of flow restriction. It must be a comparison between different models though, because of the second point:
• With two oil filters of the same construction, a lower pressure differential on one filter likely indicates greater total leakage paths. In the case of the Frams, that seems likely to be primarily from the bypass valves, but this can also be due to tears in the filter media.

The second point is more significant. If I had bought the green and blue FS3600 below and tested them in this way, I would assume that the green filter leaks significantly more than the blue. This could be an easy way to identify faulty valves or tears in media before installation.

Filters are marked red for filter 1, green for 2, blue for 3, and yellow for 4.

FE3600:
Filter 1: 23178AF - 2023, 178th day
Filter 2: 23022AF - 2023, 022nd day (longer louvers than 1 and 3)
Filter 3: 23178AF - 2023, 178th day

XG3600:
Filter 1: A40462 - 2024, 046th day
Filter 2: A40462 - 2024, 046th day
Filter 3: A41413 - 2024, 141st day
Filter 4: A40462 - 2024, 046th day

FS3600:
Filter 1: A30202 - 2023, 020nd day (holes in metal core, no wire backing)
Filter 2: A40673 - 2024, 067th day (louvers, wire backing)
Filter 3: A40673 - 2024, 067th day (louvers, wire backing)

The Carquest was tested at a later date, so at least part of the lower pressure differential is because of that.

Next up was testing of the ADBVs. I used an air compressor at low pressure to pressuring the outlets of each valve and checked for tests.

I was surprised to see that most leaking was from the joints between the cans and the face plates.

XG1 - Very bad leaking from the ADBV, to the point where I couldn't test for rim leaks
XG2 - Bad rim leak
XG3 - Bad rim leak
XG4 - Medium rim leak
FS1 - Minor rim leak
FS2 - Bad rim leak
FS3 - Bad rim leak
FE1 - Fully seals
FE2 - Tiny rim leak
FE3 - Minor ADBV leak
Carquest - Fully seals

I also tested ADBV leaking with about 25 ml of oil in each can, stored upside down for a while. The results lined up pretty well with the air testing, although all of the FS and FE3 leaked more in this test than what their air pressure tests would indicate.




Finally, I tested bypass valve leaking by putting a small amount of oil into the outlet of each filter. The oil quickly disappeared from the valve seat area on every filter, but I'm not sure if this is entirely due to leaking, or if maybe wicking up the filter media played a part. I will have to test more once I cut one open.

 

[gfh77665]

Lots of good work. I applaud this.

I am also very glad Project Farm didn't do this, some people here would have had a STROKE!

 

[Chris142]

I'm not sure if vacuum testing something that is designed to operate under pressure is the correct way to test it..But i'm no engineer.

 

[Hatsuwr]

I'm not sure if vacuum testing something that is designed to operate under pressure is the correct way to test it..But i'm no engineer.

The best way to test an oil filter is in the range of expected operating conditions. These tests were specifically meant to be non-destructive to look into their usefulness for assessing a filter (to some degree) without making it unusable.

Regarding just the matter of applying a vacuum to the outlet vs pressurizing the input, there will be little difference between the two. The pressure differential is the primary factor.

 

[ZeeOSix]

This could be an easy way to identify faulty valves ...

Finally, I tested bypass valve leaking by putting a small amount of oil into the outlet of each filter.

To clarify, it's not the actual bypass valve. If there's a leak path, it's the air gap seen between the leaf spring and end cap metal-to-metal "seal".

Next up was testing of the ADBVs. I used an air compressor at low pressure to pressuring the outlets of each valve and checked for tests.

I was surprised to see that most leaking was from the joints between the cans and the face plates.

If you say air leakage there, then those would most likely seep oil during actual use. How were you trying to detect the leakage at the can to base plate joint?

 

[twX]

An oil filter that is less restrictive to air will likely be less restrictive to oil.

When using air, there may be a difference in terms of which part of the filter is dropping most of the pressure.

With oil, most of an oil filter's restriction usually comes from the filter media. In these compressed air tests, most of the restriction could be from the inlet holes and ADBV, especially since they're such tall filters with a lot of media. The amount of inlet hole flow area on these filters correlates with the results of the restriction test. The Carquest and FS have the least restrictive inlet holes, followed by the FE and the XG.

Consider an air intake system on a car. The air filter might have 3 times more media area than these oil filters, while its intake piping will generally have over 100 times more flow area than an oil filter's inlet holes. I'm thinking that with air, the media restriction of these oil filters might be pretty insignificant compared to the canister restriction. This would also mean that a leaf spring seal leak wouldn't have any obvious effect on the test results.

 

[ZeeOSix]

With oil, most of an oil filter's restriction usually comes from the filter media. In these compressed air tests, most of the restriction could be from the inlet holes and ADBV, especially since they're such tall filters with a lot of media.

He's drawing a vacuum through the filter ... not using compressed air to determine the flow restriction.

 

[Hatsuwr]

To clarify, it's not the actual bypass valve. If there's a leak path, it's the air gap seen between the leaf spring and end cap metal-to-metal "seal".


If you say air leakage there, then those would most likely seep oil during actual use. How were you trying to detect the leakage at the can to base plate joint?

Added oil to the filter, then slightly pressurized. The leak is inside the seal though, so would just be a potential drain back issue.

And I should mention that, while I think zero leaking from the ADBV and face plate junction should be the standard, only the first XG3600 (and maybe the ones I marked as having bad leaks) seemed to leak enough to likely be an issue for typical shut off periods.

I'm thinking that with air, the media restriction of these oil filters might be pretty insignificant compared to the canister restriction. This would also mean that a leaf spring seal leak wouldn't have any obvious effect on the test results.

I agree with the principle of the of the concern, but considering that there was almost zero overlap between the grouping of results between identical filters, there is clearly a measurable difference in restrictions between identical looking filters made on the same day at the same location.

The most significant possible contributors, in my opinion, are the media (and glue spillage on them), leakage paths (I assume these are mostly going to be bypass valve related), and variations in the ADBV (some of the filters had very noisy resonance from the ADBVs when testing with air).

Personally I'd expect the media and ADBV to be more consistent in the restriction they introduce, at least compared to unintentional rippling of the leaf springs, but I wouldn't bet much at all on that.

 

[ZeeOSix]

Added oil to the filter, then slightly pressurized. The leak is inside the seal though, so would just be a potential drain back issue.

And I should mention that, while I think zero leaking from the ADBV and face plate junction should be the standard, only the first XG3600 (and maybe the ones I marked as having bad leaks) seemed to leak enough to likely be an issue for typical shut off periods.

That makes more sense. Yes, if it's leaking air from there, then that could be interpreted as the ADBV leaking when you checked for that. That's been seen on some other filter brands too, and may be the cause when someone reports their engine rattles on cold start-up ... even though the actual ADBV valve is sealing 100%, the filter still has a drain-back path though the seam leak.

 

[CharlieBauer]

To summarize;

1) 4 Ultras manufactured 2024 have an average of a Bad rim seal / adbv leaking (1 Medium RS, 1 Very Bad ADBV, 2 Bad RS)

2) 2 Titaniums manufactured 2024 both have a Bad seam leak

3) 1 Titanium manufactured Jan 2023 has a Minor seam leak

4) 3 Champ 2023 manufactured Endurance had only 1 that fully seals. 1 had minor adbv leak and 1 had tiny rim seal leak

5) Carquest (manufactured date not mentioned) no leak

Defect Rate

1) 2024 Fram Ultra / Titanium: 100% & Bad

2) 2023 Fram Titanium: 100% & Minor

3) 2023 Champ Fram Endurance: 67% & Tiny - Minor

4) Carquest: 0%

My own oil leak testing of Fram Ultra showed;

2021: Not leaking
2022: Slow leaking
2024: Fast leaking

Conclusion

Fram has steadily become the worst quality major oil filter manufacturer since First Brands acquired them and there are defects throughout every filter they manufacture.

 

[Glenda W.]

I tested three FS3600, three FE3600, four XG3600, and because of the disappointing results, added a Carquest Premium 85516.

I went into this testing expecting Fram to perform very well, fully planning to use them in my vehicles. Based on the results of these tests and Fram's failure to respond to my messages for weeks now, I will not be using Fram products anymore. I will run the 3rd XG3600 for a brief time for an engine flush and then cut it open for more testing, but the others are going to be returned and I'll be using the Carquest until something better comes up.



First test was to attach a vacuum to the outlets of the filters with a debris catching filter in between (I tried both cotton and cellulose). I wasn't expecting as much as there was from the Fram's, otherwise I would have used a better filter (maybe media from an air filter). I did not find anything from the Carquest. The following pictures are all from the Fram filters.

My takeaway from this, (besides maybe not getting Fram and maybe looking more into alternatives to fiberglass/microglass media), is to take a few seconds and vacuum out any new filter before installing. Very little effort, and there is clearly benefit at least in the case of these filters.

View attachment 245934
View attachment 245939
View attachment 245941View attachment 245942



The next test was pressure differential, mostly to identify leaks. This was again done with a vacuum. I believe there are two things that can be gained, if we assume that leakage paths are a greater cause for differences in pressure differential than construction variance within a model:

• An oil filter that is less restrictive to air will likely be less restrictive to oil. In the absence of particle size and efficiency information (which takes priority in my opinion), and if we assume no leaks (at least, no leaks for the lower pressure differential filter), a filter with a lower pressure differential is likely better than a filter of a DIFFERENT MODEL with a higher pressure differential in the context of flow restriction. It must be a comparison between different models though, because of the second point:
• With two oil filters of the same construction, a lower pressure differential on one filter likely indicates greater total leakage paths. In the case of the Frams, that seems likely to be primarily from the bypass valves, but this can also be due to tears in the filter media.

The second point is more significant. If I had bought the green and blue FS3600 below and tested them in this way, I would assume that the green filter leaks significantly more than the blue. This could be an easy way to identify faulty valves or tears in media before installation.

Filters are marked red for filter 1, green for 2, blue for 3, and yellow for 4.

FE3600:
Filter 1: 23178AF - 2023, 178th day
Filter 2: 23022AF - 2023, 022nd day (longer louvers than 1 and 3)
Filter 3: 23178AF - 2023, 178th day

XG3600:
Filter 1: A40462 - 2024, 046th day
Filter 2: A40462 - 2024, 046th day
Filter 3: A41413 - 2024, 141st day
Filter 4: A40462 - 2024, 046th day

FS3600:
Filter 1: A30202 - 2023, 020nd day (holes in metal core, no wire backing)
Filter 2: A40673 - 2024, 067th day (louvers, wire backing)
Filter 3: A40673 - 2024, 067th day (louvers, wire backing)

The Carquest was tested at a later date, so at least part of the lower pressure differential is because of that.

View attachment 245945



Next up was testing of the ADBVs. I used an air compressor at low pressure to pressuring the outlets of each valve and checked for tests.

I was surprised to see that most leaking was from the joints between the cans and the face plates.

XG1 - Very bad leaking from the ADBV, to the point where I couldn't test for rim leaks
XG2 - Bad rim leak
XG3 - Bad rim leak
XG4 - Medium rim leak
FS1 - Minor rim leak
FS2 - Bad rim leak
FS3 - Bad rim leak
FE1 - Fully seals
FE2 - Tiny rim leak
FE3 - Minor ADBV leak
Carquest - Fully seals

I also tested ADBV leaking with about 25 ml of oil in each can, stored upside down for a while. The results lined up pretty well with the air testing, although all of the FS and FE3 leaked more in this test than what their air pressure tests would indicate.




Finally, I tested bypass valve leaking by putting a small amount of oil into the outlet of each filter. The oil quickly disappeared from the valve seat area on every filter, but I'm not sure if this is entirely due to leaking, or if maybe wicking up the filter media played a part. I will have to test more once I cut one open.

Great work! Thanks

 

[Ronn]

Fram has steadily become the worst quality major oil filter manufacturer since First Brands acquired them and there are defects throughout every filter they manufacture.

I will continue to use the Endurance with no reservations. I expect a typical response from the “Hysterical” crowd.

 

[CharlieBauer]

The Carquest was tested at a later date, so at least part of the lower pressure differential is because of that.

Any chance you can do this test under the same conditions with the Carquest and one of each of the Frams to see what the difference is? At the moment, it is suggesting that a properly manufactured Carquest has less resistance than a leaky Fram.

Finally, I tested bypass valve leaking by putting a small amount of oil into the outlet of each filter. The oil quickly disappeared from the valve seat area on every filter, but I'm not sure if this is entirely due to leaking, or if maybe wicking up the filter media played a part. I will have to test more once I cut one open.

As mentioned earlier, in the Fram, we are looking for a leak path between the end cap and the leaf spring. In the Fram Ultra / Titanium design, the oil can't all wick out through the media without some of it travelling down the slope of the leaf spring and into the gap between the end cap hole and the leaf spring towards the end cap hole edges. Some oil will flow down into the gap without leaking past the end cap so when I tested, I topped up after all the oil disappeared.

As mentioned, only a 2021 Ultra held oil indefinitely.

As to the Carquest, it seems you said it leaked? I think in that design there is a chance that the oil would wick out. We'd need to examine some pictures to understand if it all would. Otherwise, it would be worrying if a PG filter leaked through the bypass valve.

 

[ZeeOSix]

Where'd the horse in a body cast come from?

 

[bbhero]

I will continue to use the Endurance with no reservations. I expect a typical response from the “Hysterical” crowd.

At the beginning of that movie Cobra from 1986 with Sylvester Stallone in it playing Cobra in that grocery store …..

Crazy guy : I gotta bomb… I’m going blow this whole place up ! !

Cobra : Go ahead… I don’t shop here anyway…


That sums up what I think about this….

 

[Hatsuwr]

Any chance you can do this test under the same conditions with the Carquest and one of each of the Frams to see what the difference is? At the moment, it is suggesting that a properly manufactured Carquest has less resistance than a leaky Fram.

As to the Carquest, it seems you said it leaked? I think in that design there is a chance that the oil would wick out. We'd need to examine some pictures to understand if it all would. Otherwise, it would be worrying if a PG filter leaked through the bypass valve.

Yea it's hard to say much about the bypass leak testing without knowing more about the construction. When I get around to cutting some of these open I think I'll have a better idea of how I can do it better. I don't think the Carquest was leaking through the bypass, probably just passing through the media and maybe wicking up it a bit.

I went ahead and retested the 3rd XG3600 that I am keeping, and results were very close to its original testing, so I would agree that the Carquest is less restrictive.

 

[CharlieBauer]

I went ahead and retested the 3rd XG3600 that I am keeping, and results were very close to its original testing, so I would agree that the Carquest is less restrictive.

There are 2 independent tests of oil filter restriction for Fram and PG that I am aware of:

1) 2021 Ascent Filter testing. At 22.5lpm / 5.94gpm

a) Fram Ultra OG: 3.38psi
b) Royal Purple: 4.38psi

2) 2023 Brand Ranks: At 6gpm

a) STP XL: ~3.5psi
b) Fram Ultra (First Brands): 4psi
c) Fram Endurance: ~3.7psi

STP XL is a PG / Carquest Premium so your result correlates.

 

[ZeeOSix]

^^^ And the flow vs dP difference between most oil filters isn't going to matter at all on normal street cars in normal driving conditions. A + or - of 1 PSI isn't going to matter at all.

 

[fantastic]

Fascinating home brewed testing. Appreciate the time you took to post this. Only thing I would've liked to see is testing a couple more Carquest filters to even the parameters. I'm going to vacuum my Fram filters now.

 

[malibuguy79]

Interesting results for sure