Thank you for that post!! I’ll get on those links when time permits.
Filter Efficiency: when small differences aren't small
- Thread starter Hohn
- Start date
- Joined
- Oct 3, 2023
- Messages
- 865
This is why I always preferred to run a larger 3600 sized filter over a stock sized 4967 if it fits my engine. More media surface area is always good and also less differential pressure under the same flow conditions. The only downside to running a larger filter is that it needs to be pre-filled prior to install to avoid long oil pressure delays on first start as the pump primes the engine with new oil.
Machinery Lubrication online magazine talks about the benefits of running a larger filter in industrial applications to maximize service interval hours. https://www.machinerylubrication.com/Read/29114/dirt-holding-capacity
Fun fact, Donaldson is one of the few companies that make both stock size and oversized filters for same HD diesel engine and hydraulic systems applications under their BLUE / Duramax Synteq line of filters should space permit for one.
Here's another good Machinery Lubrication article, with some info about filter efficiency and wear. The take-away from Figure 1 is that with tighter journal bearings, it's more important to have better oil filtration. Modern engines do have pretty tight journal bearings, and the film thickness in journal bearings varies with oil viscosity, engine RPM and load, and therefore the bearing film thickness can get pretty small at times. Keeping as many particles out of the oil down to around 10u is beneficial. A filter that's rated at 99% @ 20u is going to be better for this than a much less efficient filter.
https://www.machinerylubrication.com/Read/31963/how-to-identify-and-control-lubricant-contamination
https://www.machinerylubrication.com/Read/31963/how-to-identify-and-control-lubricant-contamination
Hate to be a curmudgeon here, but...of course as gearheads and techies, we want the absolute best, or at least the best available. Have there been any rigorous studies that prove conclusively there is a performance difference between a say, 95% vs 99% efficiency in a filter? Is there noticeable wear? Does the engine start burning oil sooner?
I believe there have been studies and they've concluded that the smaller molecules do the most wear over time . No I can't point to where the articles are at
The difference between 95% @ 20u and 99% @ 20u isn't going to make a huge difference in wear, but a difference of say 50% @ 20u or 99% @ 40u compared to a 99% @ 20u filter over the long run would allow much more wear debris in the oil. The longer the OCI, the better it is to use a high efficiency oil filter. If the OCI was done every 1000 miles, the filter efficiency wouldn't matter as much.
trueI wish we knew the real numbers for the Subaru and Toyota filters , we know pretty much what the boss is and we like to tease about it . Millions and millions of Toyota's running around lasting quite a while on filters that are considerably less efficient than the boss .
I can’t see a thing wrong with that filterI have had time to relect on my poorly chosen words on the Fram TG. I of all people should have known better. My apologies to all involved and although I will not be using up my stash of Fram TGs I will concede that they do in fact have a much better flow rate than the competition.
I stand corrected.
View attachment 265437
Your post got me thinking of LSJr’s latest video on the GM recall and the switch to 0w40. It seems to me that thicker oil is in order with these newer vehicles that are geared really tall.
My accord doesn’t get past 2000 rpm in 10th gear until nearly 80mph. It will happily cruise at 1500rpm in 10th gear at 60mph. How thick is that oil film in the bearings at such low RPM? It seems to me that even a 30 grade might be a bit thin under those conditions since the engine will hit nearly 20psi of boost at such low RPM. That’s a LOT of stress on rod and main oil films to have so much cylinder pressure at such low RPM. (Heck this is why larger diesels all run 40 grade).
I’m kind of wondering now if I shouldn’t be trying some of this 15w40 HPL PCMO I have instead of the 5w30 Valvoline Euro Xl.
Yes, the combination of low RPM and high rod loads as you described is going to make the minimum film thickness suffer compared to higher RPM conditions. And throw in some possibly fuel dilution that further reduces the oil viscosity and it adds another negative factor to the bearing MOFT. No engine suffers from a highter HTHS viscosity oil, unless one thinks losing 0.1 MPG is "suffering".
Very good point. Regarding flow vs filter efficiency, if you have an oil filter rated at 99% efficiency @ > 20 microns, and it's not made of synthetic media, obviously the flow may decrease due to the higher restriction/efficiency.
Would that lower flow affect the engine's oiling system flow rate in normal passenger car engines?
A more flow restrictive filter is only going to effect the oil pressure after the PD pump goes into pressure relief. And even so, the reduction of oil pressure is only going to be a few PSI difference. PD oil pumps are designed to over feed the engine with oil flow if the designer did it right, so cutting a few PSI of oil pressure at high RPM is still going to provide adequate oil flow to the oiling system. If a few PSI difference in oil pressure from different oil filters being used was so critical, you'd see damaged and blown up engines all over the place.
If you consider that OE filters for Japanese and German vehicles seem to be 99% efficient "only" at the 40 to 50 micron range, then at least we can say there is a consensus that you don't want particles larger than 50 microns getting through the filter.
Thus the problem with a Fram that leaks 5% to 15% of oil flow is that it is only 85% to 95% efficient above 50 microns which is worse than the "rock catchers" that are sufficient for Japanese vehicles to last hundreds of thousands of miles.
Reviewing literature of the OE filter manufacturers for the German brands, you will overall see an emphasis on sealing / preventing bypass and holding capacity (to prevent bypass). The integrity of filtering is very apparently the focus.
In air filters, both integrity and efficiency are stressed.
^^^ If a non-leaking filter was 99% @ 20u, and then had an internal leak of 20% past the media, then the efficiency would basically be 99% - 20% = 79% @ 20u. I think that would be a worse efficiency than 99% @ 25u.
Just look at the incoming particle count vs leaving particle count, and what's the resulting beta ratio. If there are 100,000 20u particles going in, and if 20% of them get past the media due to leakage, then 100,000 x 20% = 20,000 leaked past. Add that to the 100,000 x 0.01% that got through the 99% efficient media (1,000) = 21,000 total particles going out (vs 100,000 going in) went past the filter. Beta ratio of that is equal to 79% @ 20u.
Just look at the incoming particle count vs leaving particle count, and what's the resulting beta ratio. If there are 100,000 20u particles going in, and if 20% of them get past the media due to leakage, then 100,000 x 20% = 20,000 leaked past. Add that to the 100,000 x 0.01% that got through the 99% efficient media (1,000) = 21,000 total particles going out (vs 100,000 going in) went past the filter. Beta ratio of that is equal to 79% @ 20u.
Last edited:
Doesn't quite work that way. An internally leaking filter will shift the whole efficiency vs particle size curve down, like shown in the example below. If the filter was 99% @ 20u and had a 15% internal leakage the resulting efficiency would be 84% @ 20u. The loss in efficiency decreases slightly as the particle size decreases because the media itself is less efficiency as the particle size decreases.
I think you may have misread my post because your graph is pretty much confirming the point I made.
You said in post 74: "Thus the problem with a Fram that leaks 5% to 15% of oil flow is that it is only 85% to 95% efficient above 50 microns".
You referenced "above 50u", but my example says if the filter is 99% @ 20u then with a 15% internal leak it's going to be 79% efficient down to 20u, particles well below the "above 50u" you referenced. That's what I was pointing out.
Re: my comment of "doesn't quite work that way" meaning it's not going to be like your claim of "only 85% to 95% efficient above 50 microns" with a 5% to 15% internal leak.
It's going to be 85% to 95% down to 20u if it was 99% @ 20u to start with. That's the point I'm clarifying. It's not going to be "only 85% to 95% efficient above 50 microns" as you claimed.
Here's another example showing how the whole efficiency curve is shifted down with a 15% internal leak on a non-leaking filter that's 99% @ 40u and 91% @ 20u without any internal leakage.
Last edited:
