UOA "Insolubles %" vs ISO Particle Count

ZeeOSix

$100 site donor 2022
Joined
Jul 22, 2010
Messages
40,069
Location
PNW
Many times someone will say ... "Hey, the oil filter must have been doing a good job because the UOA "Insolubles %" was low". But it's also been mentioned that the "Insolubles %" is an insensitive and bad parameter to measure the oil cleanliness to reflect oil filtering effectiveness.

So I went and found 4 UOAs on BITOG from Blackstone that included the ISO Particle Count data. The data was compared and graphed - see attachment below. Note that the "Insolubles %" doesn't really correlate well with the ISO Particle Count data. I know this is only 4 data points, but I'm thinking if you did this with many more UOAs the same non-correlation would hold true. So in order to really judge how well an oil filter functions based on a UOA, the ISO PC data is more telling than the "Insolubles %".

Attachment shows up smaller than it was, so the board must resize attachments. Tried getting a larger one to show up, but no-go.

Oil Filter Particle Count Comparision (Filter A thru D).JPG
 
Originally Posted by CharlieBauer
Can't conclude anything with 4 data points.


Yeah, I basically said that. But I'd bet with more data points it will still show that the "Insolubles %" is a useless measurement and has no good correlation to actual oil cleanliness ... that's the message here.

Sometimes it only takes but a few data points to see a trend. Since these tests were all done by Blackstone in the same manner, it gives the data a little more confidence. I'll add data as it shows up and build on the trend.
 
Here's another data point. It's the cleanest PC of all of them ... Call it "Filter E".

Blackstone UOA Data
ISO 16/15/12
Insolubles % = 0.2

So Filter E showed the cleanest PC by far, yet the insolubles was the same (0.2%) as the filter that had the worst PC (Filter A). No "PC to Insolubles" correlation at all.


Filter E (Insolubles at 0.2).jpg
 
Here's a Blackstone sample I took of some transmission fluid. Although the insoluble % was zero, the fluid particle count told another story and recommended I change it.

J24133.jpg
 
Originally Posted by A310
Here's a Blackstone sample I took of some transmission fluid. Although the insoluble % was zero, the fluid particle count told another story and recommended I change it.


That's a good particle count considering it's a 50 to 100 micron transmission filter. Very similar results to a Fram Ultra.
27.gif
 
Originally Posted by CharlieBauer
Originally Posted by A310
Here's a Blackstone sample I took of some transmission fluid. Although the insoluble % was zero, the fluid particle count told another story and recommended I change it.


That's a good particle count considering it's a 50 to 100 micron transmission filter. Very similar results to a Fram Ultra.
27.gif



I wish this were true, but truth be told I was running the fluid through an Amsoil Dual remote full flow/bypass filter and a MotorGuard M100 in parallel with the Amsoil setup. I'm going to have to start over on this little experiment as said transmission is be transplanted behind my Detroit 4-53T. I've drained the trans of all fluids, so will start with a clean system and go from there.

IMG_2711.JPG


IMG_2712.JPG


IMG_5516.JPG


IMG_5526.JPG
 
Originally Posted by CharlieBauer
Originally Posted by A310
Here's a Blackstone sample I took of some transmission fluid. Although the insoluble % was zero, the fluid particle count told another story and recommended I change it.

That's a good particle count considering it's a 50 to 100 micron transmission filter. Very similar results to a Fram Ultra.
27.gif



Don't know which Ultra PC you're looking at CB, but here's a Fram Ultra PC ... yes, one data point. Insolubles was 0.1% on this UOA.

Much better than the PC of the transmission filter that A380 posted - 6 vs 91 particles at 21 microns.

Fram Ultra PC.JPG
 
Originally Posted by CR94
So how can high insolubles not show up in counts? Outside the range of sizes counted?


Don't know exactly, but sure seems like there's a big disconnect between ISO PC data and the insolubles numbers. It's been mentioned many times that the insolubles number is very insensitive measurement and can't really be a good indication of oil cleanliness. The data I've seen pretty much shows that so far.

I'd sure think the standardized ISO Cleanliness Test (ISO 4406) would be more accurate data.
 
I have an oil sample from my V10 with 25,000 miles on it I have send off yet. I'm thinking of getting a particle count on this sample to check against % insolubles and see how they compare. The last sample I had done said the insolubles were .2% after 20,000 miles.
 
Originally Posted by ZeeOSix
Originally Posted by CR94
So how can high insolubles not show up in counts? Outside the range of sizes counted?


Don't know exactly, but sure seems like there's a big disconnect between ISO PC data and the insolubles numbers. I'd sure think the standardized ISO Cleanliness Test (ISO 4406) would be more accurate data.


Insolubles are solids and liquid. Also subject to measurement error. I believe Blackstone "estimate" it.

Particle counts are also subject to measurement error. The best way to measure is with in line particle counters.
 
I'd sure think the well known in industry ISO 4408 test is more accurate than an "estimated" insolubles test.
 
Originally Posted by ZeeOSix
Originally Posted by CharlieBauer
Originally Posted by A310
Here's a Blackstone sample I took of some transmission fluid. Although the insoluble % was zero, the fluid particle count told another story and recommended I change it.

That's a good particle count considering it's a 50 to 100 micron transmission filter. Very similar results to a Fram Ultra.
27.gif



Don't know which Ultra PC you're looking at CB, but here's a Fram Ultra PC ... yes, one data point. Insolubles was 0.1% on this UOA.

Much better than the PC of the transmission filter that A380 posted - 6 vs 91 particles at 21 microns.


What mileage was on the Ultra that you picked for the comparison?
 
Originally Posted by ZeeOSix
Originally Posted by CharlieBauer
What mileage was on the Ultra that you picked for the comparison?


Looks like 10K miles ... he used it for 2x 5K mile OCIs.

https://www.bobistheoilguy.com/forums/ubbthreads.php/topics/4753215/Toyota_2GR-FE_TGMO_with_Partic


An ISO code of 23/17/12 is a little weird. 23 is a very poor result, especially after 10k miles, 17 is slightly worse than average and 12 is slightly better than average.

It's a pity he changed the filter at 10k miles because it was likely on the cusp of achieving the kind of superior real world results that the 99% media is purchased for. I believe the higher capacity of the Ultra means it needs more caking and therefore more miles before it achieves superior results.

For example, there was a recent toughguard that achieved an ISO code of 21/14/10 after just 5k miles:

https://bobistheoilguy.com/forums/ubbthreads.php/topics/4107645/Honda_NC700X_Red_Line_10W30
 
Compare the ISO codes for the new oil out of the bottle, and the oil that had been filtered by the TG for 5000 miles--it was cleaner at 5K. That is some impressive filtering going on. (21/14/10 vs. 22/20/16 new). Those sets of numbers don't look like blackstone extrapolations either--they have different patterns.

Originally Posted by CharlieBauer

For example, there was a recent toughguard that achieved an ISO code of 21/14/10 after just 5k miles:

https://bobistheoilguy.com/forums/ubbthreads.php/topics/4107645/Honda_NC700X_Red_Line_10W30
 
Last edited:
Originally Posted by CharlieBauer
Originally Posted by ZeeOSix
Originally Posted by CharlieBauer
What mileage was on the Ultra that you picked for the comparison?

Looks like 10K miles ... he used it for 2x 5K mile OCIs.

https://www.bobistheoilguy.com/forums/ubbthreads.php/topics/4753215/Toyota_2GR-FE_TGMO_with_Partic

An ISO code of 23/17/12 is a little weird. 23 is a very poor result, especially after 10k miles, 17 is slightly worse than average and 12 is slightly better than average.

It's a pity he changed the filter at 10k miles because it was likely on the cusp of achieving the kind of superior real world results that the 99% media is purchased for. I believe the higher capacity of the Ultra means it needs more caking and therefore more miles before it achieves superior results.


If you look at the PC curves in the data I plotted in my first post, all the curves have the same basic shape, so an ISO code of 23/17/12 doesn't really seem weird to me. The Ultra that was used for 10K miles is Filter D in my original post.

Well, there's a whole separate wrinkle and discussion that could go on about if an oil filter actually becomes more efficient with use. I'm sure you've probably seen this in a few threads. If you know the way the ISO 4548-12 efficiency is calculated (per the procedure), the filter in the example below would have an ISO efficiency of 75% at 20 microns - the average of the beginning and end measured efficiencies: (90%+60%)/2.

So a filter with a high ISO efficiency of 99% at 20 microns means it doesn't lose hardly any efficiency with use, so I don't think it's really going to get any better than 99% at 20 microns unless it's about to clog and the delta-p increases to where the efficiency spikes up near the end of the graph in the example below. Or maybe the Ultra was starting to clog up some and slough off some debris due to the delta-p getting higher even though it was only used for 10K miles? Just because a filter is "rated" for 20K doesn't mean it will be ideal to run it 20K miles depending on how dirty the engine is.

[Linked Image]


Originally Posted by CharlieBauer
For example, there was a recent toughguard that achieved an ISO code of 21/14/10 after just 5k miles:

https://bobistheoilguy.com/forums/ubbthreads.php/topics/4107645/Honda_NC700X_Red_Line_10W30


Yep, I've seen that one too. Didn't use it because there was no insolubles data given. The TG is rated at 99% @ 20 microns also ... and up to 10K miles of use.
 
Originally Posted by ZeeOSix
If you look at the PC curves in the data I plotted in my first post, all the curves have the same basic shape, so an ISO code of 23/17/12 doesn't really seem weird to me.


All ISO codes follow the same general pattern so the curves will look similar, especially if you plot them logarithmic-ally thus altering the scale!
 
Originally Posted by CharlieBauer
Originally Posted by ZeeOSix
If you look at the PC curves in the data I plotted in my first post, all the curves have the same basic shape, so an ISO code of 23/17/12 doesn't really seem weird to me.

All ISO codes follow the same general pattern so the curves will look similar, especially if you plot them logarithmic-ally thus altering the scale!


ISO cleanliness particle count data is pretty much always plotted on a logarithmic scale ... it's easier to compare as shown in my graphs in the first post. Plus the y-axis isn't gigantic in size, and it gives very good graph resolution near zero when logarithmic. It would be pretty unreadable below 2000 if it wasn't logarithmic - the data points for 14 and 21 microns will not show up readable on the graph.. See the same data plotted below without a logarithmic y-axis.

The ISO 23/17/12 is Filter D, which is that Ultra discussed a few posts above. Looking at the family of curves, the Filter D curve isn't out of bed with the rest of the filters ... it's just showing a cleaner level of oil across the plotted particle range - mostly in the 14 and 21 microns particle counts, so 23/17/12 doesn't really seem weird to me.


Non-Log Y-Axis.JPG
 
Back
Top