Data Challenge!! UOA Trends 7yr / 100k

Joined
Oct 12, 2010
Messages
931
Location
GA
Hi folks,

I've been collecting UOAs on my 1999 4Runner with 3RZ engine since 2018. I now have data for 14 (!!) UOAs to sort through. There should be enough information here to help guide me on making decisions for future oil changes. My criteria here are to minimize ppm wear in Fe, Al, and Cu as I think that's the best proxy for "overall wear."

I drew some tentative conclusions already but want to see what you think about the data first. The variables I've looked at are:

Oil type (Rotella T6, M1 Turbo Diesel Truck, and Rotella Gas Truck)
Oil weight (30wt and 40wt)
Interval length
Miles on engine
Filter changed before interval or not (I change every other OCI)

Screenshot 2025-04-23 at 7.59.51 AM.webp


Here is a google sheet with ALL the DATA

Please feel free to treat the Iron = 14ppm as an outlier -- this OCI had hundreds of miles of desert offroad and is not representative of normal service for this vehicle.

Looking forward to see what you think! And thank you!
 
Hi folks,

I've been collecting UOAs on my 1999 4Runner with 3RZ engine since 2018. I now have data for 14 (!!) UOAs to sort through. There should be enough information here to help guide me on making decisions for future oil changes. My criteria here are to minimize ppm wear in Fe, Al, and Cu as I think that's the best proxy for "overall wear."

I drew some tentative conclusions already but want to see what you think about the data first. The variables I've looked at are:

Oil type (Rotella T6, M1 Turbo Diesel Truck, and Rotella Gas Truck)
Oil weight (30wt and 40wt)
Interval length
Miles on engine
Filter changed before interval or not (I change every other OCI)

View attachment 275082

Here is a google sheet with ALL the DATA

Please feel free to treat the Iron = 14ppm as an outlier -- this OCI had hundreds of miles of desert offroad and is not representative of normal service for this vehicle.

Looking forward to see what you think! And thank you!
Great effort, but the data clearly shows all of the oils have protected your engine equally well. Even with the outlier, it’s likely that your lab has a +/- 1ppm resolution, which means all you can really tell so far is… there’s not much to tell. As proof, you didn’t change the filter at the time of your 14ppm Fe reading, but yet your next sample was, to that point, your lowest iron reading yet. You’ve got a good engine, but IMO all your data confirms is that all three of those oils are sufficient to keep your engine parts separated and lubricated.

At this rate, 400k should be easy-peasy!
 
I agree with your points, but I want to see if I can differentiate "good" from "better" in this data, in the quest to hit 500k ;)
 
I think I may see the Cu slightly increasing but it's so small of a change. Good data and you've kept everything pretty consistent. These oils seem to be serving you well. If you want to make it interesting just to do something else try a 15w-50 IF you live in a warm climate.
 
I agree with your points, but I want to see if I can differentiate "good" from "better" in this data, in the quest to hit 500k ;)
You can't, the tool simply lacks the resolution. The error bars are wider than the difference you are trying to see. @SubieRubyRoo explained it quite well. Your equipment is healthy. Some of the information that would have been interesting to see would be viscosity and TBN retention.
 
I agree with your points, but I want to see if I can differentiate "good" from "better" in this data, in the quest to hit 500k ;)
Without any specific metric then the answer is no. What constitutes “good” in those spectrographic analyses?

For a fully-formulated motor oil of reasonable and sufficient HT/HS, wear isn’t going to be a determining factor. As already noted, they are all good from that standpoint which is to be expected. Such an analysis doesn’t make quality terminations for things like oxidation resistance or deposit formation or things like ring sticking, etc.
 
Last edited:
Without any specific metric then the answer is no. What constitutes “good” in those spectrographic analyses?

For metrics my thought was to minimize ppm/mi for the key components Fe, Al, and Cu
 
Hi folks,

I've been collecting UOAs on my 1999 4Runner with 3RZ engine since 2018. I now have data for 14 (!!) UOAs to sort through. There should be enough information here to help guide me on making decisions for future oil changes. My criteria here are to minimize ppm wear in Fe, Al, and Cu as I think that's the best proxy for "overall wear."

I drew some tentative conclusions already but want to see what you think about the data first. The variables I've looked at are:

Oil type (Rotella T6, M1 Turbo Diesel Truck, and Rotella Gas Truck)
Oil weight (30wt and 40wt)
Interval length
Miles on engine
Filter changed before interval or not (I change every other OCI)

View attachment 275082

Here is a google sheet with ALL the DATA

Please feel free to treat the Iron = 14ppm as an outlier -- this OCI had hundreds of miles of desert offroad and is not representative of normal service for this vehicle.

Looking forward to see what you think! And thank you!
Learned conclusion. The reports look great and you wasted your UOA money.
 
Sorry, just have to poke a little when I can. Engineering topics but this isn't work :cool:

I had the thought to compare the different variables (oil weight, interval length, etc) vs Total wear PPM / interval miles

- Looks like 40wt oils give lower wear
- Total wear PPM counts are independent of interval length -- indicates longer OCI should be beneficial
 
Sorry, just have to poke a little when I can. Engineering topics but this isn't work :cool:

I had the thought to compare the different variables (oil weight, interval length, etc) vs Total wear PPM / interval miles

- Looks like 40wt oils give lower wear
- Total wear PPM counts are independent of interval length -- indicates longer OCI should be beneficial
But it’s not wear and it’s not statistically significant. I think people are trying to tell you that and the wider lesson not really an oil selection method.

Pick one and run it consecutively and watch your general engine health using UOAs if you desire
 
But it’s not wear and it’s not statistically significant. I think people are trying to tell you that and the wider lesson not really an oil selection method.

Pick one and run it consecutively and watch your general engine health using UOAs if you desire

Pablo, can you help me understand what you mean by "it's not wear"? Are the ppm counts of the main wear metals not the best way to determine the amount of wear happening during an OCI?

Points taken about measurement error / error bars, and dataset only has n=14. Only so much solid statistics one can do with it.
 
I had the thought to compare the different variables (oil weight, interval length, etc) vs Total wear PPM / interval miles

- Looks like 40wt oils give lower wear
- Total wear PPM counts are independent of interval length -- indicates longer OCI should be beneficial
That’s the major problem here. Gross wear is one thing. Attributing it to the oil used isn’t valid from this type of analysis. Measuring wear differences from oil is a much more difficult test because it isolates the oil as a single variable.

You have multiple and equally (if not more) significant variables that affect the spectrographic analysis results. You cannot eliminate those variables. The oil is likely to be in the noise here unless it’s of such inadequate HT/HS that it’s causing major wear.

These $35 UOA are a whole lot less deterministic than is often imagined.
 
Pablo, can you help me understand what you mean by "it's not wear"? Are the ppm counts of the main wear metals not the best way to determine the amount of wear happening during an OCI?

Points taken about measurement error / error bars, and dataset only has n=14. Only so much solid statistics one can do with it.
It’s not all wear caused by lubricants in use. We are taking parts per million- and even some of those low amounts are metal sloughing from any surfaces the oil merely contacts for example.

Kschan says it even more clearly
 
Last edited:
Back
Top Bottom