4WD
$50 site donor 2025
Platinum Ultra Ointment will fix it …
Last edited:
Lake Speed Jr. Reviews Several 0w-40's
- Thread starter fantastic
- Start date
Kinda messed up that folks here bust on his face....just me.
Exactly, We live in an imperfect world but some folks can't handle that.
A bit of an update on my UOA dataset. Based on this post and the discussion of LSJR's lastest YouTube video regarding M1 0W40 oils and statistics of what is meaningful or not in a UOA, here are some quick stats on my dataset as I probably have one of the more robust ones here in the BITOGosphere (21 to-date). I just calc'd stadard deviation of each variable and looked at anything above/below as being worth a note/look.
A few notes:
I just used my best judgement on what to exclude - the orange highlighted cells. These are when the engine was new and the intial drop of the various analytes hadn't been "flushed out" yet, roughly 20K miles but you can see some had dropped to close to average levels after the first change so I kept those in. I didn't include the first change in the viscosity analysis b/c it's a 30 grade best I can tell from the factory. I'm primarly looking at values higher than one standard deviation from the average as something important (red highlight/"bad") but for example cSt 100 being high is better than low so those get a green highlight/"good". Flashpoint being lower than average is "bad" so gets a red lightlight. TBN is related to OCI so left it out. Finally, the HPL oils pose a bit of an issue as they VOA I have done shows ~10ppm Si so I backed out 8 from each (except the 5/8 sample as this was a custom blend David was working on with me and the it only had 8ppm so just backed out 4..again, judgement call). I know some folks are seeing a bit of Al in their VOAs but mine had none - I *think* these are only in the no-VII VOAs? @OVERKILL showed 3 I believe in his Supercar 5W40 which is the same as my Euro but I just left that alone, they seem reasonable to me as-is but maybe a few ppm could come out. Won't change anything.
This is about what I would have flagged without the quick stat analysis and jives with the notes of what would be expected with those issues/repairs/etc. To the question about LSJR being concnered with a slight up-tick in wear metals switch oils/two changes to flush the previous oil out, yeah, 4 ppm variance without a lot of data isn't valuable at all/can't say much to compare two oils...it's just dataset noise. With that said, few will ever have 30 UOAs on a vehicle (30 being a good number you can find for a sample population to use for this kind of analysis) so at some point, you just look at what you have and make a call. When I look at geological exploration data (that's what I do for a living - geologist so plenty of experience with dataset and stats) we look at anything beyond 3 standard deviations (high or low) on certified reference material (CRM) QA/QC samples from the lab as being a possible issue. It does bring up the question does Blackstone provide their internal QA/QC data to verify the results? What type of CRMs do they use/calibrate with? The question I always ask/pose when I give training on dataset analysis and QA/QC...."when the lab tells you sample X is value Y of some analyte (use gold), how do you know it's Y and not ABC?" The answer is CRMs/SRMs tell you this.
Here, you'd have to be careful b/c on quick glance, you could say my engine doens't like the HPL oil....all of the red flags for wear metals are using HPL! But of course there are those notes indicating why it's likely just from use (track) and repairs or other causes.
A few notes:
I just used my best judgement on what to exclude - the orange highlighted cells. These are when the engine was new and the intial drop of the various analytes hadn't been "flushed out" yet, roughly 20K miles but you can see some had dropped to close to average levels after the first change so I kept those in. I didn't include the first change in the viscosity analysis b/c it's a 30 grade best I can tell from the factory. I'm primarly looking at values higher than one standard deviation from the average as something important (red highlight/"bad") but for example cSt 100 being high is better than low so those get a green highlight/"good". Flashpoint being lower than average is "bad" so gets a red lightlight. TBN is related to OCI so left it out. Finally, the HPL oils pose a bit of an issue as they VOA I have done shows ~10ppm Si so I backed out 8 from each (except the 5/8 sample as this was a custom blend David was working on with me and the it only had 8ppm so just backed out 4..again, judgement call). I know some folks are seeing a bit of Al in their VOAs but mine had none - I *think* these are only in the no-VII VOAs? @OVERKILL showed 3 I believe in his Supercar 5W40 which is the same as my Euro but I just left that alone, they seem reasonable to me as-is but maybe a few ppm could come out. Won't change anything.
This is about what I would have flagged without the quick stat analysis and jives with the notes of what would be expected with those issues/repairs/etc. To the question about LSJR being concnered with a slight up-tick in wear metals switch oils/two changes to flush the previous oil out, yeah, 4 ppm variance without a lot of data isn't valuable at all/can't say much to compare two oils...it's just dataset noise. With that said, few will ever have 30 UOAs on a vehicle (30 being a good number you can find for a sample population to use for this kind of analysis) so at some point, you just look at what you have and make a call. When I look at geological exploration data (that's what I do for a living - geologist so plenty of experience with dataset and stats) we look at anything beyond 3 standard deviations (high or low) on certified reference material (CRM) QA/QC samples from the lab as being a possible issue. It does bring up the question does Blackstone provide their internal QA/QC data to verify the results? What type of CRMs do they use/calibrate with? The question I always ask/pose when I give training on dataset analysis and QA/QC...."when the lab tells you sample X is value Y of some analyte (use gold), how do you know it's Y and not ABC?" The answer is CRMs/SRMs tell you this.
Here, you'd have to be careful b/c on quick glance, you could say my engine doens't like the HPL oil....all of the red flags for wear metals are using HPL! But of course there are those notes indicating why it's likely just from use (track) and repairs or other causes.
If supercar 0W-40 became API SP, they will all be gone.
Only if you're outside of europe duh!
I wasn't busting on his face. I sincerely hope that he has had that thing looked over by a competent doctor. I know one guy that has had much of his mouth and cheek cut away due to cancer and another coworker that is dead from throat cancer. Both of them chewed tobacco,
It's come up multiple times in multiple LSJR posts, I mean, he has a thing on his lip, I'm quite sure he knows whether this is something that is important to remove or not.
Why?
Must have missed them.
The Supercar 1SN isn't compatible with SP, and it seems to me, the base oil mix used for SN needs the stronger addpack from SN to offset the polarity of the added esters.
So if supercar becomes SP, it won't be Ester based like what happened with the others ( M1 FS, M1 ESP x3).
dnewton3
Staff member
I have always appreciated the efforts you put into logging your lube experiences. Much more dutiful than the typical person would achieve.
The issue I have with your data (not that there's anything "wrong", but would be misleading to others) is that for any one variable (lube), the max you have is 4 samples each (LM and the HPL). While you have slightly over 20 samples, you don't have anywhere enough of any one lube. So you do have data which could be compared/contrasted to macro data, but it's not credible for micro analysis. The stdev you calculate is not going to be relevant to any one lube, but the mass of them.
If someone wants to know how a single lube performs, in their application, they have to run 30 samples (minimum) of that one lube. Using multiple lubes induces multiple variables; that precludes drawing conclusions on any one variable. This is why most folks just abandon the attempt; it's expensive and time consuming.
Again - not picking on you, because I understand the dedication it takes to put this kind of effort forth. I think we're in agreement, if I've understood what you intend to convey.
Another limitation that affects us all is how data is reported to us. We see UOA data in whole numbers, but then we have to "average" values and end up rounding out the tenths (or hundredths). Any proper system of analysis requires at least a magnitude of measurement finer than the reported value. None of us have the ability affect changes in this regard, so we just have to acknowledge the problems inherent in the processes we use.
I look at my UOA data in this way - for the wear metal results, this is measuring how my engine is wearing, not oil to oil comparison or how well one oil is performing vs. another (which is what is preached by yourself and others w/r to UOA use here ad-nauseaum). BS reports universal averages for the analytes for the engine, not the oil brand/type in that engine. My dataset is working for wear metal/analyte trends and seems to follow expected results well even though that isn't always valid to conclude on lower-population datasets b/c coincidence does happen. The viscosity retention is primarily related (as long as I don't have egregious fuel dilution) to the oil not the engine and so for that variable, these data are looking at oil to oil and yes, ideally you'd have many more data points for each but this is about as good as it's going to get in the context of a vehicle's life and my dataset and basic analysis are useful for me (and likely others here), even though not perfect due to lack of data - I have 7 different oils if I am counting correctly in my table. You often accept less-than-ideal datasets and analyses in real-life situations like this to avoid analysis paralysis and understand that if you actually did do 30 of each oil, the averages and conclusions may look different to your points. Maybe Molygen is better than HPL at viscosity retention, we just don't know right? So my comment overall is...why bother then with UOA beyond coolant in your oil? For all the folks that post amazing long-run UOAs...they are actually meaningless then?
Last edited:
Agree. Remember, he has to earn a living. Ask this…who pays him?
Well here, YouTube!
I doubt he's being paid by anyone. He has an oil analysis business and is advertising it.
In a perfect world the growth on his lip should not matter but he presents his face to the public in every video. It is like watching or talking to someone with a large hairy mole on their face or a lazy eye, it will attract your gaze regardless. It does not help that it is a natural inclination of all humans to look at someones lips when they talk.
Of course, but here, it's a technical forum and this guy is presenting some technical data, and folks comment about his lip, seriously? Kind of goes with the whole thing you mom used to tell you...."if you don't have anything nice to say don't say it".
Similar threads
