Mobil 1 has highest film strength

[Shannow]

RazorsEdge,
you gave me the impetus to give google another crack at the secret test procedure that outshines the ASTM...

Originally Posted By: 540 Rat
Some folks complained that this “Lab” testing was basically worthless because the testing wasn’t performed in an actual running engine. The uninformed should be aware that “Lab” testing is the gold standard for countless companies in many, many industries. It shortens development time, and allows them to design and build a quality product in a timely manner. And of course time is money, in industry. In fact, many companies simply could not compete or even exist without lab testing.

Here’s one example. You simply cannot test a new jet aircraft in actual flight, to see at what point its wings will break off. So, you MUST perform static “Lab” testing on the ground, to determine that information.

It is similar with this oil testing, in that not all testing is practical in a running engine, even though in a perfect world that would be ideal. Try testing 44 different oils in an engine, with the "exact same" conditions every time, and doing all that in a timely manner. That simply would NOT be possible. So, that’s were oil “LAB” testing comes in. And wouldn't you rather select an oil that "Lab” tests well, rather than one that does not? Or even worse, have no real idea what oil to select, and simply have to go by your gut feeling, or have to go by advertising hype, or have to pick the prettiest bottle, or to have to simply guess?

I choose to use technical data for determining which oil to select for myself. So, I invested a lot of time and money to perform motor oil “Lab” torture testing, at a representative oil temperature, in order to generate that data. Because you simply cannot find this data anywhere else.

The tester's psi numbers are generated from a non-rotating test specimen simply being “gradually” pressed down onto a spinning ring, that's it. And that gradual application of the load, allows time for the local friction point to heat up to so that the heat activated zinc can become effective. No engine parts are designed that way. Testers and engines are NOT the same, and are NOT intended to be the same. Lab testing speeds up the evaluation process so that you don't have to wait 100,000 miles to see what happened.


So it's not 3 ball, four ball, two ball with electric currents...it's the Timken test machine...aka the 1 armed bandit.

LOLercoasters...

http://speedtalk.com/forum/viewtopic.php?f=1&t=30596

Garak, you'd like that thread.

 

[Garak]

Originally Posted By: Shannow
Assuming Rat's secret protocol is similarly constructed, anything from 85,000 to 115,000 is statistically the same, and 111,123 is nonsense.

If he's got the errors down to 0.001 as his use of the number suggests, the ASTM test that arises will make him a rich man.

Most rudimentary part of his test is measuring the area of the wear scar...he's already introduced 0.2% error if he measures the scar at room temperature versus 100C in the machine...there's the last 3 sig figures gone...no idea how accurate the scar measurements themselves are.

That, in a nutshell, demonstrates the problem. Now, if he's an engineer, I'd like to meet the lab instructor that would have allowed data like he posts. If I had submitted data without the error bar, or something that is logical horse manure based upon misuse of significant figures, the lab would have been a failure, and then I would have failed the class due to having incomplete lab work, and so forth.

Originally Posted By: Shannow
Garak, you'd like that thread.

That thread is hilarious. He does a drive by post there and bolts. Did he forget his password to that forum or were the questions Kevin Johnson posted a little too uncomfortable?

There's a certain math and science forum I visit once in a while. He should post there. There's a certain chemist there that would exhibit a lot less restraint in pointed out his testing flaws than we have been. Heck, he should post here himself and answer the questions.

 

[bbhero]

Maybe I'm in a better "math" understanding mood this am or what... But I can clearly see why rats testing results are impossible. To CLEARLY differentiate results down to 6 figure psi is just not very likely at all. Similar to computer model weather forecasting. Which I can read them without the champions on TV having to tell me what they mean or what the weather forecast will be. There are times in modeling forecast where the models just cannot get the 850mb 0°C line and 540 decameter line correct to within a mile tolerance. It's impossible. The snowstorm in my area just a couple of weeks ago it was really close in terms of distance. 30 miles made a big difference. No way model forecasting can get it quite that precise. Just like rats testing CANNOT either in terms of 6 figure psi results. Not happening.

 

[Garak]

Yep, you're correct. One can very easily "collect" the data to six significant figures. Whether it actually means something on its own is another matter altogether. Canadian gas pumps report fuel purchases to three decimal places, which is a millilitre of fuel. Now, given Canadian weights and measures standards, this is laughable, and I hope no one believes that Canadian gas pumps accurately measure fuel to the nearest millilitre. One has to be careful to do that in the lab. It's not happening at a gas pump.

 

[kschachn]

Originally Posted By: turtlevette
Everything else being equal, it certainly doesn't hurt to have a higher film strength.


Even if that might be true, where is it being measured?

 

[turtlevette]

Originally Posted By: kschachn
Originally Posted By: turtlevette
Everything else being equal, it certainly doesn't hurt to have a higher film strength.


Even if that might be true, where is it being measured?


You go off and think about it then give your best shot at an answer. You never told us what you do at the airport, grasshopper.

 

[kschachn]

Originally Posted By: turtlevette
You go off and think about it then give your best shot at an answer. You never told us what you do at the airport, grasshopper.


Actually what I do is none of your business, right? But I'll tell you what... If you describe how Rat's test is measuring film strength, I'll tell you what I do.

Oh and it's not at an airport.

 

[53' Stude]

LOL!

 

[Lex94]

Looks like you made the resident "finger wag" angry. aww.

 

[Shannow]

It's a valid point, what is ACTUALLY being measured ?

In order to have a wear scar to measure the area and calculate the "psi" to 6 significant figures, by definition there's a wear scar.

wear scar meaning that the part has worn, and his stationary surface has reached some conformance with the rotating ring.

Surely the data that you would want to be gathering is the point BEFORE the wear starts happening, when it's essentially point contact.

 

[BrocLuno]

Ummm, good point - best yet

 

[69GTX]

Originally Posted By: DrRoughneck
According to these measurements , the film strength of Mobil 1 5W30 advanced synthetic is 117,799 psi.

The study measured the film strength of 174 different oils and 5W30 Mobil 1 came on top.From what I understand film strength is measured by squeezing 2 ball bearings together with a film of oil in between. The force with which the balls are squeezed is gradually increased until current passes through, meaning there is metal-to-metal contact.
...


If the test has any validity then 540Rat needs to answer these types of questions:

1. In the 2013 testing 30 grade oils were the strongest. Yet this time around a 20 grade (QSUD 5w-20) heads the list at the #3 spot at 121K.

2. Why are Quaker State oils all of a sudden much higher up on the rankings, even well above the rankings of the SOPUS flagship PP or PU?

3. Why do 40 grade oils test out like pure [censored]? Those are the same oils carrying top level mfg certs like MB, Volvo, BMW, Audi, Dexos1, Porsche, etc. How many 5w-20 and 5w-30 grade oils meet those mfg approvals? Numerous boutique oils from Amsoil, RP, Joe Gibbs, Shaffer, and others seem to be at the bottom of the list. Rat would conclude they must be selling smoke.

4. If QSUD 5w-20 is the best oil tested, then it appears that the oil producers and engineers have failed in their jobs if a lesser brand is at the top of the list. Apparently, they don't know what they're doing and which oils are the best. Why is only a single 20 grade so highly tested? Next 20 grade on the list is Castrol Edge with Titanium at 99K (#31 ranking). It makes no sense that there are only two 20 grades in the top 40 get one of them leads the list at #3 ranking. Not consistent at all.

5. Mobil 1 doesn't have the highest film strength if their flag ship 0w-40 oil with numerous mfg approvals, tests way down the list at #89 (82K psi). This is probably their best overall protective oil, yet ranks miserably...and it the highest rated of all the 0w-40 grades.

6. Prolong Engine Treatment (and Extreme Concentrate) seem to improve the performance of most oils, some dramatically. Take a "decent" oil like Castrol Conventional GDX 5w-30 (95K) or PU 5w-30 (92K) and PET somehow transforms them to the 2 best rated oils ever at 130K-136K. If PET is that powerful, why aren't oil producers or boutique shops using it? Why can't those results be achieved in all the other oils? Could similar improvements been achieved using hair shampoo?

7. Since the PU SM synthetic, Pennzoil has gone backwards in the ratings. Yet there's QSUD going much higher on the list. How come the SOPUS designers didn't pick up on this? If they know about it why would they improve QS while lessening the PP/PU?

540Rat testing creates more questions than answers. I could give some validity to this testing if there was decent correlation throughout it...which there doesn't appear to be.

 

[turtlevette]

Originally Posted By: Shannow
It's a valid point, what is ACTUALLY being measured ?

In order to have a wear scar to measure the area and calculate the "psi" to 6 significant figures, by definition there's a wear scar.

wear scar meaning that the part has worn, and his stationary surface has reached some conformance with the rotating ring.

Surely the data that you would want to be gathering is the point BEFORE the wear starts happening, when it's essentially point contact.


So your side of the argument says the one arm bandit has no relevancy but the seq iv4 does. I could say a Nissan OHC sewing machine cam follower test has limited relevancy to my hi po pushrod small block Chevy. We discussed many times convenience and time of tests. Accelerated extreme pressure wear testing has some correlation to cam wear. Its a pretty decent test once you take into account the time and expense. Almost nothing. So we could weight the time and expense of testing vs relevance.

Accelerated bench tests are generally accepted in industry.

Your side says ep and scuff tests are not necessary for motor oil, but Daimler is working on one for motor oil.

 

[Shannow]

"My side"...???

I'm saying, and continue to say that the 4 ball and timken machines (and the Rat testing, as he's stated it's analogous to the timken machine) are not relevant in internal combustion engines...if they were, and they are cheap, they would be being used instead of expensive engine tests.

For convenience and time, I could rub the oils between my fingers...again, not much to do with an engine, but clearly, the most cost effective (i.e. decent) test out there.

 

[turtlevette]

Originally Posted By: Shannow
"My side"...???

I'm saying, and continue to say that the 4 ball and timken machines (and the Rat testing, as he's stated it's analogous to the timken machine) are not relevant in internal combustion engines...if they were, and they are cheap, they would be being used instead of expensive engine tests.

.


I don't think there are any absolutes. It has some relevance.

 

[BrocLuno]

Originally Posted By: 69GTX

6. Prolong Engine Treatment (and Extreme Concentrate) seem to improve the performance of most oils, some dramatically. Take a "decent" oil like Castrol Conventional GDX 5w-30 (95K) or PU 5w-30 (92K) and PET somehow transforms them to the 2 best rated oils ever at 130K-136K. If PET is that powerful, why aren't oil producers or boutique shops using it? Why can't those results be achieved in all the other oils? Could similar improvements been achieved using hair shampoo?



Ah yes, well these additives include chlorine based compounds which are not really appropriate for the inside of an internal combustion engine (and RAT says use with caution).

That's why when I parsed his old list, I removed all cases with additives and just listed the straight oils. It removes some of the confusion.

I'd never run either of those additives in a daily driven street car. Maybe in a Bonneville record attempt car that was going to be on the absolute ragged edge, and going to get the oil changed between runs or something ...

 

[Garak]

Originally Posted By: 69GTX
If the test has any validity then 540Rat needs to answer these types of questions:

Many of these hint at my concerns about repeatability. Considering he won't tell us about his methodology, I am completely entitled to assume he runs as few trials as possible, and have no evidence that the entire testing regime is nothing more than statistical noise. Next test might see the number one oil go to the bottom and vice versa, even if they're the same batch number he previously tested, for all I know.

 

[Shannow]

As to the ranking, without the error analysis suggested by Garak, the order of merit is meaningless...as per the 4 ball test, 85,000 to 115,000 "psi" are essentially the same number, so splitting hairs between them with the 6 figure accuracy is silly.

Have been reading more on the timken test that he uses, and per my previous comments, the timken OK load is the load on that is applied BEFORE the wear occurs...i.e. the lubricant is still lubricating.

Wear scar is used for calibrating. Mineral unadditised oil of around 68cst is used, and the wear scar of 1.7 to 2.9mm with a repeatability of 0.79mm is "calibrated...Still don't know how to get 6 sig figures from that range of variance.

 

[Jetronic]

Originally Posted By: Garak
And foisting six significant figures upon me makes me question his mathematical skills.


I think it's a result of averaging a number of test results? that's all I can think of. But if that'sthe case, he should post the lowest and highest test result aswell... you'd have an idea about repeatability at least.

 

[Garak]

You're being too kind. It's not acceptable to average a bunch of trials and round them to six significant figures, unless the error analysis backs up six significant figures, and you're not going to get six significant figures in this case. You do your averaging and then you round to appropriate significant figures and show your error bars, all as part of the error analysis. And when you do list the rounded data that has been subject to an error analysis, you make sure it makes sense.

157,923 psi +/- 10,000 psi, for example, is extraordinarily sloppy as a final, reported value, and will get you slapped down hard even in a first year physics lab, and I'm not talking about the use of non-SI units, either.