A1/B1-A5/B5 "Low Vis" (vs) A3/B3-B4 LL01-MB 229.5

[GMorg]

I realize that Red Line oils are not rated A3 nor A5, but they do offer measures at cold temperatures that are not presented as limits.

5W40 5500cP at -30C (HTHS 4.6cP)
and
5W30 6000cP at -30C (HTHS 3.8cP).

At minus 30C the 5W30 is thicker than the 5W40. However, the calculator "shows" the 5W40 as about 40% thicker than 5W30.

 

[Audi Junkie]

Originally Posted By: GMorg
Audi Junkie, if the issue is A3 vs A5 cold viscosity, then why doesn't the graph above that includes GC (A3) and Edge 10W30 (A5) demonstrate the opposite of your position. Doesn't it show the A3 oil as thinner than that the A5 oil via the calculator?


Ding.

Someone is actually paying attention to the material.

Good jerb.

 

[Audi Junkie]

Originally Posted By: BritGerCarLuvr
Here we are, and I've made a lot of changes & I wish to thank many members of this board for their great advice. I do, however how one more issue for which I need some clarification.
I will keep my question simple.

Q) If your car called for either oil "type", but preferred low vis A5, knowing that the A3 are higher HT/HS and Vis, would you run them in your car, and why?

Is there an advantage to running high HT/HS high Vis in a car that runs just fine with the A5 stuff.(long life/ Better starts? Wear more/Less/)

Now running A5 in both Jags and the High HT/HS A3 in the BMW and the Benz.

 

[GMorg]

AJ, I'm not looking for a gold star. I am trying to figure out why you won't admit that one cannot know the cold viscosity without an actual measurement and simultaneously use "evidence" opposite to your position. I had assumed that you were arguing in good faith. If you actually understand the weakness in your position, then I have foolishly wasted my time.

Live and learn I suppose.

 

[Big O Dave]

This is an AWESOME thread! Great entertainment, with some information, too. Even Doug and Ed got involved.

Excellent - keep the fur flying!

 

[Audi Junkie]

You want me to admit Edge 5w-30 is thinner than TDT 5w-40?

Ok.

Looks like >10% deviation to me.

 

[d00df00d]

Quote:
I don't need "accuracy" to compare two oils.

This should have been the end of the discussion. It's futile to argue logic with someone whose only coherent point is that he (or she) doesn't actually care about the truth.

Can we move on, please?

 

[JHZR2]

Originally Posted By: GMorg
I realize that Red Line oils are not rated A3 nor A5, but they do offer measures at cold temperatures that are not presented as limits.

5W40 5500cP at -30C (HTHS 4.6cP)
and
5W30 6000cP at -30C (HTHS 3.8cP).

At minus 30C the 5W30 is thicker than the 5W40. However, the calculator "shows" the 5W40 as about 40% thicker than 5W30.


Excellent. The argument starts to fall apart.

Im really not trying to destroy AJs argument, rather make it robust enough that it has a leg to stand upon. Posting the same few plots doesnt do anybody any good, and then data like this comes along and completely blows the argument apart.

Just doing those error analyses that I proposed would help us to see the +/- range where we could fall and make anything else more palatable. Now we get data like this and instead of still giving consideration to the other, instead we have REAL data that just invalidates other claims unless something else real is proposed.

 

[jmac]

Originally Posted By: JHZR2




Just doing those error analyses that I proposed would help us to see the +/- range where we could fall and make anything else more palatable. Now we get data like this and instead of still giving consideration to the other, instead we have REAL data that just invalidates other claims unless something else real is proposed.


As well doing the sensitivity analysis would only show the sensitivity of the basic, linear curve equation. Linear here being used to describe a simple curve with no non-linear asymptotic deflection points in the region. Things that introduce non-linearities like pour-point depressants at unknown ratios would still not be accounted for.

 

[Audi Junkie]

Originally Posted By: d00df00d
Quote:
I don't need "accuracy" to compare two oils.

This should have been the end of the discussion. It's futile to argue logic with someone whose only coherent point is that he (or she) doesn't actually care about the truth.

Can we move on, please?


lol, you really don't get it.

If a boulder falls and crushes you, does it matter if it was 10 tonnes or 10 tonnes and six ounces?

Asserting that I have no logical basis and such is exactly what the Wiki links on Logical Fallasies is about.


Maybe I spent too much time in Zoltan's class...

http://hss.sas.upenn.edu/people/zoltan-domotor

or not enough....


http://en.wikipedia.org/wiki/Indeterminacy_problem


Indeterminacy problem
From Wikipedia, the free encyclopedia
Jump to: navigation, search
For other uses, see Indeterminacy (disambiguation).

The indeterminacy problem is posed as a kind of paradox in the study of the sociology and history of science. It is often used as an argument against the rational value of scientific thought.

The argument is as follows:

1. There are an infinite number of possible theories,
2. There can only be a finite amount of experimental evidence,
3. Therefore it is impossible to disambiguate between all viable theories.

While this is true in this abstract form, in practice, parsimony is used to limit the number of theories which are considered useful, and likewise, to limit the number of directions to take experimental research. Occam's Razor, seeking the simplest explanation, forms much of the basis of modern scientific thought; a theory which accounts best for all observed phenomena, does not predict disproven phenomena, and does not introduce unobserved phenomena is to be preferred.[1]

Moreover, data obtained by one experiment can be applied to more than one hypothesis, and proven hypotheses can be applied to more than one theory. For example, the theory of gravity has implications for innumerable other theories. There is not an infinite number of possible theories, since the available theories are limited by the ideas which have actually been imagined by scientists.

 

[Quattro Pete]

Originally Posted By: Audi Junkie

If a boulder falls and crushes you, does it matter if it was 10 tonnes or 10 tonnes and six ounces?

The example you have given here shows a high level of accuracy. Difference between 10 tons and 10 tons and six ounces is only about 0.002%.

What I and others here are saying is that these viscosity calculators may be quite a bit more inaccurate than this as you move away from 40C temperature.

For all you know, they may be off by as much as 100% or even 1000%. Nobody knows.

If a rock falls on you, does it matter if it weighs 6 ounces or 60 ounces? Sure it does.

 

[A_Harman]

Originally Posted By: Big O Dave
This is an AWESOME thread! Great entertainment, with some information, too. Even Doug and Ed got involved.

Excellent - keep the fur flying!

You must also enjoy watching train wrecks.

 

[Trajan]

Originally Posted By: A_Harman
Originally Posted By: Big O Dave
This is an AWESOME thread! Great entertainment, with some information, too. Even Doug and Ed got involved.

Excellent - keep the fur flying!

You must also enjoy watching train wrecks.

Especially when the engineer has an ego that keeps him/her from admitting they're on the wrong track. Even in the face of the oncoming train.

 

[Big O Dave]

Originally Posted By: A_Harman

You must also enjoy watching train wrecks.

As a veteran BITOGER, I find these "wrecks" part of the allure of this site.

 

[JHZR2]

Originally Posted By: jmac
Originally Posted By: JHZR2




Just doing those error analyses that I proposed would help us to see the +/- range where we could fall and make anything else more palatable. Now we get data like this and instead of still giving consideration to the other, instead we have REAL data that just invalidates other claims unless something else real is proposed.


As well doing the sensitivity analysis would only show the sensitivity of the basic, linear curve equation. Linear here being used to describe a simple curve with no non-linear asymptotic deflection points in the region. Things that introduce non-linearities like pour-point depressants at unknown ratios would still not be accounted for.


Excellent point. However it would be interesting to do the exercise with, say, redline where we put in the specs, make the curve, and see how it hits at the cold temperatures (using the cSt to cP relation I provided early), then see how things change with very minor variations at 40 and 100C. I agree the math is wrong, but what AJ is trying to do is find a decent enough approximation. I get and can appreciate that. The issue is that AJ has never worked out any indication that the basis of his calculations are nearly right or what kinds of errors do exist. Ive stated how, and even if the math is still slightly wrong, at least what kind of an error bar and deviation he could show to help indicate how things vary. Alas, all we get is wiki articles instead of math and attempting to put the whole story together so it is palatable.

AJ's claims may well be right, but without substantiation of the method, the claims are worthless speculation. Sad part is that it would be REALLY easy to substantiate them or at least show us an error bar with readily available data such as the RL info...

 

[azsynthetic]

Originally Posted By: JHZR2
The issue is that AJ has never worked out any indication that the basis of his calculations are nearly right or what kinds of errors do exist.


Well, I like to help AJ out since he apparently has no math back ground and does not understand that 2 points don't make a curve. So, here is a little help for AJ on "indication"

From Wikipedia, the free encyclopedia:

Indication can refer to:

1. A synonym for Sign.

2. Human interface Highlighting the single object pointed to as a cursor is moved, without any other user action such as clicking, is indication.

3. Indication (medicine). A valid reason to use a certain test, medication, procedure, or surgery.

I think AJ is using number 2......


Now, back to the topic of low visc, according to Royal Purple Product specs at -35C using ASTM test D-4684 we have:

5w20 = 22,200 cP
5w30 = 34,800 cP but
5w40 = 21,500 so it is indeed thinner than 5w20

at -30C we have:

10w30 = 18,200 but
10w40 = 14,900 which is much thinner than 10w30

 

[jmac]

Originally Posted By: JHZR2
Originally Posted By: jmac
Originally Posted By: JHZR2




Just doing those error analyses that I proposed would help us to see the +/- range where we could fall and make anything else more palatable. Now we get data like this and instead of still giving consideration to the other, instead we have REAL data that just invalidates other claims unless something else real is proposed.



As well doing the sensitivity analysis would only show the sensitivity of the basic, linear curve equation. Linear here being used to describe a simple curve with no non-linear asymptotic deflection points in the region. Things that introduce non-linearities like pour-point depressants at unknown ratios would still not be accounted for.


Excellent point. However it would be interesting to do the exercise with, say, redline where we put in the specs, make the curve, and see how it hits at the cold temperatures (using the cSt to cP relation I provided early), then see how things change with very minor variations at 40 and 100C. I agree the math is wrong, but what AJ is trying to do is find a decent enough approximation. I get and can appreciate that. The issue is that AJ has never worked out any indication that the basis of his calculations are nearly right or what kinds of errors do exist. Ive stated how, and even if the math is still slightly wrong, at least what kind of an error bar and deviation he could show to help indicate how things vary. Alas, all we get is wiki articles instead of math and attempting to put the whole story together so it is palatable.

AJ's claims may well be right, but without substantiation of the method, the claims are worthless speculation. Sad part is that it would be REALLY easy to substantiate them or at least show us an error bar with readily available data such as the RL info...


Absolutely, agree with all the points you've made now and before.

 

[Garak]

Originally Posted By: azsynthetic
Now, back to the topic of low visc, according to Royal Purple Product specs at -35C using ASTM test D-4684 we have:


I think you hit the nail on the head right there. It's probably a good generalization that at cold temperatures, 5w-30 is thinner than a 5w-40 or a 5w-50, or a 10w-30 thinner than a 10w-40. In the absence of the proper information, it's probably a safe assumption.

Your examples, however, show that it's not always the case. A 5w-30 (or any other oil) has to fall within certain specifications at certain temperatures. At extremes, different things can happen. That's further muddied by differences between synthetics and conventionals and other oil chemistry issues.

AJ is right in many points, but the methodology may be flawed. One particular 0w-30 isn't necessarily going to have better cold flow properties at every conceivable temperature than one particular 5w-30. I can certainly agree with that. However, to get the 0w or 5w and the -30, certain specifications must be met at the appropriate temperatures. Again, outside that, things can change.

I do have a problem with the viscosity calculator graph. It's a nice little tool, but as others have pointed out, it has its limitations. One cannot take two data points from a PDS and have a complete picture of the oil's performance at all conceivable temperatures the oil may face from the coldest winter ambient in the world right to the flash point of the oil.

The fact that a 5w-30 can conceivable be thinner than a 0w-30 at at least some range of temperatures demonstrates better than anything else the folly in placing too much stock in the graph.