Determining the REAL V.I. of a motor oil

[CATERHAM]

The October issue of LUBES'N'GREASES had an excellent article by Dr. Jack Zakarian, titled
"V.I. Too Resistant to Change" that was captioned and discussed in the following post:

http://www.bobistheoilguy.com/forums/ubbthreads.php/topics/3148253/1

The gist of the piece was that Jack Zakarian was quite critical of the method detailed in ASTM D2270 by which the V.I. of an oil is calculated.
He summarized as follows:
"ASTM V.I. does not accurately represent the "real V.I." of oil .... the rating method suffers from a number of inconsistencies that penalize lower viscosity oils compared to their higher viscosity counterparts. In addition, V.I. has no fundamental relationship to the true viscosity-temperature behavior of an oil, even though it is widely presumed to have one. At best, V.I. ia a very rough guide to viscosity-temperature behavior.
By understanding the deficiencies in the rating method, users will be better able to interpret and understand the actual V.I. numbers."

So I got to thinking, how can we avoid the V.I. penalty that applies to lower viscosity oils and determine the "REAL V.I." regardless of how thick or thin an oil actually is?
One method that does avoid this problem occurs when an oil's viscosity that's based on the same KV40 and KV100 spec's used to calculate it's ASTM V.I. are plotted on a graph commonly used in viscosity calculators. When you graph oils with different viscosities you are comparing the oil's "REAL V.I.".
But what we need is a "REAL V.I." number to replace the ASTM V.I. number. What better than using a ratio of KV40/KV100? We could also use the reciprocal but I think whole numbers are more insightful than fractional numbers.

For example M1 0W-40 (ASTM V.I. 185) has a KV40/KV100 ratio (75cSt/13.5cSt) of 5.56, meaning it's kinematic viscosity increases 5.56 times as the temperature drops from 100C to 40C.
M1 5W-30 (V.I. 172) has a ratio (61.7/11.0) of 5.61.
M1 0W-20 (V.I. 173) has a ratio (45.8/8.7) of 5.26.
Although the lighter M1 5W-30 has a significantly lower ASTM V.I. it's viscosity change with temperature ratio is only slightly higher and the even lighter 0W-20 actually has a lower ratio or a higher "REAL V.I." to M1 0W-40. That is an example of the V.I. penalty that applies to lower viscosity oils that Dr. Zakarian was referring to.

Just for fun, I've ranked a bunch of oils based on their KV40/KV100 ratios representing their "REAL V.I.'s" from the lowest to the highest:

Joe Gibbs XPO 0W-5, race oil, PDS spec's - ASTM V.I. 170, 11.5cSt/3.3cSt = 3.48
RL 0W-2, race oil, PDS spec's - V.I. 136, 11/3 = 3.67
Marvel Mystery Oil, PDS & VOA - V.I. 96, 10/2.64 = 3.79
Sustina SN 0W-20, PDS - V.I. 229, 32.69/7.94 =4.12
Mazda Moly SN 0W-20, VOA - V.I. 221, 35.85/8.37 =4.28
TGMO SN 0W-20, VOA - V.I. 216, 37.4/8.5 = 4.38
Motul 300V Sprint 0W-15, VOA - V.I. 155, 23.3/5.1 = 4.57
Honda (Idemitsu) SM 0W-20, PDS - V.I. 197, 40/8.5 =4.66
TGMO SN 0W-20/M1 SN 0W-20, 60/40 0W-30 blend, PDS - V.I. 203, 48.9/10.2 =4.79
Sustina SN 5W-30, VOA - V.I. 194, 51.5/10.3 =4.99
PP SN 5W-20, PDS - V.I. 169, 45.34/8.56 =5.30
Havoline Syn SN 5W-30, PQIA - V.I. 177, 53.5/10 =5.35
PP SN 0W-20, PDS - V.I. 164, 46.8/8.6 =5.44
PP SN 5W-30, PQIA - V.I. 171, 55.1/10.0 =5.51
PU Euro SN 0W-40, PDS V.I. 186 75.3/13.6 =5.53
Havoline Syn SN 5W-20, PQIA - V.I. 155, 47.5/8.44 =5.66
PYB SN 5W-20, PQIA - V.I. 155, 49.4/8.7 =5.68
Motul 300V 0W-40, VOA - V.I. 174, =5.86
GC 0W-30, PDS - V.I. 167, 72/12.21 =5.93
Castrol 0W-40, VOA - V.I. 165, 79.19/12.95 =6.11
PU SN 5W-40, PDS - V.I. 166, 80.7/13.2 =6.11
Shell T6 5W-40, PDS - V.I. 170, 87/14.2 =6.13
M1 SN 5W-50, PDS - V.I. 180, 108/17.5 =6.17
RL SN 5W-50, PDS - V.I. 186, 130/21 =6.19
Motorcraft 5W-50, PDS - V.I. 181, 108/17.5 =6.48
RL SN 10W-60, PDS - V.I. 187, 170/26 =6.54
PYB SN 10W-30, PDS - V.I. 135, 69.7/10.53 =6.62
Rotella T3 10W-30, PDS - V.I. 144, 80/12 =6.67
Mobil Delvac 1 5W-40, PDS - V.I. 151, 102/14.8 =6.89
Castrol TWS 10W-60, PDS - V.I. 174, 168/24 =7.00
Rotella 15W-40, PDS - V.I. 135, 120/15.5 =7.74
Widman Dino 20W-50, V.I. 121, 175/18.8 =9.30
Mobil Spectra Syn Elite 150 (base oil), PDS - V.I. 206, 1705/156 =10.93

If your favourite oil isn't in the above list it's easy enough to calculated it's KV40/KV100 ratio to determine it's REAL V.I. and see where it would slot-in in the above.

Light oils with low ASTM V.I.'s can have remarkably high real V.I.s while even high V.I. heavy oils don't have high real V.I.'s.
It is a basic characteristic of oil that the heavier it is the more it's viscosity will change with temperature.

Since 40 grade and heavier oils can't have high real V.I.'s it should give one pause to question whether they really need to run a 40 grade oil in the first place, and if they do to consider choosing a high ASTM V.I. oil such as M1 0W-40 which at least minimizes the fact that heavy oils are disproportionately heavy on start-up for the viscosity gains that are made at higher temp's.

 

[OVERKILL]

Question here:

This quote:

Quote:
the rating method suffers from a number of inconsistencies that penalize lower viscosity oils compared to their higher viscosity counterparts. In addition, V.I. has no fundamental relationship to the true viscosity-temperature behavior of an oil, even though it is widely presumed to have one.


Seems incredibly similar to what Shannow has been saying on here for quite some time and catching a great deal of flack for doing so

And while I find your ratio calculation quite neat, it doesn't change what he's said here, because it still uses the same numbers to calculate it.

Ultimately in order to get some form of meaningful VI we'd need some additional data points like 0C, -15C....etc

Maybe something (and I believe I recall you mentioning this in the past) based on the same test used for HTHS, but at a number of different temperature points could be employed?

 

[CATERHAM]

That's a separate argument, when one should rely more on MRV and CCS spec's vs V.I. at sub freezing temp's.
Where I believe there is no argument applies projecting down at least down to the freezing point and it is only V.I. that we can rely upon for that.

But my point is not about that, but rather on how ASTM V.I. is measured and the penalty that's applied to low viscosity oils.
The example I gave was that M1 0W-40 with it's 185 V.I. is reality no higher than a 160 V.I. 5W-20. There are no high V.I. 40 grade and heavier oils.

 

[mongo161]

How could they miss PYB 5w30 SN and PU 5w30 SN?

 

[jamesyarbrough]

The Caterham blend looks really good. I wonder how that would work out with M1 0w20 instead

 

[TrevorS]

Excellent insight CATERHAM.

This is all about how someone did not think about how the calculation of a metric would affect its interpretation.

As you say in response to OVERKILL, this is a different issue to not having more datapoints at temps that would be more meaningful to us.

Funnily enough, prior to you posting, I was graphing 0w40 vs 0w20 and thinking along the same lines as you about interpretation.

 

[OVERKILL]

Originally Posted By: CATERHAM
That's a separate argument, when one should rely more on MRV and CCS spec's vs V.I. at sub freezing temp's.
Where I believe there is no argument applies projecting down at least down to the freezing point and it is only V.I. that we can rely upon for that.


But that's only because that's the only data we have, correct? And that's what I believe the author if that piece is pointing out.

Quote:
But my point is not about that, but rather on how ASTM V.I. is measured and the penalty that's applied to low viscosity oils.
The example I gave was that M1 0W-40 with it's 185 V.I. is reality no higher than a 160 V.I. 5W-20. There are no high V.I. 40 grade and heavier oils.


VI defined as using the current definition or your new definition calculated using the same values. Appears to me that we are just looking at the same thing but painted a different colour

 

[TrevorS]

No it's not the same thing.

How a metric is calculated affects its interpretation. That's what this alternative method is about.

 

[MolaKule]

And I have stated in many posts, the VI doesn't mean a whole lot, since if the two viscosity data points are real close, verses them being far apart, affects (biases) the calculated VI.

You can use Widman's calculator to prove this to yourself.

 

[SlipperyPete]

Originally Posted By: CATERHAM
In addition, V.I. has no fundamental relationship to the true viscosity-temperature behavior of an oil, even though it is widely presumed to have one. At best, V.I. ia a very rough guide to viscosity-temperature behavior.


Why be concerned with VI if the above is true?

 

[MolaKule]

Originally Posted By: SlipperyPete
Originally Posted By: CATERHAM
In addition, V.I. has no fundamental relationship to the true viscosity-temperature behavior of an oil, even though it is widely presumed to have one. At best, V.I. ia a very rough guide to viscosity-temperature behavior.


Why be concerned with VI if the above is true?




Quote:
At best, V.I. ia a very rough guide to viscosity-temperature behavior.



We like to try and and attempt to quantify as many characteristics as possible.

 

[TFB1]

Originally Posted By: SlipperyPete
Originally Posted By: CATERHAM
In addition, V.I. has no fundamental relationship to the true viscosity-temperature behavior of an oil, even though it is widely presumed to have one. At best, V.I. ia a very rough guide to viscosity-temperature behavior.


Why be concerned with VI if the above is true?



Living in a area that seldom see lows of below 20*F I don't, better things to worry about...

 

[lonhahn]

Hello everyone. If I understand what Caterham is saying (and has said in the past), it looks like the PP 5-20 is actually lighter than the 0-20. If so, could I use all the PP 5-20 that I got on sale recently for my 2008 CR-V in my 2012 4 cyl. Accord (which calls for 0-20) as well? It is out of warranty, 56,000 miles, and it rarely gets below freezing here in Northern California.

 

[lonhahn]

I need to update my signature, I no longer have the 98 accord.

 

[OVERKILL]

Originally Posted By: TrevorS
No it's not the same thing.

How a metric is calculated affects its interpretation. That's what this alternative method is about.


Right, but its still a metric calculated off of only two data points.

 

[Shannow]

All the more reason to drop VI altogether, as it was originally a tool for comparing the "quality" of basestocks, in an era in which artificial Viscosity improvers were unheard of.

Originally Posted By: Shannow
another of those old arbitrary definitions.

Best oil of the day (Pennsylvania crude) was given a "100", and others compared to it. (like octane - 2,2, 4 trimethyl pentane is "100", while n-heptane is "0", while "0" still has knock resistance)

Calculation (as per The Lubrication Engineer's Manual 1971, 1981 reprint) is in SUS.

Take two reference oils, which has the same viscosity at 210F as the oil to be tested. One reference oil has a VI of "0", the other "100".

Measure the viscosity of the two reference oil at 100F, and the test oil at 100F.

Gives you 3 viscosities.
U = viscosity at 100F of the new oil
L = viscosity of the VI 0 oil at 100F
H = viscosity of the VI100 oil at 100F

VI = (L-U)/(H-U) expressed as a percentage.

Like most older type calcs, it was expecting the test oils to be in the range of their poor and "best" oils...


http://www.bobistheoilguy.com/forums/ubb...lat/fpart/2/q/1

100 was an arbitrary number, not an inherent property, given to their best oils, the zero was also arbitrarily placed, against an oil... all oils of the same VI have the same slope on the log chart which certainly makes sense when comparing naturally occuring oils for "quality" as defined by VI.

e.g. for the two reference oils of 18cst at 210F, the 100VI oil has a 100F viscosity of 213cst, the 0VI oil 456.3...incidentally, the 50VI was 334.6 (half way between the zero and 100 grade oils), and a VI 140 turned 115.2

So the zero certainly DOES have a ratio of viscosities,
0VI - 25.2
50VI - 18.6
100VI - 11.8
140VI - 6.4

Which makes minimal sense...and their reciprocals

0.040
0.054
0.085
0.156

make even less.

Thus I say ditch it altogether.

40C and 100C are meaningless data points for oil comparison...well the 40C at least.

Better would be something that has a cold temperature pumping viscosity limit, and HTHS...

e.g. my nomenclature...
-40W-3.6

An oil capable of being used at -40C/F, and an HTHS of 3.6, and a relatively minor education campaign, people might be able to work out stuff for themselves.

The lighter grades would be capable of demonstrating their cold weather performance by virtue of the fact that they were minus some ridiculous number at the cold pumping limit viscosity

 

[TrevorS]

Originally Posted By: OVERKILL
Originally Posted By: TrevorS
No it's not the same thing.

How a metric is calculated affects its interpretation. That's what this alternative method is about.


Right, but its still a metric calculated off of only two data points.


Many are eg EPS in finance which Warren Buffet uses while fund managers perform all sorts of analysis while 80% of them underperform the market.

The point is that calculating it the way CATERHAM did gives a different way to compare oils which for the purposes if understanding VI is arguably better than the old way.

This does not mean VI is important or the underlying data points are the best. Just that given we have that data, if were going to interpret it with a metric called VI, this method is potentially more meaningful.

 

[TrevorS]

Originally Posted By: Shannow
An oil capable of being used at -40C/F, and an HTHS of 3.6, and a relatively minor education campaign, people might be able to work out stuff for themselves.


Not in the land where people are sold to with "Good" "Better" "Best" or where people ask Walmart clerks what oil they should buy.

 

[turtlevette]

Mobil Spectra Syn Elite 150 (base oil), PDS - V.I. 206, 1705/156 =10.93

what is this? Motor honey?

The first 2 racing oils on the list have a 40C viscosity equal to the operating viscosity 100C of a 30 weight. That's why you can start "cold" and go straight to 6000rpm with no abnormal wear. Like we do in drag racing.

 

[Cujet]

Finally, some worthwhile data. Now if we can chart the oil over a much wider, but practical range of temperatures, we can develop a viscosity curve.

That viscosity curve can then be represented by tensor-based partial differential equations (just a joke) . More seriously, we could develop a method or series of numbers or characters that denote the properties at a glance. instead of 5W-30, we could use 3-5W-30-20 to denote equiv viscosity at temperatures ranging from -40C to 150C.

This could easily display HTHS viscosity in simple terms. Or we could simply revise the VI calculation.