A_Harman index - base oils

[OVERKILL]

There has been much discussion over the years and more recently, calculations that predicated on results from this Index, which was drafted up years ago by member A_Harman, to determine, roughly, the amount of VII in a lubricant. Essentially, the closer to 1 you got, the lower the VII content.

Given the recent extrapolations based on this, I thought it prudent to run the calculation on some base oils, which of course we know, have no VII in them. This proved somewhat difficult to do broadly because it would seem HTHS data for base oils isn't a parameter commonly published. Using what was available, I started with XOM Chemical's SpectraSyn line of PAO products and the results were what one would hope to expect:

SpectraSyn 8: 1.017
SpectraSyn 6: 1.077
SpectraSyn 4: 1.011

The only real stand-out being SpectraSyn 6, as it is closer to 1.1.

Then, I ran it on the SpectraSyn Plus bases, and this is where things got interesting:

SpectraSyn Plus 3.6: 0.874
SpectraSyn Plus 4.0: 0.886
SpectraSyn Plus 6.0: 0.936

So, if we were to use the A_Harman index as the basis for a VII content calculation on a lubricant with a significant component of it comprised of a SpectraSyn Plus PAO base, the perceived VII content would be artificially inflated.

Based on this, I would encourage those interested in this to see if they can find other bases that list HTHS, perhaps we can compile a list of base oil A_Harman indexes.

 

[Gokhan]

Ah, interesting attempt!

Two problems though, the first minor, the second major:

(1) I improved the density correction factor to 0.905, as opposed to A_Harman's 0.885. This is more consistent with the literature and didn't give any A_Harman index greater than 1 in finished oils.

(2) See footnote 2 in the datasheet. They say they only did a single- or two-sample measurement on HTHSV. HTHSV is notoriously difficult to measure and I am guessing that they didn't try to obtain any accuracy.

These numbers are up to 15% off from the theoretical value of 1, which is consistent from what one would get from a single- or two-sample measurement on HTHSV. Therefore, unfortunately, I don't think they mean anything.

Interestingly, they also list a temperature density correction factor but we need to figure out how to use it and then see if the scalar 0.905 I employed is accurate.

 

[OVERKILL]

Thanks for your reply, I'll respond in kind:

Per point 1, I would expect that the desired outcome would be "1" for a straight base oil, indicating zero VII content, and I believe that was the spirit of what he was trying to accomplish with the formula. Tweaking it so that it yields nothing above 1 (shifting everything down) was likely required due to "non oil" things in the finished product or, perhaps variations in the base oils themselves, which we can see based on the examples I've posted. I don't believe it should be expected that all straight bases will yield "1", but rather, there will likely be a range that may depend on the type of base as well. That's why I've encouraged seeking out data that includes HTHS for other bases, because it would allow the compilation of a chart; raw data, which can then be analyzed and potentially drawn from.

Per point 2, I expect the same is the case for a finished motor oil. Ergo, this follows through on anything being run through this calculation, it isn't exclusive to what I've done here.

Putting aside any personal opinion on the matter, I think it is quite telling that the two families; and these are families, of bases, yielded consistently different results. I really don't think that can be ignored because of the extrapolation taking place that hinges on "1" being VII-free.

Even if we consider a 15% fudge factor and that the HTHS obtained is averaged you are looking at a range of 0.99 to 1.31; which yields an A_Harman index range of 0.759 to 1, with a 0.879 average. All three SpectraSyn plus bases would need to have conveniently erred downward, whilst all the regular Spectrasyn bases just happen to be bang on the money? Doesn't smell right, and it shouldn't to you either.

I think the rather obvious conclusion one draws from the data, as flawed as it is, is simply that bases are going to vary and that we either need to factor in other traits that levels the field to provide a firmer base for what constitutes a calculated "VII free" product or work the base data as a range into any derivative calculations so that anything resulting from this would be presented as such. Say 6-8% VII for example, if the compiled base oil data yielded that amount of deviation.

 

[Shannow]

Not sure how "notoriously difficult" it is to get valid HTHS data.

Sure, back when the were injecting oil into the main bearings of test engines, yes...but Selby through the late 70s and 80s developed techniques with good repeatability and reproducability. Can buy bench test machines capable of doing 20 samples per hour. For example ASTMD5481 has repeatability of less than 2.5%, and reproducability of less than 4.2% (both maximum spread around mean value), so you can't fob off 15% as being "normal error"

Am interested in the basis of the "notoriously difficult" statement as the basis for the SPECIFICATION sheet differences...

Seriously if you were Mobil, putting out a spec sheet for the knowledgable oil blender, you'd put your best foot forward and stump up for accurate testing.

Balance of probabilities, Mobil mean what they put on the sheet.

BTW, I use this for density when messing with the H.I....

https://planetcalc.com/2834/

 

[Gokhan]

Originally Posted by Shannow
BTW, I use this for density when messing with the H.I....

https://planetcalc.com/2834/

Thanks for the formula, which is:

Density_2 = Density_1 * correction,

with

correction = exp{-alpha_1*(T_2 - T_1)*[1 + 0.8*alpha_1*(T_2 - T_1)]}.

T_1 = 15.6 C (or 15 C) and T_2 = 150 C. alpha_1 is the reported density correction factor at T_1.

I calculated it for Exxon Mobil SpectraSyn PAO and SpectraSyn Plus PAO. This is what I got:

Code
Base stock Initial T alpha Final T Correction



4.1 cSt 15.6 6.41E-04 150 0.912

5.8 cSt 15.6 6.28E-04 150 0.914

8.0 cSt 15.6 6.24E-04 150 0.914

3.6 cSt Plus 15.6 6.48E-04 150 0.911

3.9 cSt Plus 15.6 6.44E-04 150 0.912

5.9 cSt Plus 15.6 6.32E-04 150 0.913

The density correction factor I used in my latest tables is 0.905. Therefore, it's about right. Also, it's for finished oil and expected to be somewhat different than for the "neat" base stocks. Moreover, Group III and GTL base stocks would be somewhat different as well.

 

[Gokhan]

A_Harman indexes you calculated:

Originally Posted by OVERKILL
SpectraSyn 4: 1.011
SpectraSyn 6: 1.077
SpectraSyn 8: 1.017

SpectraSyn Plus 3.6: 0.874
SpectraSyn Plus 4.0: 0.886
SpectraSyn Plus 6.0: 0.936

What I calculated using better density correction factors listed above:

4.1 cSt: 0.981
5.8 cSt: 1.042
8.0 cSt: 0.985

3.6 cSt Plus: 0.847
3.9 cSt Plus: 0.859
5.9 cSt Plus: 0.906

Now, we don't expect these low-viscosity "neat" base stocks to temporarily or permanently shear significantly. It's true that their high-viscosity PAO (100 cSt or greater) shears to some degree with increasing viscosity and their 1000 cSt can shear like a VII, which is not unexpected as it's practically plastic. However, I would expect negligible shear for these low-viscosity PAO neat (no additives or VII, pure) base stocks.

Then, what is going on? Two possibilities:

(1) The Widman operational viscosity calculator doesn't reliably work.

(2) We are having a garbage in, garbage out (GIGO) situation by being provided bad data by Exxon Mobil.

Given the scatter in the numbers, could it be that (2) is the case, namely garbage in, garbage out?

 

[OVERKILL]

Originally Posted by Gokhan
A_Harman indexes you calculated:

Originally Posted by OVERKILL
SpectraSyn 4: 1.011
SpectraSyn 6: 1.077
SpectraSyn 8: 1.017

SpectraSyn Plus 3.6: 0.874
SpectraSyn Plus 4.0: 0.886
SpectraSyn Plus 6.0: 0.936

What I calculated using better density correction factors listed above:

4.1 cSt: 0.981
5.8 cSt: 1.042
8.0 cSt: 0.985

3.6 cSt Plus: 0.847
3.9 cSt Plus: 0.859
5.9 cSt Plus: 0.906

Now, we don't expect these low-viscosity "neat" base stocks to temporarily or permanently shear significantly. It's true that their high-viscosity PAO (100 cSt or greater) shears to some degree with increasing viscosity and their 1000 cSt can shear like a VII, which is not unexpected as it's practically plastic. However, I would expect negligible shear for these low-viscosity PAO neat (no additives or VII, pure) base stocks.

Then, what is going on? Two possibilities:

(1) The Widman operational viscosity calculator doesn't reliably work.

(2) We are having a garbage in, garbage out (GIGO) situation by being provided bad data by Exxon Mobil.

Given the scatter in the numbers, could it be that (2) is the case, namely garbage in, garbage out?

Seems your different density correction factor to bring us up to the HTHS temp didn't appreciably change the results, taking us back to the foundation of what we've already discussed and the range we are seeing here. Regarding #1, Molakule has already stated rather emphatically that they (visc calcs) are not accurate. So, I would be inclined to believe, as Shannow noted, published spec data from XOM before putting faith in a mechanism that a blender has already indicated, is flawed.

 

[Gokhan]

Originally Posted by OVERKILL
Regarding #1, Molakule has already stated rather emphatically that they (visc calcs) are not accurate. So, I would be inclined to believe, as Shannow noted, published spec data from XOM before putting faith in a mechanism that a blender has already indicated, is flawed.

Nah, Exxon Mobil's HTHSV data looks messed up here.

Here is the giveaway:

SpectraSyn 6 cSt PAO:
KV100 = 5.8 cSt
VI = 138
HTHSV = 2.08 cP

SpectraSyn Plus 6 cSt PAO:
KV100 = 5.9 cSt
VI = 143
HTHSV = 1.86 cP

SpectraSyn Plus is not only a thicker oil than SpectraSyn but it has a higher viscosity index. It makes no sense for it to have an 11% lower HTHSV. In fact, we would expect it to be higher. I don't buy the claim that their better PAO (Plus) thins faster with the increasing temperature than their standard PAO, despite thinning slower below 100 C according to the VI. We also ruled out temporary shear here.

Another giveaway is that the SpectraSyn 4 cSt and 8 cSt give reasonable A_Harman indexes but not the SpectraSyn 6 cSt. Again, something must be wrong with their HTHSV measurements.

The two-point (for example, using KV40 and KV100 like the Widman operational-viscosity calculator) viscosity extrapolation is based on ASTM D341. It's theoretical, yes, but I don't expect it to be useless, which the Exxon Mobil data would suggest if it were accurate.

https://compass.astm.org/EDIT/html_annot.cgi?D341+17

If the Exxon Mobil data were accurate, then the A_Harman index is unable to distinguish between TGMO 0W-20 and a virtually shearless pure (neat) base stock with no VII. I would stop short of declaring the A_Harman index as garbage in, garbage out. The A_Harman index has given reasonable results from what we have calculated for many oils. It's probably that the Exxon Mobil's in-house HTHSV tester or perhaps whoever operated it that has problems, which seems to be hinted by the footnote "single-sample or two-sample average determinations." The biggest problem would be not being able to precisely set the temperature at 150 C.

 

[4WD]

SpectraSyn Plus is 3.6, 4, or 6
SpectraSyn is 40 & 100
SpectraSyn Elite (mPAO) is 65, 150, & 300

 

[Gokhan]

Originally Posted by 4WD
SpectraSyn is 40 & 100

and 2, 2B, 2C, 4, 5, 6, 8, and 10...

 

[Shannow]

Originally Posted by Gokhan
It makes no sense for it to have an 11% lower HTHSV. In fact, we would expect it to be higher. I don't buy the claim that their better PAO (Plus) thins faster with the increasing temperature than their standard PAO, despite thinning slower below 100 C according to the VI. We also ruled out temporary shear here.


LOL...you stated that I was wrong for using commons sense to discredit your latest metric when I pointed out all the logical flaws for your conclusion that M1 5W30 had sub 2 HTHS basestocks and was packed with VII...through sensible reasoning as to why your position as flawed.

You stated that science had no place for intuition or thereabouts.

Now...it's all your version of common sense.

So which is it ?

Maybe you should offer Mobil a retainer to fix their data sheets...data sheets that blenders rely on to get their oils on spec or close to it first go (which is exactly what Chevron's BOQI patent was for...to get a new blend on spec, not to judge oils...just saying)

So...could you please explain again, how the HTHS test is "notoriously difficult" to get right...

 

[Shannow]

Originally Posted by Gokhan

It's probably that the Exxon Mobil's in-house HTHSV tester or perhaps whoever operated it that has problems, which seems to be hinted by the footnote "single-sample or two-sample average determinations." The biggest problem would be not being able to precisely set the temperature at 150 C.


More supposition to discredit evidence.

The footnote is not a hint...it's just a statement of fact, across many of the parameters on the data sheet...
* density
* Kinematic viscosity -40C, 40C, 100C...are these "notoriously difficult to measure as well ???
* Noack
* Refractive Index
* Brookfield (MRV), CCS,
* autoignition temperature....and on and on.

Basically every metric that isn't "calculated" like the density correction and surface temperature.

You are implying incompetence, in an ASTM standard that has a repeatiability between two disparate measurements of 2.3% of the mean, and a reproducibility of 3.6%...then publishing it.

Your suppositions just don't jive.

However, I would point you that in addition to the NOACK, the mobil data sheets also have 6.5 hour evaporation at 205C...could help you calibrate some of your volatility arguments by adding another point to the curve.

Re "the biggest problem" being able to precisely set the temperature...these machines are capable of dozens of back to back test per hour. They don't have IGOR twiddling dials and waiting for the dial to move.
https://www.cannoninstrument.com/en/uproduct/ravenfield-bsc-hths-tapered-plug-viscometer?id=Ravenfield%20BS%2FC%2B

A guaranteed 0.02C at 150C in an off the shelf instrument, that rules out operator incompetence, Igor dial twiddling, and the "difficulty" on temperature control in the publishing of data sheets.

 

[Gokhan]

Originally Posted by Shannow
Originally Posted by Gokhan
It makes no sense for it to have an 11% lower HTHSV. In fact, we would expect it to be higher. I don't buy the claim that their better PAO (Plus) thins faster with the increasing temperature than their standard PAO, despite thinning slower below 100 C according to the VI. We also ruled out temporary shear here.
LOL...you stated that I was wrong for using commons sense to discredit your latest metric when I pointed out all the logical flaws for your conclusion that M1 5W30 had sub 2 HTHS basestocks and was packed with VII...through sensible reasoning as to why your position as flawed.

You stated that science had no place for intuition or thereabouts.

Now...it's all your version of common sense.

So which is it ?

Maybe you should offer Mobil a retainer to fix their data sheets...data sheets that blenders rely on to get their oils on spec or close to it first go (which is exactly what Chevron's BOQI patent was for...to get a new blend on spec, not to judge oils...just saying)

So...could you please explain again, how the HTHS test is "notoriously difficult" to get right...

I have no idea what your point is, Shannow, but you don't make sense. Did you even read my post?

Are you claiming that an improved-quality PAO base stock that is thicker at 100 C and has a higher viscosity index will have a lower viscosity at 150 C than an inferior PAO base stock that is thinner at 100 C and has a lower viscosity index? Please explain without digging everything I've ever posted on BITOG.

PS: Per your other point, according to Exxon Mobil, 4 cSt PAO as HTHSV = 1.46 cP by the way. It's not to mention that M1 HM 5W-30 is mostly Group III. So, there is no contradiction in my table.

 

[Shannow]

Originally Posted by Gokhan
PS: Per your other point, according to Exxon Mobil, 4 cSt PAO as HTHSV = 1.5 cP by the way. It's not to mention that M1 HM is mostly Group III. So, there is no contradiction in my table.


Except the 5W30 in the blend guide shows the 5W30 as 43.1% PAO6 and 39.8% PAO8, and should result in an HTHS for the basestock alone of 2.27....and with a VI in the low 160s, there is NO WAY that the M1 5W30s are built the same way as TGMO...

BTW, check the Spectrasyn 6 Data sheet...
http://exxonmobilchemical.ides.com/en-US/ds244628/SpectraSyn%E2%84%A2%206.aspx?I=30156&U=1

Note 2...so which one is indicative of HTHS incompetence ?

If two data sheets have the same giveaway to incompetence, which one do YOU trust ???

 

[Gokhan]

Originally Posted by Gokhan
Are you claiming that an improved-quality PAO base stock that is thicker at 100 C and has a higher viscosity index will have a lower viscosity at 150 C than an inferior PAO base stock that is thinner at 100 C and has a lower viscosity index? Please explain without digging everything I've ever posted on BITOG.

Shannow, I asked you a simple question.

 

[Shannow]

Originally Posted by Gokhan
Shannow, I asked you a simply question.


It's really hard to keep track of your constant editing while I'm typing....how many were in that last one, four or five ?

You claimed common sense as proof of your position of Mobil incompetence...I pointed out that you discredited a common sense argument covering multiple bases...So is common sense right, or only when you want it to be ?

Are data sheets and blending guides right only when you want them to be ?

Those were my "simply" questions, that you are using the regular "look over there, a bunny" tactics to step away from.

 

[Gokhan]

Originally Posted by Shannow
Originally Posted by Gokhan
Shannow, I asked you a simply question.


It's really hard to keep track of your constant editing while I'm typing....how many were in that last one, four or five ?

You claimed common sense as proof of your position of Mobil incompetence...I pointed out that you discredited a common sense argument covering multiple bases...So is common sense right, or only when you want it to be ?

Are data sheets and blending guides right only when you want them to be ?

Those were my "simply" questions, that you are using the regular "look over there, a bunny" tactics to step away from.

No, I never edited my question. It's a very simple scientific one. None of the questions you are asking are scientific but the usual bickering from you. I will go ahead and answer though. Please answer mine.

 

[Gokhan]

Originally Posted by Shannow
Originally Posted by Gokhan
PS: Per your other point, according to Exxon Mobil, 4 cSt PAO as HTHSV = 1.5 cP by the way. It's not to mention that M1 HM is mostly Group III. So, there is no contradiction in my table.


Except the 5W30 in the blend guide shows the 5W30 as 43.1% PAO6 and 39.8% PAO8, and should result in an HTHS for the basestock alone of 2.27....and with a VI in the low 160s, there is NO WAY that the M1 5W30s are built the same way as TGMO...

BTW, check the Spectrasyn 6 Data sheet...
http://exxonmobilchemical.ides.com/en-US/ds244628/SpectraSyn%E2%84%A2%206.aspx?I=30156&U=1

Note 2...so which one is indicative of HTHS incompetence ?

If two data sheets have the same giveaway to incompetence, which one do YOU trust ???

The answers are simple:

(1) Mobil 1 is not equal to the examples in the Exxon Mobil blending guide. As we're talking about specific numbers, there is no point in comparing them.

(2) I pointed that footnote twice before in this thread before you did.

(3) I proved that the HTHSV in the SpectraSyn datasheet is not correct. You can go through my proof. It's simple science and not an issue of trust or belief.

I'm still waiting for your answer to the question I asked you.

 

[Gokhan]

Originally Posted by Gokhan
Are you claiming that an improved-quality PAO base stock that is thicker at 100 C and has a higher viscosity index will have a lower viscosity at 150 C than an inferior PAO base stock that is thinner at 100 C and has a lower viscosity index? Please explain without digging everything I've ever posted on BITOG.

I'm repeating this for the third time, Shannow. What is your answer?

 

[Shannow]

LOL, just sitting here watching your posts change in real time...with not a single edit line or reason is priceless...