Viscosity Modeling

[Hokiefyd]

In comparing various Mobil 1 products in a 30 grade...

Despite not knowing the MRV for any of Mobil 1's 5w30 oils available (at least from ExxonMobil's United States website), it seems reasonable to believe that at very cold temperatures, any of them would have thickened more than Mobil 1's own 0W-30. Yet using Widman's viscosity graphing available at widman.biz, Mobil 1's 5w30 and 5w30 Extended Performance (EP) appears to not thicken as much as the 0w30 does. At 0 deg C, 5w30 is at 490 cSt, 5w30 EP is at 483 cSt, and 0w30 is up at 525 cSt. See attached graphic.

And because this is a mathematical model, the relationship between the four plot lines remains constant. The 0w30 is always a little thicker than either of the two 5w30 products (not counting the 5w30 ESP, which was thrown in out of curiosity).

So it seems that the viscosity graphs cannot be smooth…that is, they must curve back and bend depending on the temperature. And I suppose there is no way to model this with a simple mathematical calculator using only two data points (cSt at 40 deg C and at 100 deg C). So I say all that to pose the following questions (some directly related to viscosity and some not).

1) If we assume that there is some temperature at which the 5w30 oils will be thicker than the 0w30 oil, there has to be some temperature where the shape of the graphs change. Is there any possible way to estimate the temperature at which this change in behavior occurs?

2) A follow-on to question 1 above, down to what temperature should we have relative confidence in a mathematical model? We take the data at 40 deg C and at 100 deg C to be accurate, because it's provided by ExxonMobil. In theory, we could also take any modeling result near those temperatures (say at 30 deg C or at 90 deg C) to be accurate, because they're still close, and should be well represented by a model connecting those two points. But as you start to get down to 10 deg C, 0 deg C, and below…when do we lose confidence in the model?

3) Assuming that the model is fairly accurate down to moderately cold temperatures (perhaps as low as -10 deg C), it appears that 0w30 offers little benefit over either of the 5w30 grades in terms of startup viscosity (again, excluding the ESP product). With that general understanding, how relevant is dexos1 and HTO-06 to applications not requiring them? In other words, if each of these three oils can be considered to be nearly equivalent in terms of viscosity, would the dexos1 and HTO-06 certification on either of the 5w30 products identify them as slightly more robust compared with the API SN 0W-30? For what it's worth, they all meet GM 4718M.

I know this entire discussion can likely be filed in the "splitting hairs" folder, but learning is what BITOG is all about, right?

 

[Jim Allen]

I agree that this model presents only a blurry picture that is less accurate as it drops below 40C. There are probably better ways to do it... not that I know what they are. Try comparing CCV or MRV as, I think, they will give you a better picture of the oil's cold flow performance. Also, look at the ways the winter grading takes place. Finally, look at VI. I'm not sure if there is a one snapshot model that will give you all the details on cold flow.

 

[Gabe]

Originally Posted By: Hokiefyd


1) If we assume that there is some temperature at which the 5w30 oils will be thicker than the 0w30 oil, there has to be some temperature where the shape of the graphs change. Is there any possible way to estimate the temperature at which this change in behavior occurs?


This has me wondering too...

If we know the actual viscosity formula, then we would be able to set the two equations equal to each other to determine when the 0W will begin to flow better. The shape of the graph doesn't have to change. The VI of the oil is how fast the graph rises; the viscosity at a given temperature is how far up the graph begins.

I would imagine that a basic viscosity model would be fairly accurate.

 

[bepperb]

I would say

1.) When components of the oils begin to change phases from liquid to solid the entire curve gets thrown out the window. That's why these graphs are no good for extreme cold, because there are additives that alter this considerably. I'm not sure if M1 Truck/SUV is still available but that had a pour point around -58 or so, as compared to regular M1 at around -44 (at the time) but no difference would be shown in the graph.

2.) The model is definitely out of whack at the extreme cold temps, still I would assume that it's good down to -10c.

3.) HTO-06 would only ever apply to a 5-30, so even if the 0-30 would pass (and I would assume it would now) it would be funny to list it. I assume Dexos is the same, they don't want to label anything besides the required 5-30.

Consider water using the widman graph. The graph would have no idea that at 1 degree the water is thin and at -1 it would be infinitely viscous. If a small amount of salt was added with no difference in viscosity at room temp but a huge difference at -5 degrees. The graph would be worthless as soon as phase change became a factor as it is with oil. Problem with oil is that it's made of so many different components.

 

[A_Harman]

I wouldn't trust the Viscosity Index calculation much below 0C for synthetic oils, and below 20C for conventionals. Once wax crystals start to form in subzero temperatures, the viscosity becomes unpredictable.

Taking the KV40 and KV100 values for Mobil 1 Advanced Fuel Economy 0w30, and plugging them into Widman's calculator at -40C, gives a predicted kinematic vis of 23993. Converting KV to Dynamic Viscosity by multiplying by density (.842) gives 20200 cP @ -40C. Mobil's published MRV for this oil is 13250 cP, which is about 33% lower than the value that the VI calculation predicts.

 

[CATERHAM]

The important thing to remember here is that you're graphing a kinematic viscosity measure which does not correlate well with the actual operational viscosity in an engine. The kinematic measure does not properly take into account the effect of varing polymers levels nor the pressure-viscosity coefficients of different base oil chemistries.
One should always factor in the HTHSV rating and not just the oils VI which is what the graph represents.
In this case AFE 0w30 has a low 3.0cP HTHSV although it's 166 VI is also relatively low. Nevertheless the fact remains that this oil is lighter in terms of actual operational viscosity than all the other Mobil 30wt oils at all oil temp's.

 

[Garak]

Originally Posted By: bepperb
3.) HTO-06 would only ever apply to a 5-30, so even if the 0-30 would pass (and I would assume it would now) it would be funny to list it. I assume Dexos is the same, they don't want to label anything besides the required 5-30.


I'd agree on the part about the HTO-06 spec, but dexos1 now includes other grades on the approved list, well beyond the original 5w30.

 

[Hokiefyd]

Originally Posted By: CATERHAM
In this case AFE 0w30 has a low 3.0cP HTHSV although it's 166 VI is also relatively low. Nevertheless the fact remains that this oil is lighter in terms of actual operational viscosity than all the other Mobil 30wt oils at all oil temp's.


Let me try to push this wet noodle a little further...

Let's compare Mobil 1 0w30 with, say, Quaker State Ultimate Durability in 5w30 flavor. The HTHS is given as 3.0, same as Mobil 1's 0W-30. But Quaker State Ultimate Durability has both a lower kinematic viscosity at 40 deg C compared with the Mobil 1, and has a higher viscosity index (172 vs. 166). Could one then posit that Quaker State Ultimate Durability 5w30 will be as light or lighter than Mobil 1 0w30 at most temperatures, say above -10 deg C or so?

 

[Clevy]

Well the 5w30 EP starts thicker,so yes it wouldn't thicken as much as the AFE since the AFE has a wider viscosity spread........
Am I close. That was a shot in the dark

 

[Hokiefyd]

Caterham, Quattro Pete, others: am I understanding this correctly (below)? Thanks.

Originally Posted By: Hokiefyd
Let me try to push this wet noodle a little further...

Let's compare Mobil 1 0w30 with, say, Quaker State Ultimate Durability in 5w30 flavor. The HTHS is given as 3.0, same as Mobil 1's 0W-30. But Quaker State Ultimate Durability has both a lower kinematic viscosity at 40 deg C compared with the Mobil 1, and has a higher viscosity index (172 vs. 166). Could one then posit that Quaker State Ultimate Durability 5w30 will be as light or lighter than Mobil 1 0w30 at most temperatures, say above -10 deg C or so?