18,000-mile results are in!

[3 Mad Ponchos]

http://oilstudy.spacebears.com

Strange, strange. Looks to me like we could have kept going. How is it possible for insolubles to vary so much? What the **** is going on with viscosity and TBN?

Sigh. Grumble. An unsatisfying conclusion.

Cheers, 3MP

 

[Patman]

That report looks strange to me, as a lot of things dropped down that shouldn't have. How could tin all of a sudden go down to 1ppm after being much higher for so long.

And how did iron and lead not increase at all, instead iron goes down? Not possible unless you had changed the oil filter. Copper also went down.

And there is no way the TBN went up like that. Or the insolubles going down like that.

These results just don't make any sense at all.

 

[TR3-2001SE]

Maybe samples where switched by mistake?? Any of this oil left to sample again?? Very mysterious!

 

[Patman]

I don't think the samples would've been switched, because everything is pretty close to what it was on the previous sample, so it's definitely oil from his engine.

I just thought of what it could be. Did you take the sample the same way as you have all along, or did you let the oil drain longer than usual and then take the sample? If so, it could be that your normal method (letting a little bit dribble out and then drawing the sample) shows you higher wear metals than a midstream sampling method would.

 

[Al]

It's very simple. That oil just didn't want to be drained. It wanted to continue to live

I'm no chemist here but perhaps you get up to the higher amounts of lead the error may be a higher number yet still accurate to a % amount. Also perhaps when you get to higher amounts of lead perhaps some of it may actually be in solution and actually precipitates out. so the question is: is it a sampling error or the mechanism that allows a certain amount to stay in solution.

The other scary thought is that maybe the pieces of lead worn from the bearing were so big that the filter caught them and they didn't show up in a sample

 

[Doug Hillary]

Hi,
3MP - a great test and a very interesting result

The oil really wanted to live just like the Shell Helix Ultra ( synthetic )oil I changed out of a my V8 watercooled Porsche a couple of days ago - after 12 months use

Probably the results deserve some comparison as both engines are all alloy and with broadly similar power

Thank you for sharing the data with us all

Regards

 

[3 Mad Ponchos]

Patman, same procedure as always bud.

Cheers, 3MP

 

[Drstressor]

Well it's probably too late now, but it would have been good to pull a valve cover and see if there is any sludge building up. That could be where your insolubles are.

 

[HondaRD]

I just wonder of possible lab error due to mis-calibration or human error. I never had an oil analysis done but I sometimes wonder what levels of quality control are taken at the lab. I don't understand why there can be such large variations especially when samples are compared between different labs.

 

[Patman]

I'm curious, will you be installing a Fumoto or similar type of drain plug device in order to make things way easier for you on the next rounds of tests?

Do you think Blackstone would retest your oil sample to see if the numbers show different? Without a doubt there is something amiss with this sample report.

[ October 10, 2003, 04:28 AM: Message edited by: Patman ]

 

[Pablo]

I hate to say it but IMHO all these numbers are within cumulative error range. We tend to get all interested about 11 ppm vs. 15 ppm (for example) and, well the S.D. does overlap....

One Q. 3MP: Are you still gonna do a lab study on the Amsoil swill (flush) oil?

 

[3 Mad Ponchos]

DrStressor, pulling a valve cover on an LS1 is no simple task. I don't really have time for that kind of operation anyway.

Patman, a reader said he was sending me a valve, if he does, I'll consider using it. For a re-test, I dunno, Blackstone is well aware of this project and I suspect they would have done their own re-test without my asking for it if they found the results unreasonable.

Pablo, I hadn't planned on having the flush oil analyzed. What would we gain from this?

The main thing that bothers me is the insolubles. To this point the insolubles have been pretty stable and predictable, and seemed to serve as a reliable indicator of when to change the filter. With these results, that all comes under question.

Another observation is that there seems to be pretty much no meaningful mathematical correlation between the two TBN methods. A pity.

Cheers, 3MP

 

[Pablo]

3MP many miles back you had proposed a lab comparison of your own...and I thought I read you would do it on the flush oil....

Did I dream this up???

[QUOTE.....but we'd still like to do a comparison using numbers from our own test car. So, the plan is to select one of the oil-flush intervals between oil types for drawing samples. This way each lab is receiving oil from the same batch. When we get the results from all labs, we'll post a detailed comparison here.
[/QUOTE]

[ October 10, 2003, 10:11 AM: Message edited by: Pablo ]

 

[3 Mad Ponchos]

Oh right, yeah.

Unfortunately my readers don't seem to care, as I haven't received any support for the idea. I'm not talking muted support, I'm talking none at all.

Cheers, 3MP

 

[doyall]

quote:

---

a reader said he was sending me a valve, if he does, I'll consider using it.

---

For the sake of consistency, shouldn't the sample on the Amsoil run come from the same location as the sample on the M1 run?

 

[3 Mad Ponchos]

Not to worry, doyall, the valve goes in place of the drain plug, which is where I'd been getting my samples.

Cheers, 3MP

 

[Patman]

quote:

---

Originally posted by doyall:

quote:

---

a reader said he was sending me a valve, if he does, I'll consider using it.

---

For the sake of consistency, shouldn't the sample on the Amsoil run come from the same location as the sample on the M1 run?

---

With the Fumoto valve though, the sample would still be coming from the same spot, it would just be easier for him to draw the sample since he wouldn't have to use the conventional drain plug anymore. I'm sure he often got worried when he slightly loosened his drain plug, that he might slip and drop the drain plug in the pan, and inadvertantly drain out too much oil. With the Fumoto he wouldn't have to worry about that occuring.

edit-I see I was typing this up at the same time as 3MP.

[ October 10, 2003, 10:57 AM: Message edited by: Patman ]

 

[pscholte]

This is not a comment based just on this test, but it is a building impression I have. I am beginning to suspect the reliability--over time-- of the labwork supporting our UOAs. On any given day I believe a report can be fairly accurate, but given a string of UOAs, I have my doubts.

 

[3 Mad Ponchos]

Pscholte, I've pondered this question too, and I dunno. The labs seem to run a pretty tight ship, but the variances are all too real.

My thoughts on this is that motor oil is not as uniform as we'd like it to be -- there are pockets and concentrations of contaminants that are not evenly dispersed, so there is no such thing really as a "representative" sample.

Just a working theory, could be severely flawed. YMMV.

Cheers, 3MP

 

[TooSlick]

3MP,

I'm not surprised by these results at all, when you consider the following factors:

1) It's hard to take a 4 ounce sample that is representative of the bulk oil properties

2) There is a fair amount of variability in these inexpensive lab tests - perhaps +/- 3 ppm in terms of wear metals, and +/- 0.5 in terms of TBN. This is particularly true for determining the TBN as the oil reaches the end of its' service life. For this reason I never go below a TBN of 2.0

3) You only increased the drain interval by 6%, in going from 17k miles to 18k miles. In addition you added some fresh oil after the 17k sample was taken, which affects wear metal concentrations and TBN. This can make it look like the wear metal concentrations are actually decreasing ....

This is why I have argued for sampling only every 3000 miles, so that the trends in wear rates and oil degradation are more apparent. If you look at the results from any two adjecent tests taken 1000 miles apart in your data set, it's hard to tell much of anything.

FWIW, the total solids measurement is clearly in error, unless you somehow totally depleted the detergent/dispersant additives and the solids fell out of suspension. If that had happened, I think you would have seen chunks when you drained the Mobil 1.

I still think it is a very worthwhile test, but you have to take into account both the accuracy and precision of this type of data ....

TS