THIS oughta spark some conversation:
http://neptune.spacebears.com/cars/stories/margins.html
Cheers, 3MP
http://neptune.spacebears.com/cars/stories/margins.html
Cheers, 3MP
Oil Analysis Margins of error
- Thread starter 3 Mad Ponchos
- Start date
- Status
As you discovered, the equation for SUS as a function of cSt is not easily invertable (even using symbolic math software).
I have used that function to generate a lookup table of cSt to SUS, and I've fit an aribitrary function to that table. The function is:
The relative error in cSt using this function is under 1% for cSt between 2.36 and 40.4 (which seems to include all common values of interest). A plot of the relative error is included below.
Hope this is useful to you.
Also of possible interest to you is analysis of the same oil by Schaeffer's and Blackstone.
Helluva great post!! Thanks man.
Revive this bad boy:
4 LAB Comparo
As for your PM, sorry I didn't get back at you. Just busy and occupied. I just don't have time for web site stuff. Maybe I can send you my 4 lab reports or something. Maybe scan them??? I'm extremely crappy at web site creation maintenance.
[ April 08, 2003, 10:23 PM: Message edited by: Pablo ]
Revive this bad boy:
4 LAB Comparo
As for your PM, sorry I didn't get back at you. Just busy and occupied. I just don't have time for web site stuff. Maybe I can send you my 4 lab reports or something. Maybe scan them??? I'm extremely crappy at web site creation maintenance.
[ April 08, 2003, 10:23 PM: Message edited by: Pablo ]
Interesting stuff. Nice to see how what we get from the labs does need a little "thought" applied to it before taking it as gospel.
Nice to see the blind test came out relatively close. I know I'd be OK with the minor differences between the two. None of them were make or break kind of differences, although we all know some of the frustration with TBN measurments...
Nice to see the blind test came out relatively close. I know I'd be OK with the minor differences between the two. None of them were make or break kind of differences, although we all know some of the frustration with TBN measurments...
Patman
Staff member
Interesting results!
I've always used this formula to convert SUS to cst, it's quite simple:
0.226xSUS-(195/SUS)=cst
So 60.0 SUS equals 10.31 cst
I've always used this formula to convert SUS to cst, it's quite simple:
0.226xSUS-(195/SUS)=cst
So 60.0 SUS equals 10.31 cst
MPH, your function worked spot-on! Very nicely done. PM me your real name so you can get due credit for your efforts.
Pablo, all I need from you is permission to use your data, and a few minutes of your time to provide some commentary after I've written it up. You've already done the hard part (getting the results!).
Gopher, I was fairly pleased with the results also, though I had hoped to see closer TBN numbers. The more I learn about TBN the more I find it frustrating and, to some extent, misleading. I plan to go into much greater depth about TBN later.
Patman, ASTM D2161 provides conversion tables between cSt and SUS (this is where my original equation came from). I wish I could just reproduce the entire table but if ASTM wants $40 just to BUY the thing, I shudder to think what the reprint rights must be like. Anyway, your equation is close but 60.0 SUS is actually 10.22 cSt. Knowing this, would you say your equation is close enough? I'm used to thinking in terms of SUS so I'm not sure how big of a difference you would perceive between 10.22 and 10.31.
Cheers, 3MP
Pablo, all I need from you is permission to use your data, and a few minutes of your time to provide some commentary after I've written it up. You've already done the hard part (getting the results!).
Gopher, I was fairly pleased with the results also, though I had hoped to see closer TBN numbers. The more I learn about TBN the more I find it frustrating and, to some extent, misleading. I plan to go into much greater depth about TBN later.
Patman, ASTM D2161 provides conversion tables between cSt and SUS (this is where my original equation came from). I wish I could just reproduce the entire table but if ASTM wants $40 just to BUY the thing, I shudder to think what the reprint rights must be like. Anyway, your equation is close but 60.0 SUS is actually 10.22 cSt. Knowing this, would you say your equation is close enough? I'm used to thinking in terms of SUS so I'm not sure how big of a difference you would perceive between 10.22 and 10.31.
Cheers, 3MP
OK, I updated the page, fixing a couple of typos and including mph's version of the equation so now the calculator works both ways. Schweeet.
Cheers, 3MP
Cheers, 3MP
That rule-of-thumb has good accuracy over the range of viscosities that we are interested in, but your calculation out to two places is false precision. Taken from the ASTM tables, 60.0 SUS = 10.22 cSt. mph's formulae is also a rule-of-thumb. I'd like to see a graph comparing the accuracy of the two rule-of-thumbs over the same range shown above. Nice work mph!
Here you go:
Not surprisingly, my formula performs better over the range of interest, because I used more fit parameters than the formula Patman quoted above (5 vs. 2). If I were doing the conversion by hand or with a simple calculator, I'd likely use the simpler formula, though.
One point to consider is that I'd bet (not knowing much about the subject) that the formula Ponchos uses from the ASTM document is itself a fit to the big table of lab measurements, and will itself disagree with them slightly. Without knowing the magnitude of this error, we don't know how accurately we can go in either direction. But I have no doubt that in practice, any reasonable approach will serve our purposes far better than we need it to.
(BTW, rather than coming up with a function for computing cSt as a function of SUS, one could write the calculator program to use a numerical equation-solving algorithm, such as Newton's method. This would allow arbitrary accuracy at essentially all values of input, at the expense of some programming.)
That's brilliant mph, thanks! You are right about the ASTM equations being rules-of-thumb themselves. They were, "computed by fitting a smooth curve to the original experimental data points", according to ASTM.
[ April 09, 2003, 10:57 PM: Message edited by: Jay ]
[ April 09, 2003, 10:57 PM: Message edited by: Jay ]
Patman
Staff member
I think I know another reason why my formula shows different numbers. I believe my method of conversion was for converting SUS at 212F, while the formula here is at 210F.
Either way, I love the new formula and it's easy and quick way to convert via your website Brian! I will use your automatic conversion from now on.
[ April 12, 2003, 06:46 AM: Message edited by: Patman ]
Either way, I love the new formula and it's easy and quick way to convert via your website Brian! I will use your automatic conversion from now on.
[ April 12, 2003, 06:46 AM: Message edited by: Patman ]
Should be required reading.
http://neptune.spacebears.com/cars/stories/margins.html
When looking at the fluctuation in the numbers over the course of our oil study, it's natural to ponder the margin of error for the tests. Because even though we suspend disbelief and accept that the testing found exactly 2,930 mg/Kg of calcium, deep down inside we know that no test is perfect. At best it is a very very good approximation.
So just how good is the approximation? Most people tend to think of this as a margin of error expressed as a percentage. Well, sometimes that's true, but in the world of science it's seldom that easy. What we call margin of error, science geeks call precision, and they measure it in terms of repeatability and reproducibility:
Repeatability
* Same operator
* Same laboratory
* Same equipment
* Same conditions
* Identical sample
Reproducibility
* Different operator
* Different laboratory
* Equivalent equipment
* Equivalent conditions
* Identical sample
In a nutshell, repeatability is when you make back-to-back passes at a dragstrip; reproducibility is when you let your buddy borrow your car and he races it at a different dragstrip. Based on the standards we've been reading, reproducibility has a margin of error two to four times worse than repeatability. Sometimes this margin of error is expressed as a percentage, but not always. With any margin of error statement, watch for the measure of error (percentage, parts per million, or whatever it may be).
Now, a complicating factor is that when we look at oil degradation trends, we're not actually using an identical sample as required by both measures of precision: each test point is a new sample (to continue the analogy, your buddy is using a car nominally identical to yours, rather than actually using your car). We have no basis at all to expect precision within the repeatability range over the course of this study, so we will instead hope to see precision within the reproducibility range.
To get an idea of the real-world reproducibility of oil analysis, we subjected Blackstone Laboratories to a blind test. When we drew the 6,000-mile sample, we drew a second sample immediately following the first. We submitted the sample by way of accomplice Les Carnes, who -- lucky dog -- "owns" a 2003 Corvette made of vaporware. Les sent in the sample for us and forwarded the results. Blackstone was therefore "blind" to the true origins of the sample, ruling out any potential bias at the lab. The results were quite favorable, well within reproducibility ranges for most tests.
It's worth noting that this is not a perfect reproducibility test. Though the oil samples were very similar, there is no way to ensure that they were identical. The second sample came out of the pan several minutes after the first, giving the oil more time to settle and cool. The second sample also had to travel to Texas before going to the lab; in all there is a six-day difference between tests, a point mostly of importance to TBN, which can sometimes age unpredictably. And then there is simply the random distribution of particulates, which cannot be assumed to be a uniform mixture.
http://neptune.spacebears.com/cars/stories/margins.html
When looking at the fluctuation in the numbers over the course of our oil study, it's natural to ponder the margin of error for the tests. Because even though we suspend disbelief and accept that the testing found exactly 2,930 mg/Kg of calcium, deep down inside we know that no test is perfect. At best it is a very very good approximation.
So just how good is the approximation? Most people tend to think of this as a margin of error expressed as a percentage. Well, sometimes that's true, but in the world of science it's seldom that easy. What we call margin of error, science geeks call precision, and they measure it in terms of repeatability and reproducibility:
Repeatability
* Same operator
* Same laboratory
* Same equipment
* Same conditions
* Identical sample
Reproducibility
* Different operator
* Different laboratory
* Equivalent equipment
* Equivalent conditions
* Identical sample
In a nutshell, repeatability is when you make back-to-back passes at a dragstrip; reproducibility is when you let your buddy borrow your car and he races it at a different dragstrip. Based on the standards we've been reading, reproducibility has a margin of error two to four times worse than repeatability. Sometimes this margin of error is expressed as a percentage, but not always. With any margin of error statement, watch for the measure of error (percentage, parts per million, or whatever it may be).
Now, a complicating factor is that when we look at oil degradation trends, we're not actually using an identical sample as required by both measures of precision: each test point is a new sample (to continue the analogy, your buddy is using a car nominally identical to yours, rather than actually using your car). We have no basis at all to expect precision within the repeatability range over the course of this study, so we will instead hope to see precision within the reproducibility range.
To get an idea of the real-world reproducibility of oil analysis, we subjected Blackstone Laboratories to a blind test. When we drew the 6,000-mile sample, we drew a second sample immediately following the first. We submitted the sample by way of accomplice Les Carnes, who -- lucky dog -- "owns" a 2003 Corvette made of vaporware. Les sent in the sample for us and forwarded the results. Blackstone was therefore "blind" to the true origins of the sample, ruling out any potential bias at the lab. The results were quite favorable, well within reproducibility ranges for most tests.
It's worth noting that this is not a perfect reproducibility test. Though the oil samples were very similar, there is no way to ensure that they were identical. The second sample came out of the pan several minutes after the first, giving the oil more time to settle and cool. The second sample also had to travel to Texas before going to the lab; in all there is a six-day difference between tests, a point mostly of importance to TBN, which can sometimes age unpredictably. And then there is simply the random distribution of particulates, which cannot be assumed to be a uniform mixture.
Originally Posted By: Audi Junkie
Should be required reading.
As should the date on the post to which you're replying.
Two necroposts in one morning, AJ? Next cup of coffee's on me.
Should be required reading.
As should the date on the post to which you're replying.
We at BITOG tend to get upset when wear metals increase from 4 to 10, when lab repeatability is +/- 5. The concern should be on a "WATCH" list when wear metals reach 50. This for Fe according to a DYSON professional analysis. Not positive about the 50 value but, I have it recorded somewhere. We can get over concerned about nothing-call it paranoid about something we don't have a clue about.
Does anyone know what happened to the guy(s) that did this test?
I know they cancelled doing more testing b/c they got into motorcycle riding - did they keep the car? did they do anymore testing at all?
Do they still check in to their site?
I know they cancelled doing more testing b/c they got into motorcycle riding - did they keep the car? did they do anymore testing at all?
Do they still check in to their site?
- Joined
- Sep 28, 2002
- Messages
- 39,798
You mean 3 Mad Ponchos?
I think the whole exercise proved to them that such exercises are WORK ...even when some otherz chip in to foot the bill. Just ask river_rat ..or ask me for that matter.
I think the whole exercise proved to them that such exercises are WORK ...even when some otherz chip in to foot the bill. Just ask river_rat ..or ask me for that matter.
Originally Posted By: addyguy
Does anyone know what happened to the guy(s) that did this test?
Last I know his back melted into the ice while taking a sample mid January, then re-froze - trapping him under the car. About time in August his neighbors "find" him under the car taking his 7019th oil sample.
"Necropost" that's hilarious.
Does anyone know what happened to the guy(s) that did this test?
Last I know his back melted into the ice while taking a sample mid January, then re-froze - trapping him under the car. About time in August his neighbors "find" him under the car taking his 7019th oil sample.
"Necropost" that's hilarious.
- Status