"Most wear occurs at start-up"

[glennc]

JAG, I missed your post a moment ago. Very interesting. I'll have to go back and read the whole linked thread. What I get from it is that in their tests, there was no difference in wear between synthetic and conventional oil, and no difference between different grades, but elevated wear on cold oil by a factor of up to 20-25.

The only question I have is how quickly that wear rate comes down.

If it comes down fairly quickly and you drive for an hour, then start-up wear might be only 5% or 10% of total. But if you do many short trips or if the start-up wear continues to be high for a long time (cold day, say) then you could easily have 95% of wear be start-up wear.

In that case, why don't we see orders-of-magnitude differences in wear between UOAs for short-trip driving in cold weather and UOAs for vacation driving in moderate weather? Most UOAs in general seem to imply overall wear rates per mile within a factor of two or three for a given engine type and usually less than that.

Some possibilities: 1) it is in fact true that UOAs don't effectively show start-up wear? How or why would that be? 2) The test findings don't apply perfectly to real-world driving for some reason. Again, how or why not?

Those are really the only explanations I can think of.

 

[glennc]

Originally Posted By: JAG
Real wear measurements don't use UOAs either at all or do only for supplemental information. Either the radiotracer method or tear downs is the right way.


No doubt they are the right way for precise measurement of wear, especially of specific wear surfaces. However UOA does show elemental metals in oil fairly accurately and should therefore at least be in general proportion to overall wear, should it not? Surely good enough for large variations to be apparent?

I agree if you needed to track wear precisely UOA would not be the correct tool.

Am I missing something important?

 

[FZ1]

Yeah. Highway miles are considered the easiest miles put on a car. Think about it.

 

[Brian Barnhart]

Originally Posted By: JAG
Indeed. Real wear measurements don't use UOAs either at all or do only for supplemental information. Either the radiotracer method or tear downs is the right way.


And when you read the comments from the UOA provider, they are limited to comments regarding the oil’s condition and length of use, such as “next time try extending the interval to … or next time try reducing the interval to… or TBN is still good”, or the possibility of contamination from outside elements , such as “check air filter… or possible coolant leak”. What you don't see is: “next time try brand X to reduce wear in your engine” or “next time try a higher (or lower) viscosity oil to reduce wear”. And this is from the folks who perform and see the results of many UOAs each day.

 

[MGregoir]

This may or may not fit in here, but a mechanic I know had a truck go 400,000 miles (which is hard to do in a cold climate) before he tore the engine down to rebuild it. It still looked new inside. When he got it new, he added a second block heater and an oil pan heater for winter use.

His reasoning behind that was many engines he had torn down to rebuild here had significantly less scuffing of the cylinder walls on the side of the block with the block heater, rather than the "cold" side of the block.

Not really scientific method but it was based on observation of two different scenarios and what seemed to make the difference.

 

[glennc]

Originally Posted By: Brian Barnhart
And when you read the comments from the UOA provider, they are limited to comments regarding the oil’s condition and length of use, such as...

Well you do typically see comments about wear. "Looks like your engine is wearing well," or "the high Silicon levels might be the cause of the elevated wear we are seeing," etc.

As far as I can tell UOA is an excellent tool for tracking wear: rates, trends, differences between oils and so on. Just not as high-precision a tool as is necessary for the purposes of professional research work.

I don't know why start-up wear would be any different, but I'm open to the possibility if there is some reason to think that it might be.

 

[JAG]

Originally Posted By: glennc
JAG, I missed your post a moment ago. Very interesting. I'll have to go back and read the whole linked thread. What I get from it is that in their tests, there was no difference in wear between synthetic and conventional oil, and no difference between different grades, but elevated wear on cold oil by a factor of up to 20-25.

The only question I have is how quickly that wear rate comes down.

I think that modeling the relative wear rate curve to get in the ball-park for our purposes could be done by a straight-line starting at room temp with 20 times the wear rate (value of 20) and ending at 1 at room temp. I'm sure the truth would be some form of curve with a smooth transition to a value of 1...but who likes curves?

Originally Posted By: glennc
If it comes down fairly quickly and you drive for an hour, then start-up wear might be only 5% or 10% of total. But if you do many short trips or if the start-up wear continues to be high for a long time (cold day, say) then you could easily have 95% of wear be start-up wear.

I see what you are saying that the statistic doesn't cover all cases, depending on how long the engine stays at operating temp. I have a bit of a problem with that too. I'm sure that people who state the statistic are trying to simplify the situation to increase readers' understanding.

Originally Posted By: glennc
In that case, why don't we see orders-of-magnitude differences in wear between UOAs for short-trip driving in cold weather and UOAs for vacation driving in moderate weather? Most UOAs in general seem to imply overall wear rates per mile within a factor of two or three for a given engine type and usually less than that.

Some possibilities: 1) it is in fact true that UOAs don't effectively show start-up wear? How or why would that be? 2) The test findings don't apply perfectly to real-world driving for some reason. Again, how or why not?

Those are really the only explanations I can think of.

I think it's possibility #1 and might be #2 (e.g. I doubt dirt ingestion was allowed in the SAE paper's test). Correlation between real wear rates and UOA measured wear rates isn't always good. It's not just a precision (repeatability) issue but what it is an accuracy issue resulting from particle size limitations of UOA measured wear and maybe some other reasons.

We should dig up some more data on this interesting subject.

 

[glennc]

Yes Jag the particle size thing is the one major flaw I see as obvious in UOA. If UOA is to be an effective guide to wear then the particle sizes measured by UOA must be in proportion to other particle sizes across different types of wear. Those too large to be measured or caught by the filter can not be disproportionate under different wear conditions.

My working hypothesis since this site is largely about wear as inferred from UOA results is that no major disproportionality exists in the majority of cases. To support that view I would point to the general acceptance of UOA for tracking wear in many commercial applications as well as our automotive uses.

If that is not the case, then the usefulness of UOA is greatly limited. I can accept that if it turns out to be true, but it would throw into doubt many of the things that have been "learned" on this site over the years from the vast UOA library that has accumulated.

 

[Oilgal]

Well dang. sigh

 

[glennc]

Originally Posted By: JAG
I think that modeling the relative wear rate curve to get in the ball-park for our purposes could be done by a straight-line starting at room temp with 20 times the wear rate (value of 20) and ending at 1 at room temp. I'm sure the truth would be some form of curve with a smooth transition to a value of 1...but who likes curves?


I was thinking along the same lines. I imagine it would have to be a logarithmically diminishing curve though: 20,10,5,3,2,1 or something along those lines. Another issue is that this result was using oil at room temperature. If you extrapolate that curve below room temperature you would get "winter start" wear more like 80,40,20... or somewhere along those lines. Far worse than the test discovered, in other words. The upside is with a logarithmic curve like that the highest wear rate will be of short duration and normalize fairly quickly to within a few multiples of the warm wear rate.

 

[glennc]

By the way McGregoir that was a very interesting post about the engines wearing on the cold side. I'm re-thinking my disinclination to mess with block heaters a bit.

Now I'm afraid to go out and start up my car, which is what I need to do so I can get some actual stuff done. It's like forty degrees outside, that would be like an 80 on the wear scale!

 

[JAG]

Originally Posted By: glennc
My working hypothesis since this site is largely about wear as inferred from UOA results is that no major disproportionality exists in the majority of cases. To support that view I would point to the general acceptance of UOA for tracking wear in many commercial applications as well as our automotive uses.

In commercial vehicles, with one specific oil, once several analyses are obtained, UOAs are great for showing if there are problems with air filtration, coolant leak, fuel dilution, injector problems, or many other engine problems. They are also great for determining proper OCIs. There's huge value in knowing all these things for a fleet owner who has expensive vehicles and down-time is lost money.

On the subject of heaters, my GTI is garage-kept and I still plug the oil pan heater in for 10 minutes of so before driving it when I can. It used to stay outdoors and on really cold mornings, 30 minutes or more of heating made starts significantly smoother sounding/feeling. I even sowed up a wrap-around oil filter heater which seemed to help even more but it is a pain to put on and take off so I don't use it now. I wish I had a block heater. It would allow the oil temp to reach its peak faster.

 

[Gary Allan]

Quote:
My working hypothesis since this site is largely about wear as inferred from UOA results is that no major disproportionality exists in the majority of cases. To support that view I would point to the general acceptance of UOA for tracking wear in many commercial applications as well as our automotive uses.


I think I found a way (sorta) for those who have the "Eyesgotz2nose" syndrome with UOA. I bugged Terry to find a lab that would "digest" UOA samples and then run them. This reduces all the mass metals to the particle level and therefore readable. The ppm's are in the thousands ..not the single digit realm. This SHOULD be a good parallel to "wear". You're seeing all that there is to see. It should be, in some respects, better than real world measurements since, although you can't determine (with assured accuracy) where the material came from, you surely know that it came from somewhere and that less, in this "total" inventory, is better than more. You can also run conventional UOA indications along side to see if the mass of elements correlates to indicated elements. The hole in the comparison is that filtration alters the mass results. Someone using a bypass filter will tend to show 30% less wear. So, it's not without its flaws compared to the real world measurement ..but does have some advantages too. They probably don't assess timing chain and other component losses (perhaps they do by weighing them?).

I'm going to do a test or two like this when my wallet is just a little more stout. It's not outragously expensive, but it's enough to ration the investment in interesting distractions (I have many ..and need to ration them.).

 

[Brian Barnhart]

Originally Posted By: glennc
To support that view I would point to the general acceptance of UOA for tracking wear in many commercial applications as well as our automotive uses.

In commercial applications we see UOAs used to track oil condition. In engine development we see UOAs used to access design changes, detect design issues/problems, and predict impending failures. But I've never heard of a study indicating UOAs are suitable for comparing the wear performance of different oil brands, viscosities, or type. If such a study exists, it certaily would represent a major pillar of BITOG and a link or reference to it should be promimently displayed on the home page and the first page of the UOA forum.

And so far, apparently no one knows of a study or data showing a relationship between oil viscosity to start-up wear (other than the "room temperature" start-up study that indicates no relationship between the two exists). An abundance of opinions, suppositions, and working assumptions do not make up for the absence of a proper study with statistically valid data.

 

[Oilgal]

Originally Posted By: JAG


I even sowed up a wrap-around oil filter heater which seemed to help even more but it is a pain to put on and take off so I don't use it now.


What a terrific idea! JAG you really are a genius. I bet even just an insulating material would be helpful, and you should market that.

I have a stupid question also. At what temperature is an oil pan or a block heater is needed? We are moving to Northern Vermont next week, and now I am a little worried about next winter.

 

[TallPaul]

Originally Posted By: JAG

I even sowed up a wrap-around oil filter heater ...
Could be trouble if filter drains out. Probably burn out the heater.

 

[JAG]

Originally Posted By: Oilgal
I bet even just an insulating material would be helpful, and you should market that.

I have a stupid question also. At what temperature is an oil pan or a block heater is needed? We are moving to Northern Vermont next week, and now I am a little worried about next winter.

The sewn (sp?) up wrap around oil filter heater was only in place during heating prior to driving. I thought about going commercial with the idea but didn't think enough people would bother with the hassle since block heaters and oil pan heaters are much more effective. During cold weather driving insulating the oil filter might raise oil temp just a tiny bit. It's easy enough to try it and no harm done if it's useless. Insulating the oil pan should do more good due to higher surface area.

When it's needed is an undefined question. Needed for engine to start? Needed as in a good idea? Anytime you can increase oil temps prior to driving, there is advantage from a wear and fuel consumption perspective. The longer the heater is plugged in, the slower the temperature rise is while electricity energy ($$) spent is steady. A 150 watt oil pan heater gets REALLY hot...it will burn your hand, so I'm sure it can do some oil cooking/aging if left on long enough in high enough ambient temps, which is another reason to not go overboard with heating duration. So you have to estimate an ideal heating time duration.

TallPaul, that heater was from one of those heating pads for warming people's bodies. I took the plastic wrap off and sowed it's coils into a flannel fabric.

 

[Oilgal]

Originally Posted By: Gary Allan
Quote:
The hot engine hot oil and cold engine hot oil tests showed the lowest wear levels and the cold oil (hot or cold engine) showed significantly higher wear.


Then the vital component must be hot oil (heat). Otherwise, lower visc oils would show radically less wear than heavier oils. They simulate hot oil sooner.

..and what was deemed significant?



Say it's a cold morning. brrrrrr

If you preheat a multivis oil like 5w 30, before you start your engine, wouldn't you then be starting your 'cold' engine with 30 weight oil?

Could that really be good?

 

[Gary Allan]

Quote:
wouldn't you then be starting your 'cold' engine with 30 weight oil?


Oilgal, let me provide a missing link for you. A multivisc oil is always thicker when cold. Let me use a 10w30 as an example. It (used to be) a 10 weight oil. Now it can be a 20 weight basestock due to broader VI of newer evolutions ..but originally it was just a 10 weight oil. We assign this 10W due to it's properties @ -25F (IIRC). Now you dope it with VII that only allow it to thin to a 30 weight equivalent @ 100C/212F. You now have a 10w30.

That is, a multivisc fluid appears like a 10 weight @ -25F (which is VERY THICK) because it IS a 10 weight oil ..but only thins to a 30 weight at 212F which is (relatively) very thin in comparison.

It "simulates" fluids of different viscosities at different temperatures.

You're judging things by your perceptions at room temperature.

 

[MGregoir]

I think she is talking more about internal tolerances. Nothing has expanded to size, so you have cold parts with a thin oil protecting them when clearances are largest and everything is slapping and rattling around.

Most oil pan heaters I know of don't bring oil temperatures up that high and with a block or coolant heater, the whole engine gets "heat sinked" up to a temperature, so it is not such a big deal. If you just put hot oil in a cold engine, I could see it not being that beneficial. If the whole engine is heated, it would be perfect.