Who wants to see some tests done on the material properties of OEM Subaru pistons?

[magnusonsubie]

Original Post

 

Alright guys I have a huge favor to ask. If anyone who has had their ringlands fail and have gotten a new shortblock or still has their original pistons sitting around I am taking a properties of materials lab class right now and at the end of the semester I get to choose my own project to test the properties and strengths of a material. If I could get all the pistons from the same shortblock labeled from the cylinders they came out of I might be able to break them and see if there is an inconsistency across pistons in our engines or if the pistons are just a weak material in general. Let me know if you'd be willing to help me out with this. I will gladly pay the shipping cost for the pistons to get to me but I can't really afford more then that.

 

What was found from the tests done

Abstract

Comparing the hardness of the crown across all four pistons to observe any differences between the hardness top and bottom and where the piston four had ringland failure. Also observing the grain structures in polished samples underneath a microscope the engineer is able to see how different samples have dislocations and failed points in the cast aluminum piston. By performing these tests the engineer is able to come up with theories about why the piston had ringland failure.

Introduction

From 2005-2009 Subaru produced the Legacy GT and Outback XT vehicles in the United States equipped with the EJ255 turbocharged engine. Now that some of these vehicles are 10 years old and reaching higher mileage there are more and more instances appearing of ringland failure specifically on piston four. Many people have speculated that these failures occur due to the design of the engine as cylinder 4 is the rear drivers side which has the longest exhaust runner before the turbo. Another theory that people have come up with and tuners have somewhat proved is true as far as watching air fuel ratios when the turbo is spooling up creating boost. The pistons observed were out of a 2008 Outback that was on the stock tune for roughly 95000 miles, had an ECU (Engine Control Unit) update to what is known as stage one for 15000 miles and then the next 10000 miles consisted with an upgrade to a high flow catalytic converter and another ECU update to what is known as stage two. These ECU updates are purely a means to keep the car running correctly with proper fuel and air delivery. At roughly 120000 miles on the car the pistons tested came from was when the ringland failure occured.

 

Procedure

In this lab the independent variable was the cylinders which the piston came from. For the hardness test all four pistons were tested on the piston crown on the top and bottom of the piston with respect to how it is oriented in the engine. For the microscopy the sample from pistons were from piston three and piston four to be able to compare the failed section to a non failed section on a different piston.

 

The dependent variable in this lab was the crystal structure of each metal sample observed under the microscope during the microscopy section. For the hardness testing section the dependent variable was the hardness of the different location tested.

 

The controlled variables in this test were the same tools were used to prepare all samples observed. Also using the same microscope on the same zoom setting. For the hardness test the 1/16" diameter ball indenter and the Rockwell B scale was used to measure the hardness. 

Results

Figure 1 below shows the pattern for hardness testing the pistons. The results from the hardness test follow Figure 1 in Table 1.

 

 

Test location Piston 1 Piston 2 Piston 3 Piston 4

1 54.2 55.9 54.9 44.3

2 51.2 54.7 55.4 54.1

Table 1: Hardness of Pistons.

 

(This is just a circle showing where the hardness tests were completed, Test 1 was the top of the crown on the top skirt side, 2 was the top of the crown on the bottom skirt)

Figure 1

 

Figure 2 below shows the original known ringland failure area. Figure 3 following shows the section of ringland that came apart upon creating the microscopy samples.

 

Original Failure 2

Figure 2: Original Failure

 

Piston #4 Failed section 1

Figure 3: More failure after cutting

 

Figures 4-7 show the red dye test conducted to check for any cracks in the piston cross-section we cut away on piston three and four.

Decking #3

Figure 4: Preparing piston 3 for the red dye test

Decking #4.2

Figure 5: Preparing piston 4 for the red dye test

Red Dye Test #3

Figure 6: Red dye test on piston 3 showing no cracks

Red Dye Test #4.6

Figure 7: Crack in Ringland on Piston 4

Crack in #4.11

Figure 8: Piston four ringland failure crack

Piston #3.6

Figure 9: Piston 3 microscopy

Piston #4 Cross-Section Microscopy

Figure 10: Piston 4 Microscopy Cross-section

 

Discussion of Results

Hardness tests were performed on the top, location one, and on the bottom, location two, as seen in figure 1, of the piston crowns. The original ringland failure occurred on the top side of piston four in the area visible in figure two and more failure was found and is shown in figure three. Rockwell B scale was used for the hardness tests on all four pistons to have the most accurate data.

 

Each piston had multiple hardness test conducted in the same region that were 2.5 times the size of the penetrator away from the edge of the specimen and three times the size of the penetrator away from any 2 other penetrations. The Rockwell hardness numbers that were collected on pistons one, two, and three were from locations one and two and were all comparable to one another, the hardness numbers can be seen in Table 1: Hardness of Pistons. The hardness test numbers for piston four had a dramatic difference between the three other pistons and itself. The hardness test on piston four in location one had a similar value to the other location one tests on the three other pistons, but location two hardness number had a difference of 10.7 hardness compared to the average of the other three pistons. This difference in hardness number supplied the information that the hypereutectic cast aluminum piston from cylinder four, in test location one was relatively softer compared to the other pistons. This difference in material hardness is one possible conclusion as to why the piston’s ringland failed. The cause for this change in material hardness could be due to uneven burn across the crown of the piston, injector location, or excessively hot burn due to the factory tune lean condition when the fueling system changes over from closed loop to open loop that is pressured by the EPA emissions regulations. Conclusions cannot be drawn from this test though due to the possibility that the hardness of the piston four crown could have changed post-failure rather than pre-failure.

 

A red dye test was completed on piston three and four while cutting pieces of the piston for the microscopy sample. It was found that there was a small fracture in piston four located in the ringland, this can be viewed in figure 7. The fracture was below the original ringland failure and appears to be a large fracture that would have most likely failed as well if the piston was continued to be used. One can see the cross section of piston four and the large failure in the ringland in figure 7 as compared to what a complete and functional ringland looks like in figure 6. In figure 6, piston three can be viewed which has no apparent cracks or fractures.

 

For the next metallurgy experiment we conducted, we used cross-sectional areas from the ringlands from piston three and four to make microscopy samples from. After the samples were constructed they were only sanded and not polished. This was still more than enough preparation for analysis. The hypereutectic aluminum has silicon that is used to decrease the amount of contraction and expansion during cold starts of the engine. Piston three microscopy sample shows the silicon material that is present in the metal, which is the black area shown in the aluminum. This distribution can be seen in figure 9. Also the grain structure of figure 9 shows dislocations and grain boundaries that are non-uniform, giving the piston a greater hardness. The microscopy sample for piston four seen in figure 8 shows a extremely different grain structure and silicon content. The grain boundaries are more uniformed and straight compared to piston three, which leads to the conclusion that the material in piston four was annealed in some matter during its using. The annealing process is most likely the cause for the grain structure properties that would cause the metal to have a lower hardness number which can be confirmed with the results from the hardness testing that was conducted on the material. Figure 9 also shows a lack of silicon due to the lack of and poor distribution of black regions in the microscopy sample that represent the silicon. The lack of silicon in the piston four material would cause for the metal to expand and contract more so than an area that has silicon. In figure 8, a crack is shown with material on both sides of the crack. The crack was located between the bottom ringland and the piston skirt and can be seen in figure 10. On the left side of the figure there is a medium amount of silicon as compared to figure 9, but there is still so amount of dark spots on that side of the sample. On the right side of the ringland crack, there is almost no silicon present, and the small amount that is present is extremely light in color. This difference in silicon content in these different areas of the material could have led to the reason why the material cracked, due to different coefficients of contracting and expanding. The reason why the two different areas would have such a different amount of silicon content could again be due to an uneven burn across the crown of the piston, injector location, or excessively hot burn due to the factory tune running dangerously lean when switching from closed loop to open loop fueling parameters that is pressured by the EPA emissions regulations. This correlation would lead to annealing, and dispersion of silicon throughout the metal causing the hypereutectic material strengthening mechanism to be less effective.

Conclusions

After testing, it was concluded that the most likely source of ringland failures in the EJ255 engine can be traced to the hypereutectic cast aluminum piston composition and the close proximity of the ring package to the crown of the piston. Comparison of the collected data to a brand new factory piston would allow clearer conclusions about the microscopy data to be made and would be an excellent way to further this research. Solutions to this problem are available and commonly consist of rebuilding the engine with redesigned, aftermarket, forged pistons. The simple truth of these failures is that emissions regulations have pushed automotive engineers to work with ever narrower margins of error and this leads to these unforeseen problem areas in design.

 

References

"How to Avoid Blown Ringlands in Your Turbo Subaru." Viking Speed Shop. WordPress, 20 Aug. 2015. Web. 09 Dec. 2015.

 

Tsuneishi, Scott. "The Truth Behind The Subaru EJ-Series Engines."SuperStreetOnline. N.p., 04 Feb. 2011. Web. 09 Dec. 2015.

Special Thanks

Thank you to OB2.5XT on the legacygt.com forum for generously donating his used pistons for this lab to be completed.

Stan Hitchcock for helping with the machining and red dye testing of the pistons.

Edited December 9, 2015 by magnusonsubie

[Nightmaresmk]

I have a set in my spare block. But OEM Subaru pistons aren't that weak they take some serious abuse but are to tight on ring clearances.

[Rhitter]

Gunnar, you are in luck, I have at least one of mine. I need to check if I kept all of them or just the broken one though.

 

Do you drive through Oakland / Berkeley ever?

[magnusonsubie]

Gunnar, you are in luck, I have at least one of mine. I need to check if I kept all of them or just the broken one though.

 

Do you drive through Oakland / Berkeley ever?

Sometimes I do and that isn't far from my campus so I wouldn't have a problem swinging over on a weekend or something.

[Rhitter]

Sometimes I do and that isn't far from my campus so I wouldn't have a problem swinging over on a weekend or something.

 

Yep, I have all 4 of them. PMd

[GTTuner]

I would be interested in the properties of a NEW OE and a 100,000 mile OE piston. I swear they get brittle around 80-100K miles.

[Rhitter]

New OE is pricey.

 

Old 120k mile pistons that I am giving him are free.

[monkeyposeur]

I have a set in my spare block. But OEM Subaru pistons aren't that weak they take some serious abuse but are to tight on ring clearances.

 

The ring gaps aren't much different from forged pistons. The PTW clearances are much tighter on OEM vs. forged though. .001 or so for OEMs and about .003 for forged. Different piston material for each so different expansion rates.

[GTTuner]

I have a new OE and an 80,000 mile one with a cracked ring land if he wants it.

[magnusonsubie]

I have a new OE and an 80,000 mile one with a cracked ring land if he wants it.

If you're willing to ship them to me or at least the new OE one would be awesome. I talked to my professor about it yesterday and even he is interested to see what we find.

[GTTuner]

If you're willing to ship them to me or at least the new OE one would be awesome. I talked to my professor about it yesterday and even he is interested to see what we find.

 

PM me your address I'll get one out to you. Just a new one?

[snm95ls]

As someone who is an amature met nerd, what do you plan to analyze?

 

It would be nice to see grain structure all across the cross section of the piston especially near the ring lands.

 

I ASSume mostly destructive testing?

 

I wanted to do the same thing while in college, but with different parts for a different engine.

 

[magnusonsubie]

Talking with my instructor he suggested doing brinell and Rockwell hardness tests on different sections of the pistons to see if there's a difference between where the ringland failed and the opposite side. And if there is enough material present we will machine it down and try to do a Charpy break test on it.

[snm95ls]

Talking with my instructor he suggested doing brinell and Rockwell hardness tests on different sections of the pistons to see if there's a difference between where the ringland failed and the opposite side. And if there is enough material present we will machine it down and try to do a Charpy break test on it.

 

If you are both up for it, have proper equipment and there is enough material left, try take a section, lap it and etch to reveal the grain boundaries and stick it under a microscope.

 

Basic outline of the procedure:

 

http://vacaero.com/information-resources/metallography-with-george-vander-voort/1217-metallography-and-microstructure-of-aluminum-and-alloys.html

 

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&cad=rja&uact=8&ved=0CCcQFjABahUKEwiF9-Gsp5XIAhUCCz4KHXOoBjk&url=http%3A%2F%2Fwww.researchgate.net%2Fpublictopics.PublicPostFileLoader.html%3Fid%3D53ea4a31d3df3e9e348b4611%26key%3Dd5a4c514-d3d2-46f5-b593-4ad24fa08287&usg=AFQjCNEGx9tnpiuAWH773JnKz2NNFmvhoQ&sig2=MN-y0yalkfCKbxBgo4qIpw

 

If not, the tests you have planned out ought to reveal some interesting data.

[fahr_side]

The ring gaps aren't much different from forged pistons. The PTW clearances are much tighter on OEM vs. forged though. .001 or so for OEMs and about .003 for forged. Different piston material for each so different expansion rates.

 

 

Remember there are tolerance ranges on all parts. We have seen some factory motors with really scary gaps on the top rings, way smaller than the recommended gaps on aftermarket piston kits.

 

 

Sent from my iPhone using Tapatalk

[rapture]

Subscribed

[magnusonsubie]

Alright small update this morning with definitely more to come in the next few weeks.

 

This is the pistons given to me by OB2.5XT and pictures of where he had the ringland failure.

 

http://images.tapatalk-cdn.com/15/11/20/31fb9b045e5d5929894552d36bcc91bd.jpg

http://images.tapatalk-cdn.com/15/11/20/aa964cca53466486820c07b13e23247d.jpg

This is Piston 4 and was near the top of the cylinder.

 

We've now finished all of the in class labs we were required to do and now it is time to do our final projects. I did the first part I planned to do this morning and that was to clean off the piston crown, and then Rockwell "B" scale hardness test the pistons generously given to me by OB2.5XT. I started with piston 1 and because of the design of the machine we were only able to test the crown on the areas where it could be supported by the piston skirt.

 

This is Piston 1 in the hardness testing machine.

 

http://images.tapatalk-cdn.com/15/11/20/152ebc4b07de9b58c5608bbc5b83077f.jpg

 

 

The first 3 pistons had a rough average of around 55 HRB both top and bottom. Piston 4, the one with ringland failure near the top side piston skirt, had a rough average of 45 HRB and around 55 HRB on the bottom side.

 

So far that is all that I've done but I was surprised to even have that much of a difference from a simple test like that. Next up is to cut up a section of the piston and set it into a puck and polish it to view the grain structure under a microscope to see if there are any imperfections or dislocations in the material.

[snm95ls]

Awesomesauce!

 

So glad you guys are going to examine the grain structure as well.

 

 

[mondtster]

Remember there are tolerance ranges on all parts. We have seen some factory motors with really scary gaps on the top rings, way smaller than the recommended gaps on aftermarket piston kits.

 

Which is why I've always questioned why people on this board seem to push buying a replacement shortblock from Subaru, rather than assembling an engine with tolerances more suited to a performance application in mind.

 

I suspect that Subaru knows that their ring gaps are tight. People need to remember that these engines and cars were designed with a purpose in mind, and that was to be a passenger car, not some high performance machine. Things like emissions and oil consumption are far more important to them than reaching high marks for durability (although too low will cause a loss of customers).

 

The area of the piston that breaks is where the high load bearing area is.

[Rhitter]

Which is why I've always questioned why people on this board seem to push buying a replacement shortblock from Subaru, rather than assembling an engine with tolerances more suited to a performance application in mind.

 

I suspect that Subaru knows that their ring gaps are tight. People need to remember that these engines and cars were designed with a purpose in mind, and that was to be a passenger car, not some high performance machine. Things like emissions and oil consumption are far more important to them than reaching high marks for durability (although too low will cause a loss of customers).

 

The area of the piston that breaks is where the high load bearing area is.

 

Because a lot of engine builders suck. And there were a couple high profile / high money (aren't they all?) builds that failed.

 

So the best advice this board has is stay with the OEM shortblock.

 

Look at Boxkita's head rebuild / engine whatever the **** is going on fiasco.

 

After talking with my engine builder I went forged.

 

BUUT look at Sgt.Gator he actually drives a racecar and was on the OEM short block for a long time. Might still be, don't remember how he rebuilt it.

[fahr_side]

I think it's just stubborn and foolish to install a new factory shortblock without checking the ring gaps and resetting where necessary. It's usually only the top ring that's too tight.

 

 

Sent from a device using some software.

[mondtster]

I think it's just stubborn and foolish to install a new factory shortblock without checking the ring gaps and resetting where necessary. It's usually only the top ring that's too tight.

 

That was kind of my point. There's nothing wrong with a stock shortblock, but make sure the clearances are set for the kind of use the engine will see.

[xt2005bonbon]

I think it's just stubborn and foolish to install a new factory shortblock without checking the ring gaps and resetting where necessary. It's usually only the top ring that's too tight.

 

 

Sent from a device using some software.

 

Just to make sure, one can do that without splitting the block right? I think I did see a youtube video where a guy was able to do so.

[snm95ls]

Correct.

[Zac88]

Just to make sure, one can do that without splitting the block right? I think I did see a youtube video where a guy was able to do so.

 

Link to that video?