1986 325iX 3.1L Stroker + Turbo

[digger]

I've seen this before on short skirted pistons with the wrist pin moved up to accommodate a longer rod like these. There is no excuse for that failure, though.

You did good stopping yourself and taking time on the diagnosis, and saved yourself big time in the long run.

i wonder if these were these specced as a NA piston that was later adapted to FI. they look awfully featherweight

[downforce22]

holy s..t was there any striations indicating the crack was a fatigue crack? it didnt look like it from what i saw.

that pin wall looks awfully thin for a FI engine im sure its tapered bore but still for FI dont tick the light weight wrist pin box, part of the job of the pin is to stiffen the piston, it acts like a spine because steel is 3x the modulus of the piston material

was the pin a sloppy fit on the bore on the unbroken side of the piston?

the pin fit within the piston nice and smooth, it didnt feel like it was slapping around. I couldn't tell any stress cracks or striations in the piston. If there were some, it would seem the pin would have broken through it.

Damn, that sucks but you doged a MAJOR bullet. I love this thing, but it needs OEM tails badly. Those ones need to be burned to the ground

Thanks, I was rocking the oem tails for a while but I like the look of these better, I think it fits the theme of the car a bit more.

I've seen this before on short skirted pistons with the wrist pin moved up to accommodate a longer rod like these. There is no excuse for that failure, though.

You did good stopping yourself and taking time on the diagnosis, and saved yourself big time in the long run.

Yeah I am curious to hear what JE says about the piston. It is frustrating when I pay for the pistons and it has a material failure like this. But like you said, glad it wasn't worse.

Very nice build. How is the front diff holding up? Really want to eventually turbo my ix but people seem to say they are basically made of glass when you start adding power.

I have heard the same and done some research on this. There are a few others who turbo's their iX without issue, including user QuintQ5 and ss454 whose build threads were on e30tech before it went down. Quint's car was a stock iX with a turbo making enough power to slip the clutch on an old thunderbird turbo. ss454 had a set of 3.64 diffs and didn't have problems. Right now with the stroker and turbo I am sure I am above 200 horsepower but as has been discussed, torque is what breaks things.

Here is a thread with some resources, but I also picked up a spare 3.64 front diff from ebay.de just in case I do have an issue. http://www.r3vlimited.com/board/showthread.php?t=390800

I basically did this with the 3.64 diff and plan not to launch the car in 1st gear which is what would break it.

And since the transmission is a torque multiplier, you should be able to run more power through the tcase in higher gears.

165 hp * 3.82 (1st gear ratio) * 0.33 (% of torque to front wheels) = 208 ftlbs at the front diff in 1st gear
vs

165 hp * 2.2 (2nd gear ratio) * 0.33 (% of torque to front wheels) = 120 ftlbs at the front diff in 2nd gear.

With a little math, the 2nd gear equivalent of the torque given to the driveline in 1st gear is:

X * 2.2 (2nd gear ratio) * 0.33 (% of torque to front wheels) = 208 ft lbs
some math and

X = 286 horsepower

What does this mean? Assuming you don't launch the car in 1st gear, but roll onto the power, the diff can see ~300 ftlbs of torque in 2nd gear or higher. That's been my basis on my turbo ix. Plus you don't usually see much boost in first gear so it isnt a scenario you would see unless you were trying to drag race or something.

A couple pictures of the car/ pulling the car apart.






So here are a few pictures of the different pistons I have. First is the broken piston (JE, 86 mm made for s52/m54b30 crank), and the Ross Racing forged piston (84 mm, made for s52/m54b30 crank). The Ross pistons look like there was no care put into them while the JE pistons have a few bells and whistles like the anti flutter grooves and oem style dish. The JE piston is for a 9.5:1 CR while the Ross is made for a 9.75:1 CR.









And the original pistons that came in the engine, these are the high compression m20 pistons 9.7:1 pistons on the non cat motronic 1.0 engine. Stock 135 mm rods. and it is amazing how heavy the stock pistons and rods are. It is such a large difference in size vs both of the forged pistons.





And showing off that shiny exhaust and that blacked out tail end. Stock ride height and the exhaust is disconnected so that is why the muffler is hanging down low.



I should be back boosting in about 2 weeks.

[tschultz]

I also think the taillights look good on this car. The whole theme of the car is something unique and different, almost rat rod, but not rusted on purpose or something silly like that. How often are early e30's even seen? What about 3.1L M20's? And further yet a turbocharged 325ix? 3.64 gears? I didn't even know they existed for the 325ix...

Add it all up and it is a unique car in its own right. Now put on E28 rear spoiler, BBS front valance, and now the taillights make sense.

Add those together and it isn't your typical slammed e30, it isn't trying to blend in or follow the stance bros. The tail lights suit the car in my opinion.


Regarding the pistons, is there any way you can get some confirmation or clarification from JE about the type of failure (ie material failure as opposed to a design failure)? I read online they use FEA in their design to reduce weight and improve integrity before even forging parts.


Maybe you can pitch to them that you are concerned about the long term viability of the remaining pistons considering the type of failure. If possible for peace of mind, see if you can get a simulation result of the piston in their typical loading condition. Maybe with something simple like a factor of safety plot showing the location in question. This would provide you with some validation as well as peace of mind.


It also shouldn't be something that would be difficult to obtain, especially if they already do it with their modeling software.

An interesting article:

http://image.superstreetonline.com/f/34070732+w640+h427+q80+re0+cr1/impp-1109-07-o%2bpiston-anatomy%2bfea

From another website:


And another sort of interesting article. On page 5 they talk about side loading: http://www.motoiq.com/MagazineArticl...VE-Engine.aspx

[downforce22]

I have an update on the car.

Long story short, JE's 'analysis' claimed the piston installation caused the problem. Why that would take 3000 miles and a turbo to manifest, I have no idea. I didn't think they would do anything for me, especially considering I already bought the replacement.

But i was able to get the new piston back in and finally work on my tune a bit. The car started and drove alright but I needed to work on the tune and was not used to the z3 short shifter and 6 puck clutch.One day while showing the car to my buddy I head fuel spilling onto the ground. It turned out the previous owner never replaced the fuel line eventhough he put the braided steel line over it so it was cracked and brittle due to age.

I replaced that and drove it the next day. The car did great on the highway and it really is a highway cruiser with the 3.64 diffs and the turbo. In 5th you are at the perfect rpm to move into the passing lane.

After showing the car to my brother, he pointed out that my idle pressure was at 100 kpa, which was impossible at the 5500 feet of elevation we were at. I should have thought about that sooner, but this weekend I finally took a look at it. It turns out the barometric sensor was set to the mpx4250 for a 2.5 bar reading, however when i changed the sensor to preset mpx4115, it was reading the expected 80 kpa. This of course changed my fuel and ignition maps so I adjusted them and got the car idling and running. Now the autotune feature seems to be working much better.

So now I am stuck with a 115kpa sensor and running 10 psi of boost here in colorado springs would put me near 160 kpa. i can try to go to the mpx4250, the GM 3 bar, or the MAP Daddy with barometric correction. I'm thinking the baro correction would be useful in the colorado mountains. Is it a direct swap, just by soldering the 3 new connections in? Either way, she is back on the road and boosting just in time for winter. I'm still waiting for the first snow but glad to have it close enough to daily for the time being.

[digger]

Can you cut and paste word for word what the manufacturer is claiming was the cause ?

i can only think the notch for the circlip was gouged badly during install

[downforce22]

Can you cut and paste word for word what the manufacturer is claiming was the cause ?

Component Failure Investigation

Background:

Customer states slipper breakage on one side of the pin tower. We received 1 piston, 1 pin, and 2 circlips. Customer explained that he found the failure after hearing RPM dependent ticking noise and he continued running it until he found the source of the noise.



Measurements:

All dimensions related to the lock grooves were measured as well as the locks themselves and the pin. All components were in-spec except for the lock on the failed side of the pin tower. Details of the inspection dimensions can be found attached to the request form.

Observations:

Examining the fractured surface under slight magnification shows fatigue arrest marks that indicate a crack that initiated around the area of the pick lock groove. There also seems to be an initiation point on the other side of the pick lock groove around the wire lock groove area.

We also observed a deep groove on the opposite pin tower that indicated that the pin ran significantly shifted over in the pin bore towards the side of the failure. The arrow in the picture

Conclusion:

It is my opinion that the order of events is as follows. First, the lock was ejected from the piston causing the pin to slide over in the pin bore. The pin was halted by the rail spacer and the engine continued to operate. Because the pin was only very slightly engaged on one side of the pin bore, the large amount of stress caused compressive plastic deformation in the pin bore. With every engine cycle, the pin became crooked in the pin bore and caused a large amount of stress on the very outside of the failed pin tower area where a crack was initiated at the highest stressed area (the pick lock groove). The pin tower eventually failed as seen in the received sample.

The ejection of the lock was likely caused by improper installation because there were no obvious defects seen in the manufacturing process or in the design. The customer should refer to the instructions in our installation tips document provided with all pistons for a guide on how to install the wire locks (document is in this report packet).

[digger]

Was the lock in the groove when piston was removed? It probably would have fallen out in any case.

You might be able to argue that the chunk falling out caused the lock to be ejected ?

[downforce22]

Was the lock in the groove when piston was removed? It probably would have fallen out in any case.

You might be able to argue that the chunk falling out caused the lock to be ejected ?

I'm not sure, but I'm pretty positive they won't do anything for me after talking on the phone about it.

I have spent some time working on the tune. For now I adjusted my MAP sensor and have it set for the 115 sensor which reads up to 120 kpa according to megalog viewer.

The idle settings were off a bit causing the car to have an idle that increased above 1200 rpm. Once I reset my idle positions that fixed part of the issue but i still would get a bit of an issue with the overrun fuel cut on which would cause the car to stall randomly. Adjusting the fuel cut set points fixed this issue. I adjusted the spark settings a bit but currently see the idle drop when coming to a stop that occasionally stalls the engine. I am still working on fixing it.

I also smoothed the Ve table and my target afr table and it resulted in a decently fuel efficient setup. My OBC was reading 5.7 L/100km (double that since injectors are about doubled in size) while tuning, so 13.4 L/100km or 18-19 mpg and today after some more tuning, minimal boost, and overrun fuel cut on, I am seeing 5.0 L/100km or 23-24 mpg at cruising speed. That's a 30% increase!

Here are my new maps and a datalog showing what I've got. Keep in mind the engine is 3.1L M20 with 9.5:1 cr and stock 885 head/camshaft.







Looking at these, you cn see the fuel and spark map are pretty smooth. Keep in mind I am at 6,000 ft so the atmospheric pressure is ~80 kpa. The engine is a stroker so keeping RPMs lower is ideal to reduce wear on the cylinder bores and keep piston speed within reason. The engine has a lot of low end grunt already and with the high CR it has good response. The turbo is helpful up hills and in passing lanes and is barely working too much most of the time. The stock cylinder head and cam keeps the power down low in the rpm range so it is ideal at highway speeds then the turbo crams more air in above that.

You can see the hump of the VE of the engine. What I am unsure of is how the stroker crank affects timing. Obviously I can run more advance at cruise to ensure a more complete burn, but does the longer stroke warrant more advance when at higher rpm? I want to make sure i am not knocking so I haven't gone above 31 degrees and am running 91 octane while running in cool temperatures as a safety factor. I want to get this dialed in before temps go up in the spring time. Should my spark map follow the general curve/shape of the VE table?

You can also see on the datalog some good info. The MAP sensor maxes out at 120 kpa which is registering 5.5 psi of boost. The boost comes on quick and the engine revs so fast it is barely usable. You can see manifold temperature was about 65 F and the AFR as boost hits is 12.9. I probably should richen it up to my target of about 12.5, but should I go closer to 12.0? And with it being this fast barely touching 5 psi, I'm not sure I want to see what 10 or 12 psi is like!

I think my wastegate is cracking early so I may add a manual boost controller. I also want to check the MAP sensor and add a cold air box around my air intake. Aside from that I am going to start working on the cold starts with this cooler weather and add EAE for throttle transitions. My TPS has poor resolution with the throttle cracked so I may also put in a new one of those that i pick up at one of these cyber monday sales. It would be great to use TPS based EAE but I would setting for MAP based if it helps my cold start near stalling set-offs.

Oh and here is a brief video of the sound out the back and the led brake lights in action.

[downforce22]

[Good & Tight]

Good thing you pulled the engine apart and not continue to drive it.

The timing looks aggressive in mid range. What fuel are you running?

[Nisse Järnet]

Nice! :)

[downforce22]

Good thing you pulled the engine apart and not continue to drive it.

The timing looks aggressive in mid range. What fuel are you running?

Thanks for the feedback. I am running 91 right now.

After looking at some timing maps from other turbo cars, I noticed they all drop about 3-5 degrees of timing once boost comes in above 100 kpa. I did the same, just starting at about 80 kpa due to my altitude. Instantly the engine response dropped a bit and i had to retune the fueling values. But sometimes I would hear a bit of a rattling as boost built that I thought was loose change or something.

That's gone and I now know is the sound of detonation. Another description I read was the sound of gravel hitting the firewall. That's not quite what I would call it, but still a good idea of what to listen for. I dropped the 80 kpa zone to about 25 degrees and it steadily drops on my map to about 16 degrees at 200 kpa (eventhough my map sensor only goes to 120 kpa at the moment). I read through KAMotor's turbo thread and he popped his headgasket at 17 psi and had timing at 19.5 degrees instead of 16 (with water meth injection) on a precision 6262. That gives me a rough guideline for my map that at 200 kpa I should be around 15 degrees. Withmy new settings, the car is a bit slower, but there is no detonation as boost builds up to 8 psi and it is much smoother through the rev range.

The tune needs a few other issues sorted that I can't figure out.

-I set up fuel cut which seems to be working great. However, if I want to rev the engine while idling, the engine will go into fuel cut as engine vacuum drops below 20 kpa, above 1500 rpm, tps below 1% and triggers fuel cut on the decel. Is there a way to allow this to not go into fuel cut? Maybe the solution is just to turn off fuel cut. Minor detail, sure, but if someone wants to hear the sound, it is annoying to have it cutting in. Maybe I need to adjust the delay time.

-I got the CLIdle set up good and the car will go into closed loop idle rather quickly like it should, but when coming toa stop and clutch in, sometimes the car will stall. I have bumped up the spark advance at 500 rpm and adjusted the idle stepper values to the appropriate values. This helped, Maybe I need to check that it is not related to the fuelcut not supplying fuel fast enough.

-Target idle RPM - Going along with the CLIdle values, the target rpm is always close to my target. After startup, the car idles at like 500 rpm and then gets up to like 700 after maybe 20 seconds. It shoots up high due to the cranking idle pw settings I have but instantly drops lower. It would be nice to have it idling at 900 rpm until it warms up.

-Some times when coming to a stop, the idle is closer to the target 750 rpm, but other times it is at like 650 rpm. Any ideas how I can get this closer to what I set in in the target rpm curve?

[tschultz]

Good point on the timing in 100-180kpa region, much more aggressive than I have ran yet.

This is super retarded timing in that region when I was running on low octane. I have since bumped it up in that range but haven't been able to test/tune my car recently.
Mine is 8:1 M30B34 though so probably quite different.

[downforce22]

I messed with the tune a little and found the stalling when coming to a stop disappeared when the Over-Run Fuel cut was disabled. It seems to not be bringing fuel back soon enough when I put the clutch in. I will have to mess with those settings to see if I can find a happy medium. I still need to pull the megasquirt unit apart and see what MAP Sensor I have installed. The sensor only reads 5.5 psi now but I am running 8 psi of boost so it is going off the AFR and fuel setting that I have at 5.5 psi.

To top it off, I was comparing the 5862 to garrett turbos to see if I could compare the turbo compressor map of a know Garrett to the PTE 5862 turbo I have. I found the GTX3576R has almost the exact same compressor wheel as mine. The main difference would be the Garrett is running a dual ball bearing instead of journal bearing and also has water cooling. The Garrett has the upgraded wheel and my 5862 has the upgraded billet wheel. So obviously we are just looking at the compressor side for now, but I use the BorgWarner MatchBot to calculate these values and plotted several combinations on the GTX3576 compressor map. This is particularly useful to me because I live at 6,500 feet but regularly go to 5000 feet in Denver, and can get to over 11,000 feet on the various mountain passes in the area (let alone pikes peak at 14,000 feet). I plotted the different scenarios and used a base of 8 psi at 70 F. The MatchBot allowed me to incorporate 6 different points on the map with different rpm, VE, boost, and pressure drops. To make it simple, I kept the values the same besides the altitude and one scenario where I upped the boost.







The compressor map shows me what kind of efficiency I would be at for these four scenarios. The blue line represents sea level. The green line is my current elevation in Colorado Springs. The orange line represents me going over Vail Pass at 10,000 feet. The pink line is if I were to up the boost to 12 psi here in Colorado Springs.

I put these 4 scenarios in to see what would happen as I drive to different elevations. Currently, the green line shows I hit peak turbo efficiency between 3500 and 4000 rpm. This seems accurate to what it's like in the car. I attached the MatchBot values here to see what kind of power levels I would be at during daily driving. Note, I plotted corrected airflow, which should be related to the temperature and altitude. As I travel to Vail Pass at 10,000 feet, the air density goes down and my pressure ratio increases, moving me up the map. The turbo has to move more air and therefore spin faster to keep up, moving the plot up and right. This is a good thing in this case because it moves me to a wider cross section of the peak efficiency island. If I were to drive down to sea level (the blue line), I may make more horsepower, but the 5862 doesn't have to work as hard so it moves the line down and left. Air density goes up, pressure ratio goes down. Now I am no longer in the highest efficiency! For this turbo to be ideal for this engine at sea level, I would need to pump up the boost to hit the max efficiency.

Now, one final scenario is going to 12 psi at my current elevation (pink line). With the same air density as the green line, I pump more boost into the engine and now my pressure ratio increases to 2 and I am in the 75% or better efficiency zone all the way from 4000 rpm to 6000 rpm. Notice I am taking advantage of the widest portion of the most efficient island at 76% instead of breezing through it like the 8 psi green line. I can expect to be between the orange line and blue line for all running conditions wherever I drive the car which is pretty good. The turbo is a good match for the engine (3.1L at 70 F). So about the pink line, and raising boost to 12 psi... Doing this makes logical sense to increase the boost to 12 psi.. its just optimizing the engine. :giggle: potstir

[tschultz]

Good exercise on this one.
It appears your turbo is matched pretty well and the altitude does at least help the pressure ratio increase to hit that efficiency island at lower boost pressures.

Being above 70% efficient is good regardless, so although there are gains to be had I would be real careful with tuning your car (9.5:1 CR?) to those levels without a knock sensing unit.
Keeping an eye on intake temperatures will help you to understand if you need to be more concerned with turbo sizing or simply intercooling efficiency (probably something I can improve with my setup).

Maybe it is worth talking to somebody like Lars with a turbo S50/S52 (10.5-11:1 compression ratio engine) and find out what boost levels those guys run here at altitude, along with what octane fuel. Also did he dyno tune it or street tune it? I'd feel more comfortable with a dyno friend when upping the ante on intake charge pressures.

If you look at boost levels in terms of kPa, I don't think ~12psi at altitude (170kPa) is really that crazy. But the guys at sea level usually seem to have access to 93 octane fuel (or E85) to give that margin of safety at 12+psi boost levels.