Anyone seen a gen3 piston failure like this?

[engineermike]

At least so far, it sounds to me like no one who has responded has seen a Gen3 piston fail in this manner. So maybe it was a fluke, manufacturing defect, or perhaps just on the edge of the statistical manufacturing distribution of overall strength.

As far as root cause goes, there is some very interesting data in the log. The following log is only a 2 second window containing the failure, and the channel list was limited so polling rate was very fast.

The following was the sequence of events:

46.296 (relative): Traction control intervened (this was triggered by stability control) at 7200 rpm, cut throttle to 55 deg and spark to 9 deg by 7300.

46.611: Peak rpm of 7345 reached as driver was releasing throttle and throttle was closing.

46.770: Engine speed 7226 rpm and decreasing, first indication of failure becomes evident. Driver indicated no signs of distress until after he let off throttle and log agrees. Next is debatable, but I believe the piston failed between 46.745 and 46.770 because that is when the WBO2 signals diverged. By 46.869 the STFT was reacting to attempt to correct just before going into DFCO. WBO2 response rate is only 20 ms with less than 5 ms of transport delay at that load and speed. When the piston failed it turned into a large air leak, causing the air/fuel ratio to become very imbalanced between banks. This becomes more evident after DFCO. The initial indication of this divergence is at 46.770 and subtracting the 25 ms of response time gets you to 46.745 as the actual failure time.

This seems to indicate that the failure actually occurred after throttle and timing dropped and even just after the rpm started falling.

Sponsored

 

[robvas]

Some engine builder was talking about how they most often see failures when they let off at the end of the track/pull...my YouTube memory isn't good enough to remember who it was or why it happens

 

[cbrtrx]

Some engine builder was talking about how they most often see failures when they let off at the end of the track/pull...my YouTube memory isn't good enough to remember who it was or why it happens

They might be talking about thermal shock.

 

[Gen 6 Mach1]

Mike , "Purfact Storm" Possible Casting flaw + milage on engine+ boost = purfact condition for cast piston to fail at the end of a high boosted load condition?

 

[Basspro302]

Whats the connecting rod look like? What do the rod bearings look like?

 

[engineermike]

Whats the connecting rod look like? What do the rod bearings look like?

The con rod is in a million pieces. The big end is still attached to the crankshaft and the big end bearing and rod bolts are essentially undamaged. The small end didn't fare so well, but it tried to saw the block in half using the small end of the rod so that's to be expected.

Also up for debate is whether the rod broke before or after the piston broke. My interpretation is the piston broke first because if the rod broke first, I don't think it would have pulled the pin out of the piston and left the piston neatly at the top of the bore.

 

[OldbutNew]

The con rod is in a million pieces. The big end is still attached to the crankshaft and the big end bearing and rod bolts are essentially undamaged. The small end didn't fare so well, but it tried to saw the block in half using the small end of the rod so that's to be expected.

Also up for debate is whether the rod broke before or after the piston broke. My interpretation is the piston broke first because if the rod broke first, I don't think it would have pulled the pin out of the piston and left the piston neatly at the top of the bore.

Mike, I agree with your summary about the failure. You are putting quite a bit of effort into the analysis. That said, I HAVE seen that type of failure on practically brand new N.A. Coyote engines. Yes, the piston essentially developed a fatigue point down a predictable vector between the pin bores. As Jer mentioned above, 60K miles (heat cycles), boost (Increased combustion pressures), and 7300 RPM (Piston speeds and inertia), all will contribute. Like I said above: This engine could have lasted for many tens of thousands more miles if it was left stock. The heat increase and the inertia are within limits. The combustion PSI is not. Yes, you can use cam timing to mess with that, but it's a game of diminishing returns.
I still call the Whipple 3.0 as root cause.
"When you're looking at an omelet, it's hard to tell which egg broke first." - Jack Roush

 

[Basspro302]

The con rod is in a million pieces. The big end is still attached to the crankshaft and the big end bearing and rod bolts are essentially undamaged. The small end didn't fare so well, but it tried to saw the block in half using the small end of the rod so that's to be expected.

Also up for debate is whether the rod broke before or after the piston broke. My interpretation is the piston broke first because if the rod broke first, I don't think it would have pulled the pin out of the piston and left the piston neatly at the top of the bore.

Ya it’s hard to say for sure, Im leaning towards rod failure. If the piston failed (break at the pin bore) usually you see more damage to the bore from the rod and pin beating the bore to death (I’m assuming the pictures from ben1 are the engine in question). Maybe send an email to mahle or some other piston manufacturer they may offer some better insight.

 

[engineermike]

...You are putting quite a bit of effort into the analysis.

That is true. I once did the whole "break part, install better part" routine but it didn't take me long to realize this wasn't the best way to go about things. Knock, just as an example, will break forged pistons almost as quick as it will break cast. I now know that understanding root cause and properly addressing is a better approach, and with the data we have available in modern times why not dig a little deeper?

That said, I HAVE seen that type of failure on practically brand new N.A. Coyote engines.

I would love to see pics of a stock NA Coyote piston failed in the same manner as this one. I've seen lots of broken pistons and the vast majority of them were cracked or broken ring lands, followed by melting, which is part of why I'm scratching my head.

Yes, the piston essentially developed a fatigue point down a predictable vector between the pin bores.

My issue with fatigue is that fatigue failures have characteristic beech marks and a distinct final failure location. I don't see any beech marks on this failure surface, which means it failed all at once and was not likely fatigue or damage over time.

As Jer mentioned above, 60K miles (heat cycles), boost (Increased combustion pressures), and 7300 RPM (Piston speeds and inertia), all will contribute. Like I said above: This engine could have lasted for many tens of thousands more miles if it was left stock. The heat increase and the inertia are within limits. The combustion PSI is not. Yes, you can use cam timing to mess with that, but it's a game of diminishing returns.
I still call the Whipple 3.0 as root cause.

As you state, piston speed and heat cycles are within the stock envelope, but cylinder pressure is not. I'm struggling to get on board with cylinder pressure alone causing the failure for a few reasons. His timing was lower than most with similar setups, and dropped down to only 10 deg a half second before failure, it may have happened on deceleration, and if stock NA pistons have failed in the same manner it would point to the issue not being cylinder pressure related.

"When you're looking at an omelet, it's hard to tell which egg broke first." - Jack Roush

If the piston was beat all to hell and we didn't have a data log this would be correct. But the piston was pushed to the top of the bore and stayed neatly out of the way. It didn't even separate into 2 or 4 parts until after it was pushed out of the bore and the ring removed. This preserved the fracture surface, so we have something to inspect. Not to mention we have a good data log from the failure, so I'm not willing to throw my hands up and give up so easily...yet. I tend to think that ole' Jack was talking about some other context.

 

[engineermike]

Ya it’s hard to say for sure, Im leaning towards rod failure. If the piston failed (break at the pin bore) usually you see more damage to the bore from the rod and pin beating the bore to death (I’m assuming the pictures from ben1 are the engine in question). ..

You make a good point and I'm not ruling out rod failure first, which would actually make for a "cleaner" explanation. Ford upgraded rods in the GT500 and DH, so clearly they think they're near the limit even at 500 hp.

However, my issue is that if the rod had failed first then it should leave the pin and possibly the small-end of the rod in the piston. Also, where would the force come from that split the piston in two? We saw the pin was ripped out of the piston, so if the rod failed first, there wouldn't be any force there to pull the pin out or break the piston in two. I suppose the rod failing in the middle would then allow the rotating part to swing around, imparting all sorts of damage, perhaps even knocking the pin out of the piston and breaking the piston in two. I'll have to ponder this a bit.

 

[Basspro302]

You make a good point and I'm not ruling out rod failure first, which would actually make for a "cleaner" explanation. Ford upgraded rods in the GT500 and DH, so clearly they think they're near the limit even at 500 hp.

However, my issue is that if the rod had failed first then it should leave the pin and possibly the small-end of the rod in the piston. Also, where would the force come from that split the piston in two? We saw the pin was ripped out of the piston, so if the rod failed first, there wouldn't be any force there to pull the pin out or break the piston in two. I suppose the rod failing in the middle would then allow the rotating part to swing around, imparting all sorts of damage, perhaps even knocking the pin out of the piston and breaking the piston in two. I'll have to ponder this a bit.

We need a coyote NTSB task force! It is very interesting tho. I have seen a lot of broken connecting rods and I cannot recall seeing the piston pin still in the piston. There should be evidence on the rod if it broke from being over stressed and fatigue.

 

[Mach1Racer]

Not mine, but wondering if this failure mode is common or a fluke. This occurred on e85, ~1.9-2.0 load (Whipple, 3” pulley), 20 deg timing, no knock, 7300 rpm. Compression ring lands were not broken in the traditional way.

I have been following your thread here and want to ask a few questions if that is alright.

If you know the answer to these questions…

Whipple (assuming Gen 5) with a 3.0 pulley should be making about 17-18 psi and 900rwhp (give or take 15hp).

So, was this coyote or Mustang making 900rwhp? If not, what do you think it was making?

Do you know the use of the car with this motor. DD, Rolls, Dig, Track work, 2-step, ect.

Just wondering if the way the car was used played any factor in the failure. There are camps that say 800rwhp is safe, 1000rwhp is safe, ect.

If fatigue is the consensus here from the people with the most engine knowledge, then power level, uses/treatment, tuning could also add insight to this failure.

 

[engineermike]

I have been following your thread here and want to ask a few questions if that is alright.

If you know the answer to these questions…

Questions are good!

Whipple (assuming Gen 5) with a 3.0 pulley should be making about 17-18 psi and 900rwhp (give or take 15hp).

So, was this coyote or Mustang making 900rwhp? If not, what do you think it was making?

Yes, should be around 900 rwhp, minus a little because we were limiting the timing to 20 deg and rpm to 7500. The ones with this boost level making over 900 are generally running 22 deg timing and up to 8000 rpm.

Do you know the use of the car with this motor. DD, Rolls, Dig, Track work, 2-step, ect.

Just wondering if the way the car was used played any factor in the failure. There are camps that say 800rwhp is safe, 1000rwhp is safe, ect.

Pretty much just a daily driver along with whatever fun can be had on the street. Basically never a full quarter mile pull.

If fatigue is the consensus here from the people with the most engine knowledge, then power level, uses/treatment, tuning could also add insight to this failure.

Where my head currently is, is that there is a statistical distribution of acceptable casting inclusions that align with stock loading, but if you got one of the acceptable ones on the bad side of the range, then boosting applied too much stress and it let go.

Or the rod came apart first, which aligns better with it happening at peak rpm and reversal of loading.

 

[Robottrainer]

The piston was found near the top of the stroke, still held “together” by the rings and bore. I assumed the valve contact was the valves opening against the stationary piston, which pushed it down the bore a little. Cylinder head was practically undamaged.

Rods/rod bolts tend to fail on the crank down stroke during the intake stroke. This puts the most stress on the piston pin, lower rod cap and rod bolts. What does the rod look like?

 

[Mach1Racer]

Questions are good!



Yes, should be around 900 rwhp, minus a little because we were limiting the timing to 20 deg and rpm to 7500. The ones with this boost level making over 900 are generally running 22 deg timing and up to 8000 rpm.



Pretty much just a daily driver along with whatever fun can be had on the street. Basically never a full quarter mile pull.



Where my head currently is, is that there is a statistical distribution of acceptable casting inclusions that align with stock loading, but if you got one of the acceptable ones on the bad side of the range, then boosting applied too much stress and it let go.

Or the rod came apart first, which aligns better with it happening at peak rpm and reversal of loading.

Thank for your thoughtful reply as it is sometime difficult on this forum to get what you are giving. Thanks again.

Not to be controversial since it’s clear there are certain camps on this board for and against certain people in the Coyote performance world.

Saw Street Alpha Pod cast with Joe Irwin of FFRE who works directly with a tuner people love to hate; give some insight from his experience with Coyotes while his motors are currently powering world record holding cars. Another board member posted this video on 6g.

In the video at 22:45 sec to 23:50 sec, Joe suggests 700-800 and the coyote will live for a long time. Now, he didn’t say these numbers are RWHP or Crank HP. (I would love to know the answer to that.) Joe said the rods and piston or the weak spot, and suspect the Rod will bend first!

So, just maybe 900rwhp is too much to run safely for years on end. How long was this setup in the car at 900rwhp.

I only ask these questions as I was talking to others on the board and obviously turned my setup up after a 885rwhp dyno at 95 degs and no cool down. The thinking was, “Everyone says she’ll take a 1000 anyway!”

Well, maybe she will but maybe she shouldn’t.

Sponsored