Jeez – it was just IDLING!
When I checked the cooling system this afternoon, I noticed the Midget seemed to be running a tad rough. The exhaust pulses were not quite as aggressive as they had been, so I figured I’d check the valve lash. These things are notorious for not maintaining adjustment – I’ve had issues before - and it’s dialed in a few thousandths tighter than before. Additionally, on the advice of Tom at T&T, he suggested I not crank down the adjusters as tight as I had been.
I suspected something wasn’t right, so I took off the valve cover.
This is what I found . . .
Snapped right down the middle. The keepers were still in place, and the valve was being returned by the inner spring, which actually stopped the valve from dropping out of the head.
It looks worse than what it is, but the rocker was also gnarred up through this process . . .
The roller still turns easy, and other than the scraping against the collar, the rocker appears to be usable.
Tear down, check the valve and seat, repair. Had this happened at Bonneville, I’d have put it back together and headed for the line.
But because we were at home, Mark insisted we replace ALL of the collars. His point was that these were original to a used racing head I bought 5 years ago, that we didn’t know its history before hand, and that we’d been abusing it on the dyno for better than 50 pulls, two passes at Maxton and 11 passes at Bonneville.
What turned up was what would have been the next collar to break . . . Note the stress crack at 2 o’clock . . .
So in this case, my friend Dumb Luck, once again led us down the path of preparedness and enlightenment.
I simply got really, REALLY lucky. I was at the point of putting the car on the trailer, and if I hadn’t tested the cooling system, it’s likely I’d have trashed the engine on the first pass.
And I have to confess, it still could happen. It’s the risk we take doing this – pushing things to their limit.
You can'’t count on Dumb Luck, but you can learn from him.
Parts Failure and Component Lifespan Vs. Applied Load. A short lesson on the number of cycles to FAILURE.I just want to start off by saying that most of the experienced "old hands" or "racers" who follow Midget's build diary are going to either know what I'm talking about or have seen firsthand the results of either "good judgment" or perhaps "bad choices" . . . . . . And there has been some insightful advice posted about what should be done to remedy the problem . . . . .
So, to state the obvious, engines (and racing engines as well) are comprised of hundreds, if not thousands of individual components. Some of these components are operationally stressed far more highly than some of the other components. And I don't want to invite debate about which ones are more stressed than others, because at this stage of the game for me, my mind is pretty much made up, and it would take a pretty convincing debate,
along with data, for me to change my conclusions.
Component stresses are typically analyzed based on a function of the applied load Vs the duty cycle. This is to say that as the load is increased, the number of load cycles that a part can withstand (without failure) decreases. Typically for metallic parts, this load vs # of cycles graph is inversely proportional at a logarithmic rate.
The following figure shows a typical fatigue curve for stress versus cycles to failure.
The y-axis (vertical) represents component load, normally in linear increments.
The x-axis (horizontal) represents # of cycles to failure, normally with a logarithmic scale, where the first increment would represent a value of 10, the second increment 100,
etc, usually up to a value of 10 million or 100 million cycles.
Atypically, at the intersection of the x-axis & y-axis, the x value would be 1 as opposed to the more conventional value of 0. The y-axis value remains conventional at 0.
What is important to understand about this graphical analysis is:
A/ Conventionally the y-axis load value at an x-axis value of 1, represents the yield strength or ultimate tensile strength of the part. That is:
ONE CYCLE TO FAILURE.2/ At some lower y-axis values of applied load, the x-axis number of cycles to failure is larger.
d/ Eventually, the applied load is low enough for the part to withstand 10 million or 100 million cycles. This is typically considered to be:
infinite part life.z/ Also important to note is that in high rpm racing engines, 10 million cycles might be exceeded in a fairly short time period . . . . . .
The point of all this is: Parts can be designed larger, heavier, stronger, etc,
all based on the cycle life requirements. Or the opposite can be true as well. Parts can be made smaller, lighter, weaker, etc. Obviously, in racing engines, the lightest part with adequate strength,
for the service interval, is usually the part you want to utilize. Well, except for other considerations, ie, cost, availability, suitability for the application, etc, etc. Engine engineers "walk the line" so to speak, with decisions of this nature on a fairly regular basis, and most of the time good judgments and good results occur. Once a part fails, however, re-evaluation
should take place.
SO, WHAT HAPPENED HERE?Well, it's pretty simple really. I'm hoping Chris will photograph and post up the cracked retainer halves so everyone can see how the part failed. After an undetermined number of stress/load cycles, the part failed due to fatigue from the applied load. What is scary to me is how quickly the part fractured, once the fatigue crack began. In cross section the failed part exhibits very few of the classic "beach lines" typical to a fatigue failure. So, once the crack formed, the part's strength was seriously degraded, and it failed pretty quickly. In my opinion, it really isn't important to quantify anything else here, broken is broken.
What is important to note here is that it is my opinion that there were NOT very many cycles endured after the part cracked. I base this conclusion on the underside of the rocker arm. It wasn't beat up enough to have been "cycled" very many times. Chris can call it "Dumb Luck" if he wants, I prefer to think he exercised "Good Judgment" when the engine "sounded funny" to him. It shows a "mechanical empathy" few possess.
OK, SO YOU DODGED THE BULLET, NOW WHAT?Well, some guys would replace the broken part with the intention of resuming racing. And Chris may have been willing to do just that. But fortunately he has a buddy who has been doing this for longer than I care to admit at this point in time. (And willing to make a foray into the dreaded "Sconnie Nation"/"Packerland"
) There was some "discussion", but in the finish, I convinced my friend that my racing engine rule #1 should apply in this instance. You know the one: "STOP DOING STUPID SH**!!!!!!" Unsurprisingly, (to me at least), when I started inspecting the remaining original retainers, (the very first one!!) another fatigue cracked sample was found. Discussion over!! So he was requested to go into his inventory (ahem, his used parts dungeon . . . ) to see what he could find. Fortunately, suitable replacements were in good supply and ALL the original retainers were replaced. Problem solved? Well, sort of. I would rather use some sexy titanium retainers for the application, Chris has some he bought, they just won't work dimensionally. Some specialty machined and "Tufftrided" high strength steel retainers? Yes, but we are out of time. Bottom line is: the stock retainer is highly stressed for this application, but has given a reasonable service life. And most importantly, he had them. Not the "perfect" solution, but an adequate one. Does the parts replacement guarantee success? Well, no, but I feel a lot better about it than just replacing the one that cracked.
SO WHAT'S THE TAKEAWAY HERE?In addition to engine output, engine and engine part reliability are performance factors. But, if optimum performance is your goal, reliability is compromised, to some degree. Why? Because parts that never fail are too da** heavy for ultimate performance. The bottom line becomes one where cycle life should be recorded, and where highly stressed parts are replaced at some adequate "service interval" prior to "catastrophic" failure. Part cycle life can never be "cured". (Except by overly strong/heavy parts.) Part cycle life can only be "Managed". There is an important difference, make sure you "get it". Additionally, as in this example, if one of your parts fails, you should ideally replace the others, before they "replace themselves" . . . . . . . Why? Because preventative maintenance trumps post mortem inspections, EVERY TIME.
You know I'm fond of saying that: "I make no secret of the fact I'm basically lazy". What that really means is: "I try like hell to work smarter, not harder". So should you.
Go get 'em Chris,
Fordboy
Topic: Milwaukee Midget (Read 3583716 times)