PeachParts Mercedes-Benz Forum - Rod Bender True Stories

PeachParts Mercedes-Benz Forum (http://www.peachparts.com/shopforum/)

- Diesel Discussion (http://www.peachparts.com/shopforum/diesel-discussion/)

- - Rod Bender True Stories - How it Broke and Why (http://www.peachparts.com/shopforum/diesel-discussion/217243-rod-bender-true-stories-how-broke-why.html)

The only fly in the ointment is that folks with the uprated rods have reported bending as well.

Tom W

I tried to point out that making the rod heavier is not a 1 for 1 type of gain. At speed the loads are higher and it is not clear the heavier rod is the answer - a different configuration with the same net result - higher loads and not enough margin to the fatigue limit. Some still fail. If you don't drive over 4,000 rpm, maybe the odds are far enough in your favor that they won't fail.

Bigger is not always the right answer. Jim

Fatigue is to do with crack initiation and growth - were the old rods cracked?

If the rods have bent, but have not cracked, it's not fatigue.

It's possible for struts loaded in compression to fail via a buckling mechanism - in engine connecting rods, this compressive buckling is aided and abetted by the sideways inertial force as the big end accelerates side to side, and the moment caused by bearing drag. These two extra factors can make the rod bend elastically, and then as the compressive load is applied as the cylinder fires, the critical buckling load is reduced.

This is all fairly well known info, and I would be very surprised if MB didn't take it all into account. However, did they try to stretch an existing design beyond a sensible limit - was the margin against buckling for these rods as high as on other engines?

I've never even seen one of these engines, so, a few questions spring to mind;

In which direction are the rods bent? If the block were made of glass, would you see the bend looking from the front, or from the side?

Does this engine tend to knock more than most MB engines?

Are these the same rods as used in 3 litre engines mated to a longer stroke crank and larger pistons? Was the rod section increased? - the stiffness will go up with thickness cubed, so even a small increase in the "beam" section will make a big difference.

The idea that fatigued rods would have cracks seems right. That i show forged steel fails by fatigue, right?

The momentum of six rods and flyweighted crank moving in an engine a 3500 RPMs will be hard to stop. If you can get enough fluid into the cyclinder/combustion chamber in one stroke then........

I inspected Jack's car again yesterday. There is so much oil in the exhaust system that a few drops formed at the tail pipe.

I believe once they bend the loads go down significantly. Lower compression, poorer firing conditions, more blow by and they move out of the fatigue life threatening loading environment. Thus only mass and stroke, thus the speed remain the same.

In most cases the oil consumption rate quickly rises to alarming levels and most owners stop driving them. I have yet to hear of one failing by cracking. The failure is a twising and bending, I believe, although to be honest all I have seen is the different heights of the pistons at TDC as the engine is rotated by hand with the head off.

I also do not think the space is available for thickening where it would be most effective. The added stroke in the same basic block as the prior 3.0 liter engine makes it a challenge to get the section thickness you want where you want it.

"Fatigue" may be an inaccurate engineering term, however, it seems that the phenomena takes place only on engines with a substantial number of miles, and then, if you get past about 180,000 or more, the engine is immune to the problem. This leads one to believe the problems is not a first order design failure.

The issue does seem to be related to loads and cycles. And, some of those who have seen and posted shots of the old vs. the new rods have said they are visibly stouter in the new configuration. There are rumor there have been several iterations of design changes, and it has been reported that at least one new rod design didn't fix the problem. Jim

PS: This engine was smooth as butter and had next to no knocking when it was right. When it went bad it had an obvious idle speed variation, and would rock the car, visibly. Oil dripped out the back of mine as well. Might have had something to do with a quart every 100 to 200 miles going out the exhaust as mostly vapor and then sort of condensing near the outlet.

In the background of this shot there is a beige 350sdl. The new owner got a smokin' deal (pardon the pun) and is looking for a new engine for a perfect, (I mean really perfect) body. I suppose the 3 litre from another SDL would be the best replacement. This is at my indy friend's shop. I got to start it up, yikes! She still moves, barely. You can tell from the discoloration of the back tail light how bad its gotten.

http://i251.photobucket.com/albums/g...0/DSC01453.jpg

Bad 350...This is what they run like when they're ailing...

http://www.zippyvideos.com/6161441694710296/350klocker2/

Good 350...This is what they're like when they're only burning 300 miles per qt, as opposed to 100.:D

http://www.zippyvideos.com/2773151204709756/350toprun/

Any diesel I drive receives no mercy from me.
:cool:

Yeah, nobody knows. But at least one can confirm, the 3.0 liter was a solid design.

Maybe if the timing was altered, a 300 wouldn't lag so much compared to the 350? I actually happen to have a 350 pump lying around, from the first car in the first video (block makes a nice paperweight). Planning to have it modified one day.

Eventually, the problem seems to parallel a head gasket breach...Except it gives you 10's of thousands of miles of forewarning before it needs attention. Save up some $$$$ in the meantime.

If the bending were from liquid in the cylinder the lowering of the compression ratio would lessen the likelihood of it happening again, and after wearing for a while the compression is less too.

I find it hard to believe with all of benzes experience in engineering diesels that the straight engineering problem of making the rod strong enough would be what they miss on.

The block flexing too much seems a more indeterminate problem and easier to miss on.

I would love to see the two rods side by side to see what the difference is.

Tom W

>>the straight engineering problem of making the rod strong enough would be what they miss on.

Didn't Cummins make engines with dodgy pushrods for years? (The story I was told was that they were told, in no uncertain terms, to sort it out while being considered as an engine supplier for a European truck, the Ford Transcontinental)

The more serious point I would make is that when designing parts, there are margins allowed - one of the contributing factors in this margin is uncertainty - uncertainty over the exact load the part will bear, and uncertainty over the strength of each individual part. It's entirely possible that MB had under-estimated their ignorance of either or both of these uncertainties. These errors are usually picked up during a development program, but, if these failures occur sporadically, and at higher mileage, it's entirley possible that they would be missed, even by a rigorous testing program. If the 350 was a development of an existing design, it's also likely that the development program was not exhaustive, as many of the parts could claim prior development heritage.

Buckling is actually an unusual case to design for, as the yield stress of the material is much less important than the elastic modulus. In other words, as the Young's moduli of most steels are very similar, there's no point in spending lots of money on expensive steels with high yield stress values if buckling is the only load case under consideration. Of course, a connecting rod bears loads other than the compressive or buckling case, and for these, yield stress is important.

Quote:

Originally Posted by Number_Cruncher (Post 1804452)

What is your thoughts on the flexing block? Are you an engineer? Automotive? It sounds like you have some expertise here.

If you are a rod designer, I would like to ask some questions:
1. What are the design loads for compression that is created by the compression stroke of the engine?
2. What are the design loads of the firing stroke?
3. What are the design load of the piston thrust?

Now the next questions is, if you are a rod designer, how do you increase the strength of a given rod?

My understanding of structures is from building design and we typically use a pretty large safety factor on columns, often overdesigning them by a factor of 2 or 3 simply because it costs very little and weight of the column is inconsequential.

A connecting rod it seems could be strengthened by increasing the web or the flange or both. This can be done without increasing the overall size by increasing the section toward the inside of the rod structure. I recently held 616 and 617 rods in my hand and it is not readily apparent the difference. I took them to my favorite machinist and asked him to look at them. He spotted the fact that the wrist pin is larger on the turbo pistons and the area around where the bearings fit is slightly thicker. At casual inspection they look the same. The outside shape looks the same, they are the same distance from center to center too, I think.

The extra mass in the 617 turbo rod does not seem to affect the operating rpm of the motor, IIRC.

The operating rpm of the 3.5 603 on the other hand is kept lower by design, presumably because of the heavier (presumably) weight of the rods and pistons and perhaps the crank.

I would love to see the differences between the 603 3.0 rods and the 3.5 rods and then the "upgraded" 3.5 rods as well. Does anybody have information on the differences?

Tom W

Quote:

Originally Posted by Number_Cruncher (Post 1803928)

If I understand it correctly they bend straight down and twist. The twist creates a binding situation against the cylinder walls since the top of the rod is no longer parallell to the crank. This wallows out the cylinder wall and brings poor sealing on the rings and smoking, etc.

Tom W

Quote:

Originally Posted by habraken (Post 1804178)

didn't MB mess up part numbers when manufacturing these engines or so? I opened up a few and the rods seem to be randomly installed. numbers on the fron, then on the back, all random... I think they just goofed up in stuttgart. maybe they were drunk.

I have never heard this idea before.

Tom W

Quote:

Originally Posted by JimSmith (Post 1803819)

I have posted "my" theory on this several times.

I bought one of these, a taupe 1991 350SD with a cream leather interior in pristine condition with about 75,000 miles on the odometer. I loved the car. Fit my family, was safe and got reasonable mileage. And it was a Diesel that I was certain would last forever. The rod bending phenomena was already known to MB and to the small group of owners posting on various sites on the relatively new internet car forums. I was blissfully ignorant.

At about 155,000 miles it began to have a loping idle and would blow a puff of smoke out, that would sometimes turn into a smoke ring. I took it to the dealership and assumed I was going to have to rebuild the head. I was not happy about this but figured the aluminum head demon had bitten me. The service writer at the time tells me it is more likely a bent rod. I ask why and he sort of shrugs and tells me they have seen it before. More often than failed heads or failed head gaskets.

The dealership service head cheese calls me and tells me they took the head off and I have a bent rod on two cylinders. I don't recall which two. I tell them I will be over to have them show me how they made that determination. I went over and they showed me they measure the height of the piston at TDC and compare it to the others. 4 are good, two are low. Bent rods. I ask about the head. No problem. They even suggest that if I have it repaired the head is likely good to go as is. I throw a **** fit, and ask how I could have bent a rod. I am told there is nothing I could do to bend a rod. The car won't let me. I ask how it happened and they say they don't know, but it does happen. I am an MB Club member and at the time they gave 25% on parts and 10% on labor. I get an estimate, something under $7k. I authorize the repair.

I examined the head and the head gasket while I was there. Nothing I saw, and the mechanic doing the rebuilding confirmed there was no sign of any head component or head gasket failure. I tell them I want the guides and seals changed and the valve seats reground. This adds a few hundred bucks to the repair. So it is around $7,500 with new all kinds of stuff.

I get the car back. It has a 12,000 mile, one year warranty. At about 13,000 miles it is smoking again. I now go back in and ask what, specifically they did. They try to tell me it is likely the turbo seals. I tell them until they can explain how the 7,500 dollars I just spent fixed the rod bending problem, I am not going to do anything. The service manager tries to calm me down, saying all kinds of meaningless crap about how the problem was rare and they are sure they fixed it even though he can't tell me what changed. I tell him the guy who wrote up the original repair order told me it was bent rods before the head came off and that sure makes it sound like more than a rare problem.

I go home and start searching the internet. I find MBShop.com and a link to a bunch of links about this problem. It is clear I am being screwed by an old man at the dealership who runs the service department. I go back with about 50 printed pages of stuff I gleaned off the internet about this problem, including groups forming to press MB with a class action suit, and go to the sales department. I start a loud discussion about the subject, and likelihood I would be better off buying a VW Diesel, with the sales guy I now deal with, who reminds me he did not sell me the car. He gets up and says he is going to speak to the management.

He comes back and tells me the management has agreed to give me full Blue Book value for my car (about $7k back then, before considering it was smoking and had a rod bender engine in it), the full $7,500 they charged for the repair as a trade on any new or used car in inventory. I find a silver 1998 E300D TurboDiesel. We work a "deal" out and I go to a Benz dealer in New Haven to see what kind of deal they will give me. Basically a plain white with black MB Tex car, no significant options, slightly higher asking price and they offer me $3500 for my 350SD. I took the Carriage House of New London deal.

Based on much reading I did at the time the issue, as I see it, is the rods don't have an infinite fatigue life, as built. This is due to a combination of issues, many interdependent, some not. But the production methods seem to be producing borderline rods. Not all are good, not all are bad. Due to dimensional tolerances of finished parts, their starting material properties and their as heat treated properties, all of which have tolerances. You get bad ones and the rods suffer fatigue failures at something over 65,000 miles. Mine failed in all likelihood around twice that mileage. After that the rate of failure is lower meaning if you survive the first 200,000 you likely have next to zero chance of a fatigue failure because the rods met the design specifications with a bit of margin.

Redesigning the rods to be heavier may help, until you stress them at full speed. Higher mass means higher stresses. This is not, in an already settled and now dimensionally controlling environment, a path to success.

I have heard all kinds of theories about hydrolocks, ingesting chunks of the scum that collects in the intake manifold downstream of the EGR valve, flexible blocks and so on. I see no hard evidence those are the problems. And, the displacement selection of an engine in Germany is made to comply with tax codes. You will notice all car companies have offerings at the break points. When designing a new engine the manufacturer always designs around the "bread and butter" models in the line up. 3.0 liters is a break point for turbocharged engines. Insurance is paid, by the way, on hp rating regardless of body style. They figure 200 hp has a fixed potential for damage liability, regardless of whether or not it is in a coupe or sedan or station wagon.

As for those bulletproof W123's, well, I have only ruined one mechanically (my son ran it without oil after it had well over 300,000 mles) yet I have retired several due to body rot. Better rustproofing in later cars makes them last longer. And, a 240D is so loud to drive it 500,000 miles averaging 70 mph or so means you will likely suffer a hearing impairment. A W126 or W140 is another story altogether. But I don't think you will find W140s and later cars in use in Beirut or other rough cities twenty five years after they have gone out of production, in significant numbers. The bulletproof label is one that has been earned, literally. I have never been stranded by a W123. I do not think I could say the same for the W140 - I doubt it would run without any electricity, for example. A W123 will. And an ordinary Joe like me can keep a W123 running indefinitely.

Jim

I have been pondering this more and have a couple things:

1. You don't seem to know what they did exactly when they "fixed it" the first time. They would have had to replace the two rods as a minimum.

It would follow that they would at a minimum also replace the liners, and install new pistons fitted to the block. If this is the scenario then one must conclude that they did not think that the rods were the cause of the problem. One would also have to conclude that at the time upgraded rods were not yet available.

If upgraded rods were available they would have done all of them, right?

If the upgraded rods were heavier they could not have just replaced the two bent ones because of the balance problem.

If they only replaced two of the rods then the explanation of the failure in 13K miles is because the other rods failed from being too weak. (?)

2. If they replaced all the rods with upgraded ones, then why did the engine redevelop the problem in 13K miles?

The first scenario seems possible if the rods were inherently weak but they did not know it at the time.

The second scenario would support (perhaps) the hydro lock or partial hydro lock theory. If this were the problem and they believed nothing was wrong with the head (it could have been slightly distorted above the place between two cylinders and leaked into both cylinders. There is that expansion slot between the cylinders which is open to the cooling jacket, IIRC) and they simply put the head back on it would not last long before leaking into the cylinders again. (This seems unlikely as automachine shop 101 would dictate that you always skim a head if it has been removed to assure good sealing, con't you?)

Tom W

Quote:

Originally Posted by DslBnz (Post 1804230)

Good 350...This is what they're like when they're only burning 300 miles per qt, as opposed to 100.:D

http://www.zippyvideos.com/2773151204709756/350toprun/

The acceleration and top speed look weak to me, my 300SDL is faster than that :D (no EGR, no CAT, no ALDA)

The stress/strain on a rod during operation of a diesel engine are pretty well known. By the '90s, good modal finite-element packages had become available in P-Method analysis, very reliable for determining the design, material, hardness/temper etc. of the rods. I find it hard to believe that the design of the rod was inadequate for use in this engine. It is possible that something such as a resonant frequency that created repeated osclilations in the rod during operation lead to their fatigue and was not predicted, but again IMO unlikely.

I'm not trying to punch holes in anyone's theory here, but being in on the process of designing and engineering parts for automobiles (including Mercedes-Benz) for over 20years there are some things that we go through each time iinvolving analyzing the design, including likely process variations, and creating a design and process that can operate in the worst-case tolerance stack-up. This includes everything from the clips holding vacuum lines to the brake calipers. Critical parts (determined by good sense and FMEAs) are closely controlled and often real-time X-rayed in the production.

It seems that, in theory, if the rods yeild due to fatigue, and being that it is nowhere near 100% failure rate, it cannot be a design failure or it would affect 100% (all being of the same design). The attrition being more random and a smaller percent indicates to me that it was a process variation, possibly in the material, or in the heat-treatment process, or even in hand-finishing of the part, or outside influence (hydrolock and rock theories for example).

If it were a hydrolock, IMO again, it would have other indicators. If it were a hydrolock due to a head gasket etc., are we to believe that it occurred during the cranking/starting phase? If it were, is the starter capable of creating enough force to yield the rod(s) on top of the infinite compression in that cylinder that would precede impact with the liquid or would the high-compression slow and stall the starter? Would the rod be the failure mode of a hydrolocked cylinder? I've had this problem in a gas engine, cracked pistons, didn't hurt the rods, high-torque starter in a high-performance engine (jet boat).Different engine design, but the question remains and we haven't done any analysis, anyone want to donate a bad 3.5 so we can hydrolock a cylinder and crank it to see what fails?

If the hydrolock were to occur during running, it seems it would require a lot of liquid to enter the cylinder in a very short period of time. Again, possible? If such a substantial breach were in the gasket wouldn't the car be belching white (steam) smoke from the exhaust?

Are there any other possibilities? If we had a couple of "bent" rods, a couple of good ones from the same engine, could do X-ray and metallurgical analysis and compare it to the original Engineering specifications, could run a couple of 3.5s with bad head gaskets etc. we might be able to solve it. I've spent many hours trying to fail parts in a quality lab until the lab techs were tired of me, I know how hard it can be to reproduce a failure in a lab envronment, even harder on the internet.