Common Failure Modes - Engine Builder Magazine
Connect with us
Close Sidebar Panel Open Sidebar Panel


Common Failure Modes


For years, B&G Machine had been a diesel machine shop serving Seattle and the upper Northwest part of the U.S. To address certain market conditions in the mid- to late-’90s, we began building diesel engines. In doing so, we tried to identify the engines that would best serve a captive market – markets that would maximize the equipment we had on the floor and those that presented good opportunities for better profit margins. We believe where demand is high our product is appreciated.

Click Here to Read More

Once we started building engines we found there were a lot of things that crossed over from the machine shop that benefit the life of the engine. In light of where we’ve gone with the business, knowing these things can be helpful to other engine builders.

We see different fatigue areas and failure modes.  Some of the things that we do in the machine shop help to yield a longer, better running engine.  One of our Caterpillar 3508 engines went 158% of the recommended fuel burn for the life of the engine.  This is pretty remarkable when you consider that the local CAT dealer was probably getting 80%.   We’re able to get like or better than expected results even compared to a new engine.


We’ve found that focusing on some very specific operations in the following areas has been quite advantageous for us:

• Cylinder Block Top End Sealing – Minimizing gasket-related downtime through controlled finishes

• Camshaft and Cam Follower Failures – What really causes camshaft “failures”

• Rod Fretting – Recognizing rod fretting and when to condemn a rod assembly.

• Crankshafts, Counterweights and Balancing – Is it necessary?

Sometime around 2006, Caterpillar introduced a new head gasket for the 3500 series engine. We saw rather quickly that this one was having issues.

Figure 1 shows where the liner would seat in the block. The area where the liner seats would fret so badly on every cylinder that it would literally condemn the block. The expectation for the engine was to go about 15,000 hours. We pulled one out at 6,000 hours and in places it was over .040? deep with fretting. The issue quickly became a significant problem – it was eventually recalled.


The problems were so significant that many engines had to be pulled – these are 16 cylinder engines, so it’s not a small project. When an engine has to be pulled and completely ripped down, it’s a major interruption to production.

Ironically there was a specification called out by the manufacturer that said anywhere from 0 to 125 RA was an acceptable deck surface finish. If we were to technically go by Caterpillar’s specs, we could make it as smooth as we wanted or as rough as 125 RA, which is quite rough. If we follow the gasket manufacturers’ specs, they’re somewhere around 85-120 RA.


As part of our rebuild procedure on the block, we made them all 90-95 RA, in order to land right in the middle of the gasket spec (Figure 2). The loose CAT spec. for deck surface finish allows rebuilders to keep the work in house and not have to send it out to have work done. They can look at the deck of the block – if it’s shiny, if it’s rough, they can qualify that deck and carry on and rebuild the engine. That’s what was happening – many of the engines were being built without being machined, and were having some pretty significant problems.


The point isn’t to lay blame on the gasket, but this area – the deck of the block – is an area that can often be overlooked. Putting in the right finishes and doing the right things will eliminate problems down the road.

Rebuilders have come up with some very elegant solutions. A colleague of ours uses a brass or silicon bronze brass insert material around the cylinders. It’s intended to stop the erosion that is so common around those areas.  Now a lot of people will use stainless steel inserts as well.

This is a nice repair and its application isn’t limited to 3500 series Cats – that’s the engine we see as the most popular,  but in other applications we also find erosion where the water grommets go. We machine those, put the plugs in and then machine the deck of the block. The nice thing about that is, literally, if it came down to it, we would guarantee that block free of defects in that area for the life of the block to overhaul. We will guarantee the block will never have a problem with corrosion in those areas by doing those things.


Another thing we do when we do to an engine with multiple cylinder heads (in Figure 3, it’s a 16-cylinder engine)  is match-machine the cylinder heads. Rather than mix and match old heads, we match machine the entire bank on the exhaust side, which takes care of exhaust leaks and broken exhaust studs.

We know that using all new valve train and match machining the exhaust side of the cylinder head (see Figure 4) provides a nice surface for the manifold to be bolted against. In a haul truck, leaks can be an annoyance but not necessarily a big deal – where the engine is running in an enclosed area (such as in a marine application), the problems associated with that side of the engine are virtually eliminated. I can’t recall a single exhaust leak along that area that we’ve had in the last eight years.


Camshaft and Cam Follower Failures

Almost universally and regardless of what manufacturer we’re dealing with, the engines aren’t having problems when they’re new. However, after overhaul at mid-life we start to see camshaft and cam follower failures.

First, let me explain: B&G Machine started with camshafts (a cam grinder was the first piece of equipment we purchased). We have just under 800 camshaft masters for just about any camshaft you can imagine. It’s a little redundant because half of them we don’t see anymore.  We have seen so much with camshafts over the past few years, it’s just remarkable.


We’ve found issues on the intake, exhaust or injector lobe of the camshafts. In Figure 5 you’ll see cams from late-model Cummins engines. The left picture shows a cam that would be qualified – it’s one that a dealer would most likely reuse. On the right is one in pretty rough shape and would probably get thrown out.

I like this comparison in particular, primarily because there’s a guideline that a lot of the distributors use that would qualify this as a good camshaft. And I’ll tell you that the camshafts in these applications will run to overhaul when the engine is new, but unanimously, when I ask our distributors, they say replacement cams will only make it half way.


The problem?  Again, it’s because the surface isn’t finished. The cams go about halfway through and then they all start to have problems.  

Reprofiling the cam (see Figure 6) – restoring the profile, lift and duration – gives a huge advantage to the performance of the engine. In doing so, reconditioning the camshaft means all the cam followers need to be replaced. All those riding surfaces will be just like brand new.

Don’t worry,  we only take a little bit of material off the cam lobes, usually no more than .003? or .004?. Because they’re induction hardened, camshafts have plenty of material. Taking off a little bit of material doesn’t affect the riding surface at all.  It just restores the lift, duration and profile on the lobe.


I assure you, when you profile a cam you don’t lose any lifter duration because you decrease the base circle. You don’t just remove material from one side of the cam. When you reprofile the cam, you remove material from the base circle of the camshaft and make up for it in the adjustment on the head. It doesn’t have any adverse effects as far as the adjustment on the engine and the performance of the camshaft.

However, surface finish is critical. Chatter marks look like a dirt road, and actually contribute to “skidding” and camshaft failures.


We had some OE engineers at our shop about 3 years ago.  They said there are all sorts of things you can stack up (things that cause premature wear)  – lubrication, contamination, top end adjustment issues.  One of the engineers told me, however, that most of the failures are the result of the cam follower and the cam lobe not keeping in sync with each other.

Where we see most of the issues is right on the load-bearing side of the cam and  on the ramp. And it’s attributed to the variable and complete profiles (radii) designed by today’s engine OEMs.


Assuming it’s moving at a constant speed, during one rotation of the camshaft, the cam follower has to go from zero to perhaps 4 times its speed in half a rotation and from 4 times its speed to zero in the other half. In effect, what you have is a similar action to driving a car on a dirt road. The braking part isn’t so bad because you’re coming off, but it’s just a repetition of burning out and skidding.

Compounded on that are the high unit lobe pressures on the cam with today’s engines, especially with the injector lobe. The reason it’s so important now (rather than 10 years ago) to have these fresh surfaces, is because of this phenomenon:  Roller cams, more elaborate profiles and more demand on the lobe. It just emphasizes the point of trying to put a correct profile on the cam and using all new cam followers. If you go in with a slightly worn lobe, it’ll be like an exponential failure – it starts to compound on itself causing real problems.


Connecting Rod Fretting

This is an often overlooked or unknown issue. Depending on the connecting rod setup, if you’re looking to reuse a connecting rod body and redo a custom cylinder pack, fretting might actually condemn a rod.

The Caterpillar 3500 rods pose a unique challenge because of the way the cap fits. It is very difficult to recondition the big end because you can’t reduce the cap. The cap hangs over the body and the bolts come in at angles. On each of these rods, we replace all four of the rod bolts when we’re reconditioning an engine.


But it’s the fretting where the caps come together, especially where the cap fits over the body, that can be a problem.

A lot of people don’t know about this spec and they’ll simply exchange cylinder packs rather than build a custom pack. You must account for any fretting in the area where you’re sliding the cap over (see Figure 7). In an ideal state the rod cap is flush with .007? side to side (or less) – that’s the summation of the fretting on either side. It could be .002? and .005? or .0035? and .0035? – beyond that point the rod goes. We’ve seen in-field situations where these rods get reused and that spec is an unknown thing.


With all of these components, we’re trying to address any worn surface in the engine, especially any running components. They’re going to yield better engine life. This is a prime example. If this rod goes out into an engine and it’s outside that .007?, we condemn it.  If we missed that spec and it comes loose, there goes the bottom end of your engine. When we see a rod failure or we hear a concern about a bottom end failure, on this engine we’re obviously looking here first.

Maintaining Your Balance

Our expectation is that any engine we condition will produce 100 percent life. Anything suspect in the engine and all the areas that have any fretting are fixed so that when we put that engine together the benchmark is 100 percent. If it doesn’t reach that we’re asking why and trying to understand it.


One of the processes that we do is somewhat unique – we balance the entire rotating assembly on every HHP engine. With the piston and rod combinations like they are now, it’s an important step and makes for a smoother running engine.

We’ve been balancing for years and what pushed us that direction was an update from aluminum pistons to a one-piece steel slug.  The piston design went from an articulated piston to a heavy one-piece piston body. When that happened we saw a lot of variations in the weights of the components.

We realized we were playing with something different here, so we started to balance the entire rotating assembly (Figure 8). But a smooth-running engine isn’t just a comfort thing. With good align boring down the main bore of the block, proper crankshaft grinding and other steps, we can put an engine together that won’t have any vibrational fatigue. Much of the fretting comes from vibrational fatigue in the engine. We’re trying to reduce that as well.


It’s arguable, but a lot of the fretting we see in other areas may have come from vibrational fatigue as well. We believe that balancing an engine helps reduce that fretting – it definitely helps the performance of the engine to have the bottom end of the engine and the main running components balanced.

It’s a problem on engines on which the cylinder heads are a massive unit – especially overhead cam with rocker assemblies that bolt into the heads. I would imagine this issue isn’t just specific to this engine model, but there is definitely an issue here.


One particular engine came into our shop with a broken bolt in the area where the rocker shaft bolts to the cylinder head. The customer wanted a failure analysis of what happened – obviously, the bolt broke, but the REASON the bolt broke is because that area got loose. If you’re doing engines where the rocker shaft bolts down into the cylinder head, really make sure that there’s no fretting there either. Failure to address this can result in disaster.

This turned out to be a pretty major failure and we think it’s a potential problem for any D10T, C27 or newer ACERT engine. As far as repair strategies go on this engine, the fix was to use a different bolt and more torque. However, we’re keeping an eye on this engine.



I keep going back to fretting – but I think a lot of the long term wear failures that rear their ugly on the second overhaul happen because these surfaces are not machined properly. There is so much room for improvement there.

It’s no different with crankshafts. This is a really expensive component that is often qualified visually and reused and oftentimes there are problems down the road.

At B&G we carefully clean, magnaflux, inspect, and thoroughly check every crankshaft; if we’re doing a crank with counterweight bolts they’re coming out and discarded and new counterweight bolts are going in. We check the areas where the counterweights go on for any fretting as well.


Look very carefully at the radiuses of the crankshaft. Make sure they’re properly sized and there are no stress risers or areas of concern there.

Sometimes you’ll see a failure where the crankshaft breaks between the rod and the main, just like a diagonal cut. You can almost look at the beech marks on the crank and follow them all around. Just run your finger down the radius and feel the bump – that coincides with where the crank broke and that’s because there was an imperfection in the radius.

Another issue is the hardness and surface finish. A lot of cranks now are getting down to a 4-6 Ra surface finish. The expectation is quite a bit higher now for proper performance of that crank than it was a few years ago (Figure 9).


We’re straightening the cranks as well.  We’re making sure our cranks are straight to within .002? all the way down the mains. Total indicator runout on the journals is within half a thousandth – we try to finish mid to high side of spec. When we match all these things up – the straightness of the crankshaft, the straightness of the main line, the bearing clearances (typical bearing clearance that we see is .006?) – we know we have good bearing clearance and no fatigue spots on that crank. It’s a real good riding surface for that crank to lie in the block.


After the crank is straight to within a couple thousandths, we balance the rotating assembly and the crankshaft. The results are definitely measurable. We did align bore work and crankshaft work for a customer with six Waukesha engines running off of methane in a landfill. Each engine had probes that dipped down each of the mains to check temperatures to make sure nothing was getting hot.

It was nice when he called to tell us that each main had consistent temps all the way down. That’s a direct result of a straight main line and a straight crankshaft.


Granted, in the field a lack of such precision probably won’t cause a failure and you’ll get away with it, but it definitely takes away from the reliability and performance of the engine if either isn’t straight. That causes hot spots in the bearings.

Another process that we believe produces measurable results is dyno testing – and we feel it should be done on every engine built.

We’re doing more with dyno testing than just a cosmetic, checking-for-leaks thing. We use it as a diagnostic check on the engine to be sure that the performance on the engine is good all the way through its systems. Not just checking horsepower and torque, there’s quite a bit more involved.


It may take more time and cost more money, but customers with these engines rely on us to provide maximum performance for the life of the engine. They trust that we will provide the very best that we can produce. Downtime is lost money, and your careful attention to each step of the reconditioning process is an investment in their business.

Visit B&G Machine online to learn more about the company and its capabilities at 1 following a new gasket introduction, the cat 3500 started having significant fretting issues. we found you can
</p>					</div>
									<div class=

Engine Builder Magazine