Diesel Engine Noise in Trucks: Information

[Rob45]

I have recently run across odd-order noise in heavy-duty trucks, and have traced it to the engine, and to the engine block specifically.
What I see most often is high noise levels at engine 4.5-order, although sometimes also at 3.5 order.
Oh yeah: these are in-line six-cylinder turbocharged direct-injection Diesels.
Although I've primarily seen this in "yellow" engines, other makers seem to have the problem as well.

The solution appears to be to stiffen the cylinder block by means of either a block girdle between the oilpan and the block, or with a structural (e.g., cast iron) oilpan.
This boosts the resonance period up, usually out of the complaint RPM range.

Thought perhaps someone else might find this helpful.
- Rob

 

[alexit]

Are these the yellow engines with the molded plastic/fibreglass oil pan? Or does the aluminum(?) pan engines exhibit the same noise?

 

[Rob45]

Haven't seen the plastic pans, just the aluminum ones.
This happens in C13s and C15s for sure, possibly in others as well, but it seems to be limited to the later models that meet the '02 emissions regs.
Might be related to the added weight of the compound turbochargers, but it IS cured with a blcok girdle or the oilpan that comes with the "Brakesaver" option.

 

[GregLocock]

integer+ half orders are usually caused by cylinder to cylinder variation in combustion pressure.

It sounds as though they are exciting powertrain bending (or possibly torsional) modes. I've had some success moving the engine mounts to the nodes of the mode in the past, as a cheap fix that never gets implemented!

They are also good at exciting crankshaft torsional and bending modes, so retuning the TV damper and installing a bending damper is another possibility.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[Rob45]

No, crankshaft torisonals are actually quite low on these engines; this phenomenon is due to block skirt lateral motions. We considered crank torsionals early on, and ruled them out since there was no correlation between engines having the noise and the crank torsionals on those engines.
It is (or seems to be) structure-borne, so that putting in sufficiently soft cab mounts can greatly reduce it, but mounts that soft are undesireable for other reasons: overall ride, and mount durability.

It does appear to be best attenuated with block stiffening, by, as mentioned previously, either a "bedplate" (as some call them) or with a structural oilpan.

The engine mfr. has modelled these engines, and those models show these motions, and show the motions going away when a pan rail girdle is used. Not going away, of course, but becoming smaller, and moving their mode of greatest motion to another significantly higher rpm.

 

[Rob45]

AND, Greg, I do wish I had the option of moving the mount locations!
But I'm stuck with what we have: a mount at each end of the block. Right now, the big concern is seeing to it that the turbo outlet pipe doesn't melt the mount rubber out altogether!

 

[Tmoose]

I'd look at the time wave forms of your sound data and some vibration signals on the block too. I generally end up having to play with the Time period and sample rate a bit. If 4 stroke I think it would be nice to trigger off the camshaft to reference the data capture from an entire engine cycle and more easily identify particular cylinders.

 

[Rob45]

I guess I should have made it more clear that this is a general problem, and not limited to a single particular engine.
The level of the problem varies somewhat: I have two supposedly identical trucks, with consecutive serial numbers, and in one, the 4.5-order noise is most prominent at just under 1300 RPM (corresponding to roughly 55 MPH) while the other is quiet at this speed, but has high levels of 4.5-order noise in the range 1500 - 1600 RPM, more like 65 - 70 MPH.
Other trucks of similar specification are much noisier than either of these two.
Other engine designs by the same manufacturer, and to a lesser extent, engines from another manufacturer, also exhibit this problem to varying degrees.

The engine manufacturer tells me that their structural analyses have determined the noise to be produced by transverse vibration of the engine block skirt, that portion of the sides of the block that hangs down below the crankshaft centerline. And supporting this theory is the fact that improving the block structure with either a structural oilpan or a block girdle (between oilpan and block) shifts the rpm at which the noise peak occurs upward, usually high enough that it is no longer a complaint.

 

[GregLocock]

I doubt the block resonances vary by that much - if the mass isn't changing that is a 40% change in stiffness.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[Rob45]

Yes, either a 40% change in stiffness, or there is more than one mode present.

There's undoubtedly something else going on, and the engine maker admit it's a function of injection timing, fuel rate, etc.; i.e., it's a function of combustion pressure. And there is variability truck-to-truck, for reasons we haven't yet discovered.
The only "known" at this point is that beefing up the block structure helps greatly. And that the structures boys know of this issue via their FEAs.

My problem isn't the noise itself, but documenting it and its solution sufficiently to convince management to go the extra expense of implementing a solution.

 

[GregLocock]

For reasons that I can't currently remember torsional and bending modes are often at similar frequencies, so you may just be exciting one in one case and the other in the other. Stiffening the block may help, or may just move the problem up in the speed range - that is a typical problem we see with powertrain bending in longitudinal installations - we can spend 10s of dollars on stiffening and end up with a problem at 65 mph instead of 55, which doesn't really impress anyone.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[Rob45]

Sometimes I wonder if anyone actually reads the original post...