Friday, April 13, 2012

Lobing and other Low Frequency Form Error In Aircraft Piston Engines

"Overhaul to New Limits" -- what comes to mind? dimensions, sizes, clearances. That's it, right? At least that's all I hear discussed. Lets talk about Form Errors or errors in shape. These often cannot be detected using two-point micrometers and bore gauges -- the stock in trade of engine overhaul shops -- and frequently defines the limit of their investment in measurement technology!

Connecting Rod Bushing with Form Error 
Dark horizontal lines are contact points where the piston pin contacted the bushing. The pin should be bearing on a swept area and not on "high-spots"

To get you up-to-speed on Form Errors, I've put together a short video on one of the most common type of Form Error called "Lobing"

The take-away from the video is that there can be significant errors in form that cannot be detected using two-point measurement instruments (micrometers, calipers, bore gauges).

  • Diameter =size
  • Roundness = form.

Two-point measurement instruments actually measures the diametric variation of the cross section and will not reveal variations in roundness caused by an odd number of lobes (3 in this example).
Diameter 1 inch - Shape Round
Diameter 1 inch - Shape Lobe

A good example of this is a cylinder barrel. It can measure within new limits for dimension using a dial indicator but be way out of shape due to distortion and improper honing. For example, a barrel tends to wear on one side more than the other as one side is the "thrust" side that the piston pushes against as it rotates the crankshaft. Lets take that worn barrel into the shop where they "hone" it to a "new limit" but oversize dimension. Fair enough. But what has happened to the Form? Wear on one side of the barrel has in effect moved the barrel center line. The hone follows the worn barrel just making it bigger and round but does nothing to restore center line.

There are cylinder hones "rigid" ones that restore form. It is interesting to operate one as the hone head does not float but is fixed to the proper position. Most hones, however, float or follow the existing hole which means that they may restore hole roundness but not form or position.

2.476 inches inside diameter at every point - but bad form
I like to use a Slinky to demonstrate the difference between a dimension and a shape. Suppose that this represents your cylinder barrel; run a dial indicator down it and it measures 2.476 inches at any point you measure. But, as you can see the barrel is badly distorted. If your barrel is only 10 or 20 thousands of an inch distorted you would never know. -Your piston and rings would know as they attempt to adapt to the distortion as they quickly travel up and down the barrel; most likely resulting in poor sealing with high blow-by and oil consumption. The usual (incorrect) diagnosis based on these symptoms is "poor ring seating!" Take them off and  hone to ring finish and try to seat the rings again with the same poor results. To prevent this expensive course of events one could carefully examine the wear patterns on the barrel for any tale-tell signs of wear. Based on this examination a course of action can be recommended with better results anticipated.

Another example is using a crankshaft belt sander to "polish" a crankshaft journal. Pretty typical process as long as we stay true to the purpose- polishing - and not sizing. As soon as we use the belt sander to dimension undersize (example go to 3 undersize on Lycoming due to wear) we introduce the possibility of form error. Form error also occurs when we "reface" rocker arm faces.

Observing Form Error

One of the best checks for Form is by observing wear patterns during engine tear-down -- do the parts "like" one another?
Engine Bearing with high spots
The Bearing (backside view) shows horizontal stripes caused by machining/grinding the bearing boss with a gear driven tool rather than with a belt drive. Slight impulses from the gears create minute movements in the tool. This bearing doesn't like the boss regardless of the overall dimension.

One thing to consider when we talk about dimension and size. We measure at a few discrete points on a surface with infinite discrete points. 99% of the surface is not measured! One partial solution for this problem (and resulting form error) is to measure run-out whenever possible; we rotate the part while taking continuous measurements along a line. Run-out with a shaft mounted in a Vee block is the simplest method of detecting lobing in shafts.

More Articles you might be interest in

More On the Problem of  Lobing in Aircraft Engines

No comments:

Post a Comment