What is RUNOUT (Part 2)

Today I am going to show you what it looks like to record runouts with an indicator on a real shaft. Today I measured something really small in my garage!

This is our test rotor.

It is a mandrel, or very short shaft. I am going to use it to illustrate how to measure mechanical runouts.

This is the same way, in fact, that API 687 recommends rotor/shaft inspections be performed in a shop. This is just a tabletop version of that.

Let’s cover some conventions:

When taking dimensions, performing inspections, or setting up vibration data collection equipment or balance machines, one must observe some consistency in conventions.

The world of turbomachinery has arrived at the following: measured angles will increase counterrotation.

Balance machines for instance rotate as if the shafts were spinning away from the viewer (illustration below). The same way when we measure runouts, we rotate away from the viewer.

This way all our shop measurements, dimensional inspections, balance reports, and field reports will have the same convention.

We measure mechanical runouts for two reasons:

1. Determine if the journals are round

2. Determine the shape of a rotor

   
   

1.     Determine if journals are round

As I mentioned in my last post, we need journals to be nearly perfect because they are the area where the shaft will be supported inside the bearings.

What does nearly perfect mean? Dimensional, it means four things:

• A journal must be the correct diameter

  We determine this by measuring the diameter in three positions around the circumference.

  The design engineers will determine the target size a journal must be depending on the type of machine and design intent.

  
  

• A journal must be straight

  We determine this by measuring the diameters at three axial locations along the length.

  Basically, all journals should be straight by design.

  
  

• A journal must have a smooth surface finish

  We need to make sure the surface is smooth (with a particular surface finish).For journals, API 687 recommends a Roughness Average (Ra) of 32 µin (0.8µm).

  
  

  This is usually accomplished by grinding the journals on a machine called a cylindrical grinder (what you get if a horizontal lathe had a baby with a giant tool post grinder).

  These machines will be outfitted with large grinding wheels made from Cubic Boron Nitride (CBN), which is an extremely hard synthetic material with superb thermal and dimensional stability that makes it perfect for making abrasive products.

  
  

  When you grind a journal on the setup described above, you end up with a surface roughness well within that 32µin reference.

  But if you must measure it, you would use a device called a Surface Roughness Tester.

  
  

• A journal must be round!

  Wait a minute! Some may say, didn’t we already measure the diameter? In three locations?

  Well, yes, we measured at a macro scale to determine the diameter.

  But to know how round something is, it is best to check the runout.

Remember the definition of runout in terms of geometric tolerancing?

A runout is the error of the surface as if it were rotated around its central axis.

Let’s go back and look at the journal of our shaft. 

Consider, the diameter is within the dimensional tolerance allowable; however, the shape is not perfectly circular but slightly “egg shaped”.

What will happen when we spin the shaft?

Well, let’s see! I’ve done it!

Anyone’s mind blown yet!

Do you see that needle move, 0.0001”.

Now, remember, a sheet of printer paper is about 0.005”.

THAT IS FIFTY (50) TIMES THICKER THAN THE RUNOUT!!!!

That is how you determine how round something is!

The runout on this shaft is 0.0001” TIR, or Total Indicator Reading.

The “egg shape,” that I drew in red lines before, is only 0.0005” smaller on each side.

But when you rotate the shaft, you cannot really tell which way the error is “going”.

To visualize the extreme scenarios, I am going to bias the error to each side, 0° and 180°.

In our method of measuring runouts on “v blocks” we cannot really tell. But let’s not despair about that. The journal is within tolerance and within a value that is known to be suitable for the normal function of a journal.

What this runout does is that, as the shaft rotates on the “v blocks,” the centerline is going to appear to move sideways, by an amount of 70% of the runout TIR.

If the runout were higher, the centerline would move a greater amount. And since the journals are the reference of rotation for everything else on that shaft, the greater the centerline movement, the greater the chances you experience vibration from unbalance.

So, what does this runout look like in real life?

Here is a slow-mo video of the runout of this “eggshaped” journal.

Now, let’s move past measuring the mechanical runout of Journals.

The second reason we measure runouts, is to determine the shape or eccentricity of a shaft and rotor.

2.     Determine the shape of a rotor

 

Imagine my test rotor, let’s imagine that during the manufacturing process, there is an error and some of its diameters are machine and ground “off center”.

If that center diameter was machined incorrectly, it means the “center of mass” of that part of the shaft will not be in line with the centerline of the shaft.

This means there is an eccentricity on that center section.

A displacement in the center of mass and eccentricity will produce a centrifugal force that will be perceived as an “unbalance” when we spin that shaft.

This is one of the main reasons why we perform mechanical runout measurements on a shaft. It is to determine if the shaft and its components have low eccentricity.

The next reason is to see if the shaft is actually bending. Sometimes this can happen to steam turbines or centrifugal compressors. Imagine a machine operating and it’s warm and spinning. Now imagine something makes that machine suddenly stop.

If a long rotor is not kept rotating while it cools down, it is possible that it will bow and set. Just like you imagine, the weight of the rotor will pull the midspan of the rotor down, like a banana or a hammock. Remember, the rotor is only supported by the two journals.

I will exaggerate this condition in the illustration of my test shaft, but it would look something like this.

These are the main reasons to measure mechanical runout.

Recap:

1. To ensure the journals, the only means of support for a rotor, are nearly perfect. All other measurements are made relative to the centerline defined by the journals.

   
   

2. To ensure shafts and rotors are stacked correctly, minimizing the eccentricity of fits or stacked components.

   
   

3. Finally, to determine if a rotor has bowed, particularly after a sudden stop.