What is runout? (Part 1)
- Fernando E. Romero, P.E.
- Jun 29
- 9 min read
Updated: Jun 30
Today’s article was inspired by a question from a reader who asked: Does TIR include the electrical runout?
The answer is no, but maybe, it depends on how we interpret the definition of TIR!
This is a great question, because it allows us to review several concepts, terms, and definitions and to expand on what they really are, how we measure them, and how relevant they really are.
There is a lot to unpack; so much that I tried to capture all the concepts I want to hit on a fancy mind map to help me keep track of things.
I have not been able to figure out what is the best order to start describing all these concepts and how they are related, so I will cover the most important definitions and then tie them together with practical examples.
Today I want to tackle:
The definition of runout as a geometric tolerance
How runout is measured by a dial indicator and what TIR means
How an eddy probe measures electrical runout
Let’s start with, what is Runout?
A runout in terms of engineering can be defined as a geometric condition or a dynamic condition.
Let’s look at the geometric condition first. For instance, when we are designing and defining the size and shape something should have by creating a drawing, we will use geometric tolerances.
Using tolerances is how we can “control” or “specify” how precise, accurate, straight, square, round, etc.… something should be when it is manufactured.
A designer of a steam turbine will have to specify what diameter the journal needs to be and what the tolerance for that journal needs to be.
He may say: “This journal needs to be 5 inches in diameter.”
But knowing that things cannot be manufactured perfectly or that we cannot always measure things perfectly, we must specify “tolerances”.
When you are dealing with round objects, you must deal with characteristics like:
Circularity
Runout
Surface roughness
So, in the end, the engineer may have to specify those three characteristics: a journal needs to be 5 inches +/- 0.005 inches, with a max runout of 0.001 in TIR, and a surface finish of 32µin.
His drawing will have something that looks like this:
But let’s only focus on the concept of runout today. Runout is the allowable variation of a surface feature as it rotates around a datum axis.
In the drawing above, the “datum axis” is the centerline of the shaft as shown by that long and short dash line.
In the next drawing, I show what the actual part may look like once it is made.
The runout band is shown by the red dashed lines.
This means that the actual surface of the journal is allowed to be within those red dashed lines.
---parenthesis
If you want to learn more about dimensional tolerances and, in this case, about runouts, please visit this site:
These guys at Drafter Inc are working on some revolutionary ways of automating the creation of drawings and creating some great explanations on why things matter and what they mean.
---end parenthesis
So, how do we measure this on a rotor?
We do it using a dial indicator.
In the example above, I am using a dial indicator that can measure displacements down to 0.0001inches.
A sheet of printer paper is about 0.005” thick.
Watch what happens when I slide the sheet under the indicator.
The sheet pushes the indicator arm, and the dial displays the amount of the movement.
The needle moves clockwise a total of 0.005” inches.
What is TIR?
TIR is an acronym that has evolved to have two interpretations.
Let’s look at three sources where both interpretations appear.
API Standard 612, 8th Edition – Special Purpose Steam Turbines
API Standard 617, 9th Edition – Axial and Centrifugal Compressors and Expander-compressors
API Recommended Practice 687, 2nd Edition – Special Purpose Rotating Equipment Repairs
All three of these sources imply that TIR as an acronym means:
Total Indicator Reading
Total Indicator Runout
What does this demonstrate?
That I am obsessive over details and definitions! And that these three documents don’t present an identical or exact definition for this term.
Three documents that describe the same exact concept, but with subtle differences in the presentation of the terms.
My opinion is that these variations in presentation may lead some people to have slightly different interpretations and understandings of these fundamental concepts.
I only obsess over this because I was taught by brilliant engineers both in class at university and in practice at work that, if you are going to diagnose the condition of a machine, you must first fully understand the fundamental principles that govern its behavior.
Out of the two versions of the acronym, Total Indicator Reading is the better term.
Why? Because it describes the use and interpretation of the measurement made with an indicator.
Regardless of what is being measured, if you use an indicator, the total amount measured, from the minimum to the max movement of the dial, is the Total Indicator Reading.
Finally, there is even a arguably better term: Full Indicator Movement (FIM) to represent the total movement of the needle of an indicator. Eliminating the ambiguity of TIR and its two definitions.
In the evaluation of rotating equipment, we do this while supporting the shaft on “vee blocks”.
This allows the journal to glide and to be supported by two points of contact.
When we spin the shaft, the needle moves, and it gives us an idea of the variations in distance between the contact point opposite from the indicator and the indicator.
One thing to consider is that this measurement is measuring the “error” of both sides of the journal.
This is different than the definition of geometric runout as it is defined in an engineering drawing.
Remember, in the drawing the runout is the deviation from a datum or axis of rotation.
Instead, when we support a rotor on “v blocks” we are not supporting it or holding it by its axis of rotation. The rotor is resting on the surface of the journal.
That is why we say the eccentricity or “displacement” to the axis of rotation is ½ of the TIR.
We are basically averaging the error or assuming that the dimensional deviation is in “equal parts” in each surface.
This measurement we are taking here is usually called “mechanical runout”; but remember, it is not necessarily the same runout as the one on the engineering drawing.
This is why in many shops and repair facilities this term is simply called TIR. It is to represent whatever the indicator read when the rotor runout was measured while supported on “v blocks”.
We try to control this because, for bearings to properly operate, their surface needs to have a certain roundness. To put it in non-technical terms, bearing journals needs to be nice and smooth and as perfect as possible.
The more perfect a journal, the smoother the rotor will run when the machine is put in operation.
But how good is good?
Well, that is up to the engineers that designed the machines to tell us. In reality, only the designers of the machines know exactly what the technical requirements are.
For instance, API 612 and API 617, the two documents that describe the expectations to be met when steam turbines and compressors are designed, manufactured and tested, state that the Original Equipment Manufacturer should provide drawings that include the “clearances and tolerances”.
API 612 and API 617 do not specify a range or a maximum condition of runout or dimensional tolerance.
API 687 does provide some guidance, but only in terms of what to expect when you repair a rotor:
It says the shaft radial runout (TIR) at any location should be no more than what is suggested by the OEM or 0.001” if it is unknown.
It also says that the radial runout (TIR) on the shaft and probe targets should be what the OEM suggests or 0.0001” if unknown.
So, we always defer to the OEM first. Remember, those guys designed the machine to meet a specific set of requirements.
The repair industry specifies the numbers they do for two reasons:
We know journals need to be smooth and round. Since the principle of operation of hydrodynamic bearings relies on the dynamic motion of a journal inside the bearing to push and shear the oil in order to create the oil cushion, we need a smooth and continuous surface for this.
The second reason is because the smaller the runout, the smaller the eccentricity.
When we get to discuss balancing and why rotors vibrate, this will become more apparent.But for now, the punchline is: the lower the runout (low eccentricity), the lower the unbalance forces will be.
Now, let’s ask the question:
How do you know how a rotor runs when it is in operation?
Oooof, this is getting more and more complicated! Which is exciting as well!
I will reference the two greatest sources for insight and wisdom when it comes to machinery and vibration diagnostics.
Two books, written by contemporaries and pioneers of the field of vibration measurements and machinery diagnostics.
The first and my personal favorite:
Machinery Malfunction Diagnosis and Correction Hardcover – November 1, 2015
by Robert C. Eisenmann, Sr. (Author), Robert C. Eisenmann, Jr. (Author)
And the second:
Fundamentals of Rotating Machinery Diagnostics (Design and Manufacturing), 1st Edition
by Donald E Bently (Author), ASME Press (Author)
I absolutely love the Bob Eisenmann book, because I had the privilege and pleasure to work for Bob. We travelled to many sites carrying oscilloscopes, digital vector filters, data acquisition boxes, and a few times even reel-to-reel tape recorders!
I would often find myself going to his office to ask him a question, and he would wait for me to finish before he would utter: “Did you read chapter 4?”
If you are an engineer and want to learn how to diagnose the condition of rotating equipment you should get these two books.
On the other hand, Don Bently pretty much invented or made commercially available the most reliable instrument for measuring the position of a shaft while its rotating.
The device we call the “eddy current probe”, also known as the “proximity transducer system”.
This device provides an output voltage that is directly proportional to the distance between the probe and the shaft.
And in the world of machinery diagnostics, it allows us to measure the position of a shaft and the vibrations of a shaft.
I will center my explanations referencing the Bently Nevada 3300 XL Proximity Transducer System. Simply because it is the one I have personally seen the most, installed in 99% of all the machines I’ve seen built where I work and in every single control room I’ve been to, all over the world.
The 3300 XL system will provide that linear proportional voltage when it is measuring its distance to 4140 material, which happens to be a metallurgy commonly used in rotating equipment.
This is what that looks like.
It is remarkably linear and repeatable.
Eddy probes work by emitting a magnetic field into a “target” material. As the distance between the probe and the target changes, that magnetic field changes as well.
The “transducer” measures those changes in the magnetic field and translates it to a voltage.
Usually, 3300 XL probes have a sensitivity of 200 mV per thousands (mils) of an inch.
As you can see from my experiment:
When the probe is touching the target, essentially at a 0.000” distance, the voltage out of the transducer is: 618 mV.
When the gap increases to 0.005” the voltage is still 618 mV.
This means the probe is not within its “linear range”. Basically, the probe cannot be used until you are within the linear range of readings.
When the gap is between 0.005” and 0.110,” the graph draws a nice straight line.
The slope of that line is the sensitivity of the probe.
For the test I ran, the slope of the line is: 210.28 mv/mil.
For a gap above 0.110,” the line flattens out and the voltage stays at 23.520 V.
From this experiment we can conclude that this probe is useful to measure distances or gaps of 0.105” or less.
This also means that for every one Volt change measured, the gap has changed about 0.005”.
This is enough when you are monitoring the health of your machines and want to measure both positions and vibrations.
Since the eddy probe is using magnetic fields to measure distance, it is important that the “target” is demagnetized. The recommendation made by API 687 is that rotors have less than 2 Gauss residual magnetism.
I have always been intrigued to find out what effect magnetism may have over the ERO, so I conducted a short experiment.
I set my probe at a gap of 10.01 Volts. (This translates to about 0.050” from the target.)
I moved a magnet around, being careful not to disturb anything else.
You can observe how the voltage varies from 9.99V to 10.05V, this is a 0.06V swing! At a sensitivity of 200 mV/mil. This is equivalent to 0.0003”, and, in the world of rotating equipment such as steam turbines and compressors, that can be a lot!
This is a good stopping point.
There is still a lot more to cover, especially some more myths and some possible gray areas of interpretation surrounding the concept of mechanical and electrical runout.
Comentários