If you’re involved with magnetic particle testing (MT), then you’re familiar with the term “Ketos ring,” but do you know where the word Ketos comes from? Ketos® is actually Crucible Steel’s registered trade name for the industry-standard, oil-hardening tool steel (AISI-O1). However, the term Ketos ring has been used in our industry so long that it has become synonymous with the test ring specimen that we currently use as an overall system performance indicator in the magnetic particle testing (MT) process.
Most people who are familiar with the MT process are aware of the current usage of this ring; however, not everyone is aware that this ring actually started life as a gauge intended to be used for quantifying the sensitivity of magnetic particles.
William E. Hoke has been credited with being the first to understand and describe magnetic particle testing sometime after World War I. Later, Alfred V. de Forest reviewed Hoke’s work and became the first to use high current electric fields to develop the required magnetic fields necessary for magnetic particle testing. He applied for the patent for this work on July 5, 1929. Between 1934 and 1935, de Forest applied for patents covering the particles used in the MT process, and in 1935, C. E. Betz began improving the particles and pastes used in the wet method. Soon, a methodology was needed to be able to grade the overall sensitivity of the magnetic particles used in the MT process and the Ketos ring — first known as the Betz ring was born.
As time moved forward, the ring moved out of the lab and onto the production room floor where it was used as a comparative reference standard. With this development, demand for the rings went up. As demand increased, so did the number of different ring suppliers.
This is where some of the problems started: Not all suppliers produced the rings exactly the same way, and the results turned out to be much like having multiple suppliers all producing micrometers with randomly oriented numbering on the thimbles.
As you might expect, when used as a gauge for grading the sensitivity of magnetic particles, it is highly critical that all of these rings perform in exactly the same way, otherwise there would be no perceived consistency in the particle performance even if the particles actually did perform consistently.
Ring Differences
There continues to be confusion regarding the differences between the rings in use today, so let’s start by discussing exactly what these differences are.
Both the original Ketos ring and the AS5282 ring are manufactured from the same AISI-O1 tool steel. They are dimensionally equivalent, except for the optional removal of the few deepest holes in the AS5282 and ASME versions, as well as the GE version specified in P3TF48.
The intended heat treatment is the same for both specimens; however, SAE AS5282 does go on to further clarify the details of the heat treatment methodology in a manner much better than most of the other previous documents.
The tool steel that these rings are made from is generally supplied in the annealed state (technically, a normalized state) when it is sold; otherwise, the end users would not be able to machine it. It is important to remember that it is not the steel producer’s goal to put the steel in a magnetically homogeneous state, only to make it easy and uniform for the customer to machine. This as-supplied hardness generally runs approximately 91–92 RHB. The goal of the first step in the heat treat cycle, as specified in SAE AS 5282 section 3.3, is to put the steel in a machinable state. The steel supplier almost always performs this step prior to shipment of the product; if it has not been done, the OEM must do so. The procedure is as follows:
After receipt of the normalized bar stock from the steel house, the first step is to saw cut the material into ring blanks, then all the other secondary machining operations are performed on the ring. After completion of all the secondary operations, the ring goes through the final heat treatment operation, which is intended to put the ring into a magnetically homogeneous state. The RHB hardness number generally does not vary more than about one point from the time the steel is delivered from the supplier in the normalized state, to the final fully annealed state, post machining. Other than dimensional gauging, a hardness check was the only required NDT test for the first generation of test rings (Ketos rings). Since the hardness value does not change much (generally <1 point), the customer cannot easily detect if this required final annealing operation has ever been performed.
This final full-anneal operation is detailed in SAE AS 5282. This portion of the required heat treatment is essentially no different than it was 60 years ago; the verbiage has just been clarified a bit in the AS spec to eliminate as much confusion as possible. The time and temperature requirements are based off the steel manufacturer’s recommendations for this alloy, which is true for both versions of the rings. Slight variations in the time and temperature values can be seen between Crucible Steel’s recommendations for annealing and the listed NDT specs; however, these differences are no more than what can be found between the different brands of O-1 tool steel manufactured. The requirements listed in the NDT specs do tend to be a bit more conservative comparatively.
Specification Differences
The main difference between full annealing and normalizing is that fully annealed parts are uniform in softness (and machinability) throughout the entire part, since the entire part is exposed to the controlled furnace cooling. In the case of the normalized part, depending on the part geometry, the cooling is nonuniform resulting in nonuniform material properties across the part. This may not be desirable if further machining is desired since it makes the machining job somewhat unpredictable. In such a case, it is better to do full annealing.
SAE AS-5282 came into existence as a corrective action to combat the tremendous variations in the performance of rings that have been produced and sold over the past 60 or so years. Unfortunately, these variations were further amplified when ASTM E1444-94a changed the requirement for the most critical final heat treatment from a “shall,” to a “may” condition. Shall was the intended operation to have been performed on all these rings, but some NDT documents continue to use the words “may” and “should” instead of “shall.” The next revision, ASTM E1444-01, eliminated the final heat treat requirement altogether. Elimination of that step did save the ring manufacturers money; however, the result of this action spoke for itself.
This deviation from the original intent of the documents could have been a typographical error in E1444- 94a, which then was copied from document to document, just like the words “ANSI 0-I Tool Steel,” instead of the correct AISI O-1 Tool Steel, but reviewing the history of changes, it appears that this was an intended change at the time, probably because of the lack of understanding for the reason behind the final annealing step.
Why is the question of the wording important? First, the various prime manufacturers such as Boeing, GE, and Airbus, have modeled their own standards after ASTM E1444, and they have passed-on the various mistakes from each different revision of that document. Also, do you want to be using rings that were manufactured with a “may” or a “should,” instead of a “shall” and then trying to achieve comparative test results? Furthermore, do you want your AS/SAE particle supplier grading particles using a ring that was manufactured with a may or a should, instead of a shall?
Prior to the introduction of AS- 5282, there was no “magnetic” qualification test requirement on any ring to ensure that it was in fact produced in a magnetically homogeneous state. Two of the main goals of AS5282 were 1) to introduce a methodology to magnetically qualify and quantify the test ring, and 2) make the ring individually traceable to the plotted results. The shortcomings in ASTM E1444 were never addressed by SAE.
In any magnetically homogeneous material, you should expect to see a direct relationship between the field strength and the measurement distance; in this case, it would be the increasing distance from the outer periphery to the holes. The inverse square would result in air when doubling the distance and, in any magnetically homogeneous material, you should expect something approaching this same relationship. On a good ring, regardless of the version, the results should show this type of relationship when mapped.
With a qualified magnetizing power supply (not a MT machine) and proper magnetic data logging equipment, an individualized field leakage plot of the ring may be obtained in order to quantify and qualify this variable for the finished ring.
Using the Rings
People still get confused over the proper use of the Ketos test ring, so these next few paragraphs aim to clarify their use.
These rings are generally considered a quick and easy “gauge” to determine the machine’s overall system condition prior to inspection of parts. They are to be used as one of the daily systems tests that are to be logged for each machine in use. So, you use the ring to check the system, not the system to check the ring. “Certifying” the test ring with a mag machine is working backward and although this may seem obvious, it is still done.
As most people are already aware, either version of the test ring is intended to be used on machines with a DC output waveform. That could be either half-wave direct current (HWDC) or full-wave direct current (FWDC), in a single-phase or three phase configuration. Any one of these DC waveforms should give similar, but not the same, results on the test ring, assuming the ammeters are properly scaled and calibrated. Due to the pulsating nature of the HWDC waveform, HW generally tends to show more lines than FW does at a particular given amperage. It is also interesting to note that not all ammeters that claim to be scaled to properly read HWDC are, which can be another major source of consternation when using the test ring.
With all of this being said, the number of lines you get on a qualified AS5282 ring or Ketos ring should not be a variable of the ring itself, but some other combination of the variables of the entire inspection system. If you have a good ring and are having problems developing an adequate number of lines at a given amperage, it would indicate some type of problem within the system that would need to be corrected prior to performing any actual inspection work. However, when problems do come up, it is always good practice to look at the ring’s pedigree; any ring without an individual serialized field plot should be suspect right away. Make sure you have a gauge prior to trying to perform a measurement, without one it simply cannot be done.
There are many system variables that can affect how many lines you might see on the ring, but once you have a known good standard, you can then make meaningful tests.