French (Fr)English (United Kingdom)

DPC NEWS: a website dedicated to Penetrant Testing and Magnetic Testing



visits on site since April 2008

Log in


Receive HTML?

March / April / May 2016 : 1956 to Today, the Road from MIL-I-25135 to AMS-2644

Written by Dubosc
Monday, 04 April 2016 19:26

William E. Mooz
Met-L-Chek Company 
1639 Euclid Street
Santa Monica, CA 90404
310-450-1111, fax (310) 452-4046, e-mail This e-mail address is being protected from spambots. You need JavaScript enabled to view it

If there was one thing that could be said about the change from the original MIL-I-25135 to the present AMS-2644, it is that “We have come a long way, baby!” In 1956, when the first version of MIL-I-25135 was published, it is not a very far stretch of imagination to say that perhaps colored water could have qualified as a penetrant and be listed on the QPL. In fact, Loy Sockman, the founder of Met-L-Chek bought a gallon of an approved penetrant, cut it in half with kerosene, submitted it for qualification and it was approved! The writers of the specification simply did not have the knowledge or the data to design a better document. As an example, the specification required that “The penetrant shall contain dyes that fluoresce…”, and “The fluorescent brightness and contrast shall be equal or superior to that of the standard sample.” No tests were specified to evaluate these requirements. This paper will examine some of the devices and methods that were involved in the history of developing SAE-AMS 2644, the qualifying specification that is in use today.

The failings of the early specifications were recognized, and there was a conscious effort to develop more explicit requirements across the board. Since the technology of penetrant inspection was new, one of the first subjects that needed to be addressed was the identification of those characteristics that were important and how to measure them. From June 1963 to December 1964, Monsanto Research Corporation performed work relative to developing criteria for a penetrant specification. (1) They developed several laboratory devices and made tests on water washability, kinematic viscosity, tank life, specific gravity, flash point, infrared characteristics, temperature stability, and water tolerance. Three special pieces of laboratory equipment were developed and described in their report. It was perhaps obvious that a standard for fluorescent penetrant brightness had to be established, and so a device (Fig. 1a) was constructed to measure and test this. As a standard, it used a piece of fluorescent glass. (Figs. 1c and 1d) The candidate penetrants were diluted, and then were soaked into filter paper that was dried and then tested in the device. The idea behind this was that the diluted penetrant soaked filter paper mimicked what an actual penetrant indication would look like when it was developed.

Figure 1a
Test device

Figure 1b
Test device control

Figure 1a
Glass standard

Figure 1d
Glass standard under UV

A special box was designed that fluidized a bed of dry powder developer for testing. And a device to test water washabilty also was invented.

Figure 2a
Dry developer test box

Figure 2b
Water washability tester

The devices that they designed were then written into a draft specification MIL-I-8963 (ASG) dated 13 October 1965, but were never written into MIL-I-25135. The penetrant manufacturers made each of these devices, and although they were interesting at the time, they were judged to be unnecessary and they found their way into the museum of discarded test equipment.

A glass sample that had been sandblasted with various grits (Fig. 3) was suggested for testing background fluorescence, but also died.

Figure 3 : Sandblasted glass

Other tests relevant to the adequate performance of the penetrant systems were selected using many standard laboratory procedures and included requirements for testing the corrosive properties of the penetrant materials, their flash point, bioresistance, brightness, water tolerance of water washable penetrants, surface removability, tank life, storage stability, temperature stability, and UV-A stability. These were decided upon without diifculty because they were tests that were mostly standard tests made according to standard procedures.

But a method of measuring the key parameter evaded analysts, and became the focus of the group’s efforts for a considerable length of time. This was a method of measuring the sensitivity of the penetrants, or their ability to find defects of various sizes. The initial method listed in the first edition of MIL-I-25135 was to use thermally cracked aluminum blocks that were divided into halves. (Fig. 4a) The reference penetrant was applied to one half and the candidate penetrant was applied to the other half. The block was then processed and the results compared by eye. (Fig. 4b)

Figure 4a
Thermally cracked block

Figure 4b
Penetrant comparison

This was an enormously unsatisfactory method, and it resulted in great difficulties for those who purchased penetrants, because they had no way to be sure that penetrant A was as sensitive as penetrant B. The blocks were prepared locally and they were never the same from one block to another, and the cracks formed were of different sizes and not at all representatives of the cracks sought in usual inspections. The Air Force Materials Laboratory recognized this and they actively searched for a better method over a period of years. When the government decided to get out of the specification writing business and turn it over to civilian organizations, AMS Committee K was formed and the search for a more scientific method of measuring sensitivity was launched. This resulted in a number of potential methods that were considered.

The Air Force had contracted with Ohio State University (2) in 1960 to address this question, and they had developed cracked chrome panels, and were able to produce these showing coarse, medium, or fine cracks. The Monsanto research work suggested that the Ohio State cracked panels might be used to measure sensitivity. However, they stated as follows in their report: “Crack senstivity is a value of major importance requiring improvements upon the insensitive aluminum block specimen, plated iron, ceramic, titanium, and the Ohio State nickel-chrome test specimens.”

Japanese companies were later able to refine these to the place where they could make the panels with cracks of various sizes, such as 50 or 30 microns. These panels were further developed so that the cracks on a single panel ranged from coarse to fine. (Figs. 5a, 5b, 5c, and 5d)

Figure 5a
Test Panels

Figure 5b
Panels in use

Figure 5c
Tapered Panels

Figure 5d
Tapered Panels in use

Accordingly, the Air Force awarded a contract to Paul Packman, of the University of Tennessee to develop a test piece, and the result was the test bar that has been nicknamed the “Hernia Bar”, because of its large size and heavy weight. (Fig. 5e)

Figure 5e : Packman Hernia bar

Turco, then a penetrant manufacturer, developed aluminum panels that had small circular impressions of various sizes on them that were designed to test how small a flaw could be detected. These panels were made in several variations and could be used to test a single penetrant or to make comparisons between two penetrants. (Figs. 6a, 6b, 6c, and 6d)

Figure 6a
Turco Panel
Figure 6b
Panel in use
Figure 6c
Turco Panel
Figure 6d
Panel in use

James Alburger, of Uresco, patented the meniscus method and sold the kits for making the measurements. This was an optical test that purportedly measured the dye content of a penetrant as a dark spot in the center of a glass lens. In theory the dye content was a measurement of the penetrant sensitivity. (Figs. 7a, 7b, and 7c)

Figure 7a
Meniscus Test Kit
Figure 7b
Kit Contents
Figure 7c
Black Spot

The late Frank Vicki, who chaired the SAE AMS Committee K at the time, developed a tapered cracked chrome panel in collaboration with a Japanese manufacturer, that had eight parallel cracks of decreasing size that were accurately measured, and this panel was also tested for its ability to determine the sensitivity of penetrants.

There were two problems with each of these methods. The first was that the flaws or other attributes tested did not replicate the flaws that were to be detected on real parts. The second was that the tests had no way of quantifying the results in numeric terms. The Air Force Materials Laboratory tested each of these methods and was the arbiter of the results. In addition, the Air Force had examined low cycle fatigue cracked bars that had been in use at GE and determined that these satisfied the need for cracks typical of what penetrants were supposed to locate. (Fig. 8a) At about the same time, a spotmeter (Fig. 8b) was developed that could be focused on a crack indication and measure its brightness in numerical terms.

At one point, the Air Force prepared seven penetrant samples that had been tested by the low cycle fatigue cracked bars and found to be of different sensitivities. These samples were sent to the penetrant manufacturers, who were asked to test them. Grover Hardy, of the Air Force Lab, asked the results to be arranged in order of the penetrant sensitivities. Each sample was coded by a letter, the letters being A, E, G, M, O, T, and Y. Grover’s sense of humor came across when he announced that when properly arranged from the lowest sensitivity to the highest, according to the Air Force results, the letters read “YA GOT EM”.

The development of the spotmeter and the use of the low cycle fatigue cracked bars satisfied what was sought and were chosen by Committee K, with the approval of the Air Force, for use in AMS-2644.

Figure 8a
Low Cycle Fatigue bars

Figure 8b

There was one last thing that had to be dealt with. Repeated tests with this system on a single penetrant resulted in a range of results forming a normal distribution curve. It was necessary to use this statistical information in deciding whether a candidate penetrant was equal or superior to the reference penetrant. Initially, Committee K suggested that the candidate penetrant would be judged equal to the reference penetrant if the measured sensitivity was within one standard deviation of the reference penetrant against which it was being compared. The problem with this became obvious when it became apparent that if the reference penetrant were compared with itself, it would fail this test almost 32 percent of the time. The final wording that was developed and accepted for the specification was as follows:

Acceptance Criteria: Indication brightness data from the candidate material shall be compared to the lower standard deviation curve for corresponding data from the appropriate reference material (Table 2). The lower standard deviation curve shall be generated from a minimum of three runs of the appropriate reference. The candidate material shall be acceptable when 80 % of the points that generate its curve lie above the lower standard deviation curve of the appropriate reference material. The tests shall be valid only if the data for the current run with the referenced material falls within the standard deviation established from previous runs, or if the three runs to establish the lower standard deviation are current. Reference material shall be tested periodically to check validity of tests.

This paper has described some of the various methods that were suggested and tested to qualify inspection penetrants beginning with the first such specification published in 1956. A major problem was to identify those physical characteristics that were important to include in the specification, to assign values to these, and to develop methods of testing candidate products for conformance to the levels required. A number of such characteristics were initially thought to be important but were subsequently discarded, and finally one key factor was determined to be the sensitivity level, and methods of quantifying this were developed and perfected.

We have come a long way, with the present availability of five sensitivity levels of fluorescent penetrants, four methods of penetrant removal, and four different forms of developer. It took 12 years to produce the first version of AMS-2644, but the vexing question of measuring sensitivity was finally laid to rest, at least for the moment. SAE AMS Committee K continues to work with the specification so as to insure that purchasers of penetrants can rely upon the QPL qualification as a measure of quality.



Monsanto Research Corporation contracts AF-33 (616)-8483 and 33(615)-1484, Report date July 1965

Ohio State University Research Foundation contract AF-33 (616)-7420, Report dates June 1960, November 1960, February 1963 and February 1964

DOD, MIL-I-25135 (ASG) Inspection Materials, Penetrant, Department of Defense, 6 August 1956

SAE, AMS 2644, Inspection Material, Penetrant, August 1996

DOD, MIL-I-8963 (ASG), (Proposed), Inspection Materials, Fluorescent Penetrant, Department of Defense, 13 October 1965

Last Updated ( Tuesday, 05 April 2016 23:02 )