Discussion and questions related to the course Engine Building Fundamentals
Crack testing (or NDT inspection) with magnetic particles as far as i know it can be used only for ferrous metals.
From your experience, for a cylinder head or engine block which is made from allu alloy for instance, when you are susspecting some cracks inside (reasond can vary) which are not at visible areas located, an x-ray method could be an answer to this problem? Or you can use an dye penetrant method and then to scan for cracks with an endoscope? I'm just beeing curious which method could get the best results.
I dont have much radiographic experience but have done a fair bit of MPI and DPI in my time. You are correct, MPI is only suitable for ferro-magnetic materials and will only show flaws that reach the surface. It wont show an internal void in a casting or a lap in a forging for instance.
DPI is commonly used on nonferrous automotive parts, but as you suggest it is not really suitable for areas that you cant access easily. For good detection reliability, the die really needs to mechanically cleaned (solvent wipe) before developing. So not really suitable for inside a water jacket for instance.
Xray and gamma I only have had a little exposure mostly with aerospace forgings. In that industry it was always explained to me as being good for detecting "volumetric" flaws (ie. a crack big enough that metal has fallen out, or an internal void or porosity) but it was not so good for spotting fine cracks or intergranular fracture where the metal was still tightly fitting together. Whereas MPI/DPI were both very good at detecting fine/invisible surface flaws.
Internal cracks in a cylinder head are typically detected by pressure testing in a water tank.
As Adam pointed out, different methods have different weaknesses, this is why some critical welds may have both X-Ray and magnetic testing done.
From my brief encounter with a roof brick manufacturing line (wait for it ^.^), hidden cracks can be detected by means of resonance measurements. Basically, a cracked object will have a different resonance frequency in comparison to a whole piece. Unfortunate to this cause, method requires a known good piece(s) to build base value.