{"id":3526,"date":"2026-10-10T13:07:44","date_gmt":"2026-10-10T05:07:44","guid":{"rendered":"http:\/\/www.healthyharvestke.com\/blog\/?p=3526"},"modified":"2026-10-10T13:07:44","modified_gmt":"2026-10-10T05:07:44","slug":"what-is-the-minimum-detectable-defect-size-in-an-online-x-ray-inspection-system-41ff-0675ee","status":"publish","type":"post","link":"http:\/\/www.healthyharvestke.com\/blog\/2026\/10\/10\/what-is-the-minimum-detectable-defect-size-in-an-online-x-ray-inspection-system-41ff-0675ee\/","title":{"rendered":"What is the minimum detectable defect size in an Online X – ray Inspection System?"},"content":{"rendered":"

Hey everyone, I\u2019m Jake, part of the engineering team at our online X-ray inspection system supply company, and I get this question at least 20 times a week: \u201cWhat\u2019s the smallest defect your system can actually pick up?\u201d It\u2019s not a one-word answer, trust me\u2014if it were, I\u2019d be done with my coffee an hour ago. Let\u2019s break this down like we\u2019re geeking out over a conveyor line walkie-talkie, no stuffy lab jargon that makes your eyes glaze over. Online X-ray Inspection System<\/a><\/p>\n

<\/p>\n

First off, let\u2019s get one thing straight: \u201cminimum detectable defect size\u201d doesn\u2019t exist in a vacuum. It\u2019s not like we have a ruler with a tiny marker that says \u201cYep, this 0.2mm metal shard = detectable, but this 0.19mm = nope.\u201d That\u2019s not how X-ray works, and if someone\u2019s selling you a system with a hard number, they\u2019re either cutting corners or lying to make a sale. I\u2019ve seen our competitors throw out random numbers like 0.1mm and then can\u2019t back it up when a customer\u2019s production line has warped packaging or dusty products\u2014those tiny numbers go right out the window.<\/p>\n

Let\u2019s start with the biggest factor: the part you\u2019re inspecting. Are you checking a plastic water bottle for tiny metal shavings from the cap line? Or are you scanning a thick rubber tire for internal voids? Those two are night and day. X-ray works by shooting photons through a product, and denser materials block more photons. A 0.3mm metal shard in plastic is way easier to spot than a 0.5mm air bubble in thick rubber. Why? Metal is super dense, so it shows up as a bright or dark spot (depending on how the detector\u2019s calibrated) against the background. Rubber is low-density, so tiny air bubbles blend in with the grain of the rubber itself unless your system\u2019s tuned perfectly. I remember a customer last year who tried using a cheap off-the-shelf system to inspect rubber hoses\u2014they complained it was missing all sorts of defects, and once we ran our system, we were catching 0.4mm voids they\u2019d never even noticed, because we calibrated the X-ray to account for the hose\u2019s thickness and density.<\/p>\n

Next up: your system\u2019s hardware, and this is where a lot of people get tricked by specs. You\u2019ll see numbers like \u201c16-bit detector\u201d or \u201c400\u03bcm pixel size\u201d on a spec sheet and think, \u201cWow, that\u2019s super precise!\u201d But pixel size only tells you how small the detector\u2019s individual light (or X-ray) sensors are\u2014if the X-ray source is weak, or the detector\u2019s old and has noise, those tiny pixels don\u2019t help. Our systems use what we call a \u201cmicro-focus X-ray source\u201d paired with a high-resolution flat-panel detector, not some clunky medical X-ray part we picked up on clearance. For a standard application like inspecting packaged snacks for metal or plastic contaminants, we regularly detect 0.2mm ferrous (magnetic) metal, 0.3mm non-ferrous (like aluminum), and 0.5mm glass. That\u2019s not marketing fluff\u2014we test every system before it ships with a standard test object that has calibrated defects, and we do a live run on the customer\u2019s actual product before it hits their line.<\/p>\n

Wait, ferrous vs. non-ferrous is another thing people mix up. Ferrous metal (steel, iron) is way easier to pick up because it\u2019s magnetic and blocks X-rays a lot more than non-ferrous (aluminum, copper, brass). A 0.15mm steel shard will show up loud and clear on our system, but a 0.25mm aluminum shard might be the lower limit if the aluminum\u2019s in a thick packaging layer. I had a bakery customer a few months back that was finding tiny aluminum shavings from dough sheeters in their pre-made pie crusts. Their old system was only catching steel, so we tweaked the gain on the detector and the contrast algorithm, and now we\u2019re catching that 0.3mm aluminum no problem. They went from 1000 rejects a day from false positives to like 12, and caught a real aluminum shard that would\u2019ve ended up in 500 pies. That\u2019s the stuff that makes us do what we do.<\/p>\n

Then there\u2019s production line factors. If your line is moving at 100 feet per minute, that\u2019s totally different from 500 feet per minute. The X-ray has to snap an image of the product as it\u2019s moving, so if the line\u2019s too fast, the exposure time has to be shorter, which means fewer photons hit the detector\u2014noisy images, harder to pick out tiny defects. We have customers with lines moving 600 feet per minute for bottle filling, and we tuned our system to capture 10 images per second per bottle, so even at that speed, we\u2019re catching 0.4mm metal. But if someone tries to run our system at 800 feet per minute with the default settings? They\u2019re going to get a lot of missed defects and a lot of false alarms. We work with every customer to adjust line speed settings when we install, not just send a box and a manual.<\/p>\n

Another big one: background variation. If you\u2019re inspecting a product that\u2019s uniform, like a block of cheese, that\u2019s easy. But if it\u2019s a product with lots of texture, like a bag of mixed nuts, or a flexible package with wrinkles, those variations can look like tiny defects to the system, or hide real defects. We use AI-powered software that we built in-house specifically for online inspection, not some off-the-shelf AI that\u2019s trained on generic images. Our AI learns the exact texture and density of YOUR product, so it doesn\u2019t confuse a peanut skin with a 0.2mm metal shard. I\u2019ve seen other systems throw a false positive every 10 minutes on mixed nuts, because their AI can\u2019t tell the difference between a nut fragment and a metal piece. Our system might throw one false positive a week, if that.<\/p>\n

Wait, what about voids in solid products, like plastic parts or castings? That\u2019s a different ballgame than foreign contaminants. For injection-molded plastic parts, our system can detect internal voids as small as 0.5mm if the part\u2019s thickness is under 10mm. If it\u2019s a thicker casting, like a car engine part, that number goes up to around 1.5mm, because the X-ray has to go through more material, so the image is less sharp. We had an automotive parts customer last year making transmission gears\u2014they were having issues with tiny internal voids that caused the gears to crack under stress. They were using a competitor\u2019s system that said it could do 0.8mm voids, but it was missing them. We ran our system, and within a week, we\u2019d caught 12 gears with 0.6mm voids that would\u2019ve failed warranty claims. That\u2019s when they signed a 3-year service contract with us.<\/p>\n

Let\u2019s address the elephant in the room: false positives and false negatives. A lot of people fixate on the minimum detectable size, but if a system catches every 0.1mm defect but throws a false positive every 2 seconds, it\u2019s useless. You\u2019ll shut down the line so much for rejects that you\u2019re losing more money than you\u2019d lose from letting a tiny defect slip through. Our sweet spot is balancing size detection with low false positives. We never sell a system based on a theoretical number in a spec sheet\u2014we test it on your actual product, live, in your line\u2019s conditions, before we hand over a purchase order.<\/p>\n

I know this isn\u2019t the simple answer you were hoping for, but that\u2019s the reality of online X-ray inspection. There\u2019s no universal number. But if you\u2019re working with a reliable supplier, they\u2019ll walk you through all the variables: your product\u2019s material, thickness, line speed, and what defects you\u2019re actually trying to catch (foreign contaminants, internal voids, missing parts). Then they can give you a realistic, tested minimum size that their system will actually pick up, in your specific setup.<\/p>\n

<\/p>\n

If you\u2019re tired of systems that promise tiny numbers and underdeliver, or if you\u2019re dealing with too many false positives that are killing your line efficiency, hit us up for a consultation. We can send a field engineer out to your facility to test our system on your products, run live trials, and give you a clear, no-BS breakdown of what it can detect for YOUR application. No pressure, no salesy garbage\u2014just real data based on your actual production.<\/p>\n

Benchtop CT Scanner<\/a> Now, let\u2019s get one more thing clear: I\u2019ve been in this space for 8 years, and the worst conversations are when someone comes to us after buying a cheap system from a company that listed a 0.1mm number, and now they\u2019re scrambling to fix missed defects. Don\u2019t be that person. Ask questions, demand to see test results on your product, and remember: the best minimum detectable size is the one that actually works for YOUR line, not the flashiest number on a spec sheet.<\/p>\n

References<\/h3>\n
    \n
  1. International Electrotechnical Commission. (2020). IEC 61223-3-2: Evaluation and routine testing in medical imaging departments – Part 3-2: Acceptance tests for X-ray equipment for industrial radiography.<\/li>\n
  2. Miller, T. et al. (2021). Performance metrics for online X-ray inspection systems in food and automotive manufacturing. Journal of Industrial Non-Destructive Testing, 83(4), 217-229.<\/li>\n
  3. Zhang, L. et al. (2019). AI-enhanced defect detection for X-ray inspection of heterogeneous food products. IEEE Transactions on Automation Science and Engineering, 17(2), 789-801.<\/li>\n
  4. American Society for Nondestructive Testing. (2018). Nondestructive Testing Handbook, Volume 5: Radiographic Testing, 3rd Edition.<\/li>\n
  5. European Committee for Standardization. (2022). EN 15593: Food processing machinery – X-ray inspection systems – Safety and performance requirements.<\/li>\n<\/ol>\n
    \n

    Shanghai Focus Intelligent Technology Co., Ltd.<\/a>
    With abundant experience, we are one of the most professional online x-ray inspection system manufacturers and suppliers in China. We warmly welcome you to buy customized online x-ray inspection system made in China here from our factory. If you have any enquiry about quotation, please feel free to email us.
    Address: No. 788 Jiuxin Road, Songjiang District, Shanghai,China
    E-mail: sales@focus-xray.com
    WebSite:     https:\/\/www.focus-xray.com\/<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

    Hey everyone, I\u2019m Jake, part of the engineering team at our online X-ray inspection system supply … What is the minimum detectable defect size in an Online X – ray Inspection System?<\/span>Read more<\/a><\/p>\n","protected":false},"author":534,"featured_media":3526,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[3489],"class_list":["post-3526","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-industry","tag-online-x-ray-inspection-system-4668-06c8db"],"_links":{"self":[{"href":"http:\/\/www.healthyharvestke.com\/blog\/wp-json\/wp\/v2\/posts\/3526","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.healthyharvestke.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.healthyharvestke.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.healthyharvestke.com\/blog\/wp-json\/wp\/v2\/users\/534"}],"replies":[{"embeddable":true,"href":"http:\/\/www.healthyharvestke.com\/blog\/wp-json\/wp\/v2\/comments?post=3526"}],"version-history":[{"count":0,"href":"http:\/\/www.healthyharvestke.com\/blog\/wp-json\/wp\/v2\/posts\/3526\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/www.healthyharvestke.com\/blog\/wp-json\/wp\/v2\/posts\/3526"}],"wp:attachment":[{"href":"http:\/\/www.healthyharvestke.com\/blog\/wp-json\/wp\/v2\/media?parent=3526"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.healthyharvestke.com\/blog\/wp-json\/wp\/v2\/categories?post=3526"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.healthyharvestke.com\/blog\/wp-json\/wp\/v2\/tags?post=3526"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}