Nutrunner Selection: Tightening Control, RFP and FAT/SAT for Thai Factories

The first thing to decide in nutrunner selection is not “how many Nm of tool to buy”. For each joint, you decide its criticality, meaning what happens if that joint fails, and then translate into the specification a control method, a monitoring approach, proof of capability, and a link between tightening results and the part, all matched to that criticality. Because most of the tightening torque is lost to friction, the bolt’s clamp load is not necessarily on target even when the torque is within specification. That is exactly why you need a system that looks at angle and the tightening curve.

This article is written for the production engineering and quality assurance managers of Japanese-affiliated automotive parts, motorcycle parts and electrical component factories in Thailand (and ASEAN) who tighten critical bolts with click-type torque wrenches or older pneumatic tools and protect quality through marking and re-checks. It works through joint classification, tightening methods, tool types, proof of capability and calibration, data integration, how to think about cost, the RFP, FAT/SAT, and issues specific to Thailand, in that order. Note that the model factory conditions, costs, reduction rates and target examples for acceptance tests in this article are all provisional figures set up for this article only. They are not industry averages, survey results or manufacturer prices. When you evaluate your own case, replace them with your own records and quotations.

Why Nutrunner Selection Is Being Questioned at Thai Factories Now

In the second half of 2026, several triggers for reviewing tightening processes have converged for Japanese-affiliated parts factories in Thailand.

The first is production. According to Reuters’ report of figures from the Federation of Thai Industries (FTI), car production in August 2026 was 124,646 units, up 10.93% year on year. July was also up 6.12% year on year at 117,383 units. However, citing lower exports due to the situation in the Middle East, the FTI cut its full-year 2026 production outlook from the previous 3% increase to a 3.33% decrease. Monthly figures are recovering, but the annual picture is hard to read. In this kind of situation, investment that raises the reliability of the process itself is easier to justify than adding people to thicken inspection.

The second is electrification. According to a JETRO report, HEV sales in Thailand were about 130,000 units in 2024 (up 49.9% year on year), a larger market than BEVs at about 70,000 units and PHEVs at about 9,000 units. Press reports say that EV-related projects approved under BOI investment promotion numbered 198 as of May 2026, with total investment of more than USD 4 billion. HEV and EV parts include many joints where tightening reliability is scrutinized, such as battery packs and the area around motors. When a customer audit asks “Do you have records of tightening results?” or “How do you prove the capability of your tools?”, you need to have answers ready.

The third is a revision of the guideline. VDI/VDE 2862 Blatt 1, the German guideline that sets minimum requirements for tightening systems and tools for the automotive industry, was issued in a new edition in September 2026, replacing the April 2012 edition. This obviously matters to factories that do business with European customers, but even for those that do not, the idea of dividing joints by criticality and varying the depth of control accordingly is usable as a framework for organizing in-house joint management.

Taken together, these three mean that a factory needs a well-grounded answer to the question “Are click-type torque wrenches and marking enough?” Preparation on the quality management side is also covered in “IATF 16949 2nd Edition Preparation: Build Audit Evidence in 90 Days“.

What Is a Nutrunner: How It Differs from Torque Wrenches and Impact Tools, and Its Components

A nutrunner is a tightening tool that turns bolts or nuts with a motor and stops under set conditions while measuring torque and rotation angle during tightening. The Japanese manufacturer ESTIC explains that its nutrunners consist of an AC servo motor, sensors and a controller and are built into machines and equipment, while its hand nutrunners are tools that operators hold by hand to tighten.

There are three main differences from a click-type torque wrench.

• The tool makes the judgment: A torque wrench only signals with a “click” at the set torque, and people have to manage missed tightening and double tightening. A nutrunner measures both torque and angle and gives a pass/fail (OK/NG) for every tightening
• Records remain: The controller holds tightening results (torque, angle, judgment, time) and tightening curves, and can output them externally
• It can be integrated with the line: Interlocks with the PLC can be built, such as locking the tool, ejecting the part or not letting it go to the next process when the result is NG

It is also worth being clear about the difference from impact wrenches and impulse wrenches. These tools advance the fastener intermittently by impacts and are not within the scope of ISO 5393, which defines the performance test method for tightening tools (discussed later). The tools to choose for managing critical joints are nutrunners, which apply torque continuously.

The components are as follows.

• Motor: In ESTIC’s description above, it is an AC servo motor. The scope of ISO 5393 also includes pneumatic and hydraulic tools, but this article focuses mainly on electric tools, on the premise of torque-angle control and data acquisition
• Transducer: The sensor that measures torque and angle. It is built into the tool, and its measurements are the basis for judgment and records
• Controller: It holds the tightening programs (Psets: settings such as the control method, target values and judgment windows for each joint), drives the tool, makes the judgment, and stores and outputs the results
• Peripherals: Reading of part IDs (barcodes, 2D codes), reaction arms, socket trays, and communication with the PLC and data server

Figure 1: Data flow when a nutrunner is introduced. The part ID is read, the controller selects the Pset, the nutrunner, with its motor and torque/angle transducer, tightens the joint, and OK/NG is judged. If NG, the PLC interlock rejects the part, and the results are sent to the data server/MES via Open Protocol or a fieldbus.

As Figure 1 shows, a nutrunner is less a “tool” than a “process system” that links ID reading, judgment, quarantine and recording into one continuous flow. If selection ends with a catalog comparison of the tool alone, you get stuck at the end on the linking of data and on what happens at NG. If your process is mainly about screw float or small screws, “Screw Fastening Automation 2026 — Why Torque Control Alone Cannot Stop Screw Float” is closer to your needs.

Most Torque Is Lost to Friction: Torque-Angle Tightening Control vs. the Torque Method

Correct torque does not guarantee correct clamp load

What you really want from a bolted joint is not torque but the force that presses the parts together, that is, the clamp load (preload). In a technical paper, the test equipment maker PCB Piezotronics explains, as a general rule, that 80–90% of the input torque is spent on friction in the threads and under the head, and only 10–20% stretches the bolt and creates clamp load.

What this ratio tells you is that even a small change in friction moves the clamp load a great deal. The same paper shows that when bolts from the same lot are tightened to the same torque under four conditions, “unlubricated”, “underhead lubricated”, “thread lubricated” and “both lubricated”, the resulting clamp force differs by condition. Differences in plating, how oil is applied, scratches or burrs on the bearing surface, foreign matter: these changes that happen on the shop floor alter the clamp load while barely showing up in the torque figure.

For the design calculation of bolted joints there is VDI 2230 Blatt 1 (November 2015 edition, systematic calculation of highly stressed bolted joints), and for the test conditions for measuring the relationship between torque and clamp force of fasteners there is ISO 16047:2005 (covering carbon steel and alloy steel metric threads M3 to M39). If you check how the design side assumed the friction coefficient, the discussion connects with the judgment conditions on the process side.

The torque method and the torque-angle method

The torque method stops when the target torque is reached. It is simple and sufficient for many joints. However, as described above, variation in friction translates directly into variation in clamp load.

The torque-angle method first tightens to a “threshold torque” and then turns by a specified angle from there before stopping. The PCB Piezotronics paper explains that stopping at a specified rotation angle after reaching the threshold torque results in much smaller clamp load scatter than control by torque alone. The reason is that in the elastic zone the slope of the torque-angle curve is almost constant, and the clamp load is proportional to the rotation angle from the elastic origin. That is why angle is well suited to controlling clamp load.

The caution is that the threshold torque must be set above the snug (seating) zone. If you start counting angle before the parts are in full contact with each other, the angle no longer represents clamp load. As a rule, the threshold should be decided after taking tightening curves on the actual workpiece.

Figure 2: Conceptual torque-angle curve (no values). After snug, angle counting starts from the threshold torque. In the elastic zone the slope is almost constant, and beyond it lies the yield point. The area where the torque window and the angle window overlap is the OK region.

Judge with both a torque window and an angle window

Whether you tighten by the torque method or the torque-angle method, the practical key is to judge with both the “controlled variable” and the “monitored variable”. If you stop on torque, monitor angle; if you stop on angle, monitor torque. As in Figure 2, the result is OK only when the final torque is within the torque window and the final angle is within the angle window.

With this double window, you can see abnormalities that torque alone does not reveal. For example, if there is foreign matter on the bearing surface, snug may come earlier and the angle needed to reach the specified torque may become shorter. If the bolt length is wrong, the point where the curve starts to rise may shift. If the threads are damaged, torque may rise unnaturally partway through. With double tightening, torque alone may rise with almost no rotation. All of these are defects that could pass as OK with only a torque window.

Know the behavior at the yield point too

The same paper explains that as friction increases, the clamp force at the yield point decreases and the torque needed to reach the yield point increases. If you consider a method that tightens close to the yield point to aim for high clamp load, friction control is unavoidable. The paper also gives a rule of thumb, as an initial assumption when the thread friction coefficient is unknown, that “the clamp force at yield torque is about 90% of the tensile yield load”, but this comes with a condition: it is roughly correct when the average total friction coefficient μges is 0.1. Do not apply it as is to your own joints; confirm it with your design department.

The idea of judging load-displacement waveforms against windows in press-fit processes is very similar to nutrunner curve judgment. Reading “Servo Press and Press-Fit Equipment Selection: Thailand RFP and FAT/SAT Guide” alongside this article makes it easier to standardize acceptance tests for waveform judgment across the factory.

Joint Criticality Classification: The VDI/VDE 2862 Approach, Category A and Redundant Monitoring

The three categories in the 2012 edition

The April 2012 edition of VDI/VDE 2862 Blatt 1 divided joints into three categories. Category A was danger to life and limb, Category B was breakdown/stoppage (being stranded due to failure), and Category C was non-critical. Its purposes were to classify joints into categories, to define minimum requirements for tightening tools and systems for each category, and to define minimum requirements for fault detection within a category. Its scope was threaded joints throughout the automotive industry (vehicle manufacturing, component manufacturing, service and repair), and application to fields outside that scope was also recommended.

That edition treated matters such as statistical evaluation, documentation, process safety and process parameters as outside the guideline, to be ensured by the user. It also stated that categories must not be determined on the basis of consequences arising from a chain of damage (cascades).

What to check in the new September 2026 edition

The new edition issued in September 2026 is titled “Minimum requirements for the application of fastening systems and tools — Applications in the vehicle and component industry” and replaces the 2012 edition. It is 31 pages, in German and English, and comes with additional digital content in Excel format. It covers all fastening cases specified by torque, rotation angle or both, serves as a guideline for selecting and using tightening tools and systems, and uses safety, availability and environmental protection as its main criteria.

There is an important caution here. Whether the category definitions or requirements changed in the new edition could not be confirmed at the time of writing. Whenever this article explains Categories A/B/C, it means “in the 2012 edition”. When you build an in-house joint classification table, be sure to check the latest category definitions in the original text of the new edition. Note also that Blatt 2, for plant construction, mechanical engineering, equipment manufacturing and flange connections of pressure vessels, is still being developed, with publication expected in April 2029.

What Category A required (secondary source)

A 2016 industry article (CHEManager) summarized the technical requirements for Category A as of the 2012 edition, for handheld tightening systems, as follows. Note that this is a secondary-source summary, not the original text of the guideline.

• Machine capability of the tightening tool based on VDI/VDE 2645 Blatt 2
• Direct measurement of the control variable and the monitoring variable (which must be different variables)
• Provision of the tightening result
• Self-diagnostic functions and detection of faults in signal transmission
• Redundant design for acquiring the control variable or the monitoring variable

The “judge with both a torque window and an angle window” described in the previous section corresponds to this idea of “control variable and monitoring variable”.

An example of implementing redundant monitoring

As an example of implementing redundancy, AMT (Schmid Group) describes, for its ECR fixture nutrunners, a configuration in which control circuit 1 (applied torque sensor + angle sensor) and control circuit 2 (reaction torque sensor + resolver) are compared independently, and the result is judged as rejected if the values diverge. This is a vendor’s description, and other implementation methods are possible. In the RFP, a realistic approach is to have each maker explain “how redundancy is achieved” and submit it in a comparable form.

How to build your own joint classification table

The inputs to classification are critical symbols on drawings, the FMEA, and customer requirements (such as special characteristic designations). Make one row per joint, list “what happens if it fails”, “how it is tightened now”, “how it is detected now” and “past claims and rework”, and assign an equivalent A/B/C class. Ways to make FMEA work more efficient are covered in “FMEA AI Implementation: RFP and Acceptance Tests for Thai Plants“. This table becomes the foundation for everything that follows: tool selection, the RFP and acceptance tests.

Multi-Spindle, Handheld or Cordless Nutrunner: Choosing the Tool Type

The tool type is determined by the criticality of the joints, the size of the torque, the size of the workpiece, the takt time and the working posture. Comparing the main types gives the following.

Handheld nutrunners and reaction force

The challenge with handheld nutrunners is reaction force. A 1993 study (Freivalds & Eklund) related the reaction torques of six nutrunners to electromyography and subjective ratings, and reported that underpowered tools and soft joints produced larger impulses and worse ratings, and that a pulse-type shut-off mechanism produced the smallest reaction torque. The tools studied were those of the time, not an evaluation of current products, but the viewpoint that “tool capability and joint stiffness drive operator burden” is still useful.

The typical countermeasure is a reaction arm. Atlas Copco claims that its AX1 articulated arm handles tool reaction of up to 120 Nm in the standard AX1 and up to 150 Nm in the AX1.5, and that with position guidance it can also be used for poka-yoke of tightening positions. ESTIC explains that it has reduced the reaction force, which grows stronger at higher torque, through pulse control technology while maintaining tightening accuracy. Both are manufacturers’ descriptions, and whether the same effect is obtained on your own joints needs to be confirmed by testing. For systems that prevent errors in tightening position and sequence, “Poka-Yoke Devices 2026 — Types, Costs and Thai Plant Rollout” is also a useful reference.

Fixed multi-spindle nutrunners

The advantage of a multi-spindle setup is that several bolts can be synchronized and tightened with their stages aligned, from pre-tightening to final tightening. It is especially effective for joints where you want to avoid uneven tightening, such as flanges. Desoutter claims that its TWIN CVI3 controller for multi-spindle and multi-tool applications supports synchronous and asynchronous operation and can hold up to 250 tightening programs (Psets) or assembly processes. AMT gives maximum values for its ECR fixture nutrunners such as up to 1,200 Nm / 1,200 rpm with straight output. All of these are manufacturer-claimed values and do not indicate that one model is better than another.

In a multi-spindle cell, jig design determines tightening quality. Socket positioning, workpiece clamping and absorbing spindle misalignment are things that do not appear in the tool’s specifications. Please also see “Jig Design & Manufacturing in Thailand: RFQ to Acceptance“. A semi-automatic cell that stops short of full automation is another option, and that approach is organized in “Semi-Automatic Machines｜Investment Comparison with Full Automation 2026“.

Controller integration

Recently, products have appeared that move toward handling handheld, fixed and cordless tools within a single controller family. Atlas Copco explains that Power Focus 8 can connect up to 20 tools through one interface. In March 2025, Desoutter announced its AXON controller, which it says can manage all types of corded tools on a single platform. In November 2024, Bosch Rexroth announced next-generation NEXO cordless nutrunners in pistol and angle versions, citing ID code reading with a camera-based scanner, Wi-Fi 6, and the ability to keep operating in manual mode during network failures, and claims they meet VDI/VDE 2862 Category A requirements. Compliance with a category is the manufacturer’s claim; whether it applies to your own joints is something to confirm through your own classification and acceptance tests.

Proof of Capability and Calibration: ISO 5393, VDI/VDE 2645-2 and ISO 6789-1

ISO 5393: performance testing in the laboratory

ISO 5393:2017 is a method for testing the performance of power tools for threaded fasteners in the laboratory. It defines a method for measuring the repeatability (scatter) of torque over a specified range of torque rates, the torque adjustment range specified by the manufacturer, and a number of operating cycles, as well as a method for measuring the accuracy of the built-in torque measuring system. It covers tools that apply torque continuously, regardless of power source, whether pneumatic, hydraulic or electric (including battery-powered), and the torque range is 0.5 Nm to 2,000 Nm. It does not apply to tools that advance the fastener intermittently, such as impact wrenches, impulse wrenches and ratchet wrenches.

VDI/VDE 2645 Blatt 2: machine capability test (MCT)

VDI/VDE 2645 Blatt 2 (September 2014 edition) defines the procedure for the machine capability test (MCT) of tightening machines and states that it makes it possible to judge the stability and reproducibility of the “machine” among the process variables. It covers only tools that can control a target variable such as torque.

How to read the evaluation conditions in manufacturers’ documents

As an actual example of the conditions under which a manufacturer evaluates capability, there is the capability document for the EAD series of angle-head electric nutrunners that Desoutter published in March 2026. The judgment criteria based on VDI/VDE 2645-2 are “tolerance class ±7%, Cm≥2, Cmk≥1.67”, tested at 30%, 80% and 100% of target torque, with two joint stiffness conditions of 30° and 360°, a sample of 100 tightenings per test case, and three units randomly selected from production. Testing according to ISO 5393 was done with 50 tightenings per test case, three units, 0% and 100% of target torque, and 30° and 360° joints, and the document shows the torque range in which the 6σ torque scatter (the width of ±3σ as a proportion of mean torque) is less than 8%.

The point to stress here is that these are the manufacturer’s evaluation conditions, not the acceptance criteria for your own line. Even for the same model, the range shown changes if the criteria differ. In the same document, for example, the EAD20-1300 is listed with two ranges side by side: 3–18 Nm under the ISO 5393 criterion and 3–20 Nm under the tolerance class ±7% capability criterion. When you look at the “application range” in a catalog, always check under which criterion, at which joint stiffness, and with how many tightenings and how many units the value was obtained.

For your own acceptance, test on the actual workpiece or on simulated joints close to your own joints (a hard joint and a soft joint), and agree on the sample size and judgment formula at the RFP stage. The reliable approach is to have the manufacturer state in writing the differences between its test conditions and your conditions.

Manual torque wrenches and in-house periodic checks

The standard for manual torque tools such as click-type torque wrenches is ISO 6789-1:2017. It defines requirements for design conformance and quality conformance testing and for marking, and applies to indicating (Type I) and setting (Type II) tools. It does not include requirements for calibration certificates, which are dealt with in ISO 6789-2. Many factories keep using torque wrenches after introducing nutrunners, for re-torque checks or setup checks, so putting manual tools on the same register makes them easier to explain in audits.

For the nutrunners themselves, decide on transducer calibration certificates (traceable ones), in-house daily checks (such as checks on a reference joint), and the frequency of periodic capability checks. The natural approach is to set the frequency according to joint criticality and number of uses, making it shorter for tools that handle joints equivalent to Category A.

Tightening Data Traceability: Part ID, Pset, Curves and MES Integration

The flow for linking tightening data to the part ID

Written as process steps, the flow shown in Figure 1 is as follows.

1. Read the part ID (a 2D code or similar)
2. The controller automatically selects the Pset for the part number (do not let the operator choose)
3. Tighten
4. Judge OK/NG with the torque window and the angle window
5. If NG, lock the tool via the PLC interlock and eject the part (or quarantine it in an NG area)
6. Store the results (torque, angle, judgment, time, tool ID, Pset) and the tightening curve linked to the part ID
7. Send them to the MES or quality database

ESTIC explains that by outputting tightening data externally and keeping a product-specific assembly process history, it is also possible to respond to product liability (PL) law. Atlas Copco explains that ToolsNet 8 links every tightening event to a VIN or serial number and can aggregate torque curve mismatches to identify outliers. Both are manufacturers’ descriptions, but a state in which “even the curve can be retrieved by part ID” is a great help when investigating causes during a claim.

Communication methods

There are two broad methods for line integration. One is Open Protocol: in the IxB software documentation (v3.16) published on August 31, 2026, Atlas Copco states that Open Protocol is available with all licenses and explains that the behavior on client disconnection (continue as normal, lock the tool, abort tightening, etc.) can be configured. The other is fieldbus: for example, Desoutter claims that the TWIN CVI3 supports PROFINET, PROFIBUS DP, EtherNet/IP, CC-Link, DeviceNet and Modbus TCP via field modules.

The choice of method depends on the maker and network of the existing PLC, what the MES can receive, and the data volume (whether you send curves). A configuration that splits the two is also conceivable: judgment results and IDs go to the PLC via fieldbus, and details including curves go to the data server via Open Protocol or similar. The design of data collection from PLCs is covered in detail in “PLC Data Collection 2026 – Tag Count and Polling Interval Decide It, Not the Protocol“.

Operating rules to decide in advance

• Retention period and storage location: List each customer’s requirements, then either match the longest one or set it per customer
• Handling of communication drop-outs: Can the controller hold data and resend it after recovery? May tightening continue while communication is down?
• Authority to change Psets: Who can change them, and where is the change history kept?
• Handling of re-tightening: May NG parts be re-tightened, with whose approval, and how is the re-tightening result recorded?

These are factory rules, not tool specifications, but if you introduce the tools without deciding them, a “just re-tighten it and send it on” practice emerges on the shop floor. For how to build systems that do not depend on people, see also “Reduce Human Error in Manufacturing | Systems Not Slogans“.

How to Think About Nutrunner Costs and Payback (Model Calculation)

From here on is a model calculation. To repeat, all of the figures below are provisional numbers set up independently for this article; they are not industry averages, survey results or manufacturer prices. Please read them as an example of the way of thinking.

Assumptions for model factory N

• A Japanese-affiliated Tier 2 automotive parts factory in eastern Thailand. One sub-assembly line, with four critical bolts per product (2 processes × 2 bolts). Two shifts
• Currently tightened with click-type torque wrenches, and one inspector per shift re-checks the marking (two in total). Tightening results are recorded only on paper check sheets
• Tightening is done by one operator per shift (two in total)
• Annual labor cost per person (including overheads) is provisionally set at THB 280,000
• Customer claims caused by tightening defects are provisionally set at 2 per year, with handling cost per claim (sorting, transport, 8D response, etc.) of THB 400,000, giving THB 800,000 per year

Comparing the two options (mutually exclusive)

Sensitivity: when no claim reduction is assumed

If the benefit from claim reduction is set to zero, the results change as follows.

Figure 3: Simple payback in the model calculation. H (handheld nutrunners) is about 2.3 years (about 5.5 years without claim reduction), and M (fixed multi-spindle nutrunner) is about 4.3 years (about 7.1 years on the same basis). (Both are calculations based on this article’s provisional values.)

What the calculation tells you

What stands out in Figure 3 is how much the payback period moves depending on whether claim reduction is included. For H, about 2.3 years becomes about 5.5 years; for M, about 4.3 years becomes about 7.1 years. What drives the payback period is not the reduction in people but “how much you expect to reduce outflow (claims)”.

That expectation can only be backed up by a cause analysis of whether current claims really originate in tightening. If there are few tightening-related claims, the 75% reduction assumption loses its basis. Conversely, when you go back through the records, you sometimes find tightening-related cases mixed in with those that were processed as “cause unknown”. Taking stock of past claims and in-process rework records before deciding on introduction is the most direct way to improve the accuracy of the calculation.

Another thing the calculation shows is the difference in character between H and M. H has a small investment and quick payback, but the process in which operators tighten remains, and managing reaction force and tightening sequence remain as issues. M requires a large investment, but in return it can reduce sources of process variation through synchronized tightening and jig positioning. Which is better is not decided by payback alone; it depends on joint criticality, the number of product variants and the outlook for future production volume.

Cautions to avoid double counting

• H and M are mutually exclusive options. Their benefits and costs must not be added together
• M’s claim reduction of 600,000 is the same assumption as H’s, not an additional benefit on top of H
• The reduction of re-check inspectors assumes that the tool judges and records every tightening. If you keep the inspectors after introduction, do not count that portion as a benefit
• For M’s reduction of two tightening operators as well, if people are still needed to load and unload workpieces into the cell, that portion needs to be deducted from the benefit
• The 75% reduction is an assumption, not a value guaranteed by a manufacturer
• The value of handling customer audits, IATF compliance and traceability itself is not included in the monetary figures. Including it would require a separate basis, such as customer requirements or past audit findings

12 Items to Write in a Nutrunner RFP

At a minimum, the RFP (request for proposal) should contain the following 12 items. Rather than specifying a tool model first, writing down the conditions and letting manufacturers and equipment suppliers propose gives you aligned axes for comparison.

Items 1 and 4 in particular cannot be compared across proposals unless the factory prepares them first. For the joint list, the classification table from the earlier section can be used as is. For item 4, the quotation changes depending on whether you accept “the manufacturer’s standard test report” or require “additional testing on our own simulated joints”, so state it clearly at the RFP stage.

What to Check at Nutrunner FAT/SAT

The following are example evaluation items for acceptance tests. Where numbers are given, they are provisional target examples, not standard values.

FAT (before shipment from the supplier’s factory)

• Configuration check: Have the supplier submit, before FAT, the model names of tools and controllers, firmware versions, transducer calibration certificates, and the Pset list (control method, torque window, angle window and threshold for each joint)
• Capability check: Confirm tool capability on the actual workpiece or on hard and soft simulated joints. Have the supplier state the differences between its test conditions and your conditions. Use the sample size and judgment formula agreed in the RFP
• Judgment check: Create intentional defects and confirm that they are judged NG. Five types are a guide: foreign matter on the bearing surface, wrong bolt length, damaged threads, missed tightening and double tightening
• Behavior at NG: Whether tool lock, part ejection, and the rules on whether re-tightening is allowed and by whom work as designed
• Matching check: Behavior when reading the part ID fails, when the link between ID and result is missed, and when a Pset might be wrongly selected at a part number changeover. Data retention and resending during communication drop-outs
• Data check: Whether results (torque, angle, judgment, time, tool ID, Pset) and curves can be searched by part ID. Retention period and storage location
• Multi-spindle check: The sequence of stages in synchronized tightening (pre-tightening to final tightening), and the handling of the whole when one spindle is NG
• Ergonomics (handheld): Whether there is a reaction arm, working posture, and evaluation by the operators who will actually tighten
• Documents: Work procedures, NG procedures and daily check procedures in Thai and English

SAT (after installation on site)

At SAT, record the number of NGs, the breakdown of NGs, the number of reworks and the number of missed ID links during the first 30 days of operation, and compare them with the baseline values from the paper check sheet era. If NGs suddenly increase, it may not be a tool problem; defects that were previously invisible may simply have become visible. It is important to look at the breakdown and separate out problems in incoming parts or upstream processes.

Issues Specific to Thailand and ASEAN

1. Requirements differ by customer

Japanese, European and Chinese OEMs may differ in the items they require for tightening data and in retention periods. Before the RFP, list customer requirements and match them against which customers’ parts run on which lines.

2. Local support from tool manufacturers

Whether calibration, repair and spare parts can be obtained within Thailand directly affects uptime. For example, the Atlas Copco Group says it has been doing business in Thailand since 1957, has a base in Chachoengsao Province, and serves as a regional hub that also supports neighboring countries in the Greater Mekong region. ESTIC established a local subsidiary in Bangkok in March 2012, which handles sales, installation, repair and so on of nutrunners, hand nutrunners and servo presses. These are examples; other manufacturers also have local contacts. In the RFP, check the calibration frequency, the issuer of traceable calibration certificates, and replacement units during repair.

3. PLCs and fieldbuses differ by equipment maker

In Thai factories, it is not unusual for equipment from Japanese, Chinese and local Thai equipment makers to stand side by side on a single line. Since the PLC makers and networks all differ, if you do not specify the nutrunner-side interface (fieldbus type, signal assignment, whether Open Protocol can be used) in the RFP, converters or additional programming will be needed at the installation stage.

4. BOI Announcement No. 4/2569

According to a commentary article by the law firm Tilleke & Gibbins, BOI Announcement No. 4/2569, published in the Royal Gazette on March 31, 2026, is a measure for investment-promoted projects in general vehicle manufacturing (Category 3.6) and PHEV/HEV manufacturing (Category 3.8) that grants a 50% corporate income tax exemption for three years on investment in automation and robotic systems. Import duties on machinery are also exempted. If 30% or more of the value of the modification or of all the machinery is linked to automation machinery manufacturers in Thailand, the corporate income tax exemption rate is raised to 100%. Applications are said to be accepted until the end of 2027. However, this information is based on a commentary article, and the original BOI text could not be confirmed. Whether your business qualifies, and whether nutrunners and peripheral equipment are included in the eligible investment, should be confirmed individually with the BOI and specialists.

5. Training Thai operators and maintenance staff

Embed who can change Psets, and who does what when an NG occurs, through procedures written in Thai and hands-on training. If re-tightening of NGs is left to the operator’s judgment, you end up in a state where everything is OK in the data but not actually controlled. It is important to restrict authority through the controller’s login settings and to keep this consistent with the procedures.

6. Rolling out to sites in Vietnam

The joint classification table, RFP and acceptance test items created in Thailand can basically be reused on similar lines at other sites such as Vietnam. If you accept equipment against the same criteria, it also becomes easier to compare quality between sites. However, laws, subsidies and incentive schemes differ by country, so confirm them individually in each country. If you extend to automated tightening combined with robots, “Assembly Automation Robots 2026 — Part Tolerance Decides, Not Robot Accuracy” and “Robot Force/Torque Sensor Selection: Retrofit and Acceptance in Thailand” are also useful references.

Latest Developments: The New VDI/VDE 2862 Edition and the Tightening Tool Market

• New edition of VDI/VDE 2862 Blatt 1 (September 2026): A new edition replacing the 2012 edition has been issued. A draft was published in June 2025, and a correction to it was issued in September 2025. As noted above, whether the category definitions have changed needs to be checked in the original text of the new edition
• Atlas Copco second-quarter 2026 results (July 16, 2026): Group orders were SEK 50,951 million, up 27% year on year (organic growth of 26%). Orders in the Industrial Technique business area, which includes tightening tools, rose 11% on an organic basis, and the company explained that equipment grew solidly in automotive and general industry. As a new product, it cited the “EIDS series” of low-torque screwdrivers with transducers, highlighting traceability and assembly consistency
• Desoutter’s capability document (March 2026): For the EAD series of angle nutrunners, Desoutter published a document listing, by model, the applicable torque ranges under both the ISO 5393 criterion and the capability test criterion. This trend of presenting capability separately by criterion is a useful reference when specifying test conditions in an RFP
• Open Protocol documentation (August 31, 2026): In the IxB software documentation v3.16, Atlas Copco stated that Open Protocol is available with all licenses. The Power Focus 8 product documentation has also been updated to v3.16 (published August 26, 2026)

This information is based on manufacturers’ results, announcements and documents, and does not represent the movement of the market as a whole. Still, it shows that products and documentation in the tightening tool field are moving in the direction of “capturing and connecting data”. A comparison of automation including screw feeding is covered in “Automatic Screw Feeder Selection for Thai Factories | RFP and FAT/SAT“.

A 90-Day Plan for Nutrunner Selection

Days 0–30: Inventory and understanding the current state

• Identify every joint on the target line and list thread size, target torque, current tool and current detection method
• Assign an equivalent A/B/C criticality to each joint based on drawings, the FMEA and customer requirements (check the definitions in the new edition of VDI/VDE 2862)
• Compile the records of claims and in-process rework over the past 2–3 years and separate out those caused by tightening
• Check the register and calibration records of manual torque wrenches

Days 31–60: Draft policy and design

• Decide the control method (torque method or torque-angle method) and the approach to judgment windows for each joint
• Draft the tool type (handheld / fixed single-spindle or multi-spindle / cordless) and redo the model calculation with your own figures
• Design the part ID, Pset selection, behavior at NG, data storage location, and PLC/MES integration
• List each customer’s data requirements and issue the RFP (12 items)

Days 61–90: Acceptance criteria and pilot

• Compare proposals and finalize FAT items and pass criteria (simulated joints, intentional defects, sample size)
• Choose one process with high criticality and a record of claims or rework as the pilot target
• Decide how to capture the baseline values to compare at SAT (number of NGs, number of reworks, missed links)
• At the investment decision meeting, present the payback sensitivity (with and without claim reduction) and decide on the order

Frequently Asked Questions

What is the difference between a nutrunner and a torque wrench?

A click-type torque wrench is a manual tool that signals when the set torque has been reached; people make the pass/fail judgment and keep records. A nutrunner tightens with a motor, measures torque and angle, judges OK/NG for each tightening, and records and outputs the results and curves. Interlocks with the line, such as locking the tool at NG, can also be built.

How do you choose between the torque method and the torque-angle method?

The torque method is simple and can be used for many joints, but variation in friction translates directly into variation in clamp load. For joints where clamp load control is important, the torque-angle method, which turns by a specified angle from the threshold torque, is suitable. The threshold is placed above the snug zone, and with either method the basic practice is to judge with both a torque window and an angle window.

What is Category A in VDI/VDE 2862?

In the 2012 edition of VDI/VDE 2862 Blatt 1, joints were divided into Category A (danger to life and limb), B (breakdown/stoppage) and C (non-critical). Category A is the class with the strictest requirements. Because a new edition was issued in September 2026 and replaced the 2012 edition, check the latest category definitions and requirements in the original text of the new edition.

What processes are multi-spindle nutrunners suited to?

They suit processes where several bolts should be synchronized and tightened evenly in stages, such as wheel hubs and flanges. Because the position is fixed by a jig, variation in tightening sequence by operators is also suppressed. On the other hand, the investment is large, and you need to decide in advance how to handle the whole when one spindle is NG.

How much tightening data should be stored and linked?

At a minimum, store torque, angle, judgment, time, tool ID and Pset linked to the part ID. For critical joints, also storing the tightening curve helps in investigating causes during a claim. Decide the retention period by listing each customer’s requirements.

What should be checked in a nutrunner RFP and at FAT/SAT?

In the RFP, specify 12 items, including the joint list and criticality, control method and judgment windows, method of proving capability and sample size, behavior at NG, data items, and the PLC/MES interface. At FAT, actually verify NG judgment with intentional defects, ID linking and behavior during communication drop-outs; at SAT, compare NGs and rework in the first 30 days of operation with baseline values. For points related to BOI incentives or laws and regulations, confirm individually with the competent authorities or specialists.

Summary

Nutrunner selection is not the work of choosing a tool’s torque capacity; it is the work of deciding a control system matched to the criticality of each joint. Because most torque is lost to friction, for critical joints consider the torque-angle method and judgment by both a torque window and an angle window, and make the curves visible. Do not take manufacturers’ evaluation conditions at face value for tool capability; confirm it with your own joints and sample sizes. And only when part ID, Pset, judgment, quarantine and recording are linked into one continuous flow does it become traceability that can answer customer audits. As the model calculation showed, what moves the payback period is the expected reduction in claims. Starting by taking stock of past claims and rework and building a joint classification table may look like a detour, but it is the sure way forward.

You can consult TOMAS TECH even at the early stages, such as taking inventory of joints, classifying criticality, or drafting the RFP and acceptance criteria. We will work with you to sort out which process to start with, in line with the actual conditions of your factory in Thailand. Please get in touch via this form.

References

• PCB Piezotronics: Fundamentals of Torque-Tension and Coefficient of Friction Testing
• PCB Piezotronics: Tightening Strategies for Bolted Joints
• EVS: ISO 5393:2017 Rotary tools for threaded fasteners — Performance test method
• EVS: ISO 6789-1:2017 Hand torque tools — Part 1
• AFNOR: ISO 16047:2005 Fasteners — Torque/clamp force testing
• VDI: VDI 2230 Blatt 1 (November 2015 edition)
• VDI: VDI/VDE 2645 Blatt 2 Machine capability test (MCT)
• DIN: VDI/VDE 2862 Blatt 1:2012-04 table of contents and foreword preview
• DIN Media: VDI/VDE 2862 Blatt 1:2026-09
• DIN: Correction to the VDI/VDE 2862 Blatt 1 draft (September 2025)
• VDI: VDI/VDE 2862 Blatt 2 (in development)
• CHEManager: Qualitätsanforderungen hochgeschraubt (June 28, 2016)
• AMT / Schmid Gruppe: Fixture Nutrunners ECR
• Desoutter: Angle-head Electric Nutrunner Product capability performance (March 2026)
• Desoutter: TWIN CVI3 product page
• Desoutter: AXON press release (March 2025)
• Atlas Copco: Power Focus 8
• Atlas Copco: Power Focus 8 product documentation (v3.16)
• Atlas Copco: IxB Software Open Protocol (v3.16, August 31, 2026)
• Atlas Copco: ToolsNet 8
• Atlas Copco: AX1 articulated arm (July 22, 2021)
• Bosch Rexroth: Next-generation NEXO cordless nutrunner (November 25, 2024)
• EDINET DB: ESTIC Corporation annual securities report (fiscal year ended March 2026)
• Inderes: Atlas Copco Second-quarter report 2026 (July 16, 2026)
• Atlas Copco Group: Q2 2026 results presentation (July 16, 2026)
• Atlas Copco Group in Thailand
• Tilleke & Gibbins: Thailand Unveils New Incentives for Automotive and HEV/PHEV Manufacturing (April 21, 2026)
• Automotive Logistics: Thailand’s BOI attracts 198 EV investment projects (July 8, 2026)
• Capital.com (Reuters): Thailand’s August car production rises 10.9% on year (September 28, 2026)
• New Straits Times: Thailand car production rises 6.1pct on year in July (August 2026)
• JETRO: Thailand (2) — The current state of the automotive industry, global supply chains and the response of Japanese companies (November 12, 2025)
• Penn State: Freivalds & Eklund (1993) Reaction torques and operator stress while using powered nutrunners