Load Cell Selection: Load Direction, Use Case, Environment and RFP for Thai Factories

Introduction — Load Cell Selection Does Not Start with “How Many kg Is the Product”

The weighing value of a raw material hopper drifts a little every day. After piping was connected to a tank, the reading changed while the tank was still empty. You want to see the press-fit force of a bearing, but you do not know what to install. The wrong load cell was selected and it was destroyed by overload. After the cable was extended, the values stopped matching. Every machine has a different load cell and a different amplifier, and every replacement causes confusion. All of these are problems related to how load cells are selected and installed.

Here is the main point of this article up front. In this article, we take the view that load cell selection is decided by the following three points before “what capacity in kg (or kN)”.

• (1) How the load is applied: compression or tension, whether there is eccentric load or side load, and how the cell is mounted
• (2) Weighing or force monitoring: whether you weigh a static weight, or watch load changes during press-fitting, caulking or pressing at high speed to make pass/fail judgments
• (3) Environment: waterproofing, explosion protection, temperature and cable length

Capacity is calculated after that, taking dead load, the number of cells and load imbalance into account. At acceptance, the displayed value, the output value and the value read by the PLC are compared using test weights or actual loads. In addition, the eccentric load test and the change in values with piping connected are recorded.

This article is not a recommendation or ranking of specific products. Manufacturer specifications are based on each company’s published materials and are presented in the form “the company states that…”. The cost estimate and the capacity calculation example in the second half use hypothetical numbers for this article. Connecting scale values to PLCs and MES, checkweighers, 6-axis force/torque sensors for robots, servo press equipment itself, and data integration for tensile testing machines are left to their own articles, and this article links to them where relevant.

Why Thai Factories Are Revisiting Weight and Force Records Now

Capital investment activity continues in Thailand. According to the Thailand BOI (Board of Investment), there were 1,299 investment promotion applications in January to June 2026, worth USD 43.6 billion (about THB 1.47 trillion), up 37% year on year. By sector, automotive accounted for 122 applications worth USD 759.2 million, agriculture and food processing for 131 applications worth USD 1.82 billion, and machinery, automation and robotics for 82 applications worth USD 387.4 million. Applications under the “Smart and Sustainable Industry” measure, which targets machinery upgrades and automation, numbered 132, worth USD 507.6 million. All of these figures are based on applications and differ from approved amounts.

Let us also look at production figures. According to MarkLines, car production in August, announced by the Federation of Thai Industries (FTI) on September 28, 2026, was 124,646 units, up 10.9% year on year. The Manufacturing Production Index (MPI) for August 2026, reported by The Star (via Reuters) based on an announcement by the Ministry of Industry, rose 4.44% year on year. However, the ministry explains that it revised the MPI calculation method, adding new electronic products and changing the weights. Year-on-year comparisons need to be read with this revision in mind. The S&P Global Thailand Manufacturing PMI was 54.3 in September 2026, up from 53.8 in August (according to the aggregator site Trading Economics). The same site cites metals in particular as a factor behind rising input costs.

The statistics above are national figures. They cannot be used for an individual factory’s investment decision. However, if you keep records of formulation weighing values and of press-fit and caulking forces, they become material for explaining afterwards that products were made correctly. For implementation decisions, please use your own complaint records and defect records.

What Is a Load Cell — The Strain Gauge Principle

A load cell is a sensor that converts force or mass into an electrical signal. The type widely used in factories is the strain gauge type.

MinebeaMitsumi explains the mechanism as follows. Strain gauges are bonded to a part that deforms in proportion to force (the strain body), and the change in electrical resistance caused by the deformation is converted into an electrical signal. The company states that the deformation under load is “so small that you cannot see it with the naked eye”.

According to Flintec, a strain gauge is a thin wire or foil arranged in a zigzag pattern, sensitive to elongation and contraction in the longitudinal direction and insensitive in the width direction. In shear beams, the gauges are placed at 45° to the load axis to detect shear strain. The company explains that many load cells are potted with resin, and that welded hermetic sealing is used for more complete protection.

Let us look at the magnitude of the output using an example from HBM’s (now HBK) 2019 webinar material. For a steel strain body with a strain of 1000 µm/m (0.1%), a gauge factor of about 2 and a 350 Ω bridge, the change in resistance per gauge is 0.7 Ω. When the gauges are arranged in a full bridge, the output is 2 mV/V. With an excitation voltage of 5 V, the output voltage at rated load is 2 mV/V × 5 V = 10 mV. The same material shows that if this signal is read at a resolution of 10,000 digits, 1 digit is 1 µV, and at 100,000 digits it is 0.1 µV.

So a signal of a few mV is read in units of µV. For this reason, a device that amplifies the load cell output and turns it into a number is required. MinebeaMitsumi also explains that instrumentation for amplifying and displaying the signal must be connected. The same HBM material lists 50 Hz electric fields, inverters and time-varying magnetic fields as influencing factors. This is why the handling of cables and grounding affects the values. This point is covered in detail in a later section.

Note that each product has its own recommended excitation voltage. For example, each company’s materials state a recommended 12 V or less for the MinebeaMitsumi C2B1B-□-C3 and a recommended 5 V for the 100 to 500 N versions of the Unipulse SCB. Check that it matches the excitation voltage on the amplifier side.

Types of Load Cells (1) — Shapes and Applications

How shapes are classified differs by manufacturer. MinebeaMitsumi mainly divides them into four types: “diaphragm type”, “S-type”, “canister type” and “beam type”. Flintec explains them in four categories: beam, single point, compression and tension. The table below is organized by this article based on each company’s explanations.

Viewed from the application side, the correspondence is roughly as follows. For tanks and hoppers, compression types (canister or beam types) are combined with dedicated mounting hardware, or S-types are used to suspend the vessel. For platform scales and small-scale weighing, single point types are candidates. For press-fitting and pressing, candidates are the compression or tension/compression models that the manufacturer positions as being for force measurement. Tension types and S-types are used to measure tension.

Here are some size examples as well. A&D states that the LCC32 has a diameter of 20 mm, a height of 13 mm, a mass of about 23 g and a rated capacity of 200 N to 5 kN. The MinebeaMitsumi C1W1M is a center hole type with a built-in amplifier and 4 to 20 mA output, with a rated capacity of 58.84 kN to 490.3 kN. Even under the same name “load cell”, shape and size differ greatly by application.

Types of Load Cells (2) — Analog and Digital Load Cells

Load cells are also divided into two groups by the form of their output.

Analog load cells output the bridge output directly as a mV/V signal. Looking at each company’s materials, many models have a rated output in the 1 to 2 mV/V range. Because the signal is small, an indicator or a converter (amplifier) is required.

Digital load cells perform signal processing and A/D conversion on the load cell side and output values over digital communication. Each company’s explanation is as follows.

• Flintec: Digital types have built-in signal processing and A/D conversion and suit truck scales and dynamic weighing. Each load cell can be communicated with individually. Some models support CANbus
• MinebeaMitsumi (Minebea Intec) PR6204: The company states that up to 8 units can be connected over 2-wire RS485, that it supports IP68 and IP69K, and that accuracy classes C3 and C6 (25 t, 50 t) are available
• A&D LCCD11 (column type): The company states that the output is 2-wire RS-485 and that it has a built-in gas tube arrester for surge protection
• Unipulse DSCB (a high-rigidity digital load cell for force measurement): The company explains that it has a built-in amplifier, can perform hold and upper/lower limit comparison without an indicator, and can display values on a PC in real time

Going digital lets you read values and status individually for each load cell. For a tank supported by several cells, this can be an advantage when you want to isolate which one cell is behaving abnormally. Unipulse describes the DSCB as “resistant to noise”, and Flintec explains that digitalization can improve system accuracy. However, there is no material that allows us to state definitively that digital raises accuracy, or that it eliminates noise problems, in every application.

As one example of the cost difference, let us look at A&D’s catalog standard prices (excluding tax) as of May 2022. For a 100 kN column type, the analog LCC11-T010-K is JPY 80,000 and the digital LCCD11-T010-K is JPY 110,000. The difference is JPY 30,000, which works out to 37.5% higher than the analog version. Digital systems also separately require items such as the AD-4388-4 junction box (4 points) at JPY 32,000 and the dedicated cable AX-KO3217-10M at JPY 14,000. All of these are the manufacturer’s standard prices in Japan, not prices in Thailand.

In practice, the choice between analog and digital is made based on the receiving equipment (indicator, PLC, higher-level system) and on how much diagnostics you require.

The First Decision (1) — How the Load Is Applied and the Mounting Structure

MinebeaMitsumi explains that load cells “need to be selected according to the type of force to be measured (compression, tension and so on), the load range and the installation environment”. Flintec states that “it’s often said that a load cell is only as good as its mounting”, and that mounting, alignment, fasteners, mounting hardware and calibration procedures should follow the manufacturer’s recommendations. Placing how the load is applied first in this article is a way of thinking that follows the explanations of these two companies.

First, decide the direction of the load. If the load only pushes from above, the candidate is a compression type; if the object is suspended, a tension type; and if both pushing and pulling occur, a tension/compression type. Do not design a structure in which tension is applied to a compression-only model.

Next, check for eccentric loads and side loads. Flintec explains that with ordinary load cells, accuracy is lost when force is applied off center. HBM explains the following for each type of beam.

• A single bending beam changes its signal when the load point moves (lever arm effect), so it is rarely used on its own
• A multiple bending beam is significantly less sensitive to movement of the load point
• A shear beam has high stability against side forces, but secondary effects can appear in the measurement result

Some models state the allowable side load explicitly in the specification table. Kyowa states that the allowable side load for the LCX-A-ID is 15% of the rated load. For a model rated at 1 kN, that works out to 150 N. Unipulse explains that for the UIS, the effect of side loads on accuracy is small.

Single point types are designed to compensate for off-center loads. However, this does not mean there is no error wherever the load is placed. For the MinebeaMitsumi CB17, the company specifies the off-center error conditionally, as within 0.02 %R.O. when 1/2 of the rated load is applied at an offset of 50 mm on a platform of up to 200 × 200 mm. The A&D LC4101 to LC4204 are corner-adjusted, and a platform size (300 × 300 to 600 × 700 mm) is given for each model. Mounting a table larger than the platform size in the catalog takes you outside the manufacturer’s assumptions.

Also check the mounting foundation. MinebeaMitsumi explains that if the installation location sinks, correct values cannot be measured, and that an installation location with sufficient strength is needed. In the case of tanks, thermal expansion can move the brackets horizontally and apply an eccentric load to the load cells. The company introduces designs that release this movement through the mechanism of the mounting hardware.

What has been decided up to this point is the shape and the mounting method. Capacity is not decided yet. The overall picture of this order is summarized in the figure below.

The First Decision (2) — Weighing or Force Monitoring

The second fork in the road is what the load cell is used for.

Unipulse divides its product lineup, including indicators, into two groups. The company explains that “force measurement and control” is a product group that dynamically monitors load changes during press-fitting, caulking and pressing and makes pass/fail judgments on products. “Weighing measurement and control” is a product group for building automatic weighing instruments such as batch weighers, packer scales, truck scales and belt scales. Among the load cells themselves, there is also a group of models labeled “for force measurement”.

METTLER TOLEDO explains load application by dividing it into static and dynamic. Most tanks, hoppers and vessels involve static loading, and little or no horizontal shear force is transmitted to the load cells. Conveyors, pipe racks, conversions of mechanical scales, and high-powered mixers or blenders involve dynamic loading, and horizontal shear forces are transmitted, the company explains. HBK’s selection guide for single point types also places “static or dynamic weighing?” among its first questions.

The rows to look at in the specification table differ between weighing and force monitoring. Natural frequency is an easy example. Here are the values stated in each company’s materials.

Even under the same name “load cell”, the values differ by two or more orders of magnitude. If you want to see press-fit or caulking forces in detail along the stroke, check both this row and the sampling rate on the amplifier side. How many kHz is enough depends on the speed of the process and the judgment method, so this article does not give a general standard. Please decide together with the equipment maker and the load cell manufacturer while looking at actual waveforms.

The accuracy rows also look different depending on the application. Models for scales (C2B1B-□-C3, CB17, A&D LC41xx) have accuracy items in the 0.01 to 0.02% range. Models for force measurement (SCB, UIS, CMM1R) are roughly in the 0.1 to 0.2% range. This comparison is limited to the models listed above. This article does not read this as meaning that force measurement models perform worse; it interprets it as each company designing the balance between accuracy and rigidity, response and overload resistance to suit the application.

Here are related articles by application.

• If you are considering press-fit force monitoring as part of the whole equipment, Servo Press and Press-Fit Equipment Selection: Thailand RFP and FAT/SAT Guide covers force and displacement judgment and equipment-side requirements
• Caulking forces and process acceptance are covered in Riveting Machine Selection in Thailand: Radial, Orbital and FAT/SAT
• Dynamic weighing that judges the weight of each product one by one on a belt conveyor is covered in detail in Checkweigher Selection: Cutting Giveaway with the Right RFP and Acceptance Tests in Thai Factories
• To measure multi-axis forces and torques at a robot’s end effector, a 6-axis force/torque sensor is used instead of a single-axis load cell. Please refer to Robot Force/Torque Sensor Selection: Retrofit and Acceptance in Thailand
• Connecting tensile testing machine results to records is covered in Tensile Testing Machine Data Integration in Thailand: Specimen to MES

How to Decide Rated Capacity and Calculate Load Cell Capacity

Once the application, shape and environment are decided, it is finally time for capacity.

A&D explains load cell selection in three steps: “selecting the shape”, “selecting the rated capacity” and “examining measurement accuracy”. Here too, capacity comes after shape.

On how to think about capacity, METTLER TOLEDO explains that a load cell “must not break when maximum load is applied, and it must be able to make the smallest measurement requested of it”. The company’s FAQ shows the formulas for calculating capacity in weighing systems supported by multiple cells, divided into two according to how the load is applied.

The symbols mean the following.

• C: Capacity per load cell (weigh module)
• TDL: Total dead load. The weight that is always on the scale, such as the vessel, frame and agitator
• SC: Scale capacity. The maximum amount of contents to be weighed
• N: Number of load cells
• sf: Safety factor. The company states “typically 1.25”

Here is one calculation example. The numbers are hypothetical numbers for this article. Suppose a hopper with a dead load of 800 kg holds up to 1,600 kg of raw material and is supported by 3 load cells. Assume the load is applied roughly at the center and sf is 1.25.

• Sum of dead load and scale capacity: 800 + 1,600 = 2,400 kg
• Multiply by the safety factor: 1.25 × 2,400 = 3,000 kg
• Divide by the number of cells: 3,000 ÷ 3 = 1,000 kg

The calculation shows that a capacity of at least 1,000 kg per cell is needed. Note here that sf is a factor applied not to “the maximum load” but to “the sum of dead load and scale capacity divided by the number of cells”. This is a different idea from deciding with a single multiple, such as choosing a capacity that is some number of times the maximum load.

Too large a capacity also causes problems. METTLER TOLEDO explains that using load cells larger than they need to be for a given application is not recommended because accuracy suffers. It is important to stay at or above the calculated value without making the capacity too large.

In some countries, the metrology authority presents a different approach. For example, a technical note from Sri Lanka’s metrology authority (referencing OIML R76-1, issued September 26, 2024) shows an example of finding the required capacity by adding up factors such as non-uniform distribution and the zero-setting range, and of multiplying by a factor that assumes a failure in which only 2 legs of a 4-point scale touch the ground. This is a document of the Sri Lankan authority, not a Thai regulation. It is enough to keep in mind that how factors are decided differs by manufacturer, application and national regulation.

MinebeaMitsumi’s explanation only says to select with a margin over the rated capacity and does not give a multiple.

Look at the overload rows together with capacity. A specification table normally lists two values.

• Safe overload: The upper limit, indicated by the manufacturer, of the load that may be applied during use
• Ultimate overload: The load above which, according to the manufacturer, there is a risk of destruction

Details of the definitions and the notation (such as “load resistance”) differ by manufacturer, so check each company’s materials.

The differences between models are large. Each company’s materials show the following.

(%R.C. is % of rated capacity)

In press-fitting and caulking, unexpected loads can be applied momentarily because of how a jig makes contact or because of setup mistakes. For such processes, this article recommends checking the safe overload value in the specification table and designing it together with stoppers and limits on the equipment side. There is no general standard for what percentage is enough.

How to Read Accuracy — Rated Output, Non-linearity, Hysteresis, Repeatability, Creep and Temperature Effects

The specification table of a load cell has several rows related to accuracy. The main items mean the following.

• Rated output: The output (mV/V) when the rated capacity is applied. Multiplying it by the excitation voltage gives the output voltage
• Non-linearity: How far the relationship between load and output deviates from a straight line
• Hysteresis: The difference in output between increasing and decreasing the load
• Repeatability: The variation when the same load is applied repeatedly
• Creep: The change in output when a constant load is applied continuously (stated together with the number of minutes)
• Temperature effect: The change in zero point and output when the temperature changes

Here are the values by model from each company’s materials.

(%R.O. is % of rated output)

There are two cautions when reading this table.

The first is that the definitions of the indicators differ by manufacturer. A&D gives a “combined error” that groups several errors together, while MinebeaMitsumi, Kyowa and Unipulse give non-linearity, hysteresis and so on as separate items. You cannot place A&D’s combined error next to another company’s non-linearity and judge which is more accurate. If you want to compare, ask each company to provide values under the same definition.

The second is the unit of the temperature effect. MinebeaMitsumi and A&D write it “/10 °C”, while Kyowa and the Unipulse SCB write it “/°C”. For example, the zero point temperature effect of the Unipulse SCB is 0.015% R.O./°C or less, which is 0.15% when converted to a 10 °C basis. For the MinebeaMitsumi C2B1B-□-C3, the zero point temperature effect is 0.014 %R.O./10 °C and the output temperature effect is 0.012 %LOAD/10 °C. Align the units before comparing.

Also check the compensated temperature range. Each company’s materials state −10 to 40 °C for the C2B1B-□-C3, −10 to 40 °C for the A&D LC4101 series as well, −10 to 70 °C for the CMM1R and the Kyowa LCX-A-ID, and −10 to +60 °C for the Unipulse SCB. When installing in areas without air conditioning or near heat sources, measure the actual temperature at the mounting position and check whether it falls within this range.

OIML R60 Accuracy Classes (C3 and Others) and the Accuracy of the Whole Scale

Load cell model names and specifications sometimes carry “C3” or “C6”. These are accuracy classes based on R60, the load cell recommendation of the International Organization of Legal Metrology (OIML).

OIML R 60-1:2021 divides accuracy into four classes: Class A, B, C and D. For each class, the range of the number of load cell verification intervals (nLC) is defined.

“C3” means Class C with 3,000 intervals. The number of intervals is written in units of 1,000, so 3 means 3,000 and 1.5 means 1,500. Symbols for the load direction (compression or tension), a special temperature range or humidity can also be added.

The error of C3 cannot be replaced with a single figure of “± some %”. The maximum permissible error for type evaluation in R60 is determined by the verification interval v and the apportionment factor pLC. For Class C, there are three steps depending on the size of the load: pLC × 0.5 v, pLC × 1.0 v and pLC × 1.5 v. pLC takes a value from 0.3 to 0.8, and is taken as 0.7 if not stated on the certificate.

Here are product examples. MinebeaMitsumi states that the C2B1B-□-C3 is “accuracy class C3 based on OIML R60” with a maximum number of load cell intervals of 3,000. The A&D LCC11 and LCCD11 are available in C4 and C6 versions. HBK explains that the Z6 is available in OIML D1, C3, C4, C5 and C6 classes.

R60 contains an important proviso. It is a sentence stating that it is not the intention of the Recommendation to require that a load cell be of the same accuracy class as the measuring system in which it may be applied. You cannot say that because the load cell is C3, the whole scale is also equivalent to C3. The accuracy of the whole scale is determined not only by the load cell but also by the indicator, mounting, piping, temperature and so on.

On the relationship between application and class, HBM gives the following examples. For simple construction material scales for sand, cement and water, D1 to C2 is sufficient; for mixing additives and for the machinery industry (such as inspection of ball bearings), C3; for shop scales and filling machines, C3 to C6; and for checkweighers, the highest class, C6.

Let us also sort out whether OIML classes are needed for force monitoring used inside the factory. R60 is a standard for type evaluation in legal metrology. Flintec explains that legal-for-trade scales require products certified under OIML, NTEP and so on, and that C3 is typical for commercial scales. On the other hand, the company states that load cells without certification are called general purpose (GP) and are used in applications not subject to legal-for-trade regulations. For applications not used in trade, such as monitoring press-fit force, this article takes the view that the response and overload rows seen in the previous section, rather than whether a class is certified, are the core of the judgment.

Environmental Conditions — IP, Explosion Protection, Materials, Temperature and Humidity

MinebeaMitsumi explains that in harsh environments, such as those with large changes in air temperature, specifications such as temperature, dust protection and waterproofing should also be checked.

The degree of protection (IP) differs by model even within the same series. Each company states the following.

Even for the same IP68, the test conditions given by the manufacturers differ: water depth 1.5 m/100 hours for A&D, and 1 m water column/100 hours for HBK. For food factories with washdown or for tanks outdoors, this article recommends having suppliers state “the IP rating and its test conditions” in their RFP responses.

Where explosion protection is required, consider the explosion-protected model and its peripheral equipment as a set. MinebeaMitsumi states that the CCP1-□-Z has a flameproof enclosure and that the CC010 has passed the technical standards for intrinsically safe construction (class iaIICT5). There is also the PR1626 intrinsically safe interface for load cells, and the company explains that the /60 can connect up to 4 load cells and the /61 up to 8. HBK states that the Z6 is available in IECEx, ATEX and FM (USA) explosion-protected versions. Flintec states that some of its load cells have ATEX and FM certification, and gives paint booths and environments with grain flour or wood dust as examples. This article has not confirmed what is required of explosion-protected equipment in Thailand. Please check with the competent authority, the factory’s insurer and the facility designer.

Materials are also chosen according to the environment. Kyowa explains that the LCX-A-ID has a stainless steel body and is suited to equipment that requires corrosion resistance. MinebeaMitsumi states that the CMM1R is suited to use in corrosive locations. The A&D LC41xx are made of aluminum. Flintec explains that aluminum single point types are for low capacities.

Pay attention to moisture too. HBM explains that residual soldering flux changes its resistance with moisture and strongly affects the zero point. Kyowa adds “no condensation” to its operating temperature range. For temperature, compare the compensated temperature range seen in the previous section with actual measurements at the mounting position.

Cables (4-Wire/6-Wire) and Noise

Load cell cables come in 4-wire and 6-wire types.

A 4-wire cable has two wires that apply voltage to the bridge and two that take out the output. HBM’s 2019 webinar material shows an example in which, with 4-wire, the sensitivity changes because of the cable resistance. For a 350 Ω bridge and copper wire with a cross-section of 0.14 mm², the round-trip resistance for a cable length of 5 m is 1.272 Ω, and the sensitivity changes by 0.36%. Under the same conditions with a cable length of 50 m, the change in sensitivity is 3.6%. The same material states that with 4-wire the sensitivity depends on cable length and temperature, and calls for calibration with the cable used in the measurement attached.

A 6-wire cable adds two wires (sense lines) for measuring the voltage actually applied to the bridge. The amplifier side measures the bridge voltage and corrects it as needed. HBM explains that if you extend the cable, the best thing is to use 6-wire technology. METTLER TOLEDO also explains that 6-wire improves performance in applications where the indicator is far from the scale.

Whether cables may be cut or extended is described differently by different manufacturers.

• HBK Z6: The company explains that because of the 6-wire configuration, the cable can be cut to the required length. Cable length is 3 to 20 m
• METTLER TOLEDO: The company explains that load cell cables should not be lengthened or shortened in the field, and that changing the length affects the output signal. Excess length is coiled near the junction box

For this reason, whether cutting or extending is allowed depends on the model (4-wire or 6-wire) and the manufacturer’s instructions. Do not decide on general grounds that “it may be cut” or “it must not be cut”; check the instruction manual.

The cables of each company’s models (4-core or 6-core, length, detachable or directly attached) are summarized in the comparison table in a later section.

Based on HBM’s explanation, a 4-wire load cell can be considered to have its output determined with the cable length as delivered included. If an extension becomes necessary, consider changing to 6-wire or recalibrating after the extension.

For noise countermeasures, shielding and grounding are the issue. HBM describes a method of connecting the shield to the sensor housing. Because a current can flow in the shield and 50 Hz interference can occur when the sensor and the amplifier are at different potentials, the company recommends connecting the sensor housing and the amplifier housing with low resistance (a 16 mm² flexible copper wire as a recommended example).

METTLER TOLEDO explains that excess cable length should not be fixed to the live (moving) part of the scale.

Load Cell Amplifiers, Indicators and PLC Connection

Analog load cells require a device that amplifies the signal and turns it into a number. MinebeaMitsumi explains that in addition to display, some products have analog and digital outputs and functions for controlling and monitoring the weighing system. How to choose a load cell amplifier divides largely according to “what it is used for”.

As examples of indicators for force monitoring (press-fitting, caulking), Unipulse lists the following products.

• FS2000: The company describes high-speed processing at 25,000 times per second, a waveform comparison function and multi-point comparison judgment with up to 5 points
• F381A: 4,000 times per second, and two-dimensional comparison using two inputs, force and displacement
• Built-in FC400 series: High-speed A/D conversion at 2,400 times per second. Models include EtherNet/IP (FC400-EIP-FA), PROFINET IO (FC400-PRT-FA), CC-Link (FC400-CCL-FA) and RS-232C/Modbus (FC400-232-FA)

As examples of indicators for weighing, the company lists the FC500 series (high-speed A/D conversion at 1,200 times per second, fixed-quantity cut-off control functions for discharge weighing and charge weighing, CC-Link, EtherNet/IP, RS-485 (Modbus), RS-232C and so on) and the FC1000, which is equivalent to IP65.

If you want to read the value in the PLC as an analog value, one method is to use a load cell converter. Unipulse explains that the LC1111 can connect up to 4 load cells of the 350 Ω class in parallel and has 4 to 20 mA current output as standard. Other load cell converters include the Watanabe Electric Industry WVP-LCS (check the output type for each model).

There are also input units that connect load cells directly to the PLC. Mitsubishi Electric explains that the Q61LD for the MELSEC-Q series does not need an external signal converter and can be connected with 6-wire (combining the remote sense method and the ratiometric method) or 4-wire. The excitation voltage is DC 5 V, the output current is within 60 mA, and 4 load cells of the 350 Ω class can be connected in parallel. The non-linearity is stated as within ±0.01%/FS under the conditions of a rated output of 2 mV/V, 25 °C and no tare cancellation. It also has a function for calibrating by applying test weights and detection of input signal errors. This article has not confirmed equivalent products in successor series.

Regarding IO-Link, HBK explains that it offers the Z6 with IO-Link as part of its “Load Cell Measurement Chain (LCMC)”. The LCMC has a CAN/RS485 output version and 4 to 20 mA, 0 to 10 V and IO-Link output versions. This article has not confirmed specification values such as resolution and sampling rate. The idea of handling sensor values and diagnostic information together via IO-Link is explained in IO-Link Implementation: Sensor Data Integration and Acceptance.

The relationship between connection method and use is summarized in the table below as this article’s own organization.

In this article’s organization, for weighing such as batching and filling, a weighing indicator with a weighing sequence is a candidate. For monitoring press-fitting and caulking, an indicator with high-speed sampling and waveform judgment is a candidate. If you want to use values in the PLC, consider a built-in type with fieldbus or a PLC-direct unit. Which one is suitable depends on the existing PLC, the space in the panel and where the judgment is made.

If you place the amplifier outside the panel or want to consolidate wiring on the machine, Remote I/O Selection – In-Cabinet vs On-Machine, RFP and FAT/SAT for Thai Plants is also a useful reference. The design of passing scale values to the PLC or MES itself is covered in Industrial Scale PLC Integration: RFP and FAT/SAT for Thai Factories.

If you record waveforms in force monitoring, besides the pass/fail judgment you can later see trends such as the force peak gradually rising. This can be material for noticing jig wear or changes in part lots. The idea of linking trend monitoring to equipment maintenance is organized in Predictive Maintenance System 2026: Cost Layers and How to Start.

Tank and Hopper Weighing Installation — Mounting Hardware, Piping, Agitation and Vibration

In tank and hopper weighing, the handling of mounting hardware, piping and the center of gravity affects error more than the load cell itself does. Each company’s technical materials also devote considerable space to installation.

MinebeaMitsumi introduces mounting hardware with an anti-sway mechanism in four methods.

• Flexible stop method: Vibration countermeasure for tanks with agitators and similar. Also compatible with the flameproof CCP1-Z
• Steel ball method: For belt conveyors and hopper legs. Space-saving and strong against side loads
• Check rod method: For high capacities, with a high degree of freedom in arrangement. Separate load cell mounting hardware is required
• Rocker pin method: For beam types, with a low mounting height

The company’s hardware has, as design functions, prevention of lateral displacement (strength able to withstand a seismic coefficient of 0.3, that is, a moderate horizontal acceleration of 0.3G occurring at around seismic intensity 6 on the Japan Meteorological Agency (JMA) scale), prevention of overturning, prevention of eccentric loads and prevention of vibration. The company recommends installing the hardware at the height of the tank’s center of gravity, and says that if it is mounted higher or lower than the center of gravity, stay rods or check rods need to be considered. Where there is an agitator, the company explains that the horizontal load due to agitation needs to be calculated and compared with the allowable horizontal load of the hardware. It also requires installation that takes into account the direction of the tank’s thermal expansion.

HBK’s (then HBM) 2013 technical material explains the handling of piping in detail.

• Connections such as piping create a “force shunt” that diverts part of the weighed value, resulting in weighing errors. Connections should be flexible in the vertical direction
• When connecting with rigid piping, connect with the longest possible horizontal pipe section, and also provide a mechanism to absorb longitudinal expansion and contraction. Flexible couplings, hoses, elbows, metal expansion joints (bellows) and open inlets (not for pressure vessels) are given as examples
• Contents inside the piping are also included in the weight, so when weighing, keep the piping either always empty or always full
• Powder suction systems can generate negative pressure of 100 to 300 mbar. With vertical piping, a force corresponding to the cross-sectional area of the pipe directly affects the weighed value. Horizontal piping is preferable to vertical piping in every case
• A configuration in which an upright tank is supported by a single rigidly mounted load cell is not recommended
• In square hoppers at filling stations, vibration from the conveying system and impacts of coarse bulk material cause considerable shear loads, so restrain them firmly with stay rods

On the number of support points, the company’s material explains that 3-point support is statically determinate and best for tanks, with a load cell at each support point to share the load equally. With 4 or more points the system is statically indeterminate, and the load at each point must be measured and adjusted with shims or similar during installation. When only some of the supports are load cells, the company states that accuracy can be improved by an arrangement in which the load on the load cell supports is larger.

METTLER TOLEDO calls for installation to be planned by a qualified structural engineer, with reference to the installation and service manual for each model. The company explains that its handbook includes calculations of the effects of wind, earthquakes, shock, heat and load cell cables, as well as the design of piping connections. Flintec also states that alignment during installation is important for compression types, and that they are typically combined with dedicated mounting hardware.

The phenomenon “the value changed after the piping was connected” can be related to the force shunt and the handling of the contents of the piping seen in this section. This is also why acceptance tests compare values with and without the piping and agitator connected.

Note that if you only want to know how much is left in a tank, besides weight measurement with load cells there is also the method of measuring the liquid surface with a level meter. Which is suitable depends on the properties of the contents, the value required (weight or volume) and the structure of the tank. For an organization from the level meter side, please refer to Tank Level Remote Monitoring: Sensor Selection, Reorder Points and RFP in Thailand.

Representative Products from Major Manufacturers — Specifications Confirmable in Published Materials

Here we list the models for which this article could confirm published materials. This is not a ranking or a recommendation. All values are as stated in each company’s materials, and cells that could not be confirmed are marked “Could not be confirmed in published materials”.

In addition, MinebeaMitsumi’s U3S1 (S-type), CB17 (low capacity, for scales), C1W1M (center hole type) and PR6204 (digital), A&D’s LCCD11 (digital), Unipulse’s DSCB (digital), and HBK’s Z6R (20 to 200 kg, for hygienic and harsh environments, compatible with the Z6) and HLC (110 kg to 10 t) are also introduced in each company’s materials. For Flintec and METTLER TOLEDO, this article uses their technical explanation pages as sources and has not confirmed specification values for individual models.

On availability in Thailand, the facts this article could confirm are as follows. Unipulse has a Thai-language site that lists the load cell, force measurement and weighing measurement categories in Thai. METTLER TOLEDO has a Thai-language site for Thailand that lists industrial scales, weigh modules, load cells, and calibration and verification services. Minebea Intec was reported in a December 2019 trade magazine article to have opened an Experience Centre inside MinebeaMitsumi’s production site in Bangkok (product exhibition, training, product testing and technical support). This article has not confirmed the current status of this facility. We have also not confirmed the sales setup in Thailand of other manufacturers, so please contact each company or its distributors.

On prices, published prices could be confirmed only for some models in Japan. On MonotaRO, a Japanese online store, the prices as of browsing on October 13, 2026 were as follows. The A&D single point LCB03 series was JPY 11,980 excluding tax (JPY 13,178 including tax), the MinebeaMitsumi S-type U3S1 was JPY 52,980 to 62,980 excluding tax (varying by capacity), the compact compression type CMM1R was JPY 87,980 excluding tax (JPY 96,778 including tax), and the Watanabe Electric Industry load cell converter WVP-LCS-11A-1 and others were JPY 71,980 excluding tax (JPY 79,178 including tax). A&D’s manufacturer standard prices (excluding tax) are JPY 20,000 for the LC4101 series, JPY 48,000 for the LC4204-K600 (catalog PDF updated June 30, 2026), JPY 40,000 for the button type LCC32 (as of November 2025), and JPY 60,000 to 108,000 for the column type LCC11 (as of May 2022). All of these are prices in Japan, and we have not confirmed published prices at the part number level in Thailand.

Calibration, Traceability and Legal Metrology

Load cell values can be used as process records only after calibration.

The basis is calibration by applying test weights or actual loads. Mitsubishi Electric explains that with the Q61LD, the gross weight value can be calibrated by applying an actual load (test weights) to the load cell. METTLER TOLEDO explains the idea of calibration as placing a reference weight to tell the indicator (terminal) that the change in signal is proportional to the change in loading.

For 4-wire, HBM explains that calibration needs to be done in the mounting position. As seen in the earlier section, this is because the sensitivity changes with cable length and temperature.

For large tanks, it can be difficult to place test weights. Each company also offers methods that do not use test weights. METTLER TOLEDO offers a tank scale calibration service called “RapidCal” for tanks, reactors, hoppers and silos, and explains that it is traceable. A&D explains that with the AD-4389 simulator for digital load cells, simulated mass values can be supplied without using actual test weights. However, this article has not confirmed in which applications such methods are treated as equivalent to calibration with test weights. If you adopt one, decide with your quality assurance department which method guarantees what. It is safer not to assume that load cells need no calibration.

Traceability means making calibration results traceable back to national standards. In Japan, there is JCSS (Japan Calibration Service System), the measurement traceability system based on the Measurement Act. According to NITE’s explanation, registration of calibration service providers is voluntary, and the registration criteria include conformity with the requirements of ISO/IEC 17025. The registration categories include “mass” and “force”. NITE explains that JCSS signed the mutual recognition arrangement of APLAC in December 1999 and of ILAC in November 2000.

In Thailand, NIMT (National Institute of Metrology (Thailand)) holds the national standards. NIMT is a public organization established on June 1, 1998 under the National Metrology System Development Act B.E. 2540, and its calibration service categories include “Mass” and “Force and Torque”. The institute explains that it is accredited to ISO/IEC 17025. NIMT’s Force and Torque Laboratory explains that, as the primary standard for force, it maintains a standard weight generator of 0.1 to 500 kN (resolution 35 to 80 ppm) that is traceable to the SI.

Legal metrology also needs to be checked. According to Thailand’s economy report for APLMF prepared in November 2016 (approved by CBWM), the CBWM (Central Bureau of Weights and Measures), under the Department of Internal Trade of the Ministry of Commerce, is in charge of legal metrology, with 4 regional verification centers and 27 verification branch offices. The report states that measuring instruments intended for trade use and covered by the Weights and Measures Law must pass verification and be sealed with a verification mark before being put on the market or used. Non-automatic weighing instruments, automatic weighing instruments (belt conveyor scales, automatic bulk weighing systems), weights and so on are listed as covered. The law is the Weights and Measures Act B.E. 2542 of 1999, amended in 2006 and 2014. The report explains that voluntary calibration outside legal metrology is provided by laboratories accredited under TLAS (Thai Laboratories Accreditation Scheme).

The procedures required differ between weighing and force monitoring used for process control inside the factory and scales used in trade, such as when the price is determined by shipping weight. Scales for trade use can be subject to CBWM verification. However, the content above is from a report as of 2016. Please check with the authorities or distributors which of your equipment counts as trade use and how the latest rules are applied.

Because calibration intervals vary with the application, frequency of use and quality requirements, this article does not give a uniform interval. Decide it in your own quality rules and reflect it in the RFP and the acceptance records.

Requirements to Write in the RFP and FAT/SAT Acceptance Criteria

In the request for proposal (RFP), write the items covered so far in concrete terms. This lets you compare the responses of each company under the same conditions.

The record of setting values is an item that is easy to overlook. When indicators and amplifiers are replaced, they will not give the same values unless the settings are re-entered. Also decide in the RFP how the settings are saved and restored, and who prepares the procedure document for this.

Acceptance testing is divided into two stages.

• FAT (factory acceptance test): Using test weights or a reference instrument, compare the displayed value, the output value and the value read by the PLC at several loads. Check linearity, eccentric load (the difference when the load position is changed), zero point and repeatability. For force monitoring, check under conditions simulating good and defective products that the pass/fail judgment works as set
• SAT (site acceptance test): On the actual equipment, record the change in values between the states with the piping and agitator connected and disconnected. Run for a set period and compare against existing weighing records and inspection records. For force monitoring, collect the force range of good products and check that there are no false rejects or missed detections

Decide the numerical acceptance criteria (the allowable difference from the reference instrument, the eccentric load difference, the SAT period and so on) from your own process and quality requirements. In the RFP, put your own target values “as examples” and finalize them in discussion with the manufacturer and the equipment maker. There is no standard value fixed by the industry.

How to select sensors and put together an RFP is also organized from the pressure side in Pressure Sensor Selection — Use Case, Media, Surge and RFP for Thai Factories, part of the same series.

Model Cost and Benefit Estimate (Hypothetical Numbers)

All numbers from here on are hypothetical numbers for this article. Actual amounts vary greatly with factory conditions and quotations.

The assumed factory is a Japanese-owned automotive parts factory in Thailand. It has 2 press-fit stations for bearings and bushings. Currently, only post-press-fit dimensions are measured by sampling, and press-fit force is not recorded.

The current losses (baseline, per year) are set as follows. Escapes to the customer and detection in downstream processes are counted as separate events occurring on different parts. They are not double-counted.

Configuration A installs load cells and amplifiers for press-fit force monitoring at the 2 stations and passes force pass/fail judgments and waveform records to the PLC and the recording system. The breakdown of the initial cost of 600,000 THB is 2 sets of load cell and amplifier × 90,000 THB = 180,000 THB, 270,000 THB for mounting jigs, PLC programming and the recording system, and 150,000 THB for design and FAT/SAT. The annual operating cost (calibration, spare parts, settings management) is set at 40,000 THB.

(Unit: THB)

The breakdown of the calculation is as follows.

• Initial cost: 180,000 + 270,000 + 150,000 = 600,000
• Base case: 2 × 120,000 = 240,000. 600 × 150 = 90,000. 240,000 + 90,000 = 330,000. 330,000 − 40,000 = 290,000. 600,000 ÷ 290,000 ≈ 2.1 years
• Sensitivity: 1 × 120,000 = 120,000. 300 × 150 = 45,000. 120,000 + 45,000 = 165,000. 165,000 − 40,000 = 125,000. 600,000 ÷ 125,000 = 4.8 years

Here are three points on how to read the estimate.

The first is that the annual benefit is the upper limit “if all are prevented”. The actual benefit depends on how the pass/fail threshold for press-fit force is set. The threshold is decided after collecting the force range of good products in the SAT. In addition, defects caused by materials or dimensions that do not show up in the force cannot be found by force monitoring. The more such defects there are, the smaller the benefit becomes.

The second is not to count the benefit until false rejects and missed detections have been checked in the SAT. If there are many false rejects, where good products are judged defective, the shop floor will stop trusting the judgment and no benefit will be realized.

The third is that the payback period depends strongly on the frequency of events. If events are half the assumption, the payback extends from about 2.1 years to 4.8 years. Before implementation, we recommend counting the frequency of press-fit-related events from the past year’s complaint records and downstream defect records.

Frequently Asked Questions

How do you choose among the types of load cells?

First, choose by the direction of the load and the form of mounting. If the load pushes from above, compression types (canister, column, beam types and so on) are candidates; if the object is suspended, tension types and S-types; and if a single cell is to form a platform scale, single point types. Next, choose from the model groups that the manufacturer positions differently depending on whether they are for weighing or for force monitoring. Because shape classifications differ by manufacturer, it is more practical to compare by load direction, resistance to eccentric loads and capacity range than by name.

How do you calculate load cell capacity?

Calculate from dead load, scale capacity, number of cells and load imbalance, not with a single multiple of the maximum load. METTLER TOLEDO shows the formula C ≥ sf × (TDL + SC) ÷ N for tanks where the load is applied roughly at the center, and C ≥ sf × (TDL ÷ 4 + SC ÷ 2) for 4-point support where the load tends to be off-center, and explains that the safety factor sf is typically 1.25. The company also explains that a capacity larger than necessary reduces accuracy. Because how factors are decided differs by manufacturer and application, have the formula adopted written in the capacity calculation sheet in the RFP.

Which load cell amplifier should you choose?

It depends on the use. For weighing such as batching and filling, a weighing indicator with a weighing sequence is a candidate. For monitoring press-fitting and caulking, an indicator with high-speed sampling and waveform judgment is a candidate. If you only want to record values in the PLC, there are converters with analog output such as 4 to 20 mA, built-in indicators with fieldbus, and load cell input units that connect directly to the PLC. Please also check the excitation voltage and the conditions for how many cells can be connected in parallel.

Should you choose a digital load cell?

A digital load cell performs A/D conversion on the load cell side and outputs values over RS-485, CAN and so on. Because multiple load cells can be read individually, it becomes easier to isolate which one cell is behaving abnormally. On the other hand, receiving equipment, junction boxes and dedicated cables are needed, and in A&D’s example the standard price is higher than for the analog product. Digital cannot be said to be more accurate in every application, so please decide based on your existing indicators and PLC and on your need for diagnostics.

How do you calibrate a load cell?

The basis is calibration that applies test weights or actual loads and adjusts the displayed and output values. HBM explains that 4-wire load cells need to be calibrated in the mounting position, with the cable actually used attached. For large tanks, manufacturers sometimes offer methods that do not use test weights, but please decide with your quality assurance department what can be guaranteed with such methods. In Thailand, scales used in trade can be subject to CBWM verification (according to a 2016 document). Please check the latest rules with the authorities or distributors.

Can a load cell cable be extended?

It depends on the model and the manufacturer’s instructions. With 4-wire, the sensitivity changes with the cable length. In the example in HBM’s material, the change in sensitivity was 0.36% at 5 m and 3.6% at 50 m. HBM explains that if you extend the cable, it is best to use 6-wire. While HBK explains that the Z6 can be cut because it is 6-wire, METTLER TOLEDO asks that the cable length not be changed in the field. Check the instruction manual, and recalibrate if you change the length.

What is the price of a load cell?

Published prices could be confirmed only for some models in Japan. On MonotaRO, as of browsing on October 13, 2026, the A&D single point LCB03 series was JPY 11,980 excluding tax, the MinebeaMitsumi CMM1R was JPY 87,980 excluding tax, and the Watanabe Electric Industry load cell converter WVP-LCS-11A-1 and others were JPY 71,980 excluding tax. Among A&D’s manufacturer standard prices (excluding tax), the column type LCC11 is JPY 60,000 to 108,000 (as of May 2022). Other models and dates are summarized in the section “Representative Products from Major Manufacturers” in the main text. All of these are prices in Japan, and we have not confirmed published prices at the part number level in Thailand.

Summary

• Load cell selection starts, before capacity, with deciding (1) how the load is applied and the mounting, (2) weighing or force monitoring, and (3) the environment (this article’s view)
• Shape classifications differ by manufacturer. Compare by load direction, resistance to eccentric and side loads, and capacity range
• The off-center compensation of single point types is conditional on platform size and offset. Outside those conditions, the assumptions break down
• Specifications such as natural frequency differ by orders of magnitude between weighing and force monitoring models. For force monitoring, check them together with the amplifier’s sampling rate
• Calculate capacity from dead load, scale capacity, number of cells and load imbalance. Do not fix it with a single multiple, and avoid capacities that are too large
• The definitions of accuracy indicators differ by manufacturer. OIML C3 means Class C with 3,000 intervals and cannot be replaced with a single figure of “± some %”
• Check IP ratings down to the test conditions. Whether cables may be cut or extended depends on 4-wire or 6-wire and the manufacturer’s instructions
• Errors in tanks and hoppers are strongly affected by mounting hardware, piping, center of gravity and agitation (according to each company’s technical materials). Record the change in values with piping connected at the SAT
• In Thailand, scales used in trade can be subject to CBWM verification. Check the latest rules with the authorities or distributors
• Treat the benefit estimate as an upper limit, and count it only after false rejects and missed detections have been checked in the SAT

Which process’s weights or forces should you start recording, and where are today’s weighing drift and press-fit defects occurring? You are welcome to consult TOMAS TECH even at that stocktaking stage. We work with Japanese-owned factories in Thailand from organizing requirements through the RFP and FAT/SAT. Please contact us here.

References

• What Is a Load Cell? Mechanism, Types, Applications and Selection Points (MinebeaMitsumi, Japanese)
• Types and Features of Mounting Hardware with Anti-Sway Mechanism (MinebeaMitsumi, Japanese)
• Installation Conditions for Mounting Hardware with Anti-Sway Mechanism (MinebeaMitsumi, Japanese)
• Compression Types for Industrial Scales and Industrial Equipment, Model List (MinebeaMitsumi, Japanese)
• Square Beam Type Load Cell C2B1B-C3 Series, OIML Specification (MinebeaMitsumi, Japanese)
• Compact Compression Type Load Cell CMM1R Series (MinebeaMitsumi, Japanese)
• Roberval Type CB17 for Low Capacity Scales (MinebeaMitsumi, PDF, Japanese)
• Center Hole Type C1W1M (MinebeaMitsumi, PDF, Japanese)
• Digital Load Cell for Process PR6204 (MinebeaMitsumi, Japanese)
• Single Point Load Cells LC4101/LC4102/LC4103/LC4204 Catalog (A&D, PDF, Japanese)
• Compact Compression Button Type Load Cell LCC32 Catalog (A&D, PDF, Japanese)
• Stainless Steel Column Type Load Cell LCC11 Catalog (A&D, PDF, Japanese)
• Stainless Steel Column Type Digital Load Cell LCCD11 Catalog (A&D, PDF, Japanese)
• Notice of Load Cell Selection Webinar (A&D, PDF, Japanese)
• LCX-A-ID Compact Compression Type Load Cell (Kyowa Electronic Instruments)
• Load Cells, Product List (Unipulse, Japanese)
• Super Cell SCB High-Rigidity, High-Response Load Cell (Unipulse, Japanese)
• UIS Tension/Compression Type, Male Threads on Both Ends, Waterproof Load Cell (Unipulse, Japanese)
• Force Measurement and Control, Product List (Unipulse, Japanese)
• Weighing Measurement and Control, Product List (Unipulse, Japanese)
• Amplifiers (Unipulse, Japanese)
• Single Point Load Cells Buyer’s Guide (HBM)
• OIML accuracy classes explained (HBM)
• Structural Design of Tank Weighing Systems (HBM, PDF, 2013)
• Z6 Bending Beam Load Cell (HBM)
• How Does a Bending Beam Load Cell Work? (HBM)
• Webinar: Connection of load cells (HBM, PDF, 2019)
• What is a load cell and how does it work? (Flintec)
• What is a Beam Load Cell? (Flintec)
• What is a Single Point Load Cell? (Flintec)
• What is a Compression Load Cell? (Flintec)
• What is a Tension Load Cell? (Flintec)
• Digital and CANbus Load Cells (Flintec)
• Engineering Handbook, Custom Weigh Module Systems (METTLER TOLEDO)
• OIML R 60-1 Edition 2021 Metrological regulation for load cells Part 1 (OIML, PDF)
• Technical Note: Determining the capacity of the load cells required for Non-Automatic Weighing Instrument (Measurement Units, Standards and Services Department, Sri Lanka, PDF)
• Load Cell Input Unit Q61LD (Mitsubishi Electric, Japanese)
• Overview of JCSS (NITE International Accreditation Japan, Japanese)
• About us, National Institute of Metrology (Thailand) (NIMT)
• Force and Torque Laboratory (NIMT)
• Economy Report, Thailand, 23rd APLMF Meeting (APLMF, approved by CBWM, PDF, 2016)
• Minebea Intec Experience Centre to better serve customers in Southeast Asia (BakeryandSnacks.com, 2019-12-03)
• METTLER TOLEDO Site for Thailand (METTLER TOLEDO, Thai)
• Unipulse Corporation Thai-Language Site (Unipulse, Thai)
• PR No.117/2569 Thailand Secures $43.6bn 1H 2026 Investment Surge (Thailand BOI)
• August vehicle production in Thailand up 10.9%; sales up 25.6% (MarkLines)
• Thailand Car Production (Trading Economics)
• Thai August factory output rises 4.44% y/y, beats forecast (The Star, via Reuters)
• Thailand Manufacturing PMI (Trading Economics)
• Load Cell Compact Compression Type CMM1R (MonotaRO, Japanese)
• Load Cell S-Beam Type U3S1 (MonotaRO, Japanese)
• Single Point Load Cell A&D LCB03 Series (MonotaRO, Japanese)
• Load Cell Converter, Watanabe Electric Industry (MonotaRO, Japanese)