“Every molding machine has an operator stationed at it, who opens the safety gate, takes the molded part out by hand, cuts the gate and packs it into a box. We want to introduce take-out robots, but we do not know which type to choose, what size, or what to write in the specification.” This is a question we often hear from production engineering and molding staff at Japanese-owned plastic molding plants in Rayong, Chonburi, Ayutthaya and Samut Prakan, Thailand. When it comes to injection molding robot selection, people tend to compare clamping force and payload first. However, if you choose on these two figures alone, you will not reduce headcount or cycle time as much as you expected.
What determines the effect are the following two things.
- The time the operator’s hand or the robot’s gripper is inside the mold while the mold is open (the take-out time)
- Who does the gate cutting, inspection and packing after removal, and how fast
In this article, we therefore propose taking the selection in the following three steps.
- Narrow down the type based on clamping force, the molded part and the mold (2-plate or 3-plate, with or without a sprue)
- Put numbers on the take-out time and the downstream processing time as a breakdown of the cycle, decide first how many machines one person can tend, and only then write the RFP (request for proposal)
- For the signal interface with the molding machine (EUROMAP 67 and others), have suppliers specify “process interlock signals” and “safety-related signals” separately in the specification, and test them at FAT/SAT (the factory acceptance test before shipment and the site acceptance test)
Note that all seconds, shot counts, headcounts, amounts and years in this article relating to “Model Plant M” (described below) are original estimates and assumptions created for this article. They are neither industry averages nor survey results. Please read them as a “calculation template” to be replaced with your own figures.
What Is a Take-Out Robot? Comparing the Types and Clamping Force Guidelines
Four types
The equipment used to automate part removal in injection molding is easiest to organize when you think of it in four broad types.
| Type | Main role | Example applications | What to check in selection |
|---|---|---|---|
| Sprue picker | Removing sprues and runners | Small machines, processes where the runner is separated and dropped | Whether the model can also handle product removal |
| Traverse (Cartesian) take-out robot | Removing, arranging and transferring molded parts | Widely used from small to very large machines | Payload, stroke, number of axes, 3-plate support |
| Side-entry robot | High-speed removal from the side of the mold | Thin-wall parts, high-cycle molding | Floor installation space, position relative to the molding machine |
| Articulated robot, collaborative robot | Combining removal with downstream work (insert loading, inspection, packing) | Complex downstream processes, frequent product changeovers | Molding machine interface, safety design |
A sprue picker is originally a small device for removing sprues and runners. According to Wittmann’s general robot catalog (March 2026 edition), the company’s WX90 sprue picker has servo-driven, freely programmable axes, can also handle simple product removal in the tight space inside the molding machine’s safety enclosure, and is suited to molding machines with a clamping force of 35 to 150 t. However, this is a manufacturer claim for the WX90, and it does not mean that sprue pickers in general can be used for product removal.
A traverse take-out robot is mounted on top of the molding machine and descends into the mold along Cartesian axes, and it is a commonly used type. According to the lineup on the official English website of Japan’s HARMO, the company’s traverse servo robots cover the 15 to 70 t class up to the 400 to 1300 t class with the standard HRXⅢ-b series, and the 850 to 1300 t class up to the 2000 to 4000 t class with the very large TRX series. These ranges run from the lower limit of the smallest model to the upper limit of the largest model in each series, and no single model covers the whole range. On the same website, the model designation of the ARXⅢ-80, a small machine for overseas markets, is split into SW (for 2-plate molds) and GW (for 3-plate molds), which shows that the model changes depending on whether the runner is removed at the same time from a 3-plate mold.
A side-entry robot inserts its gripper from the side of the mold. According to Shini, the company’s SSE side-entry robot is designed for thin-wall products with a cycle of under 6 seconds, is suited to molding machines with a clamping force under 400 t, and is designed to stand on the floor separately from the molding machine. This is one company’s product example and not a definition shared by all manufacturers, but it is worth keeping in mind as a representative example of “thin-wall, high-cycle applications”.
Articulated robots and collaborative robots can also be used for removal if they are given a signal interface with the molding machine. According to Yaskawa (Motoman), the company’s SPI-67 is an interface that physically connects the robot to a horizontal molding machine, carries standard handshake signals such as door status, robot in operation and ejector position, can connect directly to the molding machine with dual-channel safety signals, and includes a EUROMAP 67 connector as an option. How to think about the cost of introducing a collaborative robot is covered in detail in Collaborative Robot Price: The Arm Is Under Half the Cell Cost.
Choosing the structure by clamping force (one manufacturer’s classification)
Wittmann’s catalog states that “the first thing to consider in robot selection is the clamping force of the molding machine” and divides the structures as follows.
| Clamping force range (Wittmann’s classification) | Structure considered suitable |
|---|---|
| 20 to 300 t | Fixed demolding axis |
| 300 to 1,200 t | Movable demolding axis (room for complex grippers and a lower overall height) |
| 1,200 to 5,000 t | Moving X-axis design (for large grippers and additional rotary axes) |
This is Wittmann’s recommended classification, not an industry standard. Its boundaries do not match those of other manufacturers, so rather than deciding “this tonnage means this structure”, the reliable approach is to have each candidate manufacturer write down the reasons for its proposal.
New products (K 2025 and Fakuma 2026)
At K 2025 (Dusseldorf, Germany) in October 2025, several companies announced new take-out robots. According to the trade magazine Plastics Technology, Sepro debuted its redesigned new S-Line, robots for molding machines with a clamping force of up to 900 t, at K 2025, claiming 15% higher speed than the previous generation. Wittmann premiered the Primus 118, extending the range of its Primus Cartesian robots to molding machines with a clamping force of up to 250 t. Yushin Precision Equipment had also announced that it would exhibit jointly with Wemo, its group company in Sweden, showing high-speed take-out robots and automation systems.
In addition, Fakuma 2026, the international trade fair for plastics processing, will be held from October 12 to 16, 2026 in Friedrichshafen, Germany. That is the week after this article is published. Please check information on new products in each company’s announcements after the fair.
Do Not Mistake the Catalog “Cycle Time” for the Take-Out Time
The most common misreading in a take-out robot comparison is the “cycle time” in catalogs. The specification tables in Wittmann’s catalog list a “Typical application cycle time” for each model, but a note explains that this is “the cycle time of the whole cell: molding machine + robot + peripheral automation”. In other words, this is not the take-out time during which the robot is inside the mold, but a value for the whole cell.
In the same way, always confirm what exactly was shortened when a manufacturer claims a reduction rate.
- According to a Wittmann news release (April 2021), its “SmartRemoval” function, which starts moving the robot axis into the mold before the platen has fully opened, can shorten the mold opening time by 10 to 30%. This is not a reduction rate for the whole cycle.
- According to HARMO, the EXZⅡ, a 2-axis servo swing type, shortens the take-out time by 62% and improves productivity by 15% compared with a pneumatic robot, and cuts setup adjustment time from 5 minutes to 30 seconds.
- According to the trade magazine Plastics Machinery & Manufacturing (2019), Yushin Precision Equipment claimed that its large MKA-2000S shortens take-out time by 17% and cycle time by 10% compared with the previous generation.
- The 15% speed increase of the Sepro S-Line is also a comparison with the previous-generation machine.
All of these are each company’s comparisons with its own products, and the conditions, such as the mold and molding machine, have not been published. They cannot be used as-is for comparisons with other companies’ machines or to predict the effect in your own plant. In the RFP, have suppliers present “the take-out time (from mold open complete until the gripper leaves the mold and gives the mold close permission)” and “the cycle time of the whole cell” separately, based on your own molds.
The Effect Is Decided by the Cycle Breakdown: Model Estimate 1
Assumptions for Model Plant M
From here on, we consider a fictional Model Plant M. The assumptions are as follows.
- A Japanese-owned automotive parts manufacturer in Rayong Province, Thailand. It molds interior parts on 10 injection molding machines with a clamping force of 350 t. The molds are 2-cavity (2 parts per shot)
- Two shifts (12 hours × 2), with actual running time of 20 hours a day = 72,000 seconds, excluding setups and stoppages
- Currently, one operator is stationed at each molding machine, opens the safety gate, removes the part by hand, and carries out gate cutting, visual inspection and packing. 10 machines × 1 person × 2 shifts = 20 people
- Annual labor cost per person is an assumed figure of 20,000 baht a month, or 240,000 baht a year, including overtime, bonuses, social security, and hiring and turnover costs
- Manual downstream work time (gate cutting + visual inspection + packing) is 7 seconds per part
The assumed labor cost is not the minimum wage itself. According to a DLA Piper briefing, Thailand’s minimum wage as of the July 2025 revision is 337 to 400 baht per day (400 baht in Bangkok, Chonburi, Rayong and other areas), but the model uses a total figure that includes the associated costs.
Ten seconds of take-out become two
Comparing the cycle breakdown at Model Plant M between manual removal and a take-out robot gives the following.
| Element | Manual removal (current) | Take-out robot |
|---|---|---|
| Injection, holding pressure, cooling | 26 s | 26 s |
| Mold opening and closing | 4 s | 4 s |
| Take-out (opening and closing the safety gate, time the hand or gripper is in the mold) | 10 s | 2 s |
| Molding cycle | 40 s | 32 s |
Figure 1: Cycle breakdown for manual removal and robot removal (horizontal bars, 26/4/10 s and 26/4/2 s)
With manual removal, 25% of the cycle (10 ÷ 40) is spent on take-out. With the robot, this becomes about 6% (2 ÷ 32 = 6.25%). The cycle is 8 seconds shorter, a reduction of 20% (40 s → 32 s).
Shots per machine per day rise from 72,000 ÷ 40 = 1,800 shots to 72,000 ÷ 32 = 2,250 shots. Because the molds are 2-cavity, that is from 3,600 parts to 4,500 parts. Production capacity increases by 25% (2,250 ÷ 1,800 = 1.25).
There are three points to note here.
- A robot does not shorten the cooling time. The 26 seconds for injection, holding pressure and cooling are determined by the product and the mold. What the robot shortens is only the take-out time.
- The 10 seconds and 2 seconds for take-out are assumed values for this article’s model, not manufacturer claims. Please measure them on your own molding machines with a stopwatch or the machine’s cycle log. A system that automatically collects shot records from injection molding is explained in Injection Molding Production Management System 2026.
- With manual removal, the timing of opening the gate varies from person to person, and the effective cooling time fluctuates, which can affect dimensions and appearance. More consistent take-out timing with a robot is a quality benefit, but it is not included in this article’s estimate.
The Bottleneck Moves Downstream: How Many Machines Can One Person Tend?
A take-out robot alone only cuts headcount by half
With a take-out robot, the output of one molding machine becomes “2 parts every 32 seconds”. So if people continue to do the gate cutting, inspection and packing, how many machines can one person handle?
- Option A (take-out robot only; gate cutting, inspection and packing by people): If one person handles 2 machines, 4 parts × 7 seconds = 28 seconds of work occur in 32 seconds, a load of 87.5% (28 ÷ 32). With 3 machines per person, it is 6 parts × 7 seconds = 42 seconds, a load of about 131% (131.25%), which does not work. So 2 machines per person is the limit, giving 5 people per shift × 2 shifts = 10 people.
- Option B (take-out robot + gate cutting jig station + vision inspection + automatic packing): The person’s work is reduced to replenishing cardboard boxes, carrying out full boxes and handling NG parts. One person can handle 5 machines, giving 2 people per shift × 2 shifts = 4 people.
For reference, the load on one person with today’s manual removal is (10 seconds of take-out + 2 parts × 7 seconds) ÷ 40 seconds = 60%. People are needed for manual removal not “because they are busy”, but “because the process does not run unless someone is stationed there”. A take-out robot removes this need to be “stationed”, but the manual downstream work that remains afterwards sets the upper limit on how many machines one person can tend.
Measure downstream seconds before the RFP
What this calculation shows is that if you introduce only robots, the reduction in headcount hits a ceiling set by the manual downstream work time. If the seconds per part change, the number of machines one person can tend changes with them. So before writing the RFP, measure the seconds for gate cutting, inspection and packing for each product.
If you automate the downstream processes, each of the component technologies needs its own study. For vision inspection, see Robot Vision Implementation Cost and ROI in Thailand — Supply, Accuracy and Lighting Decide It, and for packing, see Case Packing Automation Robot 2026 – Where the Line Really Stops.
Model Estimate of Investment and Payback: Slow If You Look Only at Labor Cost
Investment per machine
We assume the investment per machine at Model Plant M as follows (in baht; all figures are assumptions for this article and not market prices. We could not find published price information for take-out robots, so please confirm actual amounts with quotations).
| Cost item | Option A | Option B |
|---|---|---|
| Take-out robot unit (3-axis servo traverse type, 350 t class) | 900,000 | 900,000 |
| Safety fence, interlocks, signal connection to the molding machine | 150,000 | 150,000 |
| Drop chute, take-out conveyor | 100,000 | 100,000 |
| Installation, teaching, FAT/SAT, training | 80,000 | 80,000 |
| Gate cutting jig station | — | 400,000 |
| Vision inspection | — | 250,000 |
| Automatic packing | — | 350,000 |
| Total per machine | 1,230,000 | 2,230,000 |
| Total for 10 machines | 12,300,000 | 22,300,000 |
View 1: demand stays as it is (looking only at labor cost savings)
If demand does not change, the only effect is the labor cost saving.
| Item | Option A | Option B |
|---|---|---|
| Headcount | 20 → 10 people | 20 → 4 people |
| Labor cost saving | 10 people × 240,000 = 2,400,000 | 16 people × 240,000 = 3,840,000 |
| Maintenance cost (per year) | 30,000 × 10 machines = 300,000 | 60,000 × 10 machines = 600,000 |
| Net benefit (per year) | 2,100,000 | 3,240,000 |
| Investment | 12,300,000 | 22,300,000 |
| Simple payback | About 5.9 years | About 6.9 years |
| 5-year cumulative | −1,800,000 | −6,100,000 |
Option A pays back in 12,300,000 ÷ 2,100,000 = about 5.9 years, and Option B in 22,300,000 ÷ 3,240,000 = about 6.9 years. The 5-year cumulative figure is 2,100,000 × 5 − 12,300,000 = −1,800,000 baht for Option A and 3,240,000 × 5 − 22,300,000 = −6,100,000 baht for Option B, so neither pays back within 5 years. Against Option B’s additional investment of 10,000,000 baht, the difference in net benefit is 1,140,000 baht a year, so it takes about 8.8 years to recover the additional amount.
In other words, at Thai labor cost levels, payback exceeds 5 years if you look only at labor cost savings. Option B, which also automates the downstream processes, is a choice that is hard to justify on labor cost alone. Option B makes sense when there are reasons other than labor cost, such as “we cannot hire people” or “a quality complaint means we now need 100% vision inspection”.
View 2: demand grows by 20% (looking at spare capacity for increased production)
Next, consider the case where demand grows by 20%. This is a different assumption from View 1 (a different scenario), and it is not added to the effect of View 1. Here we also calculate Option A only.
- Required shots rise from the current 10 machines × 1,800 = 18,000 shots/day to 21,600 shots/day with a 20% increase.
- If you stay with manual removal: 21,600 ÷ 1,800 = 12 machines are needed. You would buy 2 additional molding machines; at 4,500,000 baht per machine (an assumed figure including ancillary work), that is 9,000,000 baht. Operators become 12 machines × 2 shifts = 24 people.
- If you respond with Option A: 21,600 ÷ 2,250 = 9.6 machines’ worth, so the existing 10 machines are enough (96% utilization). Operators stay at 10 people, and the load with 2 machines per person is 87.5% × 0.96 = 84%, so it works.
The baseline for comparison in this case is “increasing production with manual removal on 12 machines with 24 people”.
| Item | Option A (View 2) |
|---|---|
| Net investment | 12,300,000 − 9,000,000 = 3,300,000 |
| Labor cost difference | (24 people − 10 people) = 14 people × 240,000 = 3,360,000 |
| Maintenance cost (per year) | 300,000 |
| Net benefit (per year) | 3,060,000 |
| Simple payback | About 1.1 years |
| 5-year cumulative | 3,060,000 × 5 − 3,300,000 = 12,000,000 |
Figure 2: Cumulative cash flow (three lines: Options A and B under View 1, and Option A under View 2, years 0 to 5)
Even for the same Option A, payback takes about 5.9 years if demand stays as it is, and about 1.1 years if demand grows by 20%. What makes the difference is not having to buy 2 molding machines. In plants where demand is growing, the value of a take-out robot shows up less in labor cost savings than in the spare capacity for increased production created by the shorter cycle.
What the estimate does not include
- View 2 does not include the maintenance cost, electricity, floor space or additional molds for the 2 extra molding machines. Including them would make Option A even more favorable.
- Improvements in defect rates, lower risk of occupational accidents and more stable quality are not included in either view (this is a conservative estimate).
- BOI incentives are not included either. If they can be used, payback could become even shorter.
- The price of 4,500,000 baht per molding machine is also an assumption for this article and not a market price.
When you redo the payback calculation for your own plant, first decide “how demand will develop”, and then choose only one view that fits that assumption. If you add up the effects of the two views, you will count the same effect twice.
Signal Interface and Safety with the Molding Machine: Separate Process Signals from Safety Signals
What is exchanged over EUROMAP 67
The take-out robot and the molding machine coordinate their movements by exchanging signals. According to a briefing by ASTOR (July 2025), EUROMAP 67 is an electrical interface standard between injection molding machines and industrial robots developed by EUROMAP (the European association of plastics and rubber machinery manufacturers), and example signals include ejector back enable and ejector forward enable, confirmation of each position, mold open complete, door status, robot in operation and ejector position. Before taking the product from the mold, the robot must confirm that the molding machine has finished its cycle and that the mold is open. Some molding machines support the older EUROMAP 12 standard, so check the supported standard of each existing machine one by one.
As a newer development, the OPC Foundation has published “OPC 40079”, an OPC UA interface between molding machines and robots. According to the foundation’s document page, V1.00 is a Release Candidate dated October 31, 2021, and it covers real-time signal exchange, axis positions, enable signals, part tracking, and the exchange of production and quality data. Please ask candidate manufacturers whether any molding machines or robots have adopted it.
“Mould Area Free (MAF)” is not a safety signal
What needs the most attention here is the safety status of each signal. A briefing by Huarong (May 2026), a Taiwanese injection molding machine manufacturer, lists the main signals between the molding machine and the robot as mold open and mold close, ejector forward and back, Mould Area Free (MAF, meaning the robot is outside the hazard zone), mold close enable, two-way emergency stop, and safety gate interlocks, and then states that “process interlock signals”, which control the sequence, should be distinguished from “safety-related signals” (emergency stop, safety gate, light curtain), which prevent hazardous motion.
Universal Robots (UR), a collaborative robot maker, warned in a November 2021 safety notice about its molding machine interface products that the only safety-related inputs from the molding machine are two: emergency stop and protective stop, and that the MAF wiring is a non-safety signal and must not be used to stop the robot or the molding machine. This notice concerns UR products, but the misunderstanding of treating it as a safety basis, as in “the Mould Area Free signal is present, so the mold will not close while the robot is inside it”, can occur with any combination of manufacturers.
Therefore, in the RFP, please have suppliers submit a classification table that separates each signal into “process interlock signal” or “safety-related signal”. Process-side measures such as the one Wittmann describes, in which the robot gives the mold close enable just before it leaves the mold once vacuum monitoring has confirmed that the product is held, are effective for shortening the cycle, but they are strictly process interlocking, and should be designed and verified separately from the safety functions.
Standards to reference, and what we could not confirm
Safety design references standards on both the molding machine side and the robot side.
- EN ISO 20430:2020: The standard that sets safety requirements for injection molding machines. According to the sales page of the Estonian Centre for Standardisation (EVS), it replaced the earlier EN 201:2009. It covers machines whose platen movement is hydraulic or electric, and the mold and the exhaust system are stated not to be part of the machine.
- ISO 10218-1/-2:2025: Safety requirements for industrial robots (Part 1) and for robot applications and cells (Part 2). The EN versions became effective on April 1, 2025.
However, we could not confirm from the published abstracts where the boundary of application between the two standards lies, or whether a traverse take-out robot counts as an industrial robot under ISO 10218. Do not assume that one of the standards alone will suffice; reference the standards on both the molding machine side and the robot side, have suppliers state in the RFP which hazards are handled under whose responsibility, and have them submit a risk assessment. For decisions on application, the reliable approach is to check individually with experts in standards and certification. For how to think about safety fences, Robot Safety Fence Standards 2026 — Stopping Distance Decides is also a useful reference.
The safety guidance on horizontal injection molding machines from the US Occupational Safety and Health Administration (OSHA) lists points such as fitting the operator-side gate with an interlock so that the machine moves only when the gate is closed, guarding the part discharge opening so that hands cannot reach into the mold, and guarding the robot so that workers cannot enter its space while it is operating. This is a US government document and not a legal requirement in Thailand, but it is useful as a set of design check items.
12 Items to Include in the RFP and Quotation Request for Take-Out Robot Selection
To make a take-out robot comparison fair, it is essential to have every company make its proposal on the same assumptions. At a minimum, include the following 12 items in the RFP.
- Molding machine and mold specifications: The molding machine’s manufacturer, model, clamping force, mold opening stroke, tie bar spacing and mounting surface. Whether the mold is 2-plate or 3-plate, the number of cavities, and whether there is a sprue or runner.
- Molded part and gripper: The weight, dimensions and material of the molded part, the removal orientation, and the tolerance for scratches and deformation. State the total weight including the gripper (suction or chuck, with or without vacuum monitoring), and allow a margin in the payload.
- Reasons for the proposed type: Which of sprue picker, traverse type, side-entry, or articulated and collaborative robot is proposed, and why.
- Guaranteed take-out time: Have suppliers present the time from mold open complete until the gripper leaves the mold and gives the mold close permission, separately from the cycle time of the whole cell.
- Signal interface: Which of EUROMAP 67, EUROMAP 12, SPI, OPC 40079 and others will be used, and a classification table separating each signal into “process interlock signal” or “safety-related signal”.
- Safety: Safety fence and gate interlocks, interconnection of emergency stops, protective stop, the relationship between the molding machine’s safety gate and the robot fence, the standards referenced (EN ISO 20430, ISO 10218-2 and others), and submission of a risk assessment.
- Scope of responsibility for downstream processes: How far gate cutting, inspection, packing and insert loading are included, and the processing time per part.
- Mold changes: The gripper change time during setup, and how programs are recalled. Automating gripper changes is covered in Robot Tool Changer Selection for Thai Factories: Procurement and Acceptance Guide.
- Scope of installation and modification: The scope of installation, wiring and modifications on the molding machine side, and whether approval from the molding machine manufacturer is required.
- FAT and SAT: Test items before shipment and on site, and acceptance criteria.
- Maintenance: Inspection intervals, spare parts, engineers in Thailand and response times, and operating procedures in Thai.
- Commercial terms: Price validity, exchange rates, import duties and customs clearance, payment terms, and whether the supplier can provide the machine list and itemized quotation needed for a BOI application.
Comparing suppliers is easier if you have quotations broken down, as in the cost table above, into “robot unit”, “safety fence, interlocks and signal connection”, “chute and conveyor”, and “installation, teaching, testing and training”. In operations where programs and grippers are switched at every mold change, managing information for each mold is also important. Please also see Mold Management System Implementation in Thailand: Shot Counts and Ledgers.
What to Check at FAT/SAT
At FAT (the factory acceptance test before shipment) and SAT (the test after on-site installation), check the following points. You need to set the acceptance values yourself; the following are examples of how to set them.
- Measured take-out time and cycle time: Run continuously with the actual mold and material, for example for several hours, and compare with the guaranteed values in the RFP.
- Detection of drops and missed picks: When vacuum monitoring detects a missed pick, how the robot and the molding machine stop and how they recover.
- Safety gate and emergency stop: That both the robot and the molding machine stop when the molding machine’s safety gate is opened and when the emergency stop is pressed.
- Behavior when a process signal is lost: When a process signal such as Mould Area Free (MAF) is lost, whether the system correctly waits as part of process interlocking, not as a safety function.
- Power failure and loss of air pressure: Whether the workpiece drops, whether the gripper keeps holding it, and what the recovery procedure is.
- Changeover: The measured setup time and the gripper change procedure.
- Coordination with downstream processes: How the system behaves when the conveyor jams, a box is full, or NG parts are discharged.
- Training and procedures: Operating procedures and daily inspection procedures in Thai, and training for operators.
The basic approach is to recheck at SAT, in combination with the molding machine on site, the items already confirmed at FAT. In particular, test the interconnection of the safety gate and emergency stop at SAT without fail, because the wiring on the molding machine side is connected for the first time on site.
Issues Specific to Thailand
1. Payback calculated on labor cost alone tends to exceed 5 years
Thailand’s minimum wage as of the July 2025 revision is 337 to 400 baht per day, which is still low compared with Japan. As in View 1 for Model Plant M, a feature of Thailand is that payback calculated only on labor cost savings tends to exceed 5 years. Decisions on take-out robots should center on three points: spare capacity for increased production (whether you can avoid buying additional molding machines), the inability to hire people, and stable quality.
2. Industry scale and quality requirements
According to Krungsri Research (March 2025), there were 3,217 plastic processors operating in Thailand as of 2023, and a 2021 survey by the Thailand Plastics Institute found that 36.5% of plastic producers use injection molding. This is the share of producers that use injection molding, not the share of injection molding in production volume. In addition, according to the Federation of Thai Industries (FTI), Thailand’s vehicle production in 2025 was 1,455,569 units, of which 65.69% were for export. Quality requirements are strict in molding automotive parts with a high export share, and the value of consistent take-out timing is not small.
3. Manufacturers’ local bases and maintenance response times
A take-out robot is equipment that sees changeovers and teaching adjustments almost every day, and if it breaks down, the molding machine stops. According to Yushin Precision Equipment’s company history, the company opened a representative office in Bangkok in 1996 and established a Thai subsidiary in 2008. HARMO also announced a webinar for Thailand in 2023. This is why it makes sense to confirm in the RFP, for each candidate manufacturer, the number of engineers and bases in Thailand, response times and spare parts stock.
4. BOI’s Smart and Sustainable Industry measure
According to the Thailand Board of Investment (BOI) 2023 investment promotion guide, the machinery upgrade and automation category of the Smart and Sustainable Industry measure offers exemption of import duties on machinery and a 3-year corporate income tax exemption capped at 50% of the investment, on condition of an investment of 1 million baht or more (excluding land and working capital). For automation systems and robots, if 30% or more of the value of the machinery being upgraded comes through links with Thailand’s automation industry, the cap becomes 100% of the investment. According to the deputy spokesperson of the Thai Prime Minister’s Office, companies with Japanese shareholders applied for this measure in more than 400 cases, worth more than 55 billion baht, from 2023 to the first half of 2026.
However, according to a briefing by Alvarez & Marsal (published on Mondaq), the BOI overhauled its investment promotion measures in January 2026 and made many measures available for application in 2026 and 2027, but we could not confirm the latest requirements and deadlines for this measure. If you are considering an application, please confirm individually with the BOI. The details of the requirements are explained in BOI Automation Tax Exemption 2026 — What Decides 50% or 100%.
5. General requirements of Thai Ministerial Regulation B.E.2564
The Thai Ministry of Labour’s “Ministerial Regulation on the Standards for Administration and Management of Occupational Safety, Health and Environment in relation to Machinery, Cranes and Boilers B.E.2564 (2021)” was promulgated in the Royal Gazette on August 6, 2021. According to a briefing by Enviliance, employers are required, among other things, to prepare written operating procedures and to carry out hazard assessments of machines that pose a risk of amputation. However, we could not confirm any specific provisions on take-out robots or injection molding machines. Please check with safety experts or the competent labor authority which provisions apply to your equipment.
6. Gathering information at trade fairs
Trade fairs are also useful for comparing the latest models. Fakuma 2026 will be held from October 12 to 16, 2026. Including trade fairs in Thailand, please check the dates on each fair’s official website.
A 90-Day Plan for Introducing Take-Out Robots
If you proceed from take-out robot selection to a pilot installation in 90 days as follows, you can avoid rushing the decision without letting it stall at the study stage.
Figure 3: The 90-day plan (3 phases)
Days 0 to 30: Measurement and prioritizing target machines
- Measure the cycle breakdown (injection, holding pressure and cooling; mold opening and closing; take-out) for each molding machine
- Measure the seconds per part for gate cutting, inspection and packing for each product
- Confirm the demand outlook (staying as it is, or growing) with sales, and decide on one view for the payback
- Prioritize, starting from molding machines for products with a large share of take-out time and growing demand
Days 31 to 60: Decide the type and scope, issue the RFP, compare quotations
- Decide the type (sprue picker, traverse type, side-entry, or articulated and collaborative robot) and how far to automate the downstream processes
- Issue the 12-item RFP and compare quotations by cost item
- Have each company submit a signal classification table (process interlock signals and safety-related signals) and compare them
- Check with the BOI whether the project could be eligible
Days 61 to 90: Pilot on one machine, FAT and SAT, decide on rollout
- Run a pilot installation on one machine, and confirm the take-out time, safety and behavior during power failure at FAT and SAT
- Prepare operator training and procedures in Thai
- Recalculate the payback from the measured take-out time and the number of machines one person can tend, and decide whether to extend to the remaining molding machines
Frequently Asked Questions (FAQ)
Q1. What is the difference between a take-out robot and a sprue picker?
A sprue picker is originally a small device for removing sprues and runners. A take-out robot removes the molded part itself and also handles arranging and transferring it. Some models, such as Wittmann’s WX90, are said to be capable of simple product removal, but this is a manufacturer claim for a specific model, and it does not mean that sprue pickers in general can remove products. Whether the runner and the product are removed at the same time from a 3-plate mold is also a point that separates types and model designations.
Q2. From what clamping force is a take-out robot needed? How should I choose one?
There is no rule that “a robot is needed above a certain tonnage”. Manufacturers’ lineups range from small to very large machines; for example, HARMO’s traverse robots come in series from the 15 t class to the 4000 t class. The way to choose is to narrow down the type and size from four factors: clamping force, the molded part (weight, dimensions, removal orientation), the mold (2-plate or 3-plate, number of cavities), and the downstream processes (seconds for gate cutting, inspection and packing). Wittmann’s classification of structures by clamping force (20 to 300 t, 300 to 1,200 t, 1,200 to 5,000 t) is also one guideline, but it does not match other manufacturers’ classifications.
Q3. What does a take-out robot cost, and what is the cost of introduction?
We could not find published prices for take-out robot units from manufacturers or distributors. For Model Plant M in this article, we assumed 1,230,000 baht per machine for a 350 t class traverse type, including the unit, safety fence, interlocks and signal connection, chute and conveyor, and installation, teaching, testing and training, but this is not a market price. For actual costs, please obtain quotations by cost item, as in this article’s cost table, and compare them.
Q4. How much does a take-out robot shorten the cycle time?
Only the take-out time is shortened; the time for injection, holding pressure and cooling is not. At Model Plant M, take-out went from 10 seconds to 2 seconds and the cycle from 40 seconds to 32 seconds (a 20% reduction), but these are assumed values. Manufacturers’ claimed reduction rates (for example, Wittmann’s 10 to 30% reduction in mold opening time, and HARMO’s 62% shorter take-out time compared with a pneumatic robot) are all each company’s comparisons with its own products, and the conditions have not been published. The reliable approach is to measure the take-out time on your own molding machines and have suppliers state guaranteed values in the RFP.
Q5. What is EUROMAP 67? Can it be used as a safety signal?
EUROMAP 67 is an electrical interface standard between molding machines and handling devices such as take-out robots, and it exchanges signals such as mold open complete, ejector forward and back enable, and door status. However, not all of its signals are safety signals. In a safety notice, UR warns that the Mould Area Free (MAF) signal is a non-safety signal and must not be used to stop the robot or the molding machine. In the RFP, have suppliers classify each signal as a process interlock signal or a safety-related signal, and please check the safety design and the application of standards with experts.
Q6. Can a collaborative robot be used for part removal?
Yes. If given a molding machine interface (EUROMAP 67, SPI and others), articulated robots and collaborative robots can also be used for removal. Their advantage is that they can easily take on downstream work as well (insert loading, inspection, packing). On the other hand, you need to confirm their removal speed compared with a traverse type on your own molds, and a safety design and risk assessment that include the interconnection with the molding machine’s safety gate and emergency stop are essential. Please note that being a collaborative robot does not necessarily mean that no fence is needed.
Summary
- Clamping force and payload alone are not enough for take-out robot selection. What determines the effect is the take-out time spent inside the mold and the downstream processing time.
- The “cycle time” in catalogs is a value for the whole cell, not the take-out time. Manufacturers’ reduction rates are all comparisons with their own products under unknown conditions, so have suppliers guarantee the take-out time and the cycle time separately in the RFP.
- At Model Plant M, take-out went from 10 seconds to 2 seconds, the cycle from 40 seconds to 32 seconds, and production capacity rose by 25%. However, the cooling time does not get shorter.
- If you introduce only robots, the number of machines one person can tend hits a ceiling set by the downstream seconds (from 20 people to 10 people with Option A). Measure the downstream seconds before the RFP.
- Looking only at labor cost savings, payback is about 5.9 years for Option A and about 6.9 years for Option B. If demand grows by 20%, buying 2 additional molding machines can be avoided, so payback for Option A becomes about 1.1 years. The two views rest on different assumptions and are not added together.
- Separate signals into “process interlock signals” and “safety-related signals”, and do not use Mould Area Free (MAF) as a basis for safety. Please check the boundary of application of standards, the application of Thai ministerial regulations, and the latest BOI requirements individually with experts, the competent authorities and the BOI.
At TOMAS TECH, we are happy to talk even at the stage before you have decided how to measure the cycle breakdown and the downstream seconds, or which molding machines to start with. We also welcome questions such as wanting to reread the quotations you have in hand by separating take-out time, whole-cell cycle time and cost items, or wanting to work out how to have suppliers write the signal classification table for your molding machines. If you are considering automating the removal of molded parts, please feel free to contact us through our contact form.
References
- Wittmann (Robots and automation systems catalog, March 2026 edition): https://wittmann-group.com/sites/default/files/2026-03/robot-en-2026-03.pdf
- Wittmann (news release “Minimizing removal times”, April 2021): https://www.wittmann-group.com/sites/default/files/2021-04/witt-nr-03-2021_minimizing-removal-times.pdf
- HARMO (take-out robot product list): https://eng.injection-molding.jp/products/robot/
- HARMO (ARXⅢ-80SW/GW): https://eng.injection-molding.jp/products/robot/arx3-80sw-gw
- HARMO (EXZ series): https://eng.injection-molding.jp/exz-series
- HARMO (webinar announcement for Thailand, May 2023): https://eng.injection-molding.jp/products/news/20230502
- Plastics Machinery & Manufacturing (Yushin MKA-2000S, November 2019): https://www.plasticsmachinerymanufacturing.com/injection-molding/article/21109654/yushin-robot-targets-large-heavy-parts
- Plastics Technology (Sepro S-Line, August 2025): https://www.ptonline.com/products/sepro-launches-revamped-robot-line-and-controller
- Shini (Side Entry Robot SSE): https://www.shini.com/enus/products_i_Side_Entry_Robot_SSE.html
- Yaskawa Motoman (SPI-67): https://www.motoman.com/en-us/applications/machine-tending/spi-67
- EVS (EVS-EN ISO 20430:2020): https://www.evs.ee/en/evs-en-iso-20430-2020
- Standard Norge (NS-EN ISO 20430:2020): https://online.standard.no/en/ns-en-iso-20430-2020
- EVS (EVS-EN ISO 10218-1:2025): https://www.evs.ee/en/evs-en-iso-10218-1-2025
- EVS (EVS-EN ISO 10218-2:2025): https://www.evs.ee/en/evs-en-iso-10218-2-2025
- Universal Robots (safety notice for IMMI and EUROMAP products, November 2021): https://www.universal-robots.com/articles/ur/safety/product-alert-safety-notice-for-all-ur-immi-and-euromap-products/
- Huarong (briefing on signals between molding machines and robots, May 2026): https://www.huarong.com.tw/page/news/en/company_news/detail/194/
- ASTOR (briefing on EUROMAP 67, July 2025): https://www.astor.com.pl/en/?p=927
- OPC Foundation (OPC 40079): https://opcfoundation.org/developer-tools/documents/view/220
- OSHA (safety of horizontal injection molding machines): https://www.osha.gov/etools/machine-guarding/plastics-machinery/horizontal-injection-molding-machines/safety-tour-view-1
- Enviliance (overview of the Thai Ministerial Regulation on machinery, cranes and boilers B.E.2564): https://enviliance.com/regions/southeast-asia/th/report_4310
- Krungsri Research (Plastics Industry Outlook 2025-2027): https://www.krungsri.com/en/research/industry/industry-outlook/petrochemicals/plastics/io/io-plastics-2025-2027
- Motor Expo (FTI announcement on Thailand’s 2025 vehicle production): https://www.motorexpo.co.th/en/news/5344
- DLA Piper (Thailand minimum wage, July 2025 revision): https://www.dlapiper.com/insights/publications/2025/07/thailand-increases-minimum-wage-across-all-provinces
- Thailand Board of Investment (BOI), Smart and Sustainable Industry measure guide: https://www.boi.go.th/upload/content/Smart_and_Sustainable_Industry.pdf
- MGR Online (BOI applications by Japanese companies, August 2026): https://en.mgronline.com/politics/japan-accelerates-production-base-upgrades-in-thailand-machinery-replacement-dig/
- Mondaq (BOI’s 2026 overhaul of measures): https://www.mondaq.com/corporate-and-company-law/1739584/thailands-renewed-boi-incentives-a-strategic-window-for-growth-expansion-and-investment-20262027
- Yushin Precision Equipment (K 2025 exhibition): https://www.yushincompany.jp/en/news/archive/2025/k2025.html
- Yushin Precision Equipment (company history): https://www.yushincompany.jp/en/about/history/
- Polyformnext (Wittmann Primus 118, K 2025): https://www.polyformnext.de/automation/images/wittmann-on-the-k-2025-primus-robot-with-increased-load-capacity-1.htm
- ETMM (Fakuma 2026): https://www.etmm-online.com/30th-fakuma-trade-fair-plastics-processing-2026-a-27cd2d7cd6c4af9bea6a2f05ba7936e0
