IoT systems

IoT Initiatives at the Amata Industrial Estate

Provided by

E&H Precision (Thailand) Co., Ltd.

We would like to take this opportunity to introduce some of our IoT initiatives via this platform.

E&H Precision is a manufacturing company which uses various proprietary and cutting-edge technologies in its manufacturing processes. These include IoT systems for accumulating and analyzing various types of in-house data and manufacturing equipment operating status data, as well as data mining systems for corporate management purposes and AI-powered pattern prediction technology. Here, we will introduce some of its relevant projects.

IoT systems for real-time production status monitoring

Q Please give us an overview of your company.

Shibata: Our company was founded in Thailand in 1996. We originally produced shafts and screws for automotive audio systems, but now we also perform cutting work for integral safety components such as fuel injectors and brake parts.

At our Thai factory, we have over 650 cam-type/CNC automatic lathing machines and over 150 secondary processing machines—the most in Southeast Asia. We are currently focusing on initiatives based around using IoT in manufacturing equipment.

Q Why are you introducing IoT?

Shibata: In total, our company has around 800 production machines in Thailand, India and Mexico. However, until now there hasn’t been a method for confirming their operating statuses other than by direct visual checks of the machines themselves.

Because of this, it takes a fairly long period of time before managers can identify problems following machine stoppages. The goal of this project is to utilize the currently popular concept of IoT to construct a system capable of obtaining and analyzing real-time manufacturing data. Such a system would enhance our ability to manage the operating statuses of numerous production machines simultaneously and in turn permit us to make important business decisions in a more timely manner.

Q Implementing IoT across 800 machines used to manufacture over 1,200 types of products must have been incredibly difficult.

Shibata: First, we had to select which type of sensors would be used for measuring machine operating statuses, which we did through trial and error. The first thing we experimented with was a contact-type sensor. While this is a simple method which counts the number of times the sensor touches a product, we found that it shifted slightly at the moment where it touched the product, which made measurement errors more likely to occur. This gave us cause for concern about the potential lifespan of a solution like this.

Next, we tested a non-contact method that uses infrared technology. However, infrared radiation can produce incorrect measurements in factory environments where large amounts of lubricants are used, which was a strong disadvantage.

It made us wonder whether a method exists that would allow us to conduct measurements with fewer errors. We finally decided to utilize electric current sensors after conducting investigations into non-contact sensing methods. While there is a certain degree of error when using current sensors, stable output can be obtained via appropriate circuit design and programming.

Once able to obtain output via the sensor, the next step was to enable its visualization. At first we had no idea how or where to send the data, or what method we should use for its visualization.

We had some idea about the newly popular concept of IoT, but we didn’t know what kind of applications to use or where to start. It was then that we began researching the topic using online and text resources to gain a better understanding. It was then that we came across Microsoft Azure and decided upon its application. However, we still didn’t understand 80% of what was written on the Azure overview page, so we obtained even more online and text resources and started working out how to actually use the features of Azure.

It gradually took shape as we continued to work out which functions to use and which to not. There were still aspects that we couldn’t understand and we contacted Microsoft for help, but because this sort of work is normally done by an IT provider, they had never seen novices from the manufacturing industry trying to do it and were very interested in helping us. Since then, representatives from Microsoft have been visiting us about once a month to assist in our efforts.

Visualized web page

Q Obviously these initiatives have produced results—the first being the successful use of a nano computer attached to each machine that sends green, yellow and red signals to enable centralized online control of manufacturing equipment. Could you tell us about any upcoming initiatives?

Shibata: The second step of this project is moving towards a paperless system that sends data to mobile devices and we are gradually making progress in terms of using machine learning to analyze the data we collect. In the future, we hope to automate both data collection and analysis, with the eventual goal of achieving AI capable of detecting the causes of errors.

This system has enabled us to accurately gauge machine operating statuses, as well as provided us with hitherto unavailable insight into our manufacturing processes. There is a large amount of data hidden within the company that we have been unable to take advantage of until now and our next goal is to find ways to both aggregate and utilize this data.

As a lathing company that aims to be No. 1 in the world, our goal is to create a world-leading system that contributes to the growth of our company. There are many issues to address, but we relish the opportunity to tackle them.

AI Toilet Usage Prediction

Q Besides IoT, you have AI initiatives as well. What kind of initiatives are these?

Shibata: AI is a necessary technology for achieving our final goal of using data obtained by IoT production status monitoring systems to predict stoppages of production equipment. However, we are finding that at the current stage, we have not gathered enough data to use AI.

However, instead of just waiting for the necessary data to be gathered, we have decided to practice creating systems that use AI (machine learning) by implementing a usage status prediction system for our company’s toilet stalls.

Q What kind of setup did the toilet stall usage status prediction system use?

Shibata: In order to predict usage, we first had to obtain data on actual toilet usage. To do this, we started by working out how to obtain this data.

We were unsure of how best to obtain the data, but after noticing that we use rotating locks on the toilet stalls in our company, we decided to obtain data regarding usage status via the stall locks.

Basically, an infrared sensor detects the moment that the stall door is locked or unlocked and then transmits this data to a nano PC, which uploads it to the cloud via WiFi, where data is accumulated.

The collected data is sent to a corporate work computer or smartphone via the cloud, transmitting the usage status. The current stall usage status is color-coded, enabling which stalls are in use to be confirmed.

Q Knowing the usage status must be useful! You said it would be possible to make predictions based on this data. How is that progressing?

Shibata: We are taking advantage of the actual usage data, while also recognizing that there are various causes for toilet use, so any predictions will have to use multiple types of data—including usage history, weather conditions (automatically obtained from weather forecast websites), work calendars, etc.

At 7 AM each morning, the system produces toilet stalls usage predictions from 7 AM to 6 PM in one-hour increments.

The collected data is sent to a corporate work computer or smartphone, transmitting the usage status. The current stall usage status is color-coded, enabling which stalls are in use to be confirmed.

Q Can you tell us about any other future initiatives?

Shibata: We feel that many other companies in the Amata Industrial Estate share the same situations and problems that our company is tackling via these initiatives.

We hope that the simple IoT/AI solutions cultivated by our company can be useful for many others and are considering external sales in the future.

Thank you for telling us all about your IoT and AI initiatives in the Amata Industrial


Our Capability

  • One of the largest machining factories in Southeast Asia

    We own more than 670 automatic turning machines and more than 200 additional processing machines and process steel, stainless, brass, copper and aluminum, making us one of the largest factories in Southeast Asia.

  • Production in emerging countries make it possible to provide products at reasonable cost

    We do all the processes such as technology development, process planning, test production and production in Asia. This is the secret of our reasonable price and quick response.

  • We can do A to Z, we can do anything!

    We produce various products such as fuel injection parts, brake parts for automotive, household electrical appliances, office automation parts, fishing equipments and toys.

About Us

E&H Precision(Thailand) Co., Ltd.

We're a manufacturer of cut metal products that utilizes CNC automatic lathes at our facilities in Japan, Thailand, India and Mexico.
We can manufacture cut and machined products for various applications including automotive engines, common rail injection systems, gasoline injectors, motorcycle carburetors, audio systems, home appliances, office equipment parts, fishing equipment and toys.

700/387 M.6 T.Donhuaroh, Muang, Chonburi 20000 THAILAND

Web site

Share this service



Contact Us

Please fill out the form below if you would like to learn more about E&H Precision.


    Customer Information

    On Meeit.Biz, we support development initiatives for Amata Smart City.

    In order to support the Thailand 4.0 economic policy vision, we plan to introduce companies offering IoT solutions and companies developing initiatives for IoT.

    Contact us if your company might be interested!


    If you would like to join partner or have any questions about,

    Please contact us from the links below.


      © Copyright 2019

      We use cookies to ensure that we give you the best experience on our website. If you continue to use this site we will assume that you are happy with it. Cookie Policy