Programmable digital circuits FPGA - questions and answers
Programmable digital circuits FPGA - questions and answersDate of publication: 19-11-2025 🕒 7 min read
FPGA systems have been commercially available for many years, yet they are not very well known among electronics designers. One could risk stating that some know something about them, but usually, this knowledge is very general and rather discouraging for use in devices they are building. This text provides answers to the most frequently asked questions regarding FPGA systems and their applications, which will allow for a preliminary acquaintance with this group of systems and more informed decision-making regarding the components used.
There are many common opinions about FPGA systems that discourage their use. Meanwhile, these systems are indispensable in many fields, and it is good to have at least basic knowledge about them to make informed decisions regarding the applications being built and to realistically assess their capabilities.
- Can FPGA replace a microcontroller?
- Why are FPGA systems not very popular?
- Will FPGA systems find wider applications in the future?
- When should FPGA systems not be used?
- Can FPGA be reprogrammed?
- Do FPGA systems find applications in Industry 4.0?
- What types of memory are equipped in FPGA?
Can FPGA replace a microcontroller?
FPGA systems are used in different applications than microcontrollers, so the answer to this question depends on the specific case. The core of a microcontroller consists of many logic circuits that can also be created using hardware description languages. Additionally, FPGA systems can contain Flash memory and RAM (or have them connected externally), so one could risk stating that most cores of popular microcontrollers can be emulated using FPGA. However, one must consider the significantly higher cost of both purchasing the chip itself and implementing its hardware application and configuration.
FPGA systems and microcontrollers operate differently. The former are very fast and configured using hardware description languages. They can perform many operations simultaneously, which is a significant advantage in applications requiring high computational power. The ease of changing hardware configuration makes them excellent for testing and prototyping circuits before they are put into production. Their disadvantage is the complex application and high price of the chip itself – often many times higher than the price of a microcontroller.
The main advantage of microcontrollers is their low purchase cost and uncomplicated application. Usually, in the environment of a microcontroller, there are elements of a clock generator and those that allow it to cooperate with the environment (most often adapting the electrical parameters of interfaces to the I/O pins of the microcontroller). They are programmed using well-known high-level programming languages such as C, C++, or Python. Although they are not as fast and flexible in applications as FPGA systems, they are significantly cheaper in mass production as well as in unit applications. The hardware application is also less complicated and therefore easier to implement and much cheaper.
In answering the question of whether an FPGA can replace a microcontroller, one should not only consider the possibility itself, as it is almost certain that one can find an FPGA that can emulate the core of a microcontroller and its peripheral blocks or perform a functionally equivalent application, but rather think about the cost of such a solution. In uncomplicated applications, such as controlling devices, it is certainly more cost-effective and easier to implement microcontrollers.
Why are FPGA systems not very popular?
Presumably, the low popularity of FPGA systems is determined by their high price, difficult hardware application (for example, complex FPGA systems require complicated power circuits and voltage sequencing, as well as appropriate safeguards in case of power loss), and the complexity of programming with a relatively low familiarity with hardware description languages.
Due to their parameters, FPGA systems are often used in specialized applications, such as signal processing, telecommunications, or prototyping digital circuits, where the cost of the final solution is not as important as reliability and performance. Otherwise, when the price of the final solution starts to matter more than processing speed, microcontrollers or dedicated integrated circuits (ASICs) are more cost-effective and sufficient. Additionally, FPGA may be less energy-efficient compared to ASICs, which are optimized for use in a specific application.
Will FPGA systems find wider applications in the future?
The requirements of modern applications, where a fast microcontroller or ASIC is insufficient, may lead to a significant decrease in FPGA prices and their widespread use in consumer devices. FPGA systems certainly have enormous potential to find wider applications in the future, especially in applications requiring high data processing speed, flexibility, and reconfigurability. In professional applications, this may involve building prototypes of digital circuits and ASICs before their production, while in consumer devices, it may involve image processing for computer games and VR imaging.
In industry, the growing demand for devices that process large amounts of data in real-time in applications related to AI, machine learning, or signal processing makes FPGA systems, which can perform many operations simultaneously at great speed, increasingly attractive.
In telecommunications, with the expansion of 5G networks, IoT, or data transmission for autonomous vehicles, the need for advanced logic circuits that can be quickly adapted to specific requirements will grow, and FPGA systems may play a key role in these fields. Notably, using FPGA, one can build a hardware circuit for encrypting and decrypting communication, with the ability to reconfigure and update algorithms without the need to replace hardware. Although FPGA are less energy-efficient than ASIC, the demand for them may lead to the development of new production technologies resulting in increased competitiveness in terms of energy consumption.
When should FPGA systems not be used?
Using an FPGA in an application does not always make sense. This especially applies to applications where a much cheaper microcontroller or ASIC can fulfill the functionality. For example, if the application only requires basic functionalities, such as a controller, counter, or comparator, it is certainly more cost-effective to use a microcontroller or ASIC. However, if the counter is to serve as a timer for accurately measuring the time of a light beam reflecting off an obstacle, then such a task can only be handled by a very fast FPGA.
If the device is to be mass-produced, another reason to avoid using an FPGA may be the potentially high cost of the final solution. FPGA chips are usually many times more expensive than microcontrollers or ASICs, and additionally, the more advanced ones require complicated hardware applications. Therefore, if the budget is limited, other solutions may be more economical.
FPGA systems may be less energy-efficient compared to ASICs, which are optimized for specific applications. In applications where low energy consumption is crucial, it is better to choose a microcontroller or ASIC.
Programming FPGA systems requires knowledge of hardware description languages and may be more time-consuming than programming microcontrollers. In projects where a short time to completion is critical, microcontrollers or ASICs may be a better choice. Additionally, in applications where reliability is crucial, ASICs, which are optimized and tested by the manufacturer for specific functionality, may be a better choice.
Can FPGA be reprogrammed?
Typically, the configuration of an FPGA is stored in external memory. The contents of this memory are sent to the FPGA upon power-up. Changing the FPGA configuration requires the ability to change the contents of the external configuration memory, so as long as it is not non-volatile memory, FPGA can be easily reprogrammed. The ability to reprogram (change configuration) FPGA is used, among other things, in machine learning algorithms, where the FPGA configuration changes dynamically during the application’s operation.
Do FPGA systems find applications in Industry 4.0?
FPGA systems play a significant role in Industry 4.0 technology. Due to their high performance and reconfigurability, they are well-suited for implementing SCADA systems and PLC controllers with reconfigurability, which is used in machine learning systems. Furthermore, very fast signal analysis allows for precise real-time control of drives and the implementation of vector control algorithms.
The high performance of FPGA facilitates the handling of edge devices: processing and evaluating data, handling communication protocols with strong hardware-based encryption, integration with cloud devices and IoT systems. The advantage of using FPGA in these applications is the possibility of reconfiguration after deployment, allowing for the application of new encryption algorithms or updates to those used without the need to replace hardware.
The advantages of FPGA will also be useful in predictive maintenance involving real-time data analysis from sensors. The ability to analyze signals in real-time is useful for detecting anomalies and quickly responding to failures. Subsequently, FPGA allows for integration with AI systems and the implementation of machine learning algorithms.
Regardless of the changes associated with Industry 4.0, FPGA systems have long been used in devices for quality assessment and robotics. These are most often vision sensors operating in the visible spectrum or others, performing analysis and evaluation of images. They find wide applications in quality inspection, object recognition, positioning of grippers in automation and industrial robots. In the latest applications, these sensors are coupled with AI algorithms, allowing for even greater precision and speed of operation.
Examples:
- The ORCA-Quest vision sensor from Hamamatsu uses Intel® Arria 10 GX FPGA to process data at a speed of 18Gb/s.
- FPGA support object recognition algorithms, e.g., defect detection, shape identification, color analysis, providing integration with AI and ML through platforms like Lattice sensAI or Intel OpenVINO.
- The ability to change configuration allows for changing filtering, segmentation, or edge detection algorithms without replacing hardware. Examples may include filters: Sobel, Prewitt, Laplacian, Gaussian – used for sharpening, smoothing, and edge detection.
- The speed of FPGA operation and the efficiency of image evaluation algorithms find wide applications in security systems, ensuring the recognition of threats (fire, flooding, malfunction), people, objects, vehicles, and their license plates. In these applications, FPGA systems can act as local image processors, reducing network load and delays in responding to specific situations.
What types of memory are equipped in FPGA?
FPGA systems have various types of memory, both built-in and external, that can be utilized in applications. These include:
Built-in memories (on-chip):
- RAM (Block RAM / BRAM) – fast volatile memory, used for example as a data buffer, FIFO, cache. The configuration of the FPGA read from external memory is also stored in RAM.
- ROM – read-only memory, often used to store constant data or LUT tables.
- Distributed RAM – smaller RAM units distributed in logic blocks (CLB), used in simple structures.
- FIFO (First-In, First-Out) – a special memory structure for buffering data.
- Shift Registers (SRL) – shift registers usually with a capacity of several to several dozen bits, useful in many applications.
External memories (off-chip):
- SDRAM, DDR – dynamic RAM with high capacity and bandwidth used in signal conditioning applications, image processing, and others.
- SSRAM (Synchronous Static RAM) – fast, synchronous static memory.
- Flash / EEPROM – non-volatile memories used for data storage.
- SRAM, Flash, EEPROM, EPROM and other electrically programmable memories for storing FPGA configurations.
Digital Signal Processing requires fast access to data. Here, Block RAM (BRAM) will be used as an input/output data buffer, Distributed RAM for implementing filters and DSP functions, FIFO for synchronizing and transmitting data between DSP modules, External DDRx RAM in computations requiring higher bandwidth.
In industrial control and automation, a small amount of data is usually transmitted, but increased security of their integrity is required. In this field, the following types of memory will be used: ROM / EEPROM for storing parameters and device configurations, SRAM for storing configurations and variables, Shift Registers / FIFO for buffering data in communication interfaces, e.g., UART, SPI.
Image processing is an application that requires storing large amounts of data while ensuring fast access and ease of modification. Here, dynamic memories (RAM) are mainly used: External DDRx for buffering video frames, BRAM for local operations and LUT tables, FIFO + DMA to ensure efficient data transmission between modules.
Transfer Multisort Elektronik (TME) is one of the world’s largest global distributors of electronic components, electrotechnical parts, workshop equipment, and industrial automation. The catalog includes over 1,500,000 products from 1,300 leading manufacturers. TME’s modern logistics centers in Łódź and Rzgów (Poland), with a combined area of over 40,000 m², ship nearly 6,000 packages daily to customers in more than 150 countries.
TME also invests in the development of knowledge and skills of young engineers and electronics enthusiasts through the TME Education project, and supports the tech community by organizing the TechMasterEvent series, promoting innovation and experience exchange.
