- AS-MCUs deliver cost advantages compared to SoC- and MCU-based designs.
- S6E2D GPU offloads processing tasks like image manipulation, 2-D rendering, and text rendering.
- Customized for white-goods applications, GPU cuts memory requirements by performing many functions like blending and decompression on the fly.
AS-MCUs bring TFT HMIs to cost-sensitive home appliances
High-end home appliances are making their way into the Internet of Things (IoT). Ovens that offer recipe downloads have been on the market for several years. Some small household appliances with such functionalities are also available and ramping up in terms of volume. These high-end appliances may perform different functions, but they all have something in common: human machine interfaces (HMIs) with high-resolution, full-color thin-film transistor (TFT) screens. These types of HMIs are also of increasing interest for mid-range home appliances that may not be connected to the Internet by default. Marketing departments at white goods manufacturers see the HMI as one major potential differentiator and therefore engineers are working to also bring full-color TFT HMIs into mid-end washing machines, for example. This can be effective, but designers face challenges in achieving the desired functionality while keeping the system cost in an acceptable range.
The SoC approach for high-end appliances
High-end connected home appliances with TFT HMIs are usually based on a system-on-chip (SoC) design with enough horsepower to run the communications as well as the interface. Often, those SoCs run an ARM® Cortex®-A-class CPU with at least 400 MHz; they generally do not feature internal flash memory and only have a restricted amount of internal RAM (see figure 1). Engineers, therefore, have to add external RAM, generally DDR/mobile DDR, and external flash, and typically connected via QSPI.
Figure 1: System-on-chip (SoC) design delivers high resolution, high-performance graphics, but requires external memory and long boot times.
At system boot, the main program is copied from the external flash to the RAM and then executed from there. This means that there is a recognizable boot time inherent in such systems, particularly with interfaces like QSPI. HyperBus and HyperFlash technology, developed by Spansion (now Cypress), operate at 333 MB/s, which can speed boot time significantly.
Linux is very often used as an operating system with all the benefits of available drivers and standard libraries. This includes using hardware blocks such as a graphics processing unit (GPU) offered by the SoC device to drive the TFT display with complex graphics effects.
It would be easy to just take such systems into mid-range home appliances; the biggest reason not to do so is pure cost. SoC-based systems generally have several devices with ball-grid array (BGA) packages. This, combined with the necessary high-frequency bus to connect the DDR RAM, increases the need to go for an expensive PCB with many layers and micro-vias.
The MCU approach for mid- to low-end appliances
Home appliances are typically built in a modular way, meaning that the units to drive the motors for pumps and drums, the machine control and the HMI are distinct modules communicating over a machine internal bus system. A more economical approach to build an HMI module is to use a microcontroller (MCU). Such MCUs are frequently based on ARM® Cortex® M-class CPUs. They typically include embedded flash memory and an amount of RAM that is tailored to drive the application, but not to store the video data intended for display on the TFT screen. External RAM (VRAM), usually SDRAM, has to be added to such systems (see figure 2), therefore.
Figure 2: Simplified block diagram of an MCU-based system for mid/low-end TFT HMI modules for home appliances shows the reduction in external components.
The internal flash memory of the MCU can in some cases be enough to hold the firmware to drive the TFT HMI module as well as the bitmap data to be displayed on the TFT screen; additional external flash connected by QSPI is necessary in the case of more complex graphics content or the case in which several languages are pre-programmed into the system. There are MCUs on the market offering special direct memory access (DMA) hardware in order to speed up the data transfer to the TFT display.
From the cost point of view, the PCB can be made more cheaply because the frequencies of the bus to connect the SDRAM are lower and no BGA packages are present in the system; therefore no micro-vias are necessary in the PCB. On the down side, there is still the need for external RAM (in this case SDRAM with a parallel bus on the PCB) and the resolution of the TFT screen is restricted by the amount of data the MCU is able to shuffle into the VRAM.
The approach to firmware development is very much different compared to that of the SoC-based system. No operating system is mandatory in the MCU case; there are graphics libraries on the market that are tailored to allow full-color TFT HMI development on low-performance systems. Of course, there is a limit. A screen resolution of 480 x 272 (WQVGA) with 16-bit color is currently the sweet spot in terms of cost/performance of the displays and system cost for mid/low-end TFT HMI modules for home appliances.
Still, the external RAM does require a multilayer PCB, which adds a certain cost premium. The architecture also limits graphics performance—all blending effects have to be done in software and calculated on the CPU of the MCU. These factors limit the penetration of MCU-based TFT HMIs into the mid/low-end home appliance market. To address these problems, makers should consider application-specific MCUs (AS-MCUs).
Optimized AS-MCUs for mid- to low-end appliances
AS-MCUs combine SoC and MCU features into a single device, always aiming to get sufficient graphics performance combined with further system cost reduction, compared to standard MCU-based systems. One approach to reach this goal is to start with a standard MCU based on an ARM® Cortex®-M4F CPU with embedded flash and RAM to drive the application, plus a tailored GPU and embedded VRAM (see figure 3).
Figure 3: The S6E2D series of AS-MCUs from Cypress integrates a standard MCU with a GPU and VRAM.
This combination further reduces the PCB cost because the absence of external RAM with its parallel bus requirement enables the usage of standard dual-layer PCBs (see figure 4).
Figure 4: Simplified block diagram of an AS-MCU-based system for mid/low-end TFT HMI modules compared to a conventional MCU shows the number of integrated components.
The trick of a system based on such an AS-MCU is that it is optimized in terms of cost while offering superior graphics performance compared to a standard MCU-based system. The embedded GPU can offload operations from the main CPU, including calculating blending effects and running other graphics algorithms. In addition, the GPU in AS-MCU features special functions to reduce memory requirements.
To support designers of TFT-based HMIs for mid-class white goods, Cypress has developed the S6E2D series of AS-MCUs. This series offers the choice of embedding the graphics subsystem with GPU and 512 kB of VRAM into a capable ARM® Cortex®-M4F-based MCU. Combined with the internal VRAM, this enables makers to build TFT HMIs based on a single-chip, 120-pin LQFP package AS-MCU with only one external flash memory module connected by QSPI (or, alternatively, HyperFlash) and no additional components.
Reducing the load on the main CPU
GPUs can offload operations from the main CPU of an AS-MCU by performing functions in hardware on the graphics subsystem. Examples of such functions integrated into the GPU of Cypress’ S6E2D series include:
- Image manipulation: move, mirror, rotate, zoom/scale, overlap, alpha-blend
- 2-D rendering: draw line, draw rectangle, fill rectangle
- Text rendering: display, change size and color, anti-alias pre-rendered text
Advanced GPUs can reduce the memory needs of a system with specific functions; Cypress’s S6E2D Series features the following:
1) On-the-fly rendering functions executed by the GPU, with the results directly shown on the TFT display without the need to store them in VRAM
a. Blending of foreground and background without storing in VRAM
b. Alpha blending of images/texts with background picture and background color
c. Sequences of image manipulation and blending with background
d. Support of indirect color formats enabling the storage of every graphic element with an individually optimized color palette
2) On-the-fly decompression of pictures (bitmaps) stored in external flash without the need to store the result in VRAM, reducing both external flash memory requirements and VRAM usage.
A graphics TFT HMI system based on an AS-MCU such as the S6E2D Series can enhance graphics performance and therefore increase the value perceived by the end customer. At the same time, it enables manufacturers to reduce appliance price when compared to a standard MCU- or SoC-based system. Optimized for WQVGA display resolutions with 16-bit color, the S6E2D series provides a method to speed the development of a complete system at minimum cost. A version with external SDRAM bus in a 176-pin package is available for higher screen resolutions or growing VRAM needs. With this technology, makers can bring high-end graphics to low- and mid-level white goods for a more budget-conscious market.
Also in this issue:
Also in this issue
Conformal electronics differentiate products by improving the appearance, functionality and overall user experience.
Failure analysis methodologies like root cause corrective action give organizations tools to uncover problems and improve processes.
Also, troubleshooting flowcharts for parallel NOR and dongle detection details
Whether monitoring water levels in coffee makers or implementing smart refrigerators that can warn when that milk bottle is nearly empty, liquid-level measurement enhances performance and differentiates products from the competition.
Pre-certified modules let developers skip the byzantine process of Bluetooth qualification and focus on building better products.
Safe firmware and test routines let developers focus on their product features
Specially configured for home appliances, programmable system-on-chip (PSoC) controller simplifies the addition of functionality like displays and touch sensing for better product differentiation.
Analysis of granular operational data helps developers more easily fix the flaws in their systems