Editor's Note: When a standard LCD controller is not suitable for a particular display screen or application, OEMs must choose between custom designing their own, or hiring a company that specializes in building custom LCD controllers, such as Digital View. In this article, Digital View explains how to approach a custom LCD controller project, and why even experienced engineers sometimes prefer to outsource these designs.
For more than thirty years, major companies have been making enormous investments into development and manufacturing of advanced LCD panels. The quality, reliability, resolutions and size have continued to increase every year, while the costs have dramatically decreased. All of these issues — coupled with the benefits of low-power consumption, small physical size and the general “sexiness” of the LCD — make it very desirable to OEMs and integrators to incorporate into a wide range of products.
All LCD based products, with the exception of some embedded systems, will require a controller to provide a suitable interface and very often a range of display functions critical to a project. As this is a key part of a display system specialist companies have emerged supplying controllers thus providing companies with the opportunity to make or buy, the choice is available. To adequately make this decision, there are quite a number of issues to consider:
Decisions to be made
Among the initial decisions to be made is to define what the controller is expected to accomplish and under what conditions it has to perform – other constraints like the PCB footprint and costs will come later in the project. Some of the primary decisions to be made will include what is the maximum resolution to be displayed and what input interfaces will be needed. The supported input resolution is not necessarily limited to the displayed resolution of the LCD, as current display controllers usually support up-scaling of lower resolution data to match the LCD, and some can even support down-scaling where complex algorithms are used to reduce the resolution of the data to fit the available display format of the LCD. Supported interfaces may cover a wide gamut of choices, including support of industry-standard ARGB interfaces for PC inputs, composite and S-Video connections for low-level video support, higher definition video via analog or digital component interfaces and Digital Video via DVI interfaces.
Fig. 1 shows the major components that comprise the first 80% of the steps involved in a flat panel controller board design.
Selecting a Controller
With some idea of the basic criteria in mind, the next element to be considered is selecting a suitable “engine” to provide the intelligence to accomplish all of these tasks. Controller engines are typically high density single-chip single-purpose devices built by one of a small number of specialized developers. Primary names in the industry include companies like Genesis, Pixelworks and ST Micro. These companies each sell one or more specialized controller chips designed to be the heart of the analog controller design. Controller components vary considerably by features, flexibility, complexity, and of course cost. Considerations in making a selection must include not only reviewing the performance and features of the chip but also looking at the development environment and the support available. Many of the vendors offer a minimal reference design to speed the hardware development and a basic software development toolkit. Development toolkits typically provide basic boot-up code and a utilities package suitable to get a basic display on the screen, but considerable development will be needed to mature this into a complete product with appropriate features and stability. A final consideration here is development support. These components are developed by manufacturers envisioning sales to set-top box manufacturers, high volume desktop monitor manufacturers and the like “so, they may not be too interested in providing support for someone building a few hundred products a year. In some cases, manufacturers impose minimum purchase requirements before any support is available.
With the chosen chip set and a definition of the input ports and desired display resolutions, there is still much more work to be done. Depending on the inputs being supported, the design will need an appropriate analog-to- digital converter with performance appropriate to the application to manage the inbound ARGB signal, DVI receivers and other decoders. Decoders and receivers mean input ports to the controller, and with that come the liabilities of ESD and EMI, plus the necessity of providing protection against the ports being connected incorrectly (over voltage, reverse polarity protection etc.)
Putting It on the Panel
Having provided ports to get data into the controller, the next step is to get data out of the controller an onto the LCD panel. Big decisions now have to be made regarding the products planned for use ” now and in the future–because of the volatility of the LCD panel industry. LCD panels have very short life-spans. Because of the rapid advancements in the manufacturing technologies, LCD companies are constantly improving resolutions and viewing characteristics and interfaces. As a result, it is virtually pointless to develop an LCD controller without planning to support more than one panel as insurance against the primary choice becoming obsolete.
Supporting more than one panel now forces the design to adapt to provide switches or other mechanisms to provide selection mechanisms for alternate timing required by different panels. Selection mechanisms of any sort have now added a new requirement of some additional intelligence. It's necessary to add a microcontroller and some code to run it to make these choices possible, plus, of course, some displayed menus to allow the user to configure the system.
The panel interface is the next issue to be dealt with. Will the panel be a TTL or an LVDS interface, will it be 6- bit or 8- bit color, is it dual or single pixel, and what type of connectors are needed. There are a lot of choices to be analyzed and addressed.
Solving the Multi-Panel Puzzle
Once these issues are defined, one of the really difficult pieces of work can begin, this being solving the “multi-panel” and “multi-input” puzzle. Each LCD panel model, even if it has a similar electrical interface, will likely have a different set of timing needs to create a stable picture. To further complicate this, data received on different ports — i.e. DVI versus Video — may need to have the timing adjusted in each mode in order to display successfully. Supporting more than one panel complicates this significantly and building a general- purpose controller that can support a wide range of panel resolutions and input ports becomes extremely difficult.
With all the major parts in place, there are still a great number of issues to be addressed to ensure a reliable and effective product. Reliability is a key feature that has to be designed into the product and cannot be added later. Following good design practices and performing detailed loading analysis at each stage of the hardware design and then finally completing a detailed design proof can ensure operation with adequate and predictable margins. With a stable hardware design in place, an MTBF (Mean Time Between Failure) analysis can be completed using software tools to provide an indication of the predicted product reliability.
For a truly flexible LCD controller design, we can anticipate the need for an extremely flexible, robust and accurate power supply. LCDs require stable power supplies as any noise or ripple is immediately communicated through the display with disastrous effects. An appropriate design will accommodate a wide range of input voltages and a substantial noise immunity since there will probably be some noisy devices, like an inverter with a variable duty cycle operating on the same rail.
View Figure 2 full size
Fig. 2 shows the additional steps that must be taken to address the remaining 20% of the design, which can make or break the performance of the controller. This addresses issues like ESD and EMI, power conditioning, and termination circuitry between the LVDS transmitter and receiver.
There are many more considerations to be made in the design of an LCD controller before it can be become a viable product. These include difficult areas like EMI, RFI performance, ESD protection, operating temperature etc. and of course keeping the product compliant with new regulations like RoHS.
Is There a More Practical Alternative?
So, a more practical alternative to designing your own controller board may be to source a controller, such as the Digital View SVH-1920 (see photo), that standardizes on a single platform and supports multiple panels. Typically, a single project volume does not justify a unique design and production run; it is better to leverage off existing product appropriate to your application. And consider development support: if the volumes are low or if your company does not have a good established relationship with the component manufacturers, then support may be difficult to obtain.
And consider time to market. Sourcing from a company that specializes in flat panel controller boards with a fault-tolerant architecture and highly flexible generic design that supports multiple panels means that your solution can be available right away. Lead times can be matched to your needs.
And, of course, consider price. For a specialized, low-volume, niche application, you are not going to realize any economies of scale designing your own controller anyway.
Also, the right controller specialist offers a broad industry knowledge based on years of designing these boards, along with the kind of advanced technical expertise that results in features like an EEPROM ICs that offers timing sequences for different panels ” multiple timing libraries embedded in the chip allow it to support multiple panels ” or an LVDS transmitter that enables LVDS receivers to be implemented in more than one way on the panel.
The decision is yours. But, designing your own controller board is complicated and could mean a rocky road ahead. It always makes sense, from a performance and price standpoint, to consider outsourcing it to an expert if it provides a genuinely more practical alternative.
for more information, contact:
Digital View, Inc.
18440 Technology Drive
Morgan Hill, CA USA 95037
408 782 7773