Reuse and Cost Savings in an Airborne Display Controller Application
How FPGAs and a custom designed controller board made it possible to successfully replace a legacy CRT with an LCD controller while adding advantages over the product life cycle.
Deployed in harsh environments, military displays have rigorous operating and non-operating standards, which apply as well to circuit boards and other elements. While it’s relatively easy to procure power supplies and single board computers that meet all performance requirements, the task of finding commercial off-the-shelf LCD controllers, which typically receive the external video signal, perform video processing, and drive the LCD panel is more difficult.
Only a limited number of controllers will meet the wide range of shock, vibration, temperature, humidity and altitude requirements. What’s more, COTS controllers typically don’t include special video processing functions such as programmable line width and anti-aliasing. Controllers that accommodate less common signals may not be available at all. One example is a display that receives both raster and stroke signals multiplexed onto the same signal path as illustrated in Figure 1.
A raster only interface could take advantage of a COTS LCD Controller that automatically handles many video interface standards, but the Stroke and combined Raster/Stroke Modes create additional requirements that are not as simple to meet as synchronizing and digitizing scan lines. These include Stroke/Raster registration for accurate symbol placement, a black border around the stroked graphic items for easier identification, and the aforementioned line width control and anti-aliasing.
These constraints call for a custom video conversion and control function replacement for the analog CRT digital counterparts. IEE faced just such a challenge in designing a new 6×6 Multi-Function Display (MFD) for a CRT replacement. (See Figure 2.)
Benefits of Basing a Video Processor on an FPGA
Power, flexibility, and processing performance which compare favorably to a software-based approach are among the advantages of selecting an FPGA-based solution for custom video processing and control applications. The wide array of available Intellectual Property (IP) cores is another plus. Additional functions must be included in the design to accomplish a fully featured LCD Controller, such as a Display Data Channel (DDC) interface and an on-screen display menu, which are general-purpose display controls.
This design challenge centers around a special application, with a heavy emphasis on the raster/stroke conversion.
Raster/Stroke Video Conversion
The following describes the basic operation as it applies to a Raster/Stroke Interface Mode. Figure 3 shows a block diagram for this conversion.
The Stroke Video Processing receives the X & Y Deflection and Stroke Intensity signals that have been sampled to 14 bits at 60 MHz. The resulting data is processed for Gain, Filtering, Oversampling, Gaussian spot processing and Scaling.
The Raster Video Processing receives and identifies RS-170 and RS-343 signals that have been sampled to 14 bits locked to the Horizontal Drive signal. The resulting data is processed for Gain, Gamma Correction, and scaling.
The converted stroke and raster data is flattened to a single layer, processed for phosphor decay, and output as 24-bit RGB. Note that many CRT replacements are required to emulate the phosphor performance of the legacy display. This allows the replacement to be truly backward compatible while benefiting from high-resolution AMLCD technology.
Built in tests include SRAM and SDRAM pattern tests, Deflection, Stroke Intensity, and Raster A/D interface test, LCD I/F status check, drive signal period checks, and a Visual BIT display pattern.
While this signal type is unique for this particular application, the FPGA configuration has many elements that can carry forward to other designs.
A custom LCD Controller allows the FPGA and support components to be sized to provide just the necessary functions—an approach that makes reducing size, weight, power, and cost (SWAP-C) possible. Adopting this approach can also strengthen cyber security: The pedigree of a major logic-bearing component, including the fabrication factory location, get preserved. The hardware can be placed and routed on a circuit board outline that can better meet the requirements of mounting, cooling, harness routing, and maintainability. In the longer term, this can reduce costs and increase reliability by eliminating the need for the LCD Controller to be a separate circuit board, as the video processing and LCD control can be co-located with the microcontroller, which handles basic functions such as backlight intensity and bezel button actuation.
The resulting in-house development of the LCD Controller function provides a point of departure for more flexible and complete solutions in future products at only a delta cost. Features can be added and removed as necessary. In this example, common digital video inputs can be included as build options, and a DDC interface can be used to automatically configure the output characteristics to drive different LCD panels. FPGAs can be selected at different price points that match higher-speed interfaces or more complex functions as well as for lower-cost reduced feature sets used in other highly specific applications.
IP cores such as microprocessors and DSPs take capabilities far beyond those of traditional FPGA logic and can replace an entire single board computer in some cases. As the library of video processing functions grows, the value of the IP is repeatedly realized. As newer and more capable FPGAs are introduced, the existing code can be ported to take advantage of speed, power or cost improvements.
Extension Rather than Total Replacement
CRT form, fit, and function interface (F3I) LCD replacement is expected to continue due to ongoing military budget constraints that result in service life extension rather than total replacement. When the development of an in-house LCD Controller is called for, it’s possible to realize the benefits of reuse while adding to the organization’s intellectual property portfolio.
FPGA video processing forms the basis for legacy CRT replacement and provides the IP for low-cost variants. Coupled with the appropriate front end (analog, DVI, etc.), this approach can result in a full featured, integrated video processing capability embedded within existing display hardware without the need for a separate LCD controller. The ensuing design can address future customized applications and offers a range of cost and life cycle benefits.
Raster/Stroke Signal Characteristics
Most legacy CRTs use either a raster or stroke video input, and some can switch between modes, or even use both at once. Most engineers are familiar with these increasingly rare signals.
Monochrome raster video uses a constantly moving electron beam for which the X-Y target location on the display surface is always known. The beam moves horizontally from one edge to the other (raster line) and then moves quickly back to the next starting position (horizontal retrace). When the beam gets to the last position, it moves quickly back to the upper starting position (vertical retrace). Depending on the system the beam either paints every line or every other line. During the raster line periods the intensity of the beam determines the brightness, and in many cases, color, of the display at that point. Regardless of the intensity, starting location, raster and retrace timing, and interlaced or progressive scanning, the beam is always moving.
Monochrome stroke video does not use the horizontal and vertical line and retrace signals, but instead controls the X and Y location, and beam intensity, with separate signals. This allows the control circuit to essentially write to any part of the display at any time. Common examples include oscilloscopes and radar/IFF Plan Position Indicators (PPI).
The vertical refresh interval of the raster signal is relatively long, and some legacy displays take advantage of part of the vertical retrace blanking time to allow stroke signals to control the electron beam. This hybrid approach allows the use of stroke graphics overlaid on the raster picture and is typically used for amplifying information overlays such as symbols and text.
Figure 3 shows the notional timing diagram for a Raster/Stroke mode. In this example the X and Y deflection, gated by the Stroke Intensity, paint a box shaped object.
John Rodwig is a Director of Program Management at Industrial Electronic Engineers, Inc. http://ieeinc.com/ He has over 30 years of technical and management experience in Defense and Telecommunications, and co-holds a patent for a Radar Scan Converter. Rodwig earned a BS in Electrical Engineering from Tulane University and an MS in Engineering Management from California State University, Northridge. Rodwig can be reached at mailto:email@example.com
This article was edited from its original version that appeared in Electronic Component News (ECN), January 2017. Advantage Business Media.