I recently came across the most peculiar way to make a color CRT monitor. More than a few oscilloscopes have found their way on to my bench over the years, but I was particularly struck with a find from eBay. A quick look at the display reveals something a little alien. The sharpness is fantastic: each pixel is a perfect, uniform-colored little dot, a feat unequaled even by today’s best LCDs. The designers seem to have chosen a somewhat odd set of pastels for the UI though, and if you move your head just right, you can catch flashes of pure red, green, and blue. It turns out, this Tektronix TDS-754D sports a very peculiar display technology called NuColor — an evolutionary dead-end that was once touted as a superior alternative to traditional color CRTs.
Join me for a look inside to figure out what’s different from those old, heavy TVs that have gone the way of the dodo.
Electrical engineers depend on their oscilloscopes to see what’s happening inside a circuit. If you’ve ever tried debugging without one, you know what a difficult and frustrating experience it can be. Consequently, users expect these devices to provide accurate representations of electrical waveforms: to be an extension of their eyes. If asked to name the most important part of a scope, what would you say? The attenuator or amplifier? The timebase? The ADCs? The probes? In today’s screens-everywhere world, probably not many would say the display. But, for a generation of engineers brought up on analog scopes, the display was of paramount importance. Hence, the designers wanted the display to be as sharp as possible: any fuzziness should be due to the signal (like my bandwidth-limited wrencher), not the instrument.
Analog oscilloscopes were very sharp, and they were monochrome. This was in keeping with their original function: to plot voltage (y) vs time (t), or in some cases, voltage (y) vs voltage (x). Color doesn’t add much to such a display, although many scopes also had a z-input to modulate the trace intensity. And while dual-beam scopes existed (for displaying separate channels simultaneously), keeping three beams synchronized as they swept across the display to create a real-time color trace would have added a great deal of complexity and cost for little return.
When digital storage scopes arrived, and the display evolved from a simple graph to a full user interface, color took on new importance. Different colors could be used to disambiguate traces, visually link them to automated measurements, and distinguish text and UI elements. Later, with the advent of digital phosphor oscilloscopes (DPOs), which did a fancy digital simulation of a real-time analog scope, color could be used to reveal subtle features of the waveforms themselves. So, it made sense to add color displays.
But, if you were building high-end oscilloscopes in the early 1990s, your choices for color displays were limited: color CRTs and LCDs, and each presented problems in oscilloscope use.
The traditional way to display color images on a CRT was to use three electron beams striking an array of primary color phosphor dots (or in some systems, stripes). Several technologies were developed over the years, but they all share a common mechanism: a perforated mask of some kind is interposed between the cluster of electron guns and the screen, producing pinhole images of the three beams, hopefully with each one hitting the intended phosphor dots. In reality, this didn’t always happen, so blurriness and impure colors were common display artifacts. The use of separate color dots also meant a loss of spatial resolution: the smallest possible pixel was a cluster of red, green, and blue points.
Focusing a single electron beam into a tight dot is not a problem; scanning electron microscopes can have beam widths less than a 1 nm, although admittedly, they scan over a much smaller area relative to their size. The true difficulty is in keeping three beams aligned with each other and with the phosphor mosaic. As a result, monochrome CRTs suffered fewer image quality issues.
Not only was color CRT display quality problematic, but the devices themselves were bulky, heavy, and sensitive to vibration and shock, which could knock them out of alignment. Although some oscilloscopes were made with traditional color CRTs, the technology was far from ideal.
Some competing oscilloscopes at the time were the first to use LCD screens. But, color LCD screens from that era were a far cry from the excellent displays we enjoy today. They suffered from low contrast, poor color quality, and had abysmal viewing angles. These were the early twisted nematic (TN) displays, primitive compared to today’s version, as in-plane-switching (IPS) displays were still a few years off. Even today’s LCDs haven’t managed to fix one of the original problems: each display pixel is still composed of primary color sub-pixels. Although this can occasionally be used to advantage, for instance in sub-pixel anti-aliasing for rendering text, it can also impart a blurriness to individual pixels.
Instead of using either of the these existing technologies, the engineers at Tektronix decided to leverage their extensive experience with monochrome CRTs to create a superior display. They started with a traditional monochrome CRT with a white phosphor, then added an ingenious system of switchable colored filters to create a field-sequential color display. In this system, instead of the primary color components being distributed in space, pointillist style, they are distributed in time, with successive frames showing red, green, and blue components of an image. To the eye, the result is the same: primary color components for each pixel get blended into a perception of different colors. The marketing department over at Tektronix dubbed the technology “NuColor.”
The biggest advantage to this system over color CRTs, and even modern LCDs, is immediately apparent: each pixel remains the same tiny, single dot as rendered by the monochrome CRT, but now appearing in your choice of color. Tektronix used a 180 Hz frame rate on these displays, so the full RGB display was refreshed every 60 Hz. The downside is an occasional glimpse of the individual color frames if you look away from the display quickly. This effect can also be captured with a high-speed camera.
Compared to color LCDs of the day, the other key advantages were brightness and contrast. The NuColor display had an enormous dynamic range driven by the monochrome CRT at its heart, putting the very poor contrast of contemporary LCDs to shame.
To make this new display, Tektronix placed an electronically-controlled color shutter system in front of the CRT face. These color shutters were created using a type of liquid crystal technology, but instead of switching individual pixels, they switched the entire display at once, like today’s active 3D shutter glasses or auto-darkening welding helmets. In these LCD applications, light is passed through a neutral polarizing filter, creating polarized light. This light then traverses a liquid crystal cell which can rotate the polarization of light passing through, depending on applied voltage. Finally, the light exits through a front neutral polarizing filter. Depending on the amount of rotation that the liquid crystal cell applies, a varying amount of light is transmitted through the front polarizer.
In the NuColor display, instead of using neutral polarizing filters, Tektronix used color-selective polarizers. For instance, the first polarizer in their full-color system passed all three red, green, and blue colors of vertically polarized light, while allowing only green in the horizontal polarization. By combining three such polarizers, each with a different single-axis color, they were able to create a system that could selectively pass red, green, blue, or white light. Like DLP projectors which use a red/green/blue/white color wheel to produce brighter images than red/green/blue alone, these displays could have offered an expanded dynamic range, although this technique appears not to have been used on oscilloscopes.
Tektronix was granted at least 10 US patents on this technology. These ranged from a system that could display only red and green primaries (US Pat. 4,582,396), to one that could display three de-saturated primaries resulting in an all-pastel colored output (US Pat.4,674,841), to a full-color display with an expanded color gamut (US Pat. 4,635,051). In the patents and subsequent media coverage from the early 1990s, they touted the advantages of the new display technology in better resolution, lower power consumption, and smaller size and weight. They also estimated it would add only 2.5% to the cost of an oscilloscope, as opposed to 12.5% for a traditional color CRT.
Of course, today’s devices don’t use field-sequential color CRTs. The technology behind the NuColor display just couldn’t compete with the size and weight advantages of LCDs. Once LCD quality evolved far enough, these displays came to dominate the digital oscilloscope market just as they did nearly every other screen. Although, with OLEDs on target to outpace LCDs in smartphones this year, that may be changing.
Obviously, the NuColor displays are not being produced any longer, so some enterprising types have started providing LCD replacements for the aging CRTs in these scopes. Luckily, these scopes also sport an analog VGA output port for an external monitor.
By the way, should you happen to find yourself in possession of a 500MHz TDS-754D like this one, note that it’s field-upgradeable to a 1 GHz TDS-784D by changing a few resistors and removing a few capacitors. If you dig through the thread on the eevblog, you can figure out how. It works, although it’s nowhere near as easy as software-only upgrades like we covered for the Rigol DS1022C, Rigol DS1052E, Rigol DS1054Z, and Rigol MSO5000.
One last note. I used the osmo-fl2k code we featured back in April to render our favorite logo on the scope: originally an RF hack that turns certain cheap USB-to-VGA dongles into SDR transmitters, it also comes in pretty handy around the lab as a $5 3-channel arbitrary waveform generator.
Sharpest Color CRT Display is Monochrome Plus a Trick
Source: HackADay
