Industry Article

The Microcomputer Compensated Crystal Oscillator Is Finally Ready for Space

The latest advances in radiation-hardened digital components have enabled the development of MCXOs that, for the first time, can replace larger, high-power-consuming OCXOs in low-earth orbit (LEO) New Space applications.

Since the 1990s, the microcomputer compensated crystal oscillator, or MCXO, has been used in many applications, including military and commercial avionics, ground-based electronics, as well as undersea oil exploration. These smaller, lighter, lower-power devices can often replace bulkier and power-consuming oven-controlled crystal oscillators (OCXOs), while also providing comparable stability over a wide range of operating temperatures. But the holy grail of MCXO applications is for use in space. Almost every satellite employs a least one OCXO for precision timing, despite the big drawbacks of their high power consumption and large size. The obstacle for the MCXO is that it uses several digital devices that have been difficult to procure as space-qualified, rad-hard components. Until now.

This article compares crystal oscillator types and introduces the first MCXO that combines clever engineering design with rad-hardened digital components to qualify for NewSpace applications (Figure 1).

 

Example applications within the NewSpace realm.

Figure 1. Example applications within the NewSpace realm. Image used courtesy of NASA

 

Crystal Oscillators

All crystal oscillators, shown in Figure 2, are based on the very stable frequency vibrations of a piezoelectric quartz crystal resonator. 

 

Example diagram of a crystal oscillator.

Figure 2. Example diagram of a crystal oscillator

 

Usually, the crystals and their associated circuitry are carefully designed and crafted so that the quartz crystal will vibrate only at the one desired resonant frequency. A stand-alone crystal oscillator can hold frequency stability of less than ±50 ppm over the wide military temperature range of -55 to +125℃, which is good enough for most electronics applications. 

If a more tightly controlled stability over temperature is needed, a temperature-compensated crystal oscillator, or TCXO, adds compensating circuitry to correct for the crystal frequency’s temperature variation and can thus achieve about ±1 PPM. 

If even more stability is needed, an oven-controlled crystal oscillator uses the technique to put the crystal inside a very precise proportionally controlled oven, which can achieve about three orders of magnitude better frequency stability over temperature; however, the OCXO comes at the cost of a lot more size, weight, and power consumption. A typical OCXO draws at least a few watts of power, while the power consumption of XOs (simple crystal oscillators) and TCXOs is measured in milliwatts. Also, OCXOs typically have higher performance for other important oscillator parameters, including phase noise, jitter, and long-term stability (aging).

 

The Benefits of MCXOs

The driving purpose of the MCXO is to achieve the performance of the OCXO but with much lower power consumption and much faster warm-up (the time it takes an oscillator to reach its required stability after turn-on). The deceptively simple method the MCXO uses to do this is to operate the quartz crystal resonator at two different frequencies at the same time. 

By doing this and by manipulating the data thus created, the MCXO crystal becomes a self-sensing thermometer; that is, the crystal essentially tells us exactly what its temperature is at any given moment and to a very high degree of precision and thereby allowing the frequency to be more precisely compensated for than in a TCXO. It also uses much less power than an OCXO. 

One of the primary reasons for the superiority of the MCXO’s temperature compensation is that the self-thermometry of the quartz crystal resonator eliminates the need for a separate thermometer. 

Every TCXO and OCXO requires a separate temperature sensor to precisely monitor the temperature of the quartz crystal resonator. In the case of the OCXO, one must know the crystal temperature in order to continually correct that temperature to the desired oven temperature. In the case of the TCXO, knowing the crystal’s temperature allows the compensation circuitry to calculate the exact correction needed due to frequency-temperature variations. The difficulty is that the temperature sensor can not be mounted on the actual crystal resonator because of mass loading and contamination effects, but instead must be mounted on the outside of the crystal’s hermetically sealed package and because of the thermal time lag, the thermometer will never actually be at the exact crystal resonator’s temperature. 

The MCXO does away with this problem because the crystal reports its own actual temperature in real-time. How does the crystal do this? By making the MCXO crystal vibrate at two different frequencies at the same time. Every piezoelectric crystal can oscillate in many different modes, each with its own frequency.

 A key point of a crystal’s design is to make it prefer to oscillate in one particular mode. But with the MCXO crystal, the crystal is designed to oscillate in two modes at the same time, one being the fundamental mode of an SC cut crystal—a very special doubly rotated cut with respect to the hexagonal quartz crystal axes that give the crystals excellent temperature stability. The second mode of vibration is on the crystal’s third overtone. The crystals used in precision oscillators can vibrate on a fundamental mode, where the frequency is proportional to the thickness of the quartz blank, or on any odd-numbered overtone. 

In this case, the third harmonic overtone is used, but the third overtone frequency is not exactly three times that of the fundamental mode, but a very close 2.999. This ratio actually varies minutely with temperature, and this ratio of the fundamental mode frequency to the third overtone frequency is the most precise indicator of the exact temperature of the crystal at any moment. All of this is very carefully characterized and stored away for each MCXO crystal and is then used in real-time to calculate the exact temperature based on the ratio of the two frequencies at any moment. 

The result is that the MCXO can be made to give approximately the same performance as a good OCXO but with power less than 100 milliwatts compared to an OCXO’s three to five watts. The typical warm-up time of an OCXO is 10-plus minutes versus less than one minute for the MCXO. In other words, the MCXO can offer more than a full order of magnitude lower power consumption and faster warm-up time than the OCXO. This, for some applications, is revolutionary. 

 

Developing Space-Qualified MCXOs

When developing MCXOs in the early 2000s, the space- and rad hard-level digital components needed were very expensive, which meant an MCXO space-level product would sell for hundreds of thousands of dollars each. 

At the dawn of the era of mega satellite constellations, known as low-earth orbit (LEO), or New Space, it became possible to find microcontrollers and other digital devices that were rad tolerant and up-screenable. The use of these digital components went into the QT2020 MCXO (Figure 3), released in 2021, which has now been fully qualified for use in LEO New Space applications. 

 

The small, light (21” x 1.33” x 1.33”) low-power-consumption QT2020 MCXO.

Figure 3. The small, light (21” x 1.33” x 1.33”) low-power-consumption QT2020 MCXO.

 

The QT2020 MCXO was designed with the objective of use in satellites and other space applications, using only rad-tolerant components. The product series is available at 10, 20, 30, 40, 50, 60, or 80 MHz, with stability as low as ±10 PPB in a 2-inch by 1-inch by 1.33-inch package. And it offers the high performance of an OCXO but with less than 90mW power consumption.

The QT2020 MCXO is now a standard product that can be procured without difficulty and at a reasonable cost. Prices vary based on the stability and other options. For instance, a full, totally
RAD hard version can be developed if the application will support a higher price. 

The QT2020 MCXO has been tested for TID up to 50 kRADs without experiencing problems, and the current consumption level was “rock steady” as the radiation dosage went up – giving optimism that single-event results will be good. Now, single-event testing is being arranged.  

Figure 4 shows a simplified block diagram of the QT2020 MCXO. The signals from the dual mode oscillator get mixed down to generate beat frequency after being normalized by a frequency divider. The beat frequency is a difference between the two oscillator modes and represents the crystal temperature. It feeds the microcontroller counter to generate a digital temperature reading “N1.” Data for N1 is collected and stored in microcontroller memory. For each N1, a polynomial calculation provides a correction coefficient “N2.” A 10MHz VCXO provides the signal to one of the microcontroller counters to get compared to the Fo signal. N2 correction is applied here. A digital-to-analog converter applies a control voltage to the VCXO to keep it at the target frequency. 

 

Simplified block diagram of the QT2020 MCXO.

Figure 4. Simplified block diagram of the QT2020 MCXO

 

The chart in Figure 5 shows how the dual-mode SC cut crystal functions as a self-sensing thermometer. The dashed curved line is the frequency vs temperature curve for the crystal’s fundamental mode, and the continuous curved line is the frequency vs temperature curve for the crystal’s third overtone. When the oscillator is measured at any given temperature, the reading for the third overtone (divided by 10) is divided by the reading for the fundamental, and the resulting ratio is called the Beat Frequency and it falls on the straight line. This shows exactly at what temperature the crystal is at that time since the particular crystal in question has been precisely characterized for how its Beat Frequency varies linearly with temperature.

Major performance characteristics for the QT2020 MCXO show that it is considerably better for size, weight, power consumption, and warm-up time than any OCXO and is considerably better for frequency stability than any TCXO.

 

The QT2020 MCXO’s performance characteristics.

Figure 5. The QT2020 MCXO’s performance characteristics

 

Conclusion

The QT2020 MCXO fills a niche with much better stability than the best TCXO and offers the equivalent stability and noise to typical OCXOs. It also has extremely low power consumption, small size, fast warm-up, and is fully certified and rated up to 50 kRADs TID. Time will tell, but this rad-hardened MCXO promises to be a revolutionary, enabling technology. Already, many satellite manufacturers are placing orders, evaluating the product, and planning to fly the QT2020 MCXO in exciting New Space applications. 

 

All images used courtesy of Q-Tech

Industry Articles are a form of content that allows industry partners to share useful news, messages, and technology with All About Circuits readers in a way editorial content is not well suited to. All Industry Articles are subject to strict editorial guidelines with the intention of offering readers useful news, technical expertise, or stories. The viewpoints and opinions expressed in Industry Articles are those of the partner and not necessarily those of All About Circuits or its writers.

2 Comments
  • paulkarli May 26, 2023

    The atomic frequency standered both cesium and rubidium are used for precision timing and frequency generation. You more than likely know this type of frequency source. Made up of a atomic resonator and crystal oscillater. The oscillater uses a correction signal from Atomic Resonater. This system more than likely is not as small as a MXCO.
      I have used xtal oscillators for a very long while. Thanks for listening to me.

    Like. Reply
  • D
    David Baron May 27, 2023

    Wouldn’t rubidium based oscillators be a better choice for critical space based timing apps? They are not too large, roughly cigarette pack sized, within a decade of cesium based.

    Like. Reply