Understanding Cryocooled Infrared System Reliability

Where the ability to detect mid-wave infrared (MWIR) radiation is mission critical, readiness and the importance of long, maintenance-free infrared (IR) system operation is vital. In turn, cooled MWIR camera modules must be designed, tested, and manufactured to meet rigorous environmental and reliability requirements. This includes military temperature ranges and high shock and vibration levels. Cooled MWIR camera reliability and operational lifetime are typically determined by the operation and lifetime of the cryo-cooler within such systems.

Infrared Camera Module and System Introduction

Cryocooled MWIR camera modules come in a variety of model configurations, each incorporating similar components, but with slight variations to optimize performance versus size, weight, and power. The following components and functions are included in MWIR camera modules and systems.

Focal Plane Array (FPA): The FPA is comprised of a detector hybridized to a read-out integrated circuit (ROIC). The detector converts MWIR photons to an electrical current. The ROIC reads the current and provides an analog voltage or digital signal that is proportional to the number of photons at the detector. Presently, InSb and most HgCdTe detectors operate best at cryogenic temperatures (e.g., 77K). These low temperatures require significant cooling capacity, which translates into a larger cooler size, increased weight, and increased power. Hot MWIR Barrier Infrared (e.g., T2SL) detectors are considered high operating temperature (HOT) at approximately 120K operation and, therefore, require less cooling capacity. These higher temperatures mean smaller cooler size, lower weight, less power required, reduced time to image, and increased cooler lifetime.

Dewar: The Dewar is a vacuum package containing the FPA, cold-shield (defines the f/number and prevents stray light), and the cold-filter (determines the wavelengths of the photons at the detector). The electrical signals between the ROIC and the camera electronics are provided via hermetically sealed feed-throughs. Faster f/numbers increase performance, but also increase the size of the optics.

Cryocooler: Mechanical cryocoolers provide extended duration cooling to the IR FPA. Modern designs integrate directly to the Dewar as part of an Integrated Dewar Cooler Assembly (IDCA). They are also integrated with and controlled by camera electronics, generally through a dedicated cooler control electronics module.

Camera Electronics: The camera electronics may include multiple printed circuit board assemblies (PCBAs) to accomplish the following functionality:

• Sensor interface electronics include power, clocks, and timing logic to the FPA and, if necessary, digitization of the FPA outputs.

• User interface electronics include signal-processing for noise filtering, image enhancement, operational logic, and camera functions, including the overarching command and control interface, and video output standards for the user interface.

• Cooler controller electronics control the cooler operation to cool down and maintain the FPA/Dewar temperature chosen to optimize power and image quality.

• Optics controller electronics control the opto-mechanical lens assembly, maintain continuously athermalized focus over zoom, and support the user interface.

Continuous Zoom (CZ) Optics: A CZ optic can be integrated to provide the final imaging solution. The CZ opto-mechanical lens assembly includes the optical assembly, mechanical packaging, focus and zoom motors, and temperature sensors.

Cryocoolers Past and Present: Rotary and Linear Types

Stirling cryocoolers for tactical applications are almost always rotary or linear types. Rotary coolers use a crankshaft attached to their moving piston and displacer. This arrangement allows precise control of the relative phase angle of these components but introduces side forces from the crankshaft linkages that can impact lifetime and reliability. In contrast, linear coolers are driven by voice-coil type actuators with minimal side forces and that rely on pneumatic drive and tuning to control the phase of the displacer relative to the compressor.

Rotary coolers were the most common cooler type in early cooled sensors. Due to its precise, mechanical phase control mechanism and the fact that the phase does not vary with frequency (RPM), rotary coolers have historically had higher efficiency and faster cool-down times than linear coolers. The drawback for these coolers has traditionally been lifetime due to the increased seal wear caused by the side forces generated by their mechanical crankshaft linkages. However, modern designs have greatly improved reliability with some manufacturers claiming lifetimes of 15,000 to 30,000 hours. Another drawback for rotary coolers is exported vibration from its unbalanced crank mechanisms, which can cause jitter for sensitive systems and, in turn, can lead to an increased SWaP footprint and lower reliability if stabilization is needed. Generation of acoustic noise is also a potential issue.

Linear cryocoolers have become the most common cooler type for newer systems due to their longer lifetimes and lower vibration. However, the linear cryo-cooler mechanism does typically drive slightly larger size for a given cooling capacity and slightly longer cooldown times. Because they are tuned to operate near resonance, linear coolers are limited to a single operating frequency. Input power can be modulated through the amplitude of the compressor pistons, but the inability to increase frequency during cool-down results in longer cool-down times than what can be achieved using rotary coolers. For modern linear coolers, lifetimes of 20,000 to 30,000 hours are becoming expected.

The pulse tube cooler is a special type of linear Stirling cooler in which the moving Stirling displacer in the cold finger is replaced with an expander containing no moving parts. A column of gas replaces the physical Stirling displacer. This “gas piston” is paired with a stationary regenerator and a phase shifting mechanism to control the mass flow relative to the oscillatory pressure. This phase control mechanism is more sensitive to operating conditions than the pneumatic tuning of the standard linear Stirling cooler, leading to lower efficiency at operating points away from the design point. Cooldown time is especially impacted and is generally significantly longer than other coolers of similar capacities. The separation of the regenerator from the pulse tube, which becomes the effective displacer, leads to a larger cold finger. Pulse tubes are also prone to orientation sensitivity, particularly in high-G environments. These drawbacks have prevented pulse tube coolers from being used in most tactical applications, despite the lifetime advantage. Paired with a non-contacting flexure bearing compressor, pulse tube coolers can achieve lifetimes greater than 100,000 hours.

Cryocooler Reliability and Misconceptions

Reliability for cooled camera modules and cryocoolers is often reported in terms of either mean time between failures (MTBF) or mean time to failure (MTTF). While the terms sound similar, there are several important distinctions. Generally, MTBF is applied to repairable systems while MTTF is used for systems which cannot be repaired. More importantly, the two metrics are defined differently and do not compare directly to one another. To add further ambiguity, sometimes MTTF is referred to as mean time before failure (MTBF).

The MTBF (between) metric assumes a constant rate of failures for a component, which is a good assumption for components not subject to mechanical wear. It is equivalent to the inverse of the random failure rate of the component and is often defined as (# of hours operational)/(# of failures) for a population of interest. MTBF was often used historically for cryocoolers and cooled IR cameras. For modern cryocoolers, the failure rate prior to wear out is generally very low and the MTBF is, therefore, very high. Because it does not account for mechanical wear out of the cryocooler and is only focused on the random failure rate prior to wear out, the calculated MTBF of a cryocooler will often significantly exceed its MTTF or expected lifetime.

Cryocooler lifetime estimates are often presented in terms of MTTF and calculated using Weibull statistics. The most-used Weibull distribution has two parameters, a shape parameter indicative of the amount of wear in the system and a lifetime parameter representing the point at which 63 percent of the population will have failed. The two-parameter Weibull distribution representing the failure rate versus time is shown below.

The MTTF, defined as the time at which 50 percent of units in the population will have failed, can be calculated from a sample of units using statistical analysis software. This methodology is now used by the majority of cryocooler manufacturers, however many report the lifetime parameter of its distribution (63 percent failure) rather than the true MTTF (50 percent failure). Because this method accounts for wear out of the mechanical coolers, it provides a better estimate of its true useful lifetime. Teledyne FLIR estimates cryocooler lifetime by fitting life-test data to a Weibull distribution and calculating the MTTF (50% failure) point of the population.

While cryocooler MTTF inputs are helpful to establish IR cryocooled system reliability, there is an important caveat. Cryocoolers can be designed in as a factory-replaceable component within the system. A worn-out cryocooler can be replaced, meaning the IR system should be seen as serviceable, multiple times if necessary, thereby offering an extended infrared system operational life.

Linear Cryocooling Going Forward

Linear cryocoolers have become the most common cooler type for newer systems due to their significantly longer lifetimes and lower exported vibration. For modern linear coolers, lifetimes of 20,000 to 30,000 hours are becoming expected. For example, the Teledyne FLIR FL-100 linear cryocooler has an approximate MTTF of 27,000 hours and with ongoing product improvement is targeted to surpass 30,000 hours.

This article was written by Julie Moreira, Principal Systems Design Engineer & Tech Lead; and Ted Conrad, Sr. Principal Cryocooler Engineer, Teledyne FLIR (Wilsonville, OR). For more information, visit here .

---