Technical Papers

» LOW COST FIBER OPTIC INTERFEROMETRIC SENSORS

JEFF BUSH, CAROL A. DAVIS, FRED McNAIR, ALLEN CEKORICH, and JOHN BOSTICK
Optiphase Inc.
7652 Haskell Ave.
Van Nuys, CA 91406

Abstract

This paper outlines recent progress made by Optiphase Inc. in the development of low cost fiber optic interferometric sensors. The paper’s focus is on components under development, specific to Interferometric Fiber Sensors (IFS), which aren’t commercially available through normal telecommunications distribution channels.

 1. INTRODUCTION

Optiphase has focused solely on Interferometric Fiber Sensors for the following reasons.

• Interferometry can provide high measurement resolution and sensitivity;
• Interferometric sensors can be made accurate (if interrogated correctly);
• Many Interferometric Fiber Sensors being developed for a wide range of sensing applications require the same type of components.

To demonstrate the last point, consider the interferometric sensor configurations shown in figure 1.

Here we show configuration diagrams for a wide variety of IFS type sensors. The sensor applications are described in the figure as the underlined words. The illustrations clearly show that there are many common components in each of the sensors. Some of these common components shown are manufactured in very large volumes as they are also common to the telecommunications industries. These components are the optical sources, the fiber, the splitters, and the photodiodes used in the analog receivers. All of these components sell for relatively low cost and are suitable "as-is" for manufacturing low cost interferometric fiber optic sensors. The remaining components shown in the diagram are not telecommunications components and suppliers of these devices (at low cost) are virtually non-existent. The IFS specific devices which lack suppliers are:

• Fiber Polarizers
• Fiber Phase Modulators
• Wound Coils (for rotation rate sensing and delay lines)
• Low Cost Analog Receivers
• Interferometric Demodulators

If Interferometric Fiber Sensors are to be low cost, these components need to be low cost! This paper outlines Optiphase development activities and results on the above devices.

2. IN LINE FIBER OPTIC POLARIZERS

These devices are used in fiber optic gyros, magnetic field sensors and other Sagnac interferometers. Optiphase has developed a design approach which provides low optical loss, high polarization (extinction ratios better than -30 dB) and extremely miniature size. Optiphase routinely makes polarizers for use with single mode fiber, and is currently developing approaches for use with polarization maintaining fiber.

A photograph of Optiphase polarizers is shown in figure 2. These devices are at least 5 times smaller than the few devices commercially sold today. This small size will make them compatible with most fiber sensor packaging constraints.

A proprietary processing approach developed at Optiphase enables fabrication of these devices to be performed with low cost material and a minimum amount of labor.

The performance characteristics of the devices fabricated at Optiphase are listed below.

All In-Line Polarizers

          Operational Wavelengths                  1300 nm ± 50 nm and 1550 nm ±

          Average Loss                                   1 dB

          Dimensions                                      0.75" X 0.1" (including strain relief)

          Operational Temperature Range         -55C to + 80C

SM - SM Polarizers

          Polarization Extinction                      < - 35 dB

          Fiber Types                                      Telco, DS, 125 um

SM - PM Polarizers

          Polarization Extinction                      < - 30 dB

          Fiber Types                                      Telco, DS, PANDA,125 um

PM - PM Polarizers

         Polarization Extinction                       < - 30 dB

         Fiber Types                                       Telco, DS, PANDA,125 um

Optiphase is currently preparing to commercially produce these devices. Current expectations for pricing include SM-SM polarizers less than $100; SM-PM polarizers less than $200 and PM-PM polarizers less than $250.

3. FIBER PHASE MODULATORS

There are two generally known ways to produce optical phase modulation in fiber waveguides. One is to change or modulate the index of refraction of the waveguide and the other is to stretch the waveguide (or fiber). The index changing approaches typically use integrated optic devices. Commercially such devices aren’t found in the less than $1000 per piece price range. The lower cost approaches involve fiber stretching. Typically these involve piezoelectric or magneto-strictive techniques. Optiphase has focused on the piezoelectric approaches using thin walled cylinders (tubes) and solid cylinders (disks). The relationships governing the operation of these devices are described.

A PZT phase modulator is constructed by winding optical fiber around tube or disk PZT ceramics and applying a voltage V at a frequency w. Optical phase shift, Df (rad), for light of wavelength l traveling through one layer of fiber of length L and outer diameter dfiber wound on a disk or tube PZT of radius R and height h is:

and therefore,
The resonance frequency ¦r (Hz) and change in the radius DR of a disk piezoelectric ceramic with end face electrodes are approximated by:

where Nc (kHz-in) is the material’s radial frequency constant and d31 (m/V) is the material piezoelectric constant. Substituting into equation (3) we get a relation for amount of optical phase shift per applied voltage for an optical modulator fabricated by winding one layer of fiber on a disk PZT:

Note that for a disk fiber wound PZT modulator, the phase shift per applied voltage is only a function of disk radius and is independent of disk height. Although increasing disk height allows added fiber on the disk, the resulting decrease in DR for a given voltage cancels out the benefit of the added fiber. Therefore in disk designs, the height of the disk does not affect performance and can be made as small as practical. Multiple layer winds can be employed to increase the modulation constant while maintaining a small package size.

The resonance frequency ¦r (Hz) and change in the mean radius DRm of a tube PZT with wall electrodes are approximated by:

where Nc (kHz-in) is the material’s circumferential frequency constant, d31 (m/V) is the material piezoelectric constant, and t is the wall thickness. Substituting into equation (3) we get a relation for amount of optical phase shift per applied voltage for an optical modulator fabricated by winding one layer of fiber on a tube PZT:

Note that for a tube fiber wound PZT modulator, the phase shift per applied voltage is a function of radius, tube height, and thickness. This is because DRm is independent of tube height. Therefore, an increase in tube height does not affect DRm but does enable more fiber to be wrapped on the tube which provides more modulation. To optimize tube designs, the diameter and height should be made as large as required to achieve the desired modulation and of course the tube thickness should be minimized. Again, as with the disks, multiple layer winds can be employed to increase the modulation constant while maintaining a small package size.

Although the basic design concept of a fiber wound phase modulator is simple, the construction techniques are key to achieving good performance. Improperly wound multiple layer modulators will inefficiently transmit strains to the outer layers resulting in low modulator constants. Additionally, the wind quality as well as mounting techniques can affect the stability of the modulator constant under environmental effects. Optiphase Inc. has developed inexpensive, size efficient, fiber wound PZT phase modulators by optimizing the winding processes, bonding materials, and mounting techniques such that high modulation constants which are stable over temperature are achieved. Some standard designs using piezoelectric tubes and disks are shown below in figure 3.

The general range of performance characteristic of the these devices are listed below.

Size Range:                               0.5" to 2.5" diameter

Modulation Range:                      0.1 rad/volt to greater than 100 rad/volt (at low frequency)

Frequency Range:                      dc to >100 kHz (varies with device size)

Operating Temperature Range:     -55°C to +70°C

One of the important issues to take note of is the wide operational temperature range. Optiphase has come across many phase modulator designs (using piezoelectric devices) which are quite unstable over temperature. The packaging approaches developed, have overcome such instabilities. To demonstrate this we show temperature test results (figure 4) of several 8 layer, 1.5" diameter, fiber wound PZT (tube) phase modulators.

In figure 4, the measured modulator constant was normalized to the value predicted from equation (9). The top curve depicts the results of an early design which shows that without proper optimization of construction methods, the modulation constant is greatly affected by changes in temperature. The middle and lower curves show results from a properly optimized design. In the middle curve the modulators were left unmounted while in the lower curve, they were mounted. The multi-layer wind design requires a layer perfect wind to attain the stable performance over temperature. In both mounted and unmounted high repeatability between different modulators was attained. However, the non-mounted PZT performed slightly better over temperature.

Optiphase is currently preparing to commercially produce low cost fiber phase modulators. Current expectations for pricing will be between $100 to $250 depending on the modulator constant desired.

4. WOUND FIBER COILS

Wound fiber coils are an essential element in many IFS applications. They are used as delay lines as well as intrinsic sensors for measuring rotation rate (gyros), acoustics, vibration, acceleration, magnetic field, and some applications even include pressure measurements. Most of these applications require a precision wound coil. There are various types of coil winds (straight winds, quadrupole, trapezoid, flangeless, free standing, etc.) which are employed to meet various requirements. Optiphase Inc. has developed a semi-automated coil winder which can perform all types of precision winds.

In Figure 5, a precision thread wound delay coil is shown. Figure 6 is a quadrupole wound gyro coil using a flangeless trapezoid layer build-up. This type of wind provides a significant thermal (bias) stability improvement over traditional quadrupole wound (gyro) coils.

While the above types of wound coils provide the ultimate in performance, they will mostly be limited to the "spectacular results" sections of R&D reports (and perhaps SPIE papers) and subsequently relegated to lobby trophy display cases. The winding processes required to fabricate such coils are expensive, owing mostly to a labor intensive winding process which mandates low tension winding at slow speeds.

Optiphase has recently broken the correlation between winding tension and optical performance.

This breakthrough offers the following advantages.

• Small wound coils with better optical performance than shipping spool measurements;
• Turn perfect, layer perfect winds;
• "Hands off" winding from start to finish;
• Presently 100 to 200 RPM winding speed.

The winding apparatus which was developed to verify this new high speed approach was developed under a SBIR Phase I contract from ARPA. It is currently limited to "straight winds." In May of 1996, Optiphase was awarded an SBIR Phase II contract (from ARPA), to continue the development of this winding approach to include quadrupole winds for precision fiber optic gyros.

The high speed winding approach however, has a much broader application potential than fiber gyros made with PM fiber. It will also be suitable for other fiber types and extend to lower cost (depolarized) fiber gyros made with single mode fiber, precision delay lines, and other physical sensing applications. Optiphase intends to commercialize this new winding approach.

5. LOW COST ANALOG RECEIVERS

The majority of Interferometric Fiber Sensor designs in existence require an analog receiver with well tailored gain, phase and noise performance characteristics. Most fiber sensor designs with commercial expectations will operate with low cost optical sources designed for the telecommunications markets operating in the 1.3 and 1.55 um wavelengths. At these wavelengths, there are a number of commercial analog receiver devices sold commercially, commonly referred to as PIN-FETs. These are extremely high performance devices utilizing ceramic circuit board GaAs hybrids which come with a high performance price tag (small quantity pricing is around $500 or more).

A lower cost design approach which meets the needs of the majority of fiber optic sensors applications involves a PIN photodiode and an op-amp. As simple as this sounds, there is no known vendor for such low cost (analog) receivers who offers a simple product spread which covers the needs for the fiber optic sensing community. For most fiber sensing applications the bandwidth requirements rarely exceed a few hundred KHz. Since Optiphase has been unable to find a vendor for low cost receivers, we typically fabricate receivers (using InGaAs PIN photodiodes and Op-Amps) having performance characteristics like that shown in figure 7.

The design approach which Optiphase uses for the receiver circuitry involves a reverse biased InGaAs PIN and Op-Amp configured for trans-impedance operation.

Optiphase has developed a family of these analog receiver designs which provide operational frequency responses as low as 10 KHz and up to 10 MHz. Over this frequency range, typical NEP’s will range from 0.3 pW/rt- Hz to 2.5 pW/rt-Hz. Tailoring for a particular frequency response involves changing a single component, namely the feedback resistor.

Optiphase plans to commercialize an analog receiver "family" which covers the frequency ranges described. These devices, configured with fiber optic input connectors, are anticipated to be sold at costs below $100.

6. INTERFEROMETRIC DEMODULATORS

Optiphase has spent a considerable amount of time exploring the development of interferometric demodulators. It is common knowledge (in the IFS field) that there are a great many approaches for performing interferometric demodulation. It is also common knowledge that the type of electronics required for interferometric demodulation requires a high degree of specificity. Currently, most interferometric demodulators are custom developed for the fiber sensor being developed. This typically results in the demodulator comprising a significant cost of the fiber sensor.

If Interferometric Fiber Sensors are to be low cost, the demodulators will also need to be low cost. The approach taken by Optiphase to accomplish such an objective is to focus on demodulator approaches which are applicable over a wide range of sensing applications. This concept assumes that volume requirements for such a device will drive prices down.

Optiphase believes that the most general type of interferometric demodulator will exist as an Open Loop processing device. Further we believe that if high accuracy demodulation is to be accomplished, this demodulator will have to incorporate a means for providing a modulation (or dither) signal into the interferometer which will enable an accurate (optical) phase measuring approach. The focus of this section involves a review of these open loop demodulation approaches which employ a Phase Generated Carrier (PGC) modulation scheme. The review will examine the ability of these demodulation approaches to address diverse sensor configurations and their applicability to be produced at a low cost.

DEFINITION OF TERMS; PHASE GENERATED CARRIER

In figure 1, in the intrinsic sensor section we show a variety of interferometers all configured with phase modulators. If the phase modulators are driven with a sinusoidal dither signal they will produce a sinusoidal variation of the interferometer’s optical phase. The generalized expression for the received signal of the interferometer which employs this Phase Generated Carrier (PGC) is shown in equation 10.

P(t) = COS { b SIN(wt) - f }                                                (10)

Here we show only the portion of the received signal representing interference terms. For this expression,

f = the phase of the interferometer (includes sensed signal and offset terms);

b = the PGC modulation index, expressed in peak radians;

w = the frequency of the PGC phase modulation.

Equation 10 can be broken out into harmonic terms of the PGC "carrier" frequency.

P(t) =    J0(b ) COS{ f }                     (dc term)                                                        (11)

+          2J₁(b ) SIN(wt) SIN{f }          (1w term)

+          2J₂(b ) COS(2wt) COS{ f }     (2w term)

+          2J₃(b ) SIN(3wt) SIN{ f }       (3w term)

+          2J₄(b ) COS(4wt) COS{ f }    (4w term)

where J_n(x) is the Bessel Function of the first kind of order n and argument x.

Equation 11 shows that the resulting break out of terms produces even harmonics of w which vary with COS f and odd harmonics vary with SIN f. Typically, open loop demodulators are designed to extract these (quadrature) terms, which are used to perform the interferometric demodulation.

ANALOG OPEN LOOP DEMODULATORS

In the 1980’s, analog demodulation techniques were very popular. The simplest approach for open loop demodulation is shown in figure 8. We call this demodulator type PGC-1. It incorporates circuitry (synchronous detectors) to measure J₁(b)SIN{f }, J₂(b)COS{f }, and J₄(b)COS{f }. It also has a circuit to control the modulation depth via a servo which keeps {J₂(b) / J₄(b) } constant. We call this the Depth of Modulation (DOM) servo It also requires an additional servo which keeps the effective optical throughput constant. This is accomplished by taking the square root of the sum of the squares of KJ₁(b)SIN{f } and J₄(b)COS{f }, where K is a scaling coefficient such that KJ₁(b) = J₄(b). This demodulator doesn’t actually measure the optical phase, but rather approximates it. It really measures J₁(b)SIN{f }, and can be said to be accurate for small values of f where SIN{f } » f.

The features of this demodulator are as follows.

• Small measurement range, perhaps ± 1 radian (only really useful for Sagnac interferometers).
• Is DC coupled which makes it suitable for rotation rate sensing.
• The circuitry is rather complex, has a large part count and is a poor candidate for circuit reduction via LSI techniques.
• Is low noise and can be accurate

The key problem, outside of complexity with the PGC-1 demodulator is that it doesn’t work over a broad range of sensing applications.

A different type of open loop approach is required if we want to measure phase signals larger than ± 1 radian. One such type is shown in figure 9. We call this a PGC-2 type demodulator.

This approach was championed by NRL in the early 1980’s. Some refer to it as the "Differentiate and Cross Multiply" or DCM approach. It basically extracts the quadrature terms, and using the DCM circuit, performs an FM demodulation, followed by an integrator to extract the PM demodulated output. We have shown the PGC-2 circuit with the same DOM and optical throughput servos as in the PGC-1 circuit. The features of this approach are a bit different than those of the PGC-1 approach. The PGC-2 demodulator cannot detect dc or low frequency (phase) signals, but it is able to measure phase magnitudes to many radians.

As with the PGC-1 circuit, the PGC-2 circuit is also complex and since it doesn’t measure low frequency or DC signals, is not suited for multiple IFS applications. A third open loop analog demodulation approach is shown in figure 10. We call this the PGC-3 approach. This approach was also championed by NRL and is commonly referred to as the "phase tracker" approach. Demodulation is accomplished by using an AD639 "trig-chip", operated in a pseudo-quadrature (to the receiver output) mode. It is configured in a (electrical) closed loop mode so that changes in the receiver output signal cause input changes to the AD-639. The input to the AD-639 constitutes the demodulated output signal. In comparison to the PGC-1 and PGC-2 circuits, this is a simple circuit.

The characteristics of the PGC-3 demodulator are as follows.

• Moderate Measurement Range (up to +/- 10 radians;
• Is DC coupled;
• Less Complex;
• Higher noise floor than PGC-1 and PGC-2 circuits;
• The AD-639 limits the operational temperature range.

Due to the ± 10 radian range limit, high noise and limited temperature range, the PGC-3 demodulator isn’t well adapted for use in a broad range of IFS sensors.

DIGITAL OPEN LOOP DEMODULATORS

In the past few years, a number of digital demodulation processing approaches have been investigated. For this paper, two will be discussed. Both of these approaches use processing techniques to extract SINf and COSf from the receiver output signal. Once these quadrature terms are extracted, the general approach for determining f is shown in figure 11.

Here the SINf and COSf are digitized. An algorithm is used to determine the magnitude of both values, and their ratio is taken. An inverse trigonometric process is performed to determine the phase within the quadrant, s. The process on the right side of figure 11 involves determining the quadrant. This is performed with simple sign checking. The "offset" phase from the quadrant determination, d is determined and added to s to determine f, which is the phase (on the unit circle). The digital approach also implements a fringe counter. This is a simple algorithm which compares prior values of d to current values of d to determine if the unit circle boundary has been crossed. When a crossing is detected, the fringe count a is incremented or decremented, depending on the direction of the crossing. Optiphase has built digital demodulators based on this process and it has been found to be a reliable approach.

The first type of digital open loop demodulator reviewed is shown in figure 12. We will call this the PGC-4 type open loop demodulator. It uses a "front-end" which is similar to PGC-1. It synchronously detects the terms required for demodulation, the DOM servo and the optical throughput servo. An analog multiplexer is used to feed the signals to the A to D converter (note the digitizing process is not exactly as shown in figure 11). Through digital processing the SINf and COSf terms are determined and then the digital demodulation takes place.

The characteristics of the PGC-4 demodulator are as follows.

• Measurement range is unlimited (due to use of fringe counter)
• Is DC coupled;
• Moderately Complex;
• Is capable of low noise performance approaching that of PGC-1 and PGC-2;
• Provides both analog and digital outputs.

This is the first open loop demodulator showing a use capability for a broad range of IFS sensing applications. It does have some design complexity (which may add cost), mostly due to the analog circuitry.

A simpler digital open loop demodulator is shown in figure 13. We will call this the PGC-5 type open loop demodulator. It has eliminated virtually all of the front end analog electronics. Here the ADC is placed directly after the receiver gain control circuit. In this circuit, the synchronous detection functions are performed in the processor. The PGC-5 performance characteristics will be the same as the PGC-4 circuit.

The PGC-5 demodulator unquestionably is the most favorable approach in terms of being applicable to a wide variety of IFS applications and having the potential to being produced at a low cost.

Optiphase has developed and evaluated prototype demodulators based on the PGC-5 configuration. The test results have verified the design expectations. Optiphase plans CY 97 commercial products based on this design approach, and due to its simplicity, is expected to be low cost.

7. CONCLUSION

Major steps have been taken to reduce the cost of the IFS specific components used in fiber optic interferometric sensors. Commercial viability of this technology across a broad range of sensing applications is expected before the turn of the century.