Pin down key OTDR specs

When you specify an OTDR for field use, pay particular attention to test times, dynamic range, and resolution.

Jon Titus, Editorial Director -- Test & Measurement World, 8/1/2001

Although optical time-domain reflectometery techniques were demonstrated about 25 years ago, the advent of relatively inexpensive fiber-optic communications has pushed OTDRs out of the laboratory and into the field in the form of portable handheld instruments, often called mini-OTDRs. In fact, most OTDR manufacturers now aim their products at people who will use them for field maintenance and installation of new fibers. Most of today’s market for mini-OTDRs revolves around rugged, easy-to-use portable instruments used by people who may not know a decibel from a nanometer.

As an engineer, you may not get involved in fiber-optic installations or maintenance, but your company may rely on you to draw up specifications for the OTDRs it will purchase for field use. Before you grab a catalog or check Web sites, this quick tutorial on OTDR operations will give you a head start on understanding the capabilities of mini-OTDRs.

At its simplest, an OTDR sends a pulse of light down a fiber and then measures light reflected back from the fiber (Figure 1). These days, OTDRs work with multimode fibers that conduct light at wavelengths of 850 nm and 1300 nm and with single-mode fibers that conduct light at 1310-nm, 1550-nm, and 1625-nm wavelengths. Many portable OTDRs let buyers populate slots with modules that operate on one or two fixed wavelengths. These modules let you tailor instruments to meet specific measurement needs. Some OTDRs also can accept modules for other measurements, such as optical power.

Figure 1. A basic OTDR includes a laser-diode source and a highly sensitive photodetector. The photodetector converts reflected light into an electrical signal that the control and processing circuits measure and display on a power vs. distance plot.

Almost all the light reflected in an optical fiber arises from two phenomena: Rayleigh scattering (also called backscatter) and Fresnel reflections. Backscatter occurs to some extent in all fibers, and it stems from imperfections in the fiber caused by impurities, slight changes in the fiber materials, and other minor imperfections. Backscatter occurs almost uniformly along the length of a fiber.

On the other hand, Fresnel reflections arise from differences in the indices of refraction at interfaces between materials. You’ll observe Fresnel reflections at the end of an unterminated fiber when glass meets air. Fresnel reflections also occur at cracks, bends, and poor physical connections along a fiber. Fresnel reflections return as much as 4% of incident light, so these reflections appear to an OTDR as relatively bright “events.” (The word event has slipped into the fiber-optic lexicon to represent a specific reflection or loss that the OTDR finds.)

Time equals distance

The light reflected back to an OTDR’s detector decreases with distance, so the instrument plots the intensity of the backscatter and reflections (in decibels) vs. distance to show where events take place (see Figure 2). Connectors produce noticeable Fresnel reflections, whereas a splice simply attenuates the reflected backscatter light in a small step. An unterminated fiber end can produce a large Fresnel reflection.

Figure 2. An OTDR plot will include splices that attenuate light, connectors that produce a sharp reflection and a loss, and a Fresnel reflection that occurs at the end of the fiber or at a fiber break. A terminated fiber may not produce a Fresnel reflection.

Most OTDRs transmit more than one pulse down a fiber. By making thousands of measurements in rapid succession and averaging the results, the instrument reduces noise and offers a better indication of the actual state of the fiber. But those multiple measurements take time, in some cases, several minutes. You may see manufacturers base some OTDR measurement specs on 3-min measurement periods. (That’s the total measurement time, not the period between samples.) Although 3 min may seem speedy, it’s slow when you need to test the hundreds of fibers in newer cables. At 3-min per fiber—not counting setup time and tests at different wavelengths—testing a cable would take days. For field use, you need instruments that measure accurately and quickly—ideally only a few seconds per fiber. Unfortunately, though, OTDR manufacturers base many of their specifications, such as resolution and dynamic range, on 3-min “standard” measurement periods. That’s a lot like having car manufacturers base their gas-mileage ratings on an unrealistic test in which cars always go downhill at 60 mph. Just as drivers don’t always drive downhill—and use less gas—fiber-optic technicians don’t have the luxury to spend 3 min making a single fiber measurement.

To best gauge an OTDR’s performance, check its data sheet and examine its specs under realistic measurement conditions. Manufacturers may present excellent specs they arrived at under conditions that you can’t realistically use for field measurements. If the supplier’s specs need clarification, your best bet may involve trying the OTDR under conditions that duplicate its future work environment.

Check dynamic range

In addition to test speeds, one of the first OTDR specifications buyers look at is dynamic range. Dynamic range determines the total fiber length an OTDR can analyze. The higher the dynamic range, the longer the distance an OTDR can analyze. Unfortunately, the dynamic range specification often causes people the most problems when they choose an OTDR. First, manufacturers don’t agree on how to specify dynamic range, and second, a quick determination of dynamic-range needs often yields the wrong value.

Figure 3. A dynamic-range spec depends on how you measure it. The IEC defines a conservative specification, but many manufacturers use the slightly larger signal-to-noise (SNR) dynamic range.

A technical committee within the IEC has established dynamic range as the difference between the power of the OTDR’s signal as it enters a fiber and the noise level (Figure 3). The noise signal below the IEC limit contains 98% of the measured noise, thus this definition of dynamic range often goes by the name “IEC (98%).” You may also see dynamic range called out as “Bellcore (98%)” because Bellcore, now Telcordia, first proposed the definition. Some people also call this measure of dynamic range the peak dynamic range.

In contrast, the SNR dynamic range, also shown in Figure 3, represents the difference between the start of the OTDR’s signal and the point in the noise signal at which the signal-to-noise ratio (SNR) equals 1. Manufacturers like the SNR measurement because it provides a larger number than the IEC (98%) measurement—generally about 1.8 dB larger. To confuse people further, the deviation between SNR and IEC (98%) values varies from one manufacturer to another, from 1.5 dB to as much as 4 dB. (Ref. 1)

Even when you have dynamic-range information about instruments, it’s easy to make a quick assumption that leads you to choose an instrument with insufficient dynamic range to meet your requirements. The problem comes from equating loss in the fiber to a needed dynamic range. Say, for example, that a fiber attenuates a 1550-nm signal at the rate of 0.25 dB/km. Over a 100-km link, that amounts to a 25-dB loss, so there’s a temptation to look for an OTDR with a 25 dB dynamic range plus a slight margin. If the selected OTDR offers a slightly better dynamic range, say 30 dB, all the better. Unfortunately, this naive approach doesn’t account for the instrument’s signal-to-noise ratio (SNR), a value that factors into all measurements. You need to know how well an OTDR can accurately measure a small change in a signal.

These days, fiber splices attenuate light by as little as 0.01 dB. That’s a small difference to detect, so this type of low-loss splice presents a measurement challenge. Without a sufficient dynamic range, an OTDR won’t detect splices, and they’ll remain hidden in the instrument’s noise. If you plan to use an OTDR to detect such splices, the instrument must have a high SNR, and you must account for it in the dynamic range you specify. For the 25-dB dynamic range mentioned above, as a rule of thumb, you should add a noise margin of 10 to 12 dB for an overall dynamic range of about 35 dB to 37 dB.

If these low-loss splices occur at the far end of a fiber that needs testing, an OTDR with a reasonable dynamic-range margin still may not detect them. As a light pulse gets further from the OTDR, reflections get fainter, and eventually an instrument cannot distinguish them from noise. Figures show that with a 5-dB margin, an OTDR can have between +0.21 and -0.23 dB of noise on a signal. A 10-dB noise margin decreases the noise level to ±0.02 db. Not bad, but the instrument still won’t reveal splices that reduce a signal by 0.01 dB. (Ref. 1)

For a typical mini OTDR, most likely the manufacturer specified the 35-dB dynamic range for a standard 3-min measurement time. By averaging more samples over the longer period, the manufacturer can boost the OTDR’s SNR—essentially “buying” dynamic range at the expense of measurement speed. If you need this instrument to make rapid measurements in the field, you may end up with a lower dynamic range. You can buy instruments that offer a high dynamic range and fast measurement times, but they cost more than general-purpose instruments. Some of the latest instruments offer a 30-dB dynamic range and measurement times of 2 s or less.

Distance needs remain the same

Although the demands for better dynamic range have increased, the need for better distance resolution has changed little. Resolution divides into two requirements: the need to resolve two closely spaced events, and the need to locate a specific event.

Figure 4. Long light pulses limit the spatial resolution of an OTDR. Shorter pulses resolve closely spaced events, but they also increase the noise level on the measured signals (Ref. 2).

The first requirement forces OTDR users to look at tradeoffs. Suppose you have two splices 100 m apart in a fiber with an incidence of refraction of 1.47. That makes the speed of light in the fiber about 2 x 10⁸ m/s. In the time domain, the splices “exist” about 0.5 µs apart. If you use a 1-µs pulse to locate the splices, you’ll obtain the top waveform shown in Figure 4 (Ref. 2). Using a 100-ns pulse, as shown in the lower waveform, lets you properly determine that two splices exist.

Now comes the tradeoff. The wider pulse provides more light energy and thus improves the SNR in a wide bandwidth receiver, but spatial resolution suffers. If you use shorter pulses and a narrowband receiver, spatial resolution increases, but at the expense of the SNR, which decreases. You can see the increase in noise in the lower trace for the 100-ns light pulses.

As a general rule, pulse widths of 10 to 20 µs give you the most dynamic range while maintaining good spatial resolution. OTDRs let you select pulse widths, so you can try several settings to find the best one for your application.

An OTDR must not only distinguish between close events, it must also accurately locate events such as a splice or a break. Most OTDRs can locate an event to within 10 cm (4 in.) of its real location, and that’s good enough for most field work. But a field boss won’t send a crew out to dig up a cable 34.9813 km from the OTDR. There’s just no way to accurately know where to find that point on the physical cable.

So how would someone use an OTDR’s distance measurements to locate problems? In most cases, the OTDR provides a location that a maintenance crew can compare with other information about a cable. During installation, an operator would gather, store, and document OTDR traces to plot a cable’s route. The operator would annotate the OTDR trace to indicate junction boxes, splices, GPS coordinates, and landmarks. Based on that documentation, a boss might send a crew out to a examine a splice at a junction box on Prime Ave. in Huntington, NY, for example. The OTDR data obtained from the cable during installation would provide a documented reference. The crew would compare it with an OTDR trace from the degraded fiber to locate an area in which to narrow its search. Once at the junction box, the crew could run other tests to more accurately diagnose and locate the problem.

A good OTDR should let you annotate traces with information about the locations of optical events. Without this information, cable-service people will spend a lot of time just finding cables, let alone repairing them. Companies must ensure that current documentation reflects all changes to fiber-optic configurations, a task well-suited for an OTDR. Telcordia provides a standard that defines how OTDRs should store data (Ref. 3). Most instrument manufacturers now comply with the Telcordia specs, although some also offer proprietary data formats.

Beware of dead zones

Although OTDRs routinely detect events at long distances, they cannot always detect or measure events in what’s called the “dead zone.” When an OTDR puts out a light pulse, the pulse first encounters a coupler followed by the interface between the instrument and the optical fiber. These connections reflect light into the OTDR’s detector, saturating it and rendering it unable to detect events. During the detector’s recovery period, which manufacturers translate into a distance, the OTDR cannot detect any events.

The dead zone actually includes two concurrent periods, an event dead zone and an attenuation dead zone. The event dead zone, the shorter of the two, defines the minimum distance at which the OTDR can detect another event. The point at which the measured light intensity is 1.5 dB down from the peak defines the end of the event dead zone. The considerable longer attenuation dead zone specifies the distance after which the optical receiver has recovered to within ±0.5 dB of the backscatter signal. After this point, the OTDR can make quantitative measurements.

Typical OTDRs specify an event dead zone of as little as a few meters and specify an attenuation dead zone of from 25 to 35 m. You can decrease the dead zones somewhat by using shorter light pulses, but very short pulses gain you little increase in performance. In general, OTDRs used for field work don’t require short dead zonres, but if you need one that does, check specifications carefully to ensure the instrument will provide accurate results within the fiber length you must characterize. You want to obtain good dead zone performance using a pulse width of about 1 µs. If you must make a measurement in your instrument’s dead zone, use an OTDR to make the measurement from the other end of the fiber.

Keep it clean

Even though modern, portable mini-OTDRs look at home in a toolbox, they’re sensitive optical instruments that demand care. And in the case of an OTDR, cleanliness tops the list of ways to keep an instrument operating properly. An engineer at one OTDR supplier reported that 80% or more of the problems with OTDRs in the field center on dirty connectors and dirty fibers. Agilent Technologies provides a helpful booklet that covers techniques and procedures for keeping OTDRs clean. (Ref. 4)

Here’s a final suggestion that may help you select the best mini-OTDR for field use. Let the users help you make the choice. The people who use an instrument should feel they influenced the purchase. Their concerns, such as being able to read the display in sunlight, may trump your concerns about a specification such as battery life. The field people have to use the instrument, so after you establish the minimum technical specs, let the users put some demo units through real field trials. They’ll let you know what makes the instruments easy to use while getting the job done. T&MW

References

1. “OTDR Standards Move Towards Reflecting the User’s Real Needs,” BITS, Issue 84, Wavetek Wandel Goltermann, Eningen u.A., Germany, June 1999. p. 12. www.acterna.com/downloads/newsletters/bits/bits84/bits84.pdf   (Wavetek Wandel Goltermann is now part of Acterna. www.acterna.com).

2. Derickson, Dennis, ed., Fiber Optic Test and Measurement, Prentice Hall PTR, Upper Saddle River, NJ. 1998.

3. Optical Time Domain Reflectometer (OTDR) Data Format, Special Report SR-4731. Telcordia, Morristown, NJ, January 2000. www.telcordia.com.

4. Cleaning Procedures for Lightwave Test and Measurement Equipment, Agilent Technologies, Boeblingen, Germany, 2000. www.agilent.com/cm/rdmfg/ont/manuals/cleaning.pdf. Editor's note: This link no longer works. Please visit www.agilent.com, 10/23/03.

For more information

Calibration of optical time-domain reflectometers (OTDRs), IEC 61746, International Electrotechnical Commission, Geneva, Switzerland. www.iec.ch.

Generic Requirements for Optical Time Domain
Reflectometer (OTDR Type Equipment, GR-196-CORE), Telcordia, Morristown, NJ, September 1995 (Revised December 1997 and December 1998). www.telcordia.com.

Kerridge, Brian, “Portable OTDRs Simplify Complex Fibre Tests, ” Test & Measurement Europe, April-May 1999, p. 25.

Jon Titus has written real-time software and designed embedded systems and computer/instrument interfaces. He worked in electronics for 10 years and spent nine years at EDN magazine prior to joining T&MW in 1993. He has a BS from WPI, an MS from RPI, and a PhD from VPI.

The following company information appeared in the original print version of this article. For up-to-date information about companies, visit the Fiber-Optic/Electro-Optic Test Equipment portion of our Buyer's Guide.