WARNING: Cannot measure backscatter (OptiFiber)

                                                                                         Table 1. Attenuation Deadzones
 
The OTDR vendor may or may not specify the pulse width they are using to determine the attenuation deadzone. The OTDR vendor will typically pick a narrow pulse width, since that will yield the best attenuation deadzone. As your link increases in length, so will your need to increase the pulse width. Increasing the pulse width causes the attenuation deadzone to increase. If you leave the OptiFIber or DTX Compact OTDR in AUTO mode, the testers will select the optimum pulse width for you automatically (recommended).

Often missing from the OTDR specification is the reflectance value the OTDR vendor is using when quoting the attenuation deadzone. In the case of Fluke Networks, the attenuation deadzone is specified as:

• Multimode
  Measured at ± 0.5 dB beyond backscatter for typical connector with UPC polish (reflectance of < -37 dB multimode) using 40 ns pulse
  width (20 ns at 850 nm) at near end (100 m, which excludes dispersion).
   
• Singlemode
  For singlemode: measured at ±0.5 dB beyond the backscatter for a typical singlemode UPC connector (< -50 dB reflection using 20 ns
  pulse width).

Looking at the two examples again but side by side this time, it becomes clear that the better the reflectance, the more likely the OTDR is able to measure the attenuation loss of an event.

Figure 2

 
 
Figure 3

With a reflectance of just -26 dB, you see much more "tailing" from event "A". This results in an attenuation deadzone of more than 100 m, even with a short 20 ns pulse. The optical loss/length test using an Optical Loss Test Set (OLTS) also failed the link (when tested to ANSI/TIA-568-C.0).
 
But what you may see is that the Optical Loss Test Set passes the link, yet the OTDR gives a WARNING. If we look at Figure 1a again:

(Figure 1a)

and the corresponding optical loss/length report from the OLTS, we see a PASS:

 
Table 2

The OTDR uses a different measurement technique compared to the OLTS. The OLTS does not rely on backscatter to make its measurement. Therefore, the reflectance can be quite bad and yet the OLTS shows a PASS, as in the example above.

Is the example above (Figure 1a) a failure? As far as the standards are concerned, NO. Reflectance is not currently a test requirement in the TIA, ISO/IEC and EN cabling standards. But, the OTDR is unable to make an accurate loss measurement due to less than desirable reflectance. That results in a dilemma for Fluke Networks. We cannot report a FAIL, because reflectance is not a requirement. We cannot report a PASS, because we are unable to calculate the Overall Loss of the link. The only suitable summary for the result is WARNING.

It's worth noting at this point that most OTDRs do not calculate the Overall Loss of the link automatically.
 
 
Field polishing? 
If you are field polishing connectors, you are at an immediate disadvantage when it comes to OTDR testing. You've seen how poor reflectance can result in larger attenuation deadzones and WARNING appearing on the test report summary. If you are lucky to have a really good fiber technician on your staff, who is given adequate time to polish connectors, you can expect a reflectance value of -35 dB or better. But, even with a reflectance of -35 dB, your attenuation deadzone on singlemode fiber is going to be around 30 m. Therefore, if you have links less than 30 m, you will need to use a Receive Fiber for the OTDR to be able to measure the Overall Loss.

The problem with field polishing is that the technician only gets to see a two dimensional image of the endface connector. An example of a multimode connector endface is shown below:

                                              
                                      
                                                                                                    Figure 4



This is a good example of how the endface of a connector should look. There are no scratches, chips, resin or oxides visible. BUT, what you cannot see is the endface geometry. Let's look at the fiber connector from the side view.
 
                                         

                                                                                                    Figure 5

The only way to tell if the fiber is protruding from the ferrule (not enough polishing), or if the endface has been too aggressively polished (undercut), is to use a device called an interferometer.

Figure 6. Example Interferometer




Figure 7. Example "< 50 nm undercut"  image from an Interferometer

It is only with an Interferometer can you guarantee good endface geometry and reflectance.


The field polishing process
Most polishing sequences begin with aggressive materials, silicon carbide to remove epoxy and diamond lapping films for the beginning and intermediate polishing, that remove both the ferrule and fiber at the same rate. During the last polishing step a less aggressive material, usually silicon dioxide, is used because it attacks only the fiber. If an aggressive film is used for the final polishing step, excessive undercut will result.

Excessive Fiber Undercut is usually specified as more than 50 nm. Fiber Undercut is a condition that affects both reflectance and insertion Loss. When connectors are mated, the ferrule material surrounding the fiber compresses, which optimally allows fibers with an acceptable undercut/protrusion to make contact. Fibers that do not make intimate contact have an air-gap. An air-gap will produce unacceptable reflectance and insertion loss. Practically speaking, insertion loss is not the real victim here, reflectance is.

Figure 6
 

Figure 7