Click to Enlarge
Gain Flattening Filter Design Profile

Click to Enlarge
The typical gain of the EDDA100P amplifier without and with the GFFE15P gain flattening filter with different input powers. The blue-shaded region indicates the gain-flattening wavelength range. The filter is optimal for a -10 dBm input power; the gain flatness improves from ~3 dBm (solid red line) to <0.5 dB with the filter (dashed red line). The gain flatness at -20 dBm input power has more variation without the filter (solid green line). While gain flatness can be improved with one filter (dashed green line), cascading two filters further flattens the signal (dashed-dotted green line). The gain spectrum with 0 dBm input power (solid blue line) cannot be further improved with the filter.

Features

• Flatten Gain Spectrum of EDFA100x Series Erbium-Doped Fiber Amplifiers
• 1525 - 1570 nm Operating Wavelength Range
• Optimized for -10 dBm Input Power
• Available in SM and PM Versions

Thorlabs' Gain Flattening Filters are designed for use with the EDFA100x series erbium-doped fiber amplifiers (EDFAs). The filters are offered in SM (Item # GFFE15) and PM (Item # GFFE15P) models that are compatible with the EDFA100S and EDFA100P amplifiers, respectively. These filters are intended to be connected to the output port of the EDFA and are designed to flatten the gain spectrum of the EDFA by adding additional signal loss to the wavelengths where the gain is higher, making them ideal for applications that require equal gain across a wide spectrum. The gain spectrum of the EDFAs depends on the input power level with the gain variations increasing over the operating wavelength range as the input power is decreased. The design transmission curve of the filter, given in the upper right plot, flattens the gain of the EDFA100 series at -10 dBm input power.

One of the key specifications for the gain flattening filter is the peak-to-peak error function, which is defined as the difference between the measured transmission curve (in dB) and the design transmission curve (in dB). Each filter will have a transmission profile that deviates from the design curve to some level, and the error function captures this variation over the wavelength range. The peak-to-peak error function is a single valued specification defined as the difference between the maximum error function and the minimum error function. That value in dB is given in the specs table below.

The plot to the lower right shows the typical EDFA100P amplifier gain versus wavelength for three different input power levels (0 dBm, -10 dBm, and -20 dBm). The filters are designed to provide optimal gain flatness at a -10 dBm input power level, as seen in the gain plot to the lower right. The gain flatness at -10 dBm input power over the 1525 - 1560 nm wavelength range is improved from ~3 dB without the filter to <0.5 dB after placing the GFFE15P filter. Similarly, the gain flatness at -20 dBm input power can be improved from ~8.5 dB to <6 dB over the same wavelength range by adding the filter. Since the gain at -20 dBm shows a larger variation versus wavelength compared to the -10 dBm case, cascading two filters can further improve the flatness to <4 dB as shown in the plot. When connecting two filters together, we recommend using either a ADAFC3 or ADAFCPM2 narrow key FC/APC fiber optic mating sleeve, which are suitable for SM and PM fibers, respectively. The gain spectrum of the EDFA100P amplifier for 0 dBm input power level exhibits good flatness and cannot be further improved using these filters.