Reliability of Connectorized 10 Gbps/Lane Optical Fiber Transceivers

Live vibration and fiber pull tests configuration.

These products passed extensive qualification tests to demonstrate their robustness; including 1000 temperature cycles, damp heat, vibration, mechanical shock and thermal shock.
By: Jocelyn Lauzon, Tomasz Oleszczak, David Rolston, Robert Varano, Saïd El Kharraz, Naeem Safdari

Introduction

Reflex Photonics Inc. has developed compact 4+4 12.5 Gbps/lanes optical fiber transceivers for harsh environment applications such as Aerospace and Defense (see Fig. 1a). The LightABLE™ product series offers chip size transceivers that can operate from -40°C to 100 °C with a reach of more than 100 m on OM3 fibers with bit error rates (BER) as low as 10-15.
These products passed extensive qualification tests to demonstrate their robustness; including 1000 temperature cycles, damp heat, vibration, mechanical shock and thermal shock [1, 2]. What remained to be tested was their live reliability, through a 12-fiber ribbon cable pigtail mated to the LightABLE™ product through the proprietary MicroClip MT ferrule design (see Fig. 1b).

LightABLE rugged embedded transceivers

Fig. 1. a) LightABLE optical fiber transceivers

MicroClip MT ferrule

Fig. 1b) MicroClip MT ferrule mating between a LightABLE and a 12-fiber ribbon cable.

Test setup

In order to confirm the fiber mating reliability of this product, two different tests were undertaken: a live vibration test as per MIL-STD-883K, Method 2007.3 [1] and a live fiber pull test as per GR-468-CORE, Section 3.3.1.3.2 [2].
The test configuration is described in Fig. 2. The configuration corresponds to a loopback test, where the signals incident to the 4 transceiver detector lane of the device under test (DUT) generates the signals that drive the 4 transmitter lanes of the same DUT, that are then analyzed for error count. Before reaching the optical detector of the DUT, the 10 Gbps pseudo-random binary sequence (PRBS31) optical signal is attenuated to the sensitivity limit of the detector for a BER of 10-12 in normal room temperature test conditions (no vibration or pull test weight). Thus, if there is a signal degradation of the DUT while it is submitted to vibration or pull test weight, there will be a cumulative effect from the transmitter and receiver sections of the DUT that would impact the error count.
The live vibration test involved 5 different SR4 LightABLE units on TinLead (SN63 material) ball grid arrays (BGA) that were soldered to a printed circuit board. Having optical fiber transceivers that are connectorized rather than fiber pigtailed does allow for these products to be soldered to printed circuit boards using a standard reflow process; but then the reliability of the fiber mating associated to a connectorized configuration has to be confirmed through a process such as what is described here. During the live vibration test, each DUT, soldered to the test board, is fixed on a dynamic shaker bench and excited with single harmonic motion according to the following vibration profile: the vibration frequency is to be varied logarithmically between 20 Hz and 80 Hz and then set at a 20 g peak acceleration condition from 80 Hz to 2000 Hz; for 16 minutes in each of the orientations X, Y and Z. All 3 orientations are tested successively with the same unit (see Fig. 3a).

Live vibration and fiber pull tests configuration.

Fig. 2. Live vibration and fiber pull tests configuration.

The same test conditions were repeated for 3 pluggable DUT, having a MegArray™ electrical interface, instead of the surface mount BGA.
The same configuration was used for the fiber pull tests, for 3 DUT units, with the difference that no dynamic shaker was involved and that a weight was applied directly on the 12-fiber ribbon cable connected to the DUT, about 20 in. from the unit, while it is fixed vertically (see Fig. 3b).

LightABLE being subjected to vibration test

Fig. 3. a) Live vibration test setup for the longitudinal orientation

LightABLE being subjected to live fiber pull test
Live vibration test conditions

Live vibration test conditions

Live vibration test conditions

Vibration spectrum profile

The DUT must display, on all lanes, at all times during the tests, a BER better than 10-12 for the test result to be considered a success.

Results

For all lanes of all DUT submitted to the live vibration tests, either soldered to the test board or plugged, for all 3 orientations, the BER was better than 10-12, with most lanes from most units showing no errors during the tests.
For the live fiber pull tests, we did not measure any error on any lane during the tests, for an applied weight up to 1 kg.
This exceeds by a factor of 2 the limit set by Telcordia GR-468-CORE for the Fiber Integrity Side Pull Test [2].

 

Live vibration test results

Live vibration test results

Live fiber pull test results

Live fiber pull test results

References

[1] MIL-STD-883K, Test Method Standard Microcircuits, US Department of Defense, April 2016.
[2] GR-468-Core, Issue 2, Generic Reliability Assurance Requirements for Optoelectronic Devices Used in Telecommunications, Telcordia, September 2004.