Development of Cleanliness Specification of Receptacle Transceivers: Fiber Stub

Transcription

1 Development of Cleanliness Specification of Receptacle Transceivers: Fiber Stub Ryo Nagase : Chiba Institute of Technology Hideki Isono : Fujitsu Optical Components Ltd. Yutaka Sadohara : Sumitomo electric Photo-Electronics Components (Suzhou), LTD.

2 The history of investigation 2/ Aug : Start experiment of dust effect on receptacle device (following the investigation of SC connector by Avanex) 2. Mar : Explain an intermediate result at inemi 3. Mar : Explain a result at inemi and suggest inspection criteria 4. Nov : Start re-investigation of data, especially focus to measurement error in cooperation with inemi and IEC 5. Mar : Agreement of criteria at Core area of fiber 6. Current : Investigating Clad area

3 Selection of DUT 3/ 20 Example of Optical Data Link (SFP type) TOSA ROSA-1 ROSA-2 Special Lens Stub Ball Lens Optical Sub Assembly * Inspection of SC type fiber connector had already started. * Structure of TOSA with stub is similar to the fiber connector. It might follow the experiments of fiber connector.

4 Effect of Contamination Which characteristic is sensitive to contamination? DUT : STUB (Transmitter side of OC48-SFF) Original Contaminated 4/ 20 Parameter Original - Contaminated Difference Optical Power -6.7 dbm -> -6.7 dbm 0.0 db Pulse Mask Margin 45 % -> 43 % -2 % Spectral Width 2.3 nm -> 2.3 nm 0.0 nm ORL 58.7 db -> 33.9 db db 10 um Fine Dust (FOTP-35) ORL is the most sensitive parameter.

5 DOE-1 Outline 5/ Prepare clean sample 2. Measure ORL 3. Contaminate the STUB endface 4. Measure ORL 5. Measure the distance between Contamination and Center of fiber 6. Analyze the effect of contamination 7. Investigate the measurement error at experiment Original Contaminate 10 um Fine Dust (FOTP-35) Analyze D=7.5um Delta ORL=19.0dB Delta ORL (db) Contamination Distance (um)

6 DOE-2 Detail process flow 6/ 20 Clean fiber stub and reference cable Take picture Mate Measure ORL Demate 1-9th Clean if contaminated (stub & cable) Take pictures after 1st, 5th and 10th mating (stub & cable) Take picture Reject removables Apply contamination (stub) 10th Mate Measure ORL Clean if contaminated (stub & cable) Change if damaged (cable) Demate 1-4th Take pictures after each mating (stub & cable) 5th If ORL is not so bad, additional loop of contaminating (to make ORL worse) end

7 DOE-3 Contaminating Techniques to get the Fixed Contamination Which state is true when ORL is measured? Mating connector to measure ORL causes dusts spread. 7/ 20 Before Mating After Mating Mate the dummy connector with DUT to fix dusts. Clean the optical endface of DUT to remove movable dusts.

8 Measurement diagram of Stub ORL Precision Reflect Meter ANDO 7410B Optical Fiber DOE-4 Measurement Techniques to Reduce Measurement Error 1st period : db 8/ 20 2nd period : db 10 times Waveform 10 times DUT Time Averaging Recording Averaging (sec.) (Mating) * ORL is changing after mating. To reduce measurement error 1. Set the resolution of test equipment to high end. 2. Set DUT straight and flat. 3. Wait 30 seconds after mating and Measure. 4. Average 10 times measurements.

9 DOE-5 Analyzing Picture of Endface Contamination Size and Location is analyzed by FiberChek2 ( JDSU ) software using the function of display concentric circle on picture. 9/ 20 Enlarged distance 29um Size 11um 50 60um Contamination Distance (um) : Contamination Nearest to core Center of core Dust Size (um) : Long side of dust

10 Re-analyzing data approximating a base equation 10 / 20 5 points moving average ignorable X = exp [ - ( y ) / ] 0.92 = exp [ - ( ) / ] for ORL 27dB 3.37 = exp [ - ( ) / ] for ORL 40dB * The points in the orange square has a different trend from others, and far from target ORL. * Therefore those are ignorable when an equation of the relation between ORL and distance is approximated.

11 frequency Investigation of Measurement Error (ORL) Inspection Criteria will be guided from average line + (1) Measurement Error of ORL + (2) Measurement Error of Length N : 37 Ave : db Min : db Max : db Std dev : db Ave + 3 * sigma : db next Standard deviation of 5 times ORL measurement (db) 11 / 20 Orange line : Base equation Purple Line : Including Measurement error of ORL (3.785dB) Measurement Error of ORL (a) Base Equation y = In ( x ) (b) Considering the measurement error ( y ) = In ( x ) x = exp [ -( y ) / ] (c) Calculate the crossing to the line of -27(dB) um = exp [ -( ) / ] (d) Calculate the crossing to the line of -40(dB) um = exp [ -( ) / ] (e) Calculate the crossing to the line of -55(dB) um = exp [ -( ) / ] The next step is to find the measurement error of length around (um), (um), (um)

12 Investigation of Measurement Error (Length-1) Inspection Criteria will be guided from average line + (1) measurement error of ORL + (2) measurement error of Length Enlarged 12 / 20 DOE 1. Project endface on a monitor 2. Once defocus, after focus again 3. click Test Fiber on FiberChek2 software. 4. Fit template on monitor 5. Measure the size of dusts and distance from center of fiber.(*to simulate normal operation in the field, operator decides the position of template where to fit within several seconds.) 7. Repeat 10 times. 8. Repeated by 3 operator. Estimated measurement error is = Magnification error caused by (1) focusing, (2) equipment distortion x Measurement error caused by (3) fitting template, (4) reading template x (5) Error on size of template During this experiment, (1), (3) and (4) are variable, (2) and (5) are fixed.

13 Investigation of Measurement Error (Length-2) Table-1. Measurement result Operator Trial Number Sample Average A A A A A A A A A A A Average A Peak-Peak A Std Dev B B B B B B B B B B B Average B Peak-Peak B Std Dev C C C C C C C C C C C Average C Peak-Peak C Std Dev All Average All All Peak-Peak All Std Dev All R (Unit : um) All X-DIFF / 20 Sample-1 : Distance between fiber center and Point-A Sample-2 : Distance between fiber center and Point-B Sample-3 : Distance between fiber center and Point-C Sample-4 : Distance between fiber center and Point-D Sample-5 : Size of Point-A (clear but affected by blown out) Sample-6 : Size of Point-B (clear) Sample-7 : Size of Point-C (smallest) Sample-8 : Size of Point-D (blur) Table-2. Result of GRR study Sample d2* Operator d2* Trial d Item Symbol Study Var %SV %Tolerance Judge Equipment Variation EV % 0.0% Operator Variation AV % 0.0% Total Gage R&R GRR % 0.0% Good Part to Part PV % 100.0% Total Variation TV % 100.0% The result (1) Sample is quite different to adapt these data to GRR study. (2) Peak-Peak variation 4 um is within the twice of graduation step of template. Template will decide measurement error. Continued on the following page

14 Length (3 * sigma) (um) Investigation of Measurement Error (Length-3) y = x R² = y = x R² = Template & FBP-P5000 (280x) Length (3 * sigma) (um) y = x R² = Template & Microscope (520x) 14 / Length (Average) (um) Figure-1(a). Dependence between distance and measurement error (Stub, FBP-P5000, 3 * sigma) Length (Average) (um) Figure-1(b). Dependence between distance and measurement error (Ferrule, Microscope, 3 * sigma) (3) Blue dot and line on figure-1(a) Using template as analyzing tool, measurement error is well related to the distance (size) of defect, while distance were short such as a target range (0-50 um) that we want to find the measurement error. (4) Figure-1(b) is the result of similar experiment using microscope. Measurement error using FBP-P5000 is bigger than microscope. It will come from the magnification of equipment. As collateral evidence, all operator felt that measuring the size of small defect using FBP- P5000 was more difficult than using microscope.

15 Length (3 * sigma) (um) Investigation of Measurement Error (Length-4) y = x R² = y = x R² = Template & FBP-P5000 (280x) Length (3 * sigma) (um) y = x R² = Template & Microscope (520x) 15 / Length (Average) (um) Measurement error of length at (um) is (d) 280x, FBP-P5000, 3sigma y = * = (um) (e) 520x, microscope, 3sigma y = * = (um) Length (Average) (um) Measurement Error (3 sigma) (um) y = x Magnification Suppose that there were linear relation between Magnification and measurement error on template analyzing. Measurement error of length at 200x is (f) y = x = * = (um) Missed as at the 2010 inemi face to face meeting because of missing minus sign.

16 Guide to Inspection Criteria for ORL:-27dB 16 / 20 Inspection Criteria will be guided from average line + (1) measurement error of ORL + (2) measurement error of Length 4.0 Measurement Error (3 sigma) (um) y = x Magnification Criteria circle for ORL:-27dB is (um) : Crossing point between -27dB and approximated equation including measurement error of ORL (um) : Measurement error of Length around (um) = (um) : radius = (um) : diameter 9 (um) : diameter for easier operation

17 Appendix Measurement Error and Eyesight / 20 Length (3 * sigma) (um) y = x R² = eyesight Though sample size is only 3 (operator), there is a good relation between Eyesight" of operator and measurement error. Eyesight will be equal to Resolution of equipment. If it were true, minimum eyesight for visual inspection will be specified. It will be suitable to use the lower limit of eyesight that car driver license requires. (0.4 (= 20/50) in Japan) Dependence between eyesight of operator and measurement error (3 * sigma)

18 Interpretation of / 20 This picture is the case of fiber chipping. If chipping were at ferrule side, does it satisfy the IEC standard? Table 5 Visual requirements for PC polished connectors, single mode fibre, RL 26 db Zone name Scratches Defects A: core 2 3 μm 2 3 μm None >3 μm None >3 μm B: cladding No limit 3 μm No limit <2 μm 3 > 3 μm 5 from 2 μm to 5 μm None >5 μm C: adhesive No limit No limit D: contact No limit No 10 μm Table 1 Measurement regions for single fibre connectors Zone Diameter for single mode A: core 0 μm to 25 μm B: cladding 25 μm to 120 μm C: adhesive 120 μm to 130 μm D: contact 130 μm to 250 μm NOTE 1 All data above assumes a 125 μm cladding diameter. NOTE 2 Multimode core zone diameter is set at 65 μm to accommodate all common core sizes in a practical manner. NOTE 3 A defect is defined as existing entirely within the inner-most zone which it touches NOTE 1 For scratches, the requirement refers to width. NOTE 2 No visible subsurface cracks are allowed in the core or cladding zones. NOTE 3 All loose particles should be removed. If defect(s) are non-removable, it should be within the criteria above to be acceptable for use. NOTE 4 There are no requirements for the area outside the contact zone since defects in this area have no influence on the performance. Cleaning loose debris beyond this region is recommended good practice. NOTE 5 Criteria should be applied to all fibre pairs in the array for functionality of any fibre pairs in the array. NOTE 6 Structural features that are part of the functional design of the optical fibre, such as microstructures, are not considered defects.

19 Compare 3 specs-1 (spec) 19 / 20 Spec-1 Spec-2 Spec-3 Current Area Diameter (μm) Scratches (μm) Defects (μm) Upto 2pcs, 3 μm Upto 2pcs, 3 μm A:Core 0-15 None, > 3 μm None, > 3 μm B:Cladding Upto 3pcs, 3 μm None, > 3 μm C:Adhesive No Limit No Limit D:Contact No Limit No Limit, 2 μm Upto 5pcs, > 2 μm and 5 μm None, > 5 μm No Limit, 10 μm None, > 10 μm A:Core 0-25 Upto 2pcs, 3 μm Upto 2pcs, 3 μm None, > 3 μm None, > 3 μm B:Cladding No Limit, 2 μm Upto 3pcs, 3 μm Upto 5pcs, > 2 μm and 5 μm None, > 3 μm None, > 5 μm C:Adhesive No Limit No Limit D:Contact No Limit No Limit, 10 μm None, > 10 μm A:Core 0-25 No Limit, 2 μm None, > 2 μm None B:Cladding No Limit, 2 μm Upto 3pcs, > 2 μm and 5 μm None, > 5 μm C:Adhesive No Limit No Limit D:Contact No Limit, 5 μm None, > 5 μm No Limit, 5 μm Upto 3pcs, > 5 μm and 10 μm None, > 10 μm No Limit, 20 μm Upto 3pcs, > 20 μm and 50 μm None, > 50 μm

20 Inquiry Model : Single mode Stub (SFP OC48 Tx side) Equipment : FBP-P5000 Quantity : 700 pcs. Failure : 4 pcs. Overturned : 3 pcs. Compare 3 specs-2 (result) Ser.141 Ser.200 Ser.268 Ser / 20 *Differences in Spec. at area-a caused overturn of judgment. Spec- 1 and Spec-2 approve 2 pcs. of small defects. But Spec-3 does not approve any defects. *Because of mark on equipment, there appears same defects on every pictures. Spec-1 Spec-2 Spec Good Good Fail(areaA) 200 Good Good Fail(areaA) 268 Good Good Fail(areaA) 462 Fail(areaA) Fail(areaA) Fail(areaA)