Tony Weidberg 1

LightABLE Production Status

ID Week May 2014

ID Week May 2014

Tony Weidberg 2

LTx Project

• Aim: use commercial package for VCSEL array to improve reliability.

• Integration into fully backward compatible TX PCB – Electrical and optical

connections– Space constraints

ID Week May 2014

LightABLE OSA (Reflex Photonics)

Production Status

• 363 LightABLEs delivered (500 ordered).– Delivery late because of problems with the assembly

company. – Final batch now promised for end May.

• Two flavours of PCB produced and components mounted (except LightABLE).

• > 301 LTx assembled.– 101 wire bonded– 84 completed burn-in

• Production rate between 20 to 40/ week.ID Week May 2014 Tony Weidberg 3

Tony Weidberg 4

Power QA @ Oxford

• DC power only at this stage

• Problem is that LightABLE devices have large offset power and we only use modulation (ac coupled links).

• Better power QA @ Cambridge

ID Week May 2014

Tony Weidberg 5

Change after Burn-in

• Burn-in: 72 hours @ 70c and 10 mA drive current

• No large change observed.

• Need to re-measure one underflow

ID Week May 2014 Change (dB)

Tony Weidberg 6

Cambridge QA

• Based on existing system to test AC performance– Loopback test TX RX

• Measure BER and perform timing scan width of eye diagram

• Measure power while adjusting modulation amplitude – requires programming uProc on LightABLE– Trick to deal with large baseline power (resistors added

to DRX-12 to bleed current). – Work in progress …

ID Week May 2014

Tony Weidberg 7

Plans

• Complete production and QA for LTx• Develop BOC s/w to allow operation of LTx or

AOC.• Install LTx in one crate in June.• Check performance– Ipin– BER in physics mode tests

• Install LTx in more crates ?

ID Week May 2014

Tony Weidberg 8

Backup

ID Week May 2014

Tony Weidberg 9

Changes Since Design Review

• Essential: – Make it easier for Infineon SMC connector to

connect to LTx in USA15– Improve thermal management

• Highly desirable: Allow option of decreasing optical power from LTx in case of saturation for DORIC4A.

• Mechanical reliability, consider use of under-fill for LightABLE BGA.

ID Week May 2014

Tony Weidberg 10

Mechanical Changes• Make mate/demate with Infineon

SMC easier.– Shortened length of MT guide

pins.– Added groove near end that was

cut to help with gluing. End sticking out is the uncut end.

– Use long guide pins in jig during reflow. Ensures that guide pins will be in the correct location and pointing the correct way.

ID Week May 2014

Tony Weidberg 11

Underfill

• Decided against use of underfill for BGA– Mismatch in CTE.– Very little insertion force required.– 100 mate/demate cycles with monitoring of power every

10 cycles no changes– Thermal cycling: no change in power.

ID Week May 2014

Tony Weidberg 12

Cooling• Thermal path from LightABLE to front panel

via Thermal Gap Filler (3 Wm/K)• OSA analysis DT from junction to ambient

=20°C ( next slide).

ID Week May 2014

Tony Weidberg 13

OSA Analysis

• Measure Peak wavelength lp• Scan Mark: Space Ratio (MSR)

and measure lp• Plot lp versus MSR• Fit slope and extrapolate to

50%.• Use calibration dlp/dT DT• Fit Results variable in quality

but reasonable consistency

ID Week May 2014

Tony Weidberg

AtLinks LightABLE- uProc

We can use a microprocessor to drive the LightABLE Engine’s I2C inputs, and the SPI lines currently used for the MDACs can be used to drive the µProc

But the MDACS we must still give the right output current for the interface circuit• This could be done using the µProc’s SPI lines• It could be carried out when the µProc is initialised• The MDAC_CLR command (one of the BOC_SPI signals) could be

used to force a uProc initialisation

This is what it would look like:

ID Week May 2014 14

Maurice Goodrick

Tony Weidberg

AtLinks LightABLE- uProc Connections

ID Week May 2014 15

Z PCB

Maurice Goodrick

Tony Weidberg 16

QA: Optical Power

ID Week May 2014

1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 More0

5

10

15

20

25

30

35

40

45

LTx

Power (mW)

Freq

uenc

y

-2.4 -2 -1.6 -1.2 -0.8 -0.4 0 0.4 0.8 1.2 1.6 2 2.40

5

10

15

20

25

30

35

40

LTx

After - Before Burn-in (dB)

Freq

uenc

y

<Optical Power> @ 50% duty cycle,after burn-in

Change with burn-in (72 hours @70C)

Tony Weidberg 17

LTx in SR1

• LTx optical power.• No T correction• Initial decrease

~1%.– No burn-in

preformed for this array probably ok but should run longer

Steve McMahon

Opti

cal P

ower

(mV)

Time (days)ID Week May 2014

Tony Weidberg 18

Steve McMahonO

ptica

l Pow

er (m

V)

Time (days)ID Week May 2014

Tony Weidberg 19

Summary

• Improvements following design review have been implemented.

• Production well underway.• Populate one crate in USA-15 in May• Need to continue long-term optical power

monitoring in SR1 and USA-15.

ID Week May 2014

Backup Slides

ID Week May 2014 Tony Weidberg 20

Tony Weidberg 21

OSA Thermal Studies• New metalwork for front panel of BOC

– Takes heat through thermal pad metalwork outside crate• OSA measured for 12 channels from first LTx with micro on board.• OSA spectra taken at MSR settings of 0, 10, 20 and 24 (maximum

value which gave clean waveform).• Very crude peak finding

– Found lmax corresponding to maximum power

– Found l1 and l2 corresponding to peak power -3dB

– Best estimate lpeak= (l1 + l2)/2

– Used difference between lpeak and lmax as estimate of error.– Error on MSR set to 1% (guess)– Fit lpeak vs MSR next slides

ID Week May 2014

Tony Weidberg 22

Fit Results variable in quality but reasonable consistency

ID Week May 2014

Tony Weidberg 23

Check this point (see next slide)

ID Week May 2014

Tony Weidberg 24ID Week May 2014

Tony Weidberg 25

Check Channel 8 MSR=20• Spectra looks ok• Programme found reasonable results• 3 sigma effect doesn’t change conclusions

855 855.05 855.1 855.15 855.2 855.25 855.3 855.35 855.40.00E+00

5.00E-03

1.00E-02

1.50E-02

2.00E-02

2.50E-02

3.00E-02

3.50E-02

ch 8 MSR=20

ch 8 MSR=20

From programme:Max power @ 833.330 nmPeak @ 833.327 nmLooks ok?

ID Week May 2014

Tony Weidberg 26

Summary

• Fits to each channel for slope• Use Carlos’s analysis for temperature

sensitivity– dl/dT=0.0673 nm/K

• Extrapolate to 50% duty cycle• DT = 20 ± 4 C• Improved thermal performance compared

to previous study DT = 29 to 43 C. – Improved metalwork takes heat through

thermal pad to outside of crate.

channel slope

0 0.0273

1 0.0235

2 0.0242

3 0.0296

4 0.0241

5 0.0251

6 0.0260

7 0.0251

8 0.0445

9 0.0271

10 0.0280

11 0.0259

ID Week May 2014