12–16 Sept 2005
Heidelberg
Europe/Zurich timezone

Final Results of the Industrial Production of CMS Tracker Analog Optohybrids

13 Sept 2005, 15:30
25m
Heidelberg

Heidelberg

Germany

Speaker

Dr Markus Friedl (HEPHY Vienna)

Description

Approximately 15,000 analog optical transmitter modules with 2 or 3 channels each will be installed in the CMS experiment to read out the Silicon Strip Tracker. These Analog Optohybrids were produced in Austrian and Italian industries from mid- 2003 to mid-2005. After assembly, each unit was thoroughly tested for electrical and optical properties and all results are stored in the CMS Tracker database, which allows a wide range of statistical analyses. We will discuss statistical distributions of important parameters and their impact on the system. Among minor issues, three significant problems occurred during the industrial production. Those will be described together with the corrective actions taken and the lessons learned.

Summary

The main components of the CMS Tracker Analog Optohybrids are pigtailed laser
diodes and the Linear Laser Driver ASIC, which allows flexible adjustments for gain
and bias current of each channel such that sample spread as well as temperature and
radiation effects can be compensated.

The production of optohybrids is exceptional to industry for several facts. First
of all, the number of units is rather small and hence does not allow cost-effective
automation such that a high degree of manuallabor is involved. Nevertheless,
specialized equipment like a bonding station and non-standard optical devices are
required. Thus, particular care was taken when selecting potential manufacturers
and prototyping was started with promising candidates to get an impression of their
abilities. Finally, the contract was mainly awarded according to this experience,
since all bids were quite similar.

Despite all precautions, three major problems appeared unexpectedly during the
series production. First of all, the package of the Linear Laser Driver ASIC
slightly changed but unfortunately caused shorts on some optohybrids. Since this
happened soon after starting the production, it was decided not to attempt repair
or workaround procedures but rather re-design the PCB to avoid such potential
shorts.

A few months later it turned out that the chosen fiber buffer can develop small
cracks and ruptures which do not affect functionality but carry a mechanical risk
in the long term. Several investigations were performed and finally, three actions
were taken: The buffer type was changed from acrylate to polyethylene which does
not show such ruptures.
Because of the staged pipeline production process, more than 5000 optohybrids were
still to be equipped with the old fiber type. Hence, each fiber was visually
inspected prior to assembly and the glue curing temperature was lowered to the
minimum possible value which yielded fewer buffer ruptures.

After a third of the Austrian production was completed, it was discovered that the
company had used other SMD capacitors than specified for all optohybrids produced
so far. The assembled type was not tested for radiation tolerance. Thus, the
production was halted and a re-qualification was performed in comparison to the
previously specified brand, comprising characterization before, between and after
gamma and proton irradiations at high statistics. Fortunately it turned out that
the used type was equally suitable as the one originally specified and
manufacturing was resumed without change.

During the whole production and those problems in particular, we recognized the
importance of keeping close contact to the company, and geographical proximity
turned out to considerably ease such matters.

Since the production will be finished in mid-2005, we can extract statistical
properties from the complete data set as stored in the database. The electrical to
optical gain is the key figure since it needs tuning for each channel. Values
should be as close as possible to the target gain while retaining headroom for the
Laser Driver settings to account for individual adjustments and radiation effects.

Moreover, the power consumption, depending on the accumulated radiation level, can
be predicted from bias and supply current statistics. Other properties such as the
analog bandwidth are achieved by design; yet the spread and shape of such
distributions give an indication of quality and uniformity of the devices.

Author

Dr Markus Friedl (HEPHY Vienna)

Presentation materials