1CMS week 15/03/05, Module Production Meeting, Dean White - UCSB

Status of Sensor Bias Connection

Dean White

For the US CMS Group

2CMS week 15/03/05, Module Production Meeting, Dean White - UCSB

Sensor Bias Connection• In December 2004, while building 28 TOB

qualification modules, we noticed problems with the silver epoxy bias connection.

One module had an open connection and others had high resistance.

• After finding the first open sensor, we started to check the HV connection resistance in all modules before they were wire bonded

Of 51 TOB modules built, we found 29 sensors with 1K to over 40M ohms higher than normal (11.3K ohms) resistance in the bias connection.

We then stopped module production until we got a better understanding of what was happening and how to correct it.

Silver epoxy between frame and sensorSilver epoxy between frame and sensor

3CMS week 15/03/05, Module Production Meeting, Dean White - UCSB

Sensor Bias Connection

• But by increasing the amount of silver epoxy used so that a small amount squeezes out from under the sensor, we were able to find that the higher resistance was always between the epoxy and the sensor, not between the epoxy and the gold pad on the frame.

• During the production of these 51 modules, different procedures were tried to produce a consistently good bias connection. Added strictly defined mixing procedures for the Tra-Duct 2902. Increased the amount of silver epoxy used. Applied epoxy immediately before sensor placement. Switched to using Epotek 129-4 silver epoxy.

• None of these methods corrected the bias connection problem.

4CMS week 15/03/05, Module Production Meeting, Dean White - UCSB

Sensor Bias Connection

• ARCS testing of high resistance modules We wire bonded 4 modules with significantly high resistance in at least one

sensor (~75K, ~600K, ~2.3M & =>35M ohms) to see if they would be noticed in the ARCS test. All 4 modules ARCS tested fine.

• Even though the modules with high resistance tested fine, it seems this high resistance is an indication of a poor quality HV connection. Over time (2 weeks), we saw some of the initially poor connections improve, get worse and stay the same. Initially good connections (11.3K ohms) that we have monitored have not changed.

• We next received some HPK test structures from Manfred. Using the following methods we were able to reproduce the high resistance bias connection using the test structures and some aluminum test pieces.

5CMS week 15/03/05, Module Production Meeting, Dean White - UCSB

Sensor Bias Connection

We found that the way to duplicate the high resistance we saw in the modules was by placing the sensor test structures in the same controlled manner as the gantry (see set-up on to the right). Test pieces placed by hand always conducted fine.

At the right is a typical test strip with one aluminum piece pried off. The five aluminum pieces are taped to the granite surface. The tape simulates a 100um glue gap. Drops of epoxy are placed on the aluminum pieces near the edge and the sensor test strip is than lowered on to the aluminum pieces.

6CMS week 15/03/05, Module Production Meeting, Dean White - UCSB

1st Solution• It was found that removing the aluminum oxide layer on the backplane of

the sensor where the silver epoxy bias connection was made greatly reduced the connection resistance – usually less than 2 ohms. A stainless steel brush was used to abrade the sensor in the bias connection area.

SensorSensorabrasionabrasionSensorSensor

abrasionabrasion

7CMS week 15/03/05, Module Production Meeting, Dean White - UCSB

Stored Module Bias Resistance

Change in Sensor Bias Resistance (144 sensors)From May to July 2005

0

10

20

30

40

50

60

70

80

90

-250

-150

-50

-30

-10

-8 -6 -4 -2 0 2 4 6 8 10 30 50 150

250

ohms

Frequency

8CMS week 15/03/05, Module Production Meeting, Dean White - UCSB

Backside Bonding

Tec backside bondingTec backside bondingTOB backside bonding tabTOB backside bonding tab

• Solution to ever increasing bias connection resistance was to wire bond the bias connections• All previously built TOB modules have been retrofitted with

bonding tabs and then bonded.• All new TOB and TEC modules have bonded bias

connections.• All modules also have the original silver epoxy connection.

• Solution to ever increasing bias connection resistance was to wire bond the bias connections• All previously built TOB modules have been retrofitted with

bonding tabs and then bonded.• All new TOB and TEC modules have bonded bias

connections.• All modules also have the original silver epoxy connection.