New Results (11/01/02)

Features

• 12 x ABCD3TA (VDC and DORIC4A onboard BUT not yet bonded).
• Readout chips are bonded to 2 x Wedge detectors (12cm strip length).
• Module has only been tested using the Redundant Clock/Command inputs with Spydata outputs.
• Module is on a vertically mounted cooling frame within a Faraday cage.
• Module cooling frame.
• Module mounted on cooling frame.

Module Configuration

• All data taken with Edge Detect Off and data compression set to mode 1 (X1X) Level.
• Module trimmed at 1fC,  Trim Range 0
• FEBias = 220uA    FEShape = 30uA.
• VDD = 4.0V          VCC = 3.5V
• Detector Bias = +100V, ILeakage = 600nA

2 Test Configurations:

• Power Cables are shielded ribbon cable (approx. 3m in length), not shown in pictures above. The shields are grounded at the VME crate and the module cooling frame, which is connected to DGND via the support card.
• SCT LV-3 Power Supply used with external filter at the support card end.

A) Module mounted directly onto the cooling block (shunt-shield short-circuited?)

• Hybrid Temperature = 35 degrees (Chiller Temperature = 14 degrees).
• The module response (Computed from the Three Point Gain Macro), can be found here.
• Noise Occupancy for the the module (10^6 events per step).
• Noise Occupancy Scurves:  Link0 and Link1

B) Module electrically insulated from the cooling block using a Thermal pad (0.12mm thickness)

• Hybrid Temperature = 39 degrees (Chiller Temperature = 14 degrees).
• The module response (Computed from the Three Point Gain Macro), can be found here.
• Noise Occupancy for the module (10^6 events per step).
• Noise Occupancy Scurves: Link0 and Link1

For both test configurations the 0fC and 1fC values are 55mV and 101mV respectively.

Comments/Obversations

• Both configurations have similar Input noise and Gain characteristics, ~1500e ENC and 51mV/fC respectively.
• The Noise Occupancy at 1fC for Configuration A) is approx 5^10-4 whilst for Configuration B) is approx 10^-3.

• Both show instability in the Scurves for the Noise Occupancy scans BUT this occurs in the threshold range of 40mV - 60mV, which is close to the 0fC setting of 55mV.
• Chip5 on Link0 does not perform as well as the other chips. Because of this problem the noise occupancy is higher than expected. It is not yet understood why this chip does not perform as expected.
• For Noise Occupancy scans without shielded power cables the Scurves where very broad, this implies that there is a common-mode pick-up problem which could be perculiar to the setup here in Liverpool.
• The Scurves for Link1 tend to be better than that for Link0, maybe this could be due to the close proximity of this side to the cooling plate (which is held at DGND potential)?

Latest Results (10/01/02)

• Module mounted in VALENCIA cooling box. Picture of a module mounted in cooling box, the Support Card is screened with Al Foil.
• Module is directly mounted onto the cooling block (electrical contact made).
• Cooling box connected to AGND (wire link from support card).
• Power Cables are shielded ribbon cable (approx. 3m in length).The shields are grounded at the VME crate end only.
• SCT LV-3 Power Supply used, Jumper J31/J41 added, with external filter at the support card end.

Module Configuration

• All data taken with Edge Detect Off and data compression set to mode 1 (X1X) Level.
• Module trimmed at 1fC,     Trim Range 1
• FEBias = 220uA                FEShape = 30uA.
• VDD = 4.0V                      VCC = 3.5V
• Detector Bias = +100V,     ILeakage = 290nA
• Module Temperature = 38 Degrees (Chiller Temperature = 12 Degrees).

Results

Module Trimming Characteristics.

Module Response can be found here.

Noise Occupancy for the module (10^6 events per step).

Noise Occupancy Scurves: Link0 and Link1

The 0fC and 1fC mean values are 81.3mV and 126.6mV respectively.

Module configured with Shunt-shield

Setup configuration same as above BUT this time a thermal pad (0.12mm thickness) has been added between the cooling block and the module which acts as an electrical insulator thereby allowing implementation of the shunt shield.

Results (Using the same trim file as that for the module without the shunt shield)

Module Response.

Noise Occupancy for the module (10^6 events per step).

Noise Occupancy Scurves: Link0 and Link1

The 0fC and 1fC mean values are 82.5mV and 127.3mV respectively.

Noise Occupancy comparison for SCT LV-3 with/without the VME Gnd Jumper added (J31/J41): Link0 and Link1. The top plot shows the Noise Occupancy with the Jumper added, the bottom without 

Noise Occupancy comparison with/without the external filter. The top plot shows the Noise Occupancy with the filter the bottom without: Link0 and Link1 

Observations

All but 2 channels trimmable 

Module Input noise ~1500e ENC and Gain of ~51mV/fC at a hybrid temperature of 38 Degrees 

Noise Occupancy of  <1^10-4, though Shunt Shield implemented gives better performance 

Noise Occupancy Scurves look good, no evidence of instability. But there does appear to be a slight deviation in the Noise Occuapncy at 0.4fC for the module with Shunt Shield 

Module is sensitive to the environment, pick up. Maybe this is perculiar to the Liverpool setup