'K4-208 Forward Middle Module'

02/08/01 Latest Results for module

Features

12 x ABCD3T, VDC, DORIC4A
Module has been tested using the Redundant Clock/Command inputs with Spydata outputs and Pseudo-Opto readout.
Module mounted on Liverpool Cooling Plate (Cooling block and Detector mounting pin on chilled Al plate).
The power/grounding scheme is as follows:

            VDD/VCC supplied from SCT LV-2 with unscreened ribbon cable.
            The screen connection of the power cable at SCT LV-2 end is connected to the VME chassis with a 100nF capacitor.
            The screen connection at the support card end is connected to the Cooling Plate.
            The module (hybrid) is grounded to the cooling box at the cooling point only.
            Support card electronics uses an external bench supply BUT their ground is linked to the module DGND.
            Module is mounted within a Faraday cage which is electrically connected to the Cooling Plate.

There is a single 'screened fan-in' on this module. This is located for channels linked chips M8, S9 and S10.

Module Configuration

All data taken with Edge Detect Off and data compression set to mode 1 (X1X).
FEBias = 220uA FEShape = 30uA.
VDD = 4.0V VCC = 3.5V
Data taken with module trimmed.
Tests and Results

The module IV characteristic is ok upto 175V detector bias (I = 1.5uA at 170V). The leakage current then runs away at bias >175V. It is not yet understood why this happens.

The module response (Computed from the Three Point Gain Macro), at a temperature of 75C, can be found here.
The module exhibits oscillation at low threshold, approx 50mV. The magnitude of oscillation increases with decreasing injected Qcal. The following Scurve plots, for various injected Qcal, show this effect:

                Scurves for injected Qcal = 2.0fC: Link0 and Link1
                Scurves for injected Qcal = 3.0fC: Link0 and Link1
                Scurves for injected Qcal = 4.0fC: Link0 and Link1

A 'Wishbone' was then attached to the module. This was added to to see if the module stability could be improved by evenly distributing the electrical connection of the hybrid to the frame (instead of previously this connection being made at the centreof the hybrid). There was no obvious improvement in stability but there is a reduction in the Input noise, as shown in the module response. It should be noted that the temperature of the module had dropped to 42C.

The Analogue and Digital Ground bond wire connections where changed as follows:

The AGND - DGND jumper bond pads on the top of the module had their 17um wire bonds replaced with 25um wire.
The respective bond pads on the underside retained the 17um wire but also had additional 25um wire bonds added in parallel.
Scurves for their respective links, Top and Underside, for Qcal = 3.0fC

A 'Star Ground' was implemented by linking the AGND and DGND on the Wiggly Kapton at the module end and then REMOVING the AGND - DGND wire bonds on the top side only, the underside stays as above.

Scurves for injected Qcal = 3.0fC: Link0 and Link1
Scurves for injected Qcal = 2.0fC: Link0 and Link1

Data was taken but this time with all AGND - DGND wire bonds removed from the hybrid:

Scurves for injected Qcal = 3.0fC: Link0 and Link1

The module becomes more unstable with all ground bonds removed and only maintaining a single AGND - DGND connection on the Wiggly Kapton adjacent to the hybrid connector.

Conclusions

Return to my homepage
For comments etc. please contact Ashley Greenall
Last modified: 14/06/2001