RLT4

12 ABCD2T chips (second batch) on thin film hybrid with Cu ground plane

* Hybrid * Hybrid Results * Comparison with Chip Test Results * Module * Module Results *

Hybrid Tests

Trimming was performed in Level Mode to optimize threshold uniformity for charge injection corresponding to VCAL = 10mV.  (This is supposed to give 1fC but evidence is amounting that in fact less charge is injected).

Shaper current is 30microA and bias current is 267 microA;  Vcc is 3.50V and Vdd is 4.00V as measured on the support card.


Hybrid Results

Trim: 10mV VCAL  = 100mV VTHR
 
HYBRID RESULTS in LEVEL MODE/X1X
Chips M0-E5
 
 
Chips M8-E13
 
 
Threshold scan 1fC
.ps (~1.8MB)
.fit
.sum
.ps (~1.8MB)
.fit
.sum
Threshold scan 2fC
.ps (~1.8MB)
.fit
.sum
.ps (~1.8MB)
.fit
.sum
Threshold scan 3fC
.ps (~1.8MB)
.fit
.sum
.ps (~1.8MB)
.fit
.sum
Threshold scan 4fC
.ps(~1.8MB)
.fit
.sum
.ps (~1.8MB)
.fit
.sum
Response Curve
.ps (271kB)
.rcdat
 
.ps (271kB)
.rcdat
 
 
HYBRID RESULTS in EDGE MODE/01X
Chips M0-E5
 
 
Chips M8-E13
 
 
Threshold scan 1fC
.ps (~1.8MB)
.fit
.sum
.ps (~1.8MB)
.fit
.sum
Threshold scan 2fC
.ps (~1.8MB)
.fit
.sum
.ps (~1.8MB)
.fit
.sum
Threshold scan 3fC
.ps (~1.8MB)
.fit
.sum
.ps (~1.8MB)
.fit
.sum
Threshold scan 4fC
.ps(~1.8MB)
.fit
.sum
.ps (~1.8MB)
.fit
.sum
Response Curve
.ps (271kB)
.rcdat
 
.ps (271kB)
.rcdat
 
NB. Level mode data was taken before fanin bonding, edge mode data was taken after fanin bonding.


Comparison with Chip Test Data:

 
Position on
Hybrid
Chip ID
nDead
(wafer test)
noTrim
(wafer test)
Gain
(wafer test)
Gain @2fC
(on hybrid)
Noise
(on hybrid)
Offset by
Response Curve
Offset by Noise Occupancy
nDead
(trim= 0)
noTrim (11.25mV VCAL => 120mV VTHR)
M0
3-6,17
 
13
58.9
 
 
 
 
 0
15
S1
3-6,10
 
23
60.2
 
 
 
 
 0
24
S2
11-6,12
 
30
55.7
 
 
 
 
 1
23
S3
11-6,13
 
26
60.0
 
 
 
 
 1
30
S4
11-6,14
 
25
60.0
 
 
 
 
 0
18
E5
11-6,16
 
32
60.9
 
 
 
 
 0
 27
M8
11-2,3
 
0
60.1
 
 
 
 
 0
0
S9
11-2,5
 
1
58.6
 
 
 
 
 1
1
S10
11-2,6
 
1
59.4
 
 
 
 
 0
1
S11
11-2,7
 
0
59.1
 
 
 
 
 0
2
S12
11-2,11
 
2
59.1
 
 
 
 
 0
1
E13
11-2,12
 
1
59.2
 
 
 
 
 0
1
 
x,y
 
 
mV/fC
mV/fC
ENC
mV
mV
 
 
 


Module Tests

The module has been retrimmed in Level Mode to optimize threshold uniformity for charge injection corresponding to VCAL = 11.25mV.  (This should correspond to approximately 1fC charge.)  Shaper current is 30microA and bias current is 267 microA.   The module operates in a standard QMW box, linked to module AG by the patch card.

Low Voltage

Vcc and Vdd are as tabulated below,  having been adjusted to give the correct voltages to the input of the first hybrid.  All measurements have been made with respect to the potential of the metal QMW module box.
 
 
 
Power
Supply
Support
Card Input
Top Hybrid
Input
Top Hybrid
Output
Bottom Hybrid Input
Bottom Hybrid Output
Vcc
3.610
3.597
3.512
3.461
3.426
3.407
AG
-0.041
-0.030
0.011
0.014
0.026
0.033
Vcc-AG
3.651
3.627
3.501
3.447
3.400
3.374
Vdd
4.079
4.065
4.006
3.964
3.939
3.925
DG
-0.058
-0.046
0.000
0.016
0.033
0.033
Vdd-DG
4.137
4.111
4.006
3.948
3.906
3.892
 
V
V
V
V
V
V
 

High Voltage

Here is the IV characteristic of the module, measured at 22 Celcius, 33% RH.
 
 
V
I
25
40
50
230
75
350
100
440
125
510
150
570
V
nA
 
 


Module Results

Trim: 11.25mV VCAL = 120mV VTHR

Preliminary!!!

  • List of Bad Channels.  These 150 channels have been masked from the subsequent plots.  Of these channels 141 are on the upper face of the module, which has the poorer quality chips, and only 9 are on the lower face.  The latter 9 channels comprise the 6 channels which did not meet the trim criteria, one dead channel plus two channels which are shorted together.  A total of 3 channels have been added manually to the list, the rest of the list was output by the trimming algorithm.
  • Noise Occupancy, Level mode, link0 and link1
  • Noise Occupancy, Edge mode, link0 and link1
  • Response curve, Level mode, link0 and link1
  • Noise by Charge Injection, 100V detector bias, Level mode link0 and link1
  • With the present trimming, the module is stable down to and including 97.5mV threshold in Level Mode and 100.0mV threshold in edge mode.  Hysteresis is less than 5mV in either mode.  If one chooses to ignore the low end of the response curve due to uncertainties with regard to the performance of the charge injection circuit at the low end, and projects a straight line back from calibration data taken with higher charges, this level of performance is consistent with the module being stable above 0.5fC threshold.  If one chooses to believe the response curve, the module is stable only above 0.9fC threshold.  It is my belief that the lower of these two values is nearer the truth.  It must be a priority to understand what is going on with the calibration circuit.

    By charge injection method this module returns the following noise figures:  1450 ENC and  1500 ENC for links 0 and 1 respectively.


    Modified 18.01.00
    Comments / questions / suggestions: please contact Peter W Phillips