FED 9U Analog Characterization Stefanos Dris CERN & Imperial College.

FED 9U Analog FED 9U Analog CharacterizationCharacterization

Stefanos DrisStefanos DrisCERN & Imperial CollegeCERN & Imperial College

CMS Tracker Stefanos Dris 20/7/2004

MotivationMotivation• Analog performance of the FED front end electronics.• Match on-FED performance with lab measurements of

bare ARx12 modules. • Check functionality of ARx12 matching capacitor switch

settings-i.e., do the FED’s front end analog electronics affect what capacitance setting should be used on the ARx12s?

MotivationMotivationTest SetupFED front end electronicsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

Test SetupTest Setup• For electrical probing

measurements, pulse generator set to trigger automatically and continuously.

• For pulse reconstruction on the FED, TSC provides synchronous triggers to both the pulse generator and the FED.

• FED clock skewed (0-31) in order to reconstruct the incoming pulses.

MotivationTest SetupTest SetupFED front end electronicsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

Trigger Setup Trigger Setup

30002500200015001000Time (ns)

Trigger to FED Trigger to Pulse Gen Optical Pulse

• Triggers to FED and pulse generator synchronized.

• Pulse generator set to output a pulse 2μs after receipt of trigger.

• FED capture length = 7μs, hence pulse appears approximately in the middle of capture.

MotivationTest SetupTest SetupFED front end electronicsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

FED Front End Analog FED Front End Analog ElectronicsElectronics

• Probing at:• 100Ohm ARx12 output resistor• Differential driver output• Differential input to ADC

MotivationMotivationTest SetupFED front end FED front end electronicselectronicsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

AOH Input and OutputAOH Input and Output

• Input pulses to AOH:• 800mV differential• 1.25V common mode

MotivationTest SetupFED front end electronicsResultsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

ARx12 OutputARx12 Output25

20

15

10

5

0

Settli

ng T

ime (

ns)

140012001000800Capacitance Setting (fF)

14

12

10

8

6

4

2

0

Rise Time (ns)

Settling Time Rise Time

4.393.70 3.36 3.13

16.3

7.35

20.29

25.58

• Receiver output is similar to what we see in production testing.

MotivationTest SetupFED front end electronicsResultsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

ARx12 Batch R4ARx12 Batch R4350

300

250

200

150

100

50

0

Coun

t

15014013012011010090NGK Bandwidth (MHz)

• Receivers mounted on the FED used for these tests comes from ARx12 batch R4.• From lot acceptance testing, and NGK bandwidth data, R4 modules perform best

@1000fF.• A few of the first modules of the batch have much lower bandwidths, but these were

used for other purposes (mounted on OFEDs).

Settling times for all channels of the 10 production test modules of batch

R4NGK supplied bandwidths for all channels of all 420 modules of

batch R4MotivationTest SetupFED front end electronicsResultsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

Differential Driver OutputDifferential Driver Output30

25

20

15

10

5

0

Settli

ng T

ime (

ns)

140012001000800Capacitance Setting (fF)

14

12

10

8

6

4

2

0

Rise Time (ns)

Settling Time Rise Time

4.73 4.29 4.00 3.81

27.9425.61

20.78

10.86

• Settles better at 800fF, as opposed to 1000fF. Overshoot in step response of the differential driver IC?

• Rise time only slightly worse than ARx12 output pulses.

MotivationTest SetupFED front end electronicsResultsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

Differential Driver OutputDifferential Driver OutputMotivationTest SetupFED front end electronicsResultsResultsLinearity test setupLinearity ResultsConclusions

• Settles better at 800fF, as opposed to 1000fF. Overshoot in step response of the differential driver IC?

• Rise time only slightly worse than ARx12 output pulses.

30

25

20

15

10

5

0

Settl

ing

Tim

e (n

s)

140012001000800Capacitance Setting (fF)

14

12

10

8

6

4

2

0

Rise Time (ns)

4.734.29

4.00 3.81

27.9425.61

20.78

10.86

Settling Time Rise Time

CMS Tracker Stefanos Dris 20/7/2004

TrimDAC OperationTrimDAC OperationMotivationTest SetupFED front end electronicsResultsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

ADC InputADC Input

• Differential input to ADC almost impossible to probe properly.• Strange offsets seen between measurements on the same C

setting, as well as slight changes in amplitude.

MotivationTest SetupFED front end electronicsResultsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

FED Reconstructed PulsesFED Reconstructed Pulses40

30

20

10

0

Settli

ng T

ime (

ns)

140012001000800Capacitance Setting (fF)

25

20

15

10

5

0

Rise Time (ns)

Settling Time Rise Time

12.26

31.94 33.2536.28

6.97 6.20 6.15 5.97

• Best capacitance setting is 800fF.• Thus, for an ARx12 module showing 1000fF best capacitance

setting in the tracker opto lab, the effect of the FED’s front end electronics is to change this best setting to 800fF.

• Rise time is a lot worse than in previous stages (~2ns more than at ARx12 output).

MotivationTest SetupFED front end electronicsResultsResultsLinearity test setupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

LinearityLinearity• Two non-inverting

outputs of the pulse generator used.

• Period set to 500ns, 1% duty cycle.

• FED set to acquire 280 points, 25ns apart.

• Pulse Generator controlled by LabView VI, changes each output’s level to obtain required resultant value.

• Level change and FED acquisition done manually, 31 steps from -0.5V to 0.71V.

• Measurement took less than a minute.

1.5

1.0

0.5

0.0

-0.5

Ampli

tude (

V)

100806040200Time (ns)

Pulse Generator Output A Pulse Generator Output B Difference

MotivationTest SetupFED front end electronicsResultsLinearity test Linearity test setupsetupLinearity ResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

LinearityLinearity

• Linearity better than 1% for an input range of more than 800mV was achieved, despite the dodgy method.

• With a more precise method, the linearity should be at least as good as the result shown here.

MotivationTest SetupFED front end electronicsResultsLinearity test setupLinearity Linearity ResultsResultsConclusions

CMS Tracker Stefanos Dris 20/7/2004

ConclusionsConclusions• Best C-setting should be selected taking into account the

FED front end analog electronics (up to ADC), not just the optical link components. In this case, the receiver should be set to 800fF, not 1000fF as the production tests would suggest.

• More FED channels need to be tested in this manner to see if the ‘change’ in best C-setting is a constant effect. That is, should all ARx12s with best C-setting of 1000fF in the production tests be set to 800fF when mounted on FEDs?

• What is the effect of selecting the wrong capacitance setting on the performance of the Tracker? In the last beam test, we wanted to check this by constructing Landau curves for all different capacitance settings, as well as for different sampling points.

MotivationTest SetupFED front end electronicsResultsLinearity test setupLinearity ResultsConclusionsConclusions

CMS Tracker Stefanos Dris 20/7/2004

Impossible is NothingImpossible is Nothing