Study of the crosstalk in T0 signals for calibration runs

A. Sukhanov

3/26/2003

 

The following measurements have been taken: N2[A1] ,N2P0:9[A2] , N7[A3] , N7P0:3[A4] . Before each of the series we took measurements with all enabled channels to prove that the temperature drift is small.

 

Worksheets are here: old html, Excel

Results.

 

Supporting reports

03/11-16/03 P5, rate dependence, correlations P1-P0, N6-P0, coaxial-coaxial

03/16/03 Jitter of L0. CP, CN need to be delayed.

03/18/03 Signal distribution for P0, P4, P9 runs

03/19/03 Oscilloscope snapshots of the T0 ADCs

03/27/03 Oscilloscope snapshots of the T0 TDCs

 

Summary:

1)      The weird behavior in all T0 channels. There are significant shifts (from 10 to 60 counts!) between N2Pi and N2. All these pairs belong to same TDC but different long cables and discriminators.

2)      This is not a TDC problem. The same behavior is seen between N7Pi and N7, in this case both TDCs and cables are different.

3)      The signal TC70 which is connected to the same TDC as N7 (The TC70 is an ordinary TOF channel) does not shift when we enable/disable all T0N or T0P.

4)      This behavior is not due to parameters drift between the measurements. The measurements of N2, N2P0:3 taken next day have shown the same tendency.

 

Correlation between ADC noise and TDC noise in T0

4/12/03

Summary:

  1. There is no correlation between ADC noise and TDC noise on T0. See fig. 1a-6b.
  2. The ADC noises on different channels of the same side of T0 are highly correlated. See fig. 7a-9b.
  3. The ADC noises on different sides of T0 are highly anti-correlated. See fig. 10a,10b

 

Pileup (possible) effect in TDC

4/12/03

Summary:

The clue: For 10 lower positions (TDC_TC70<7000) the signals for channels 48:63 of this TDC09 (T0 channels) falls into the peak at 0 where the individual stops comes before the common start.

This is probably the evidence of the TDC non-linearity in the case when common start comes too soon after the previous stop.

This pileup condition can be selected using “471<TDC_L0_TDC”.

 

Strange T0N-T0P correlations in PP run

4/16/03

Summary:

1)      1)      The central blob in the Fastbus T0N vs T0P plot has diagonally striped structure. The stripes are separated by ~15 channels. See Fig. 2.

2)      2)      There is no such structure in the VME T0N vs T0P. See fig. 3.

3)      3)      The same structure exists in VME T0N vs Fastbus T0P (fig. 4.) and do not exist in VME T0P vs Fastbus T0N (fig. 5.)

4)      4)      Correlations VME T0N vs Fastbus T0N looks good (fig. 7.), but the same for T0P looks suspicious (fig.6)

5)      5)      The same correlations Fastbus T0N vs T0P for dAu runs does not have the diagonally striped structure fig. 8. The VME TDCs T0N vs T0P looks beautiful.

 

Conclusion

The striped diagonal shape in the Fastbus T0N vs T0P for PP run is due to a cross talk from T0N to T0P in the ribbon delay cable (180 ns) connecting NIM-ECL(B7) to TDC08. This cross talk is seen only in PP run because in that case both signals are coming at the same time. In the dAu run they were separated by 12 ns due to different vertex position.

 

Summary 2: Even with twinax cables the position of the T0P when only T0P0 is enabled depends on which channels in T0N are firing.

Compare runs 12177 and 12181 to runs 12178,79,80. The latter shifted down by 0.5 ns.

This is clear evidence that T0P is affected by T0N. From other hand the T0N seems to be not affected by T0P.

 

Phobos Bi-weekly meeting

05/08/03

Summary

 

Crosstalk in the SpectraStrip twisted pair cable

05/23/03

Summary

The same signals have been applied to the neighboring pairs.

A possible explanation could be due to different wire lengths of pairs 2 and 3 relative to pairs 4 and 5 but the directly measured difference in delay between talking pairs is not detectable.

 

No crosstalk between twinax cables

6/3/2003. Workbook

Summary:

  1. In the initial setup when the signal from the twinax goes following way:
    3m of SpectraStrip -> ECL/NIM -> 100 ns coaxial ->NIM/ECL -> 2m of SpectraStrip -> TDC
    The TC pulse on T0N5 gets shifted by 1.2 ns when it is coincides with the leading or the trailing edge of the T0P0. See run 12523.
  2. When the twinax is connected to the TDC using shortest possible way (2m of separate twisted pair) then no crosstalk observed. See run 12526.
  3. Adding 18 ns of SpectraStrip cable results in crosstalk of 250 ps. See run 12828

 

Crosstalk in discriminators

6/21/2003

Summary

  1. The discriminator LeCroy 3412 has no crosstalk between channels (see run 12558).
  2. The discriminator LeCroy 4413 has significant crosstalk ~125 ps (see run 12559)!
  3. The TDC 1875A has ‘acceptable’ crosstalk (amplitude ~100 ps, very narrow trouble zone). The crosstalk disappears when distance between channels is more than 2 neighbors.
  4. The crosstalk does not depend on the amplitude of the input signal.
  5. Using 3m of SpectraStrip results in doubling the crosstalk. E.g. the crosstalk from the neighboring channel, when disc. 4413 is connected to TDC via 3m of SpectraStrip is 250 ps. The crosstalk is 125 ps when using the separate wires. Actually, (see study 6/23/03) the 3m of SpectraStrip completely masks out the internal crosstalk of the discriminators.

 

Interlaced grounding

6/23/2003.

Summary.

1.      The crosstalk between the pairs of 3m long SpectraStrip cable with interlaced grounding is negligible.

2.      The crosstalk at the existing cables is so big that it make impossible to analyze data in the presence of neighboring hits.

3.      For long cables the interlace-grounded cable is only slightly better than the standard one. There is no sense to implement these cables.

 

 [A1]Enabled only T0N2 in LEDI

discriminators.

 [A2]10 measurements where two channels were enabled: T0N2 and one of the T0P

 [A3]Enabled only T0N7 in LEDI

 [A4]4 measurements where two channels were enabled: T0N7 and one of the T0P