Wiener-Crate noise report
Post on 23-Feb-2016
27 Views
Preview:
DESCRIPTION
Transcript
Wiener-Crate noise reportF. Gonnella - Laboratori Nazionali di FrascatiPhoton Veto Workng Group, CERN 04/06/2013
Francesco Gonnella - LNF 2
Study of noise in Wiener crate
4 June 2013
Francesco Gonnella - LNF 3
Common mode noise setup
2 high impedance probes connected with resistor rheophores A 20 Gs/s 1GHz BW oscilloscope AC coupled
4 June 2013
Francesco Gonnella - LNF 4
Noise with crate unloaded
Crate unloaded Noise is ~1 V (peak-peak)
4 June 2013
+7V
-7V1.6 μs
Francesco Gonnella - LNF 5
Noise with loaded crate
Crate loaded with 4 LAV FE boards + 1 Tel Board Noise is ~0.5 V (peak-peak)
4 June 2013
+7V
-7V
Francesco Gonnella - LNF 6
Noise at 20MHz bandwidth
Crate loaded with 4 LAV FE boards + 1 Tel Board Noise is ~30 mV (peak-peak)
4 June 2013
+7V
-7V
Francesco Gonnella - LNF 7
Noise vs current (full bandwidth)
0 2 4 6 8 10 120
100
200
300
400
500
600
700
800Noise amplitude on power line (full BW)
PositiveNegative
Current [A]
Am
plit
ude
[mV
]
4 June 2013
Francesco Gonnella - LNF 8
Noise vs current (20 MHz BW)
0 2 4 6 8 10 120
5
10
15
20
25
30
35
40
45
50Noise amplitude on power line (20 MHz BW)
PositiveNegative
Current [A]
Am
plit
ude
[mV
]
4 June 2013
Negative power supply seems to perform much better than positive one
Francesco Gonnella - LNF 9
Common mode noise filter setup
4 June 2013
We used the setup recommended by Wiener with a Common Mode Filter and coupled the signal in AC
Francesco Gonnella - LNF 10
Filtered vs. not filtered (full BW)
4 June 2013
Without coil
With coil
Francesco Gonnella - LNF 11
Filtered vs. not filtered (20 MHz)
4 June 2013
Without coil
With coil
Francesco Gonnella - LNF 12
Frequency spectrum: full BW, no CM filter
4 June 2013
20 MHz
Francesco Gonnella - LNF 13
Frequency spectrum: full BW, CM filtered
4 June 2013
20 MHz
Francesco Gonnella - LNF 14
Frequency spectrum: 20 MHz, no CM filter
4 June 2013
20 MHz
Francesco Gonnella - LNF 15
Frequency spectrum: 20 MHz, CM filtered
4 June 2013
20 MHz
Francesco Gonnella - LNF 16
Consequences on LAV Front-end electronics
4 June 2013
Francesco Gonnella - LNF 17
Noise in LAV FEE test We performed two different runs for each test
No signal: all final conditions but pulser is Off Signal: pulser On (100mV signal, ~10mV threshold)
Board is moved outside the crate and powered it with linear PS Analogue sums are perfectly clean Test with TDC showed that the situation is noiseless
4 June 2013
Francesco Gonnella - LNF 18
Noise in LAV FEE inside crate
4 June 2013
No signal 100mV signal
Rate in Hz extrapolated using the total effective acquired time:Tacq = 3.5E3 * 0.5E-6 * 7 = 1.2E-2sThe noise rate goes to ~Hz if the nominal Thr is higher than 7mV (nominal)
Francesco Gonnella - LNF 19
Noise in LAV FEE linear PS (no crate)
4 June 2013
The board has been moved out from the crate to a table and the ±7.5V power is provided by an external linear power supply.
No signal 100mV signal
Francesco Gonnella - LNF 20
Noise rate in LAV FEE in Wiener
7.5 8.5 9.5 10.5 11.5 12.5 13.50.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
3.00E+05
3.50E+05
4.00E+05
4.50E+05
5.00E+05
CH 57CH 60Ch 61Ch 62
4 June 2013
Rate Hz
Threshold mV
Francesco Gonnella - LNF 21
Analogue sum: crate loaded
4 June 2013
Sum 1-4
Sum 28-32
Francesco Gonnella - LNF 22
Analogue sum: crate unloaded
4 June 2013
Sum 1-4
Sum 28-32
Francesco Gonnella - LNF 23
Time resolution on LAV FEE boards
4 June 2013
“Bad” channels are all in the last part of the board
“Bad” channels are the noisy channels previously shown
Time resolution in “good” channels is around 100 ps
In “bad” channels is around 200 ps
Francesco Gonnella - LNF 24
Conclusions (1/2) Crate specification are 20 mV differential p-p in 0 – 20 MHz
bandwidth; No specification is provided at higher frequency No specification is provided on common-mode noise! Without any filter we measured up to 0.5 V noise, peaked at
250MHz
The crate specs are respected if we measure the noise as defined in the specs With common-mode filter and bandwidth limitation to 20MHz Specifications on how the measurement should be done filter out
completely high frequency and common-mode noise
4 June 2013
Francesco Gonnella - LNF 25
Conclusions (2/2) Common-mode noise is there and has bad consequences
on LAV FEE: Last 12 channels are noisy up to ~10 mV threshold Minimum threshold due to crate instead of FEE performance
Further measurements must be done in ECN3 a.s.a.p. We are planning to perform a cosmic-ray test using A11
(rotated for this purpose) with Wiener crate and LAV FEE
Differential FEE should be immune to common-mode noise e.g. CEDAR did not experience any noise issue
No problem is foreseen for threshold higher than 15 mV e.g. CHOD and MUV should not experience problem (also using
LAV FEE)
4 June 2013
26
Thank you for you attentionfrancesco.gonnella@lnf.infn.it
4 June 2013
top related