Understanding the Poor Resolution from Test Beam Run. aah

Transcription

1 Understanding the Poor Resolution from Test Beam Run aah 1

2 2004 Straw Test beam results! Doc # 3308 v#3 by A. Ledovskoy " Using Data from 2004 Test Beam " Used triplet method for beam nominally perpendicular to Straw Planes, w=(r 1 +r 3 )/2 +r 2 =2 mm " Resolution in a restricted region (actually a sweet spot) was ~160 µm. We expected better than that. Thought it would be trivial to beat 150 µm averaged over all straw radii. I have tried to analyze a small subset of data, but come to same conclusion. " Efficiency near edge is worse that we expected Could be due to non-round straws giving gaps between straws in module. Also if gas gain/thresholds worse than thought r 1 r 2 r 3 2

3 Straw Resolution Requirements! Single straw resolution set as 150 µ.! Overall momentum resolution dp/p <1% up to 100GeV/c " Seems to be actually Single Straw Resolution of ~250µ From TDR want dp/p <1% at 100GeV 3

4 3e- thresh 1e- thresh Look at Possible contributions to lousy Single Straw resolution (even if ok for dp/p requirement)! We know that the straw resolutions depend upon the thresholds in the detector. " ~3 clusters/mm, <#e> cluster ~ 3 " The lower the threshold, the better. We detect the first cluster(s) over threshold, and call that r. r " Note that tendency in measurement is to OVERestimate r Simulated signal into ASDQ discriminator due to a track Analog Signal Modeled on ASDQ pulse shape includes cluster creation, reflections, attenuation, etc. See doc #369 &1443 notice time slew of threshold from 1e- to 3e- ~10ns or ~400 µm. 4

5 Understanding the Test Beam results! Want to find a way to quantify the entire straw system " Gas Gain We have measured the Gas Gain in ArCO2 by using a special setup using a single straw, with the cathode at -HV and the anode at ground. Using an 55Fe Source (226 e- cluster), we have integrated the total cluster signal on a cable, and by using the know cable/input capacitance, and input resistance, have calculated the gas gain. BTeV Doc 448 " Preamp Gain, and effective charge integration Up to now have used ASDQ Documentation " Discriminator charge Thresholds vs input voltage! Resolution depends upon all three items! Realized that our plateau measurements, when combined with the gas gain measurement, allows an absolute determination of the straw system cluster threshold versus straw HV and ASDQ threshold inputs. 5

6 Gas Gain in BTeV Straw (from Doc 448) 6

7 Parameterization of Straw Plateau with 55Fe Source (see Doc#1945--M.Morales SIST report) Two plateaus are due to full energy (5.9keV Mn x-ray--lower edge) and Argon escape peak (2.9keV--higher edge), Assune straw! function response is gaussian 1 2. x' µ + 1 ', ) 2- # * (( x) = e 2$# x CDF( µ,# ) = % (( x) dx '& Fit = " * CDF1 +! * CDF 2 =µ 1 =" 7

8 How to find the effective Straw Thresholds (including response of electronics)! Use Plateau Curves taken at end of Test Beam Run (2004) " Both 300mV and 500mV thresholds We had noise problems, during the test beam run which forced us to run with the 500mV threshold. Files plateautest1_300_fix.dat, plateautest1_500_fix.dat Data taken with 55 Fe source placed opposite straw end (1m away) from Electronics package. 8

9 Variation of Edge Voltage (µ( 1 ) for straws within a single module (Doc#1945--M.Morales SIST report)! µ 1 is a good parameter to study gain variations across the length of the straw.! Strong dependence with the gain and the threshold of the system.! Values for the straws in one module in the straw prototype detector ranged between 1090 and 1110 V.! Want to interpret variation with respect to gain µ 1 µ 1 for all 48 Straws in module #1 9

10 Gas Gain at µ 1 Voltage (Doc#1945--M.Morales SIST report) Using fit from previous slide (for Ar-CO2) gain(v) = 0.009e V! Obtained from previous experiments on the Single Straw Prototype.! Measures the gain needed by the signal to reach the threshold.! Gain values range from 1800 to

11 DETAILS: µ 1 Voltage vs Source Position (Doc#1945--M.Morales SIST report) Simulation with 0cm Preamp/disc 1 m straw * Source Unterminated end Simulation with 100cm Note that plateaus were made at ~80-90cm, while test beam data were taken near cm, or a difference of ~18V 11

12 Results! Correcting for difference between 55Fe and testbeam positions (~18V)! From the 300mV threshold runs " Edge at 1083 V + 18V = 1101V for 226 e - cluster Gas Gain = 2040 using formula from earlier slide! From the 500mV threshold runs " Edge at V = 1154V for 226 e - cluster Gas Gain =3693! Gas gain needed for arbitrary cluster size n " GasGain(n e -, thresh)=(226/n) *GasGain(226e -, thresh) " Or GasGain(1e - ) =226* GasGain(226e - ) (Thresh used in later run) Cluster threshold gas gain (300mV Thresh) HV for 300mV Thresh gas gain (500mV Thresh) HV for 500mV Thresh 1 e - 4.6e5 1584V 8.4e5 1638V 2 e - 2.3e5 1523V 4.2e5 1576V 3 e - 1.5e5 1486V 2.8e5 1540V gas gain (350mV Thresh) 5.8e5 2.9e5 1.9e5 HV for 350mV Thresh 1605V 1543V 1505V 12

13 From 2004 Test Beam Run! Typical runs used 1400V and 1450V " 1400V, gas gain ~5.8e4 8 e - thresh for 300mV thresh 14 e - thresh for 500 mv thresh " 1450V, gas gain ~1.0e5 4.6 e - thresh for 300mV thresh 8.4 e - thresh for 500 mv thresh " These are pretty high cluster thresholds, which might explain the relatively poor resolution 13

14 Results from NEW short Test Beam run Feb 24, 2005! Setup Straw Detector to run 120GeV p at " 1400V, 1450V, 1500V, 1525V All with 350 mv thresholds. Beam located at ~0.5m from preamps (roughly half length of straw) " Formed triplets over beam profile to calculate w= (r 1 +r 3 )/2 +r 2 Used t(r, HV) relationship from Garfield calculation from Tom Coan Adjusted t offset so that w centered at 2mm. See profiles, next slide # w ~! 1.5 * # straw ignoring the real correlations in # 1, # 2, and # 3 HV 1400 V 1450V 1500V 1525V # straw (measured - averaged over triplets) 264µ 210µ 170µ 149µ 10 e e e e - r1 Approximate Cluster Threshold (calculated) r2 r3 14

15 Representative w profiles (red lines are a fit to data) w profiles from 3 adjacent w profiles from 3 adjacent 15

16 Comments on ASDQ Performance! From Dave Ambrose (4/23/02) " Threshold(mV) = 65mV/fC*Threshold(fC) +100mV fc is amount of effective charge seen by shaping of Straw Signal in ASDQ. From ASDQ specs, the ", ratio of the effective charge /total charge is ~7% Taking into account stray capacitance and the division of the signal toward and away from preamp total charge =gas gain*#electrons in cluster, collected over µs effective charge is charge in first ~8ns or so Note ASDQ spec assumes signal terminated at far side, so no reflection Q eff (300mV Thresh) = 3.1fC Q eff (500mV Thresh) = 5.1 fc " Using the Threshold formula, and calculating the effective charge in the case of the straw prototype Qeff="*gasGain*(226 e -) * 1.6e -19 C/e (using Fe55 sourc) Some details for calculating "" average over the two thresholds of 300mV and 500mV corrected for attenuation (0.82) and presence of reflected pulse (0.5) in the straw plateau data Get " = 3.3%, ~factor of 2 WORSE than nominal ASDQ specs--why?? 16

17 Conclusions! Poor resolution from 2004 Test Beam run was primarily due to running at too low a gas gain and/or too high an ASDQ discriminator threshold.! Issues remaining: " Why do we seem to get factor 2 less charge into ASDQ discriminator than ASDQ nominal specs imply Stray Capacitance too high? Impedance mismatch between straw (~300 $) and COT board? Straw fast response different than ASDQ design? Prototype HV Network not optimal? " Does the ASDQ really have enough gain? The prototype detector seemed to run in Test Beam ok at 1525 V (gas gain ~2e5), but was pretty unstable at 1550 V. Maybe a real problem for actual BTeV running requirements. Also aging might be fierce at these high gas gains. Earlier calculations assumed 1e5 gas gain (which we thought would be on high side) " Noise levels need careful attention in the Straw design! These would all have been issues to consider before ordering ASDQ s. 17