2010/04/21:Lynn Dietz Brief History of NCSN Coda Termination Definitions _______________________________________________________________________________ Fred Klein's Hypoinverse document and Jerry Eaton in talking about the history of coda termination for NCSN both describe the "standard" NCSN station as an L4C velocity sensor, sent thru the NCSN telemetry electronics black box (ending with a discriminator) at 15 dB attenuation (this comes from looking at the distribution of attenuation settings in the network, which was peaked with an even number of stations at 12 and 18 dB, giving you the average of 15 dB) and recorded in Menlo Park by something. Originally, that something was a Develocorder, which was replaced by Rex Allen's RTP, CUSP Tustin 12-bit digitizers and finally Earthworm National Instruments 12-bit digitizers. All pick and coda constants were originally determined for this "standard" NSCN station. Coda termination was originally defined by Willie Lee on the develocorder viewer as the point where signal reached 1 cm peak-to-peak amplitude. When NCSN started analyzing the signals with a computer algorithm (Rex's RTP), Rex Allen, Jerry Eaton, Willie Lee, and others conducted a brief study to determine a new coda termination method that would be consistent with the original Develocorder definition. This new coda termination was defined as the point where the discriminator output reached an average absolute amplitude of 60 mV (over a 2-second window). The parameters used by the picking algorithm are specified in digital counts. Rex's RTP system: coda term = 60 counts (1 mV/count) CUSP & Earthworm A/D: coda term = 49.14 counts (5V/4096count = 1.22 mV/count) For a standard 15dB station, the coda termination of 49.14 counts on the Earthworm A/D is equivalent to a ground motion of: EWcodaterm * EWsensitivity@15dB * L4Cmotorconstant = endcoda-groundmotion 49.14count * 0.3519uV/count * 1cm/s/10e6uV = 1.729e-5 cm/s To go from coda duration to coda magnitude, hypoinverse uses a gain correction factor (0 for the standard 15dB station). The correction factor takes into account the differences in the hardware from the sensor output to the digitizer output; in essence it corrects for the recording system sensitivity. OK, that's the NCSN's history of processing analog-telemetered seismic data. But now we need to process seismic data from sensors other than L4Cs recorded by various digital instruments. We need to come up with scaling factors between the "standard" NCSN analog Earthworm-recorded station and these new recording systems so that we can once again define consistent picking and coda termination parameters. I want to choose parameters that terminate the coda at the above 15dB ground motion value because we should not have to use any gain correction in the duration magnitude calculation. Let's determine the digital counts per unit ground velocity for the various digitizer/sensor combinations. That will give us the scaling factors to use for our picking and coda constants. Scaling between "Standard NCSN station" and other NCSN recording systems _______________________________________________________________________________ The Standard NSCN station (@15dB): Earthworm/L4: 2841716 count/cm/s = 1.0EW ( 1count/0.3519uV * 10e6uV/cm/s ) Other digitizer/sensor pairs processed at NCSN: DST/L4C: 145666 count/cm/s = 0.05EW ( 1count/6.865uV * 10e6uV/cm/s ) HRD(old)/L4C: 964320 count/cm/s = 0.34EW ( 1count/1.037uV * 10e6uV/cm/s ) HRD(new)/L4C: 784314 count/cm/s = 0.28EW ( 1count/1.275uV * 10e6uV/cm/s ) HRD(old)/HS1: 432015 count/cm/s = 0.15EW ( 1count/1.037uV * 0.448*10e6uV/cm/s ) HRD(new)/HS1: 351373 count/cm/s = 0.12EW ( 1count/1.275uV * 0.448*10e6uV/cm/s ) HRD(higain)/HS1: 35137255 count/cm/s = 12.36EW ( 1count/.01275uV * 0.448*10e6uV/cm/s ) TRI(2V-PP)/HS1: 3584000 count/cm/s = 1.26EW ( 1count/0.125uV * 0.448*10e6uV/cm/s ) TAU(2V-PP)/GS11D-O: 2552000 count/cm/s = 0.90EW ( 1count/0.125uV * 0.319*10e6uV/cm/s ) TAU(2V-PP)/GS11D-S: 2376000 count/cm/s = 0.84EW ( 1count/0.125uV * 0.297*10e6uV/cm/s ) TAU(2V-PP)/SM6-O: 2304000 count/cm/s = 0.81EW ( 1count/0.125uV * 0.288*10e6uV/cm/s ) TAU(2V-PP)/SM6-S: 1456000 count/cm/s = 0.51EW ( 1count/0.125uV * 0.182*10e6uV/cm/s ) TAU(40VPP)/STS-2: 6000000 count/cm/s = 2.11EW ( 1count/2.5uV * 15*10e6uV/cm/s ) HRD(o)/40T: 7714561 count/cm/s = 2.7EW ( 1count/1.037uV * 8*10e6uV/cm/s ) HRD(n)/40T: 6274510 count/cm/s = 2.2EW ( 1count/1.275uV * 8*10e6uV/cm/s ) HRD(n)/40Tmod: 62745098 count/cm/s = 22.1EW ( 1count/1.275uV * 80*10e6uV/cm/s ) HRD(n)/3ESPC: 11764706 count/cm/s = 4.14EW ( 1count/1.275uV * 15*10e6uV/cm/s ) HRD(n)/STS-2: 11764706 count/cm/s = 4.14EW ( 1count/1.275uV * 15*10e6uV/cm/s ) Rftk72/40T: 4195070 count/cm/s = 1.5EW ( 1count/1.907uV * 8*10e6uV/cm/s ) Rftk130/L4C: 629129 count/cm/s = 0.22EW ( 1count/1.5895uV * 10e6uV/cm/s ) Rftk130/SS-1: 2139037 count/cm/s = 0.75EW ( 1count/1.5895uV * 3.40*10e6uV/cm/s ) Rftk130/STS-2: 9436930 count/cm/s = 3.3EW ( 1count/1.5895uV * 15*10e6uV/cm/s ) K2/L4C: 3355705 count/cm/s = 1.2EW ( 1count/0.298uV * 10e6uV/cm/s ) Makalu-gain1/SS-1 1447026 count/cm/s = 0.51EW ( 1count/2.3842uV * 3.45*10e6uV/cm/s ) Makalu-gain2/SS-1 2894053 count/cm/s = 1.0EW ( 1count/1.1921uV * 3.45*10e6uV/cm/s ) Qterra/STS-1: 9600000 count/cm/s = 3.4EW ( 1count/2.5uV * 24*10e6uV/cm/s ) Qterra/STS-2: 6000000 count/cm/s = 2.1EW ( 1count/2.5uV * 15*10e6uV/cm/s ) Q680/? at CI: 9663864 count/cm/s = 3.4EW Qterra/GS-11: DST24/L4C: 838926 count/cm/s = 0.30EW ( 1count/1.192uV * 10e6uV/cm/s ) Q80/CMG3-NSN: 5810200 count/cm/s = 2.04EW ( 418000count/V * 13.9V/cm/s ) Instrument constants from Jim's notes (also in Klein's Hypoinverse Document) _______________________________________________________________________________ Recording System Sensitivity (Fred's F): 1 - NCSN analog = 0.9954 microv/count @ 24dB (Klein's Hypoinv doc) eworm A/D = 0.4977 microv/count @ 18dB (by math: 24dBvalue/2) = 0.3519 microv/count @ 15dB (by math: 18dBvalue/sqrt(2)) 2 - NCSN DST16 = 6.865 microv/count 3 - RefTek72 = 1.907 microv/count 4 - NMX HRD old = 1.037 microv/count (old stations-Bay Area mostly) NMX HRD new = 1.275 microv/count (all new stations & old Episensor channels) NMX HRD hg = 0.01275 microv/count (high gain at Mammmoth Deep Hole) NMX TRI 40V = 2.5 microv/count (Tri305 set at 40V PP, software gain=1) NMX TRI 2V = 0.125 microv/count (Tri305 set at 2V PP, software gain=1) NMX TAU 40V = 2.5 microv/count (Taurus set at 40V PP, software gain=1) NMX TAU 2V = 0.125 microv/count (Taurus set at 2V PP, software gain=1) 5 - Quanterra = 2.5 microv/count 6 - K2 = 0.298 microv/count (from K2 manual: +/-2.5V, 24bits, gain=1) Makalu-gain1= 2.3842 microv/count (from KMI email: 2^23/20 = 419430 count/v at gain=1) Makalu-gain2= 1.1921 microv/count (from KMI email: 2*2^23/20 = 838860 count/v at gain=2) 7 - NCSN DST24 = 1.192 microv/count (as per Gray Jensen 7/28/2000) Reftek130 = 0.7947 microv/count (+/-5V) for internal MEMS on ch1-3 Reftek130 = 1.5895 microv/count (+/-10V) unity gain (1) for external sensor on ch4-6 Reftek130hi = 0.0497 microv/count (+/-10V) high gain (32) for external sensor on ch4-6 Qterra Q80 = 418000 count/V NOTE: These sensitivities are for microvolts of sensor output/digital counts Sensor Motor Constants: 1 - L4C = 1 v/cm/s 2 - FBA-23 = 1.25 v/g = 0.001278 v/cm/s/s 3 - CMG-40T = 8 v/cm/s (normal Guralp CMG-40T) CMG-40Tmod = 80 v/cm/s (modified Guralp CMG-40T) 4 - HS1 = 0.448 v/cm/s 5 - Wilcoxin = 10 v/g = 0.01022 v/cm/s/s 6 - Dilatometer = 1 v/0.000001 strain 7 - Streckeisen STS-1 = 24 v/cm/s 8 - Streckeisen STS-2 = 15 v/cm/s 9 - Episensor = 5 v/g = 0.005112 v/cm/s/s 10 - Oyo GS-11 = 11 - Guralp CMG3-NSN = 13.9 V/cm/s (vertical) 12 - S-13 = 3.65 V/cm/s 13 - Ranger SS-1 = 3.45 V/cm/s (as per KMI) 14 - SM6-open = 28.8 V/m/s = 0.288 V/cm/s (as per SM6 spec sheet) SM6-shunt = 18.2 V/m/s = 0.182 V/cm/s (as per Gray's calculations) 15 - CMG-3ESPC = 1500 V/m/s = 15 V/cm/s (Guralp CMG-3ESP Compact) 16 - GS11D-open = 0.319 V/cm/s (as per Oyo spec sheet 4.5hz,350ohm) GS11D-shunt = 0.297 V/cm/s (as per Taka-aki from Oyo chart for 0.5 damping)