Changes for page ANU Seismic Data Loggers

Summary

• Page properties (1 modified, 0 added, 0 removed)
• Attachments (0 modified, 1 added, 0 removed)
  • TC120_ANU_vs_CENTAUR.png

Details

Page properties

Content

...	...	@@ -14,7 +14,7 @@
14	14
15	15	Both the LPR-200 (or "Low Power Recorder" 200) and TerraSAWR are designed to use as little power as possible, and more or less use the same amount of power.
16	16
17		-At 100 Hz and with a GPS cable connected these loggers draw about 220 mW of power once the screen is off. Adding a sensor (e.g. a Trillium Compact 120) increases this to approximately 400 mW, or 0.4 volt-amps. So, in theory 7 Ah battery should last about 10 days without a solar panel, but in practice it seems to be a bit closer to 8 which may be due to variability in power drain while in getting GPS locks.
	17	+At 100 Hz and with a GPS cable connected these loggers draw about 220 mW of power once the screen is off (higher sample rates draw more power but only marginally, < 5mW). Adding a sensor (e.g. a Trillium Compact 120) increases this to approximately 400 mW, or 0.4 volt-amps. So, in theory 7 Ah battery should last about 10 days without a solar panel, but in practice it seems to be a bit closer to 8 which may be due to variability in power drain while in getting GPS locks.
18	18
19	19	For very sunny environments (latitudes < 30) a 20V 10 Watt solar panel should have no issue keeping these loggers alive over the summer months, and assuming unobstructed skies should also be fine over winter. However there is no harm in using 20 or even a 40 Watt panel, especially for high latitudes, coastal regions, or areas without a full sky view. In theory up to a 60 Watt solar panel is fine, but we don't recommend anything over 40 Watts and that amount of power is already overkill.
20	20
...	...	@@ -165,21 +165,18 @@
165	165
166	166	= Instrument Response =
167	167
168		-Both the TerraSAWR and LPR-200 use the same ADS1281 analog-to-digital converter chip and are designed to have identical instrument response. Depending on the output sample rate (e.g. 100 Hz, 250 Hz, 1000 Hz) amplitude response is consistently flat up to ~~100 Hz but phase response can vary above 1 Hz at 100 Hz (or 10 Hz at 250 Hz).
	168	+Both the TerraSAWR and LPR-200 use the same ADS1281 analog-to-digital converter chip and are designed to have identical instrument response. The ADC (analog to digital) chip in both loggers originally samples at 1024000 Hz and downsamples towards the output data rate via a 5th order SINC filter, then another four FIR filters. If the output is below 250 Hz, a final "pure" /5 decimation is done without any sort of FIR filter.
169	169
170		-The user can choose to apply a 2nd stage "sensor gain" by selecting an instrument type in the setup menu. This effectively selects a 10 Vpp (e.g. short period sensors), 20 Vpp, 40 Vpp (most broadband sensors) regime to match the sensor's sensitivity. This has the effect of doubling amplitude from 10v to 20v, or quadrupling from 10v to 40v. If you have set your sensor correctly (and the signal isn't clipped!) you can "correct" this by simply multiplying your data by 0.5 etc.
	170	+{{info}}
	171	+The Stage 3 SINC coefficients (600+) during the initial 1024k > 16k decimation were left off as they slowed down the process x10 and contribute very little (< 0.3 db, < 0.31 ms) to the end result
	172	+{{/info}}
171	171
172		-Another important thing to note is that the group delay associated with late stage FIR filters is **automatically applied in the logger**, hence there is no need to apply this in the response. These tend to max out at 0.124 seconds for most output sampling rates (0.062 s for 100 Hz).
	174	+The user can choose to apply a 2nd stage "sensor gain" by selecting an instrument type in the setup menu. This effectively selects a 10 Vpp (e.g. short period sensors), 20 Vpp, 40 Vpp (most broadband sensors) regime to match the sensor's sensitivity. This has the effect of doubling amplitude from 10v to 20v, or quadrupling from 10v to 40v. If you have set your sensor correctly (and the signal isn't clipped!) you can "correct" this by simply multiplying your data by 0.5 etc. This gain manifests itself in stage 2 in the response information.
173	173
174	174	Instrument response can be downloaded from IRIS [[Nominal Response Library>>https://ds.iris.edu/ds/nrl/]] if need be, or [[directly from us>>http://auspass.edu.au/data/logger_response]] , or by downloading the response of an equivalent sensor at AusPass (e.g. get_stations(level='response') ).
175	175
176		-The response info from IRIS-NRL is the "full" version which (in theory!) perfectly describes the data logger's bias on the data. However this is in many way overkill and at the cost of 1) increased metadata size and, more importantly, 2) increased CPU demand in the response removal process. Testing has shown that for signals below 100 Hz, the "full" response offers little to no benefit and can increase the time it takes to remove the response for a 1 hour window of 100Hz data by a factor of x20 or more. For earthquake arrival data this is often negligible, but for data intensive tasks like ambient noise cross-correlations this can be a severe hindrance. Thus we have created a parallel version of this response which removes the SINC and FIR filters completely. These are labelled "fast" in our [[local response archive>>http://auspass.edu.au/data/logger_response]] and essentially truncate response stages 3 onwards into a "fake" decimation step from 1024000 Hz to the desired output samplerate with no filtering whatsoever.
	178	+[[Huddle test comparing a Trillium Compact 120 + TerraSAWR vs a Trillium Compact 120 + Nanometrics Centaur (M8.AUANU)>>image:TC120_ANU_vs_CENTAUR.png||data-xwiki-image-style-alignment="center"]]
177	177
178		-In the below we show both responses applied to a test signal with a frequency range of 1000 seconds to 100 Hz. The maximum discrepancy in signal is less than 0.01% (1.0001) which is far below what you should expect from the mechanical inconsistencies intrinsic to the sensor itself. Thus, we strongly advise users employ the "fast" version of this response information and it is what we use for our networks by default. If you are recording at 1000 Hz, or care deeply about signals above 100 Hz (so recorded at 250 or 1000 Hz), please use the full response. Any questions, please ask!
179		-
180		-[[Testing the "full" and "fast" versions of the ANU data logger response on synthetic 250 Hz data from 1000 seconds to 100 hertz. For all intents and purposes, they are identical.>>image:full_vs_fast.png||data-xwiki-image-style-alignment="center"]]
181		-
182		-
183	183	= ANU TerraSAWR (Gen 3, FW 3.5a, 2017- current) =
184	184
185	185	Not sure there's much left to say
...	...	@@ -215,6 +215,9 @@
215	215
216	216
217	217
	215	+
	216	+
	217	+
218	218	(% class="box" %)
219	219	(((
220	220	= TerraSAWR Specs =

TC120_ANU_vs_CENTAUR.png

Author

...	...	@@ -1,0 +1,1 @@
	1	+XWiki.robert

Size

...	...	@@ -1,0 +1,1 @@
	1	+57.0 KB

Content