miniSVS
Page 20
© 2017 Valeport Ltd
5. Appendix 1: FAQ’s
Why is the Data Different From My Old CTD Data?
Quite simply, the Valeport SV sensor is more accurate than anything else that has
previously been available. The CTD formulae (Chen & Millero, Del Grosso etc.) all have
errors in them – they were after all based on observed data taken over 30 years ago
using the best technology available at the time. The Valeport SV sensor simply highlights
those errors. This does raise an interesting point – if it is more important to you that your
data is consistent with old data, rather than accurate in its own right, then you are
possibly better off using a CTD (we would suggest a Valeport CTD, naturally).
How is it so Accurate?
Several reasons. Primarily, we use an advanced digital signal processing technique that
removes virtually all noise from the data, tells us the precise moment that the sound pulse
is both transmitted and received, and allows us to measure the time of flight with a
resolution of 1/100th of a nanosecond (10-11 seconds). Secondly, we have developed a
carbon composite material that doesn’t expand or contract with temperature, so our
“known distance” is a constant. Technically, the material will expand and contract
minutely, but over the operating temperatures of the probe, it is an almost immeasurably
small amount, and any change is included in our overall error budget. Finally, our
calibration method removes virtually all of the error sources associated with other
techniques.
But Don’t you Just Calibrate it Against Chen & Millero?
No we don’t – that would defeat the purpose. While the seawater formulae (Chen &
Millero, Del Grosso etc.) have inherent errors that are accepted as being at best
±0.25m/s, we use a different formula to calibrate the sensor. Del Grosso also published a
formula for speed of sound in pure water (with Mader, 1972), which is much more
accurate. In pure water, the only variable that can affect sound velocity is temperature
(assuming that you are at atmospheric pressure in a laboratory environment), rather than
both temperature and Salinity with the seawater equations. The Del Grosso & Mader
formula therefore has an error of just ±0.015m/s. By calibrating against this rather against
the error-filled seawater equations, we can achieve significantly better performance.
Is a Pure Water Calibration Valid?
Absolutely – the purpose of a calibration is just to compare (and adjust) the sensor output
against a known standard – it doesn’t really matter what that standard is, as long it is
precisely defined. Our standard happens to be pure water because it is the most
accurately defined standard available.
How Often Does a miniSVS Need Calibrating?
The SV sensor itself is remarkable stable. Since the entire timing system is digital, it is not
subject to the drift that analogue components often exhibit over time. The only part of
the system that can drift with time is the timing crystal itself. This is typically less than
±0.005m/s in the first year, and less than ±0.002m/s in subsequent years. We quite
confidently say that the SV sensor should remain within specification for several years.
However, the temperature and pressure sensors (if fitted) do exhibit greater drift with
time. It is our experience that in the majority of cases, performance can be maintained
by recalibrating at 2-yearly intervals. However, we are aware that many operators’ own
QA requirements state annual recalibration, and it is true that most instruments are
returned to us on a yearly basis.
