Depth sensor evaluation¶
This reports details the measurements and analysis performed on the depth sensor.
It looks at the signal both as acquired by an oscilloscope as well as the platform itself. (Pico W)
Executive Summary¶
The sensor can provide an accuracy of 5%, if compensated for zero offset. The Pico W will not require an external voltage reference or external ADC for improved measurements.
1. Introduction¶
1.1 Purpose and Scope¶
The purpose of this report is to provide an analysis on the feasability of the selected level sensor; Gravity pressure level sensor.
It covers a zero offset error, and a single measurement depth error.
1.2 Background¶
The sensor works by measuring the pressure at the depth compared to the ambient pressure. The sensor outputs a current between 4 mA @ 0 mm and 20 mA @ 5000 mm. The depth can then be worked out using the simple formula
An current-to-voltage converted is used. Conversion is simply done as
1.3 Objectives¶
The report aims to answer these questions:
- What is the zero level offset?
- How accurate is the sensor?
- What is the (approximate) response time?
- Is the Pico W default ADC sufficient for sampling?
1.4 References¶
2. Methodology¶
2.1 Step response¶
The output signal of the voltage converter was sample with an oscilloscope (sample rate of 200 Hz) and the Raspberry Pico W. (Sample rate of 10 Hz)
The probe was quickly dropped into a bucket filled with 208 mm of water. (To the bottom)
The two signals were synchronized in post-processing.
The measured signal was converted to depth using the above forumlas.
The signal acquired by the Pico W was filtered using a simple moving average
with a window size of N=10. (1 s)
2.2 Long term variation¶
For evaluating the long term variation in samples, two sets were collected. The first set was collected with the probe at ambient pressure. The second set was collected with the probe at a depth of 208 mm of water.
Both sets were collected for two hours. The distribution of the two sets were plotted. The set collected at 0 mm was used to zero-offset the two sets.
The sets were collected with about 30 minutes between.
As the zero level had been measured on two separate days (one for step response, one for long term variation) the two zero levels could also be compared for longer variations.
3. Results¶
The zero-level offset was measured to 158 mm.
The level measured at 208 mm depth was 356 mm. Compensated for the zero-level offset, this equals an error of -10 mm.
The step response is approximately 300 ms.1
There is no overlap between a 0 mm depth measurement and a 208 mm depth measurement, over a 2 hour period. The standard deviation was measured to
- ±21.2 mm @ 0 mm depth
- ±20.7 mm @ 208 mm depth
Comparing the measure zero level from two separate days, it differed by 5 mm.
4. Discussion¶
There is a significant zero-level offset that needs to be compensated for. It should be evaluated whether this shifts over time, or if it is constant.
The relative accuracy is well within requirements. The measured error with a filtered signal is 5% of measured value, which is sufficient to meet the user needs.
A response time of 300 ms is well beyond what is needed, considering the environment it is placed in.
The data also shows that the default ADC configuration of the Pico W is sufficient. Adding a voltage reference or external ADC is not required.
Author: Oscar Aurin | Date: 2025-08-07 | Revision: A
-
This is likely not the sensor itself, but the duration it takes to bring the probe to the bottom. ↩