Home        Store        Docs        Blog

Bar30 Pressure Sensor - Drifting Absolute Value

(Trevor Harrison) #1

Hi Blue Robotics,
I have a question regarding the performance of the Bar30 pressure/depth sensor. I’ve implemented the sensor as part of a small profiling float. I performed a test where the float performed a cycle in a swimming pool, starting the surface, descending to the bottom, and rising again, repeating four times. The test duration was 8 minute and in that time, the water surface ‘height’ changes 0.45 m (45 mBar), a rate of approximately 0.05 m/min (5 mBar/min). Ultimately, I’d like to perform depth control based on the Bar30 pressure sensor, but this amount of drift may be problematic.

From reading the TE Sensor Datasheet, and this forum discussion, I understand that the absolute pressure may be off by ±50 mBar or ±0.5 m (for these testing depths).
I did perform a benchtop test to assess longer term drift, but the test was not performed in water (I’ll probably be setting up this test later this week). There was a (nonlinear, tapering off towards end of test) drift from 0.03 m (1016 mBar) to 0.18 m (1032 mBar) over 50 minutes, average rate of 3.2mBar/10min. Quite a bit smaller than the pool test, but still considerable.

I noticed you did have a case of a faulty sensor, but the drift exhibited there was quite a bit more than what I’m seeing. Given the ‘long-term’ stability advertised by TE is ±30 mbar/year, even the observed drift seems strange.

Any advice would be much appreciated.

Bar30 Pressure Sensor - Datasheet doubts
Bar30 drift and out of bounds measurement
(Adam) #2

Hi Trevor,

Its true that the MS5837-30BA sensor in the Bar30 does have a certain, potentially noticeable amount of drift. I’ve done bench top tests, and have noticed very similar behavior to that present in your results. Some of this can be explained by ambient pressure changing slightly, but it is likely mostly the sum effect of variations in supply voltage,temperature, and long term sensor drift. I would say that the drift stated in the data sheet is only part of the story, but is mostly accurate- your results are certainly typical. Unfortunately, if this performance is not acceptable then the Bar30 might not be a good fit for your application.

This normally isn’t a problem for how most of our customers and we use it though. Usually with an ROV, any sort of movement of the water above the vehicle will be far more than any sort of sensor error, which is negligible anyway when just needing a depth to about the closest tenth of a meter. We’ve never had a problem with it during real world ROV flying.

However, we have made some custom “Bar02” sensors before with the MS5837-02BA pressure sensor. It can only read a depth up to about 20M, but it is more accurate with much less drift. It sounds like something like this may work better for you- if there is enough interest in it, we can look into making it a full product for high accuracy low depth use cases.

Is this something anyone else would fine useful as well?


(Bart Arts) #3

Hi, @Adam, I have a question.

You say the Bar02 will read up to 20m, but your product page says 10m. I see in the datasheet it’s up to 2000 mbar, so that would be 20 m in water pressure, ignoring the air pressure. How does this translate to real usage? I would like a range from 0-20m under water, and the bar30 is not accurate enough for me.

(Adam) #4

Hi @bartarts,

My mistake, the Bar02 pressure sensor can read up to 10m water depth, not 20m. This is 2 bar absolute with air pressure taken into account.


(Rory Gallaugher) #5

Hi there, I have an interest in a sensor with higher accuracy, but for a slightly different purpose.

I’m developing equipment for hydrometric monitoring which in my area requires an accuracy of 2mm. I’m looking to the world of robotics as the sensors are much more affordable to those sold in the scientific realm.

Why I’m here reading up on the forms is that I mistook the 2 mm resolution advertised for the Bar30 to be accuracy, so now, like others, I have a sensor that doesn’t fit my project requirements. I’ve looked at the Bar02 and my go with that as it seems that the accuracy is ~2cm in standard conditions, though if there were a sensor with 2mm accuracy you may open yourself up to a larger market.

Keep up the good work.