Wrong reading of LDDS 75

Hello, I’m not sure if this is the right corner of this forum but I hope so.

I installed a Dragino LDDS75 ultrasonic sensor a few days ago. At first, everything worked fine but now I have the problem that it sends weird distances.
But let me start at the beginning.
I have a water tank which needs to be filled by a pump every day. So I have to check how much water is in the tank.
I wanted to do this with a Dragino LDDS75 sensor which is pointed towards the water. From the sensor to the Flor there are about 2.3 Meters.
For about 2 days everything worked fine and the sponsor gave me good measurements about the water level.
But since yesterday evening every few uplinks the sensor sends the exact distance of 718mm which is nowhere near the real water level.
I was thinking it could be the temperature but then it wouldn’t be the same misleading measurement every time right?

Do you have any ideas what the problem could be?

Thank you very much in advance! !

image|750x1000
image
image

image

This picture is an picture of the wrong readings.

image

This is a picture one day bevor when everything worked fine.

Is there something in the tank at that approx distance which is now visible to the sensor compared to when you set up? I had a '75 monitoring a water butt and the sensor moved slightly to then start picking up occationally on the edge of the butt vs the actual water level, similarly one concrete tank got to a low level and then ‘stalled’ on its readings - when physically inspected we found a root from a neighbouring tree had penetrated through a weakness in the concrete casting and was setting a fixed distance that was then being picked up by the sensor!

Good idea but I don’t think so…
But I could double check if it’s not 100% 90 degrees to the wall an slightly pointed agains the wal.

image

I’d point it a degree or two away from the wall and if you really care multiply the received reading by the cosine of the angle. Better yet, mount it on an arm so it’s more in the center of things, or a piece of strapping bridging the opening.

That wall is rough, and no emitter’s output is a perfect beam, there is going to be some signal reflection from the wall, and if it does or does not meet a threshold to count as a reading may be a situation that changes depending on factors you don’t want to be influenced by.

A better sensor firmware could probably sort this out (go for the strongest reflection not the first), but you may not have access to the raw signal evaluation algorithm.