We have lots of these installed. No particular problems, but do avoid direct exposure to sunlight (plan on covering sensor) and respect the prescribed installation distances.
We have installed several units. They are reliable in a wide variety of conditions (in both mixed and unmixed tanks, Venturi flumes, open tanks, etc.) However, as stated by @WWSteev , you must respect the installation parameters: height to be measured, minimum and maximum distance of the measurement, and the position of the sensor. For example: do not place it in the center of a tank—instead it should be in an asymmetrical position precisely to avoid interference echoes. These specifications apply to any sensor, regardless of the brand. The nature of ultrasound technology is the reason for these "vulnerabilities". We have over 30 probes and have replaced two (which were more than 8 years old) over a three-year period.
I would advise you to change technologies. A water-level bubbler is a reliable and precise system without the drawbacks of ultrasonic devices.
What type of sensor?
Because by downloading the documentation from Endress, you can determine the beam angles. For example, with an FDU90 (range up to 3 meters), I observed the same phenomenon as you (max height to measure 70 cm); because too close to the wall, the sensor would sometimes detect it and signal an overflow. The FDU91 (range up to 10 meters) works great because the angle is less important. An easy fix is to put the sensor in a tube (diameter indicated in Endress manual) to orient the echo. I tried mapping, but it didn't work. It did though with the tube and with an FDU91.
It's a piece of cake. Watch out for foam though—that's all. Great product.
Regarding interference echo and positioning, the FMU90 has a convenient echo learning curve function. In a nutshell, with an empty or near-empty tank, this function allows you to record echoes and to disregard them during normal level capture. This eliminates the need for "perfect" positioning of the sensor. In my case, I installed a sensor holder near the wall, on a 6-meter deep tank with 4 large pumps and a vertical screen (i.e., a crowded space).
To start, download a copy of the operating instructions—not to be confused with the technical information they send with a quote—directly from the Endress website. Go to the section on echo suppression and follow one of the different methods described. For my part, I first emptied my tank and then used the "check distance" feature. I then set my interference echo suppression parameters manually. Once the range is defined (6 meters in my case), I launched "start mapping," making sure to keep my water level at its lowest. Once completed, the reliability of the sensor was flawless and no problem since (before it was every two weeks). Apparently you can even generate an echo curve to provide a visual of the interference and the actual level measurement.
Make sure you clearly define your operational needs in terms of the sensor itself. The technical information clearly indicates that you have to take into account the electromagnetic space around the sensor in order to properly size it. In general, for wastewater (and even drinking water), the FDU90, 91 and 92 do the job. Moreover, they are resistant to occasional immersion and infiltration of oxidizing gases (H2S, chlorine, etc.) In my case, the FDU91 is sufficient. So, to sum it up, the Prosonic S FMU90 sensor is (in my opinion) a very good product, easy to program and very reliable. NB: Carefully measure your tank and let Endress know if you require additional transmitters or if you want to use them for flow measurement. The same goes for programming the units. I plan to test them soon on a reactor that has to tackle grease and fibers that sometimes completely solidify on its surface. With the right calibration, it should work perfectly.