Just brainstorming here, but is it possible that being under constant pressure is somehow altering the piezo property of the transducers and resulting in loss of sensitivity, accompanied by resistance reduction, over time? (I haven't researched this and I agree it's unlikely).

 

I am am wondering if a capacitive probe system may be a better choice for sensing the fluid level in the container. That could have the sensor be a rod totally enclosed by a suitable plastic cover, with the tank being the other side of the capacitor. The change in capacitance would be very proportional to the level of beer in the tank, and if the plastic tube were Teflon or similar it would not become coated with the contents. And quite possibly the same drive signal would work, fed through a high resistance, and the average voltage developed would reflect the level of the contents. This is just a suggestion as to another option, I think such systems are available but not cheap. But an examination of available products could provide a good understanding of how they work.

 

I am am wondering if a capacitive probe

Funny you should mention that.... The only time I can remember using a sonic level detector was for a moist wood chip bin and it didn't work due to wood dust and caking. I replaced it with several capacitance probes mounted horizontally a few feet apart. When bin was full the fill conveyer bypassed the bin and the bin level was displayed in a control room for operator monitoring.

 

Not a conclusion and you are doing your homework and actually have more experience with them than I do. I'm just trying to point out possibilities to look into. I will suggest trying to find a vendor in your area and discuss with them your application and see what input they can help you with.

Just to be clear -- I very much appreciate your inputs. I just don't have the resources to really follow all of your recommendations. Not only am I not an industrial process engineer, I don't have an industrial process engineer on staff. I am the staff, haha. You've raised some great points, and I've got a lot of homework to do to figure out some smarter answers here. Thanks for the help.

 

I am am wondering if a capacitive probe system may be a better choice for sensing the fluid level in the container. That could have the sensor be a rod totally enclosed by a suitable plastic cover, with the tank being the other side of the capacitor. The change in capacitance would be very proportional to the level of beer in the tank, and if the plastic tube were Teflon or similar it would not become coated with the contents. And quite possibly the same drive signal would work, fed through a high resistance, and the average voltage developed would reflect the level of the contents. This is just a suggestion as to another option, I think such systems are available but not cheap. But an examination of available products could provide a good understanding of how they work.

It's definitely an option, and one that I looked in to. Part of the issue is, as you mention, the expense. Early on, I thought that an ultrasonic system would be much simpler (ha, what did I know?!), so lower-cost. There's also, nominally, no parts that are continually immersed in fluid, which means less cleaning burden. The bigger problem with a capacitive-sensing rod is that it would need to be approximately the length of the container. But not all kegs are created equal -- there are half a dozen different manufacturers of these things, and all of them vary in shape. The only consistent thing is the lid. (shrug) So I ended up going down a path where I put sensors in a modified lid, and use that to tell me the range to the fluid surface, which I use to derive volume. Keg shapes are accounted for with math, which is easier than accounting for different heights with a different rod.

One of the other interesting ideas was to use the resonant frequency of the head space. So when the keg is full, there's only a little air in the top, and the resonant frequency will be high. When the keg is near empty, there's a lot of space, and it'll resonate better at lower frequencies. I thought it was a very cool idea, but had no idea how to even start down that path, so I abandoned it. Haha.

 

Another way of sensing contents is by weight, using a load cell.

 

is it possible that being under constant pressure is somehow altering the piezo property of the transducers and resulting in loss of sensitivity

A brief bit of googling came up with this research paper, which indicates that the answer is 'yes'. However, the pressures involved in the research were orders of magnitude greater than 10psi .

 

It's definitely an option, and one that I looked in to. Part of the issue is, as you mention, the expense. Early on, I thought that an ultrasonic system would be much simpler (ha, what did I know?!), so lower-cost. There's also, nominally, no parts that are continually immersed in fluid, which means less cleaning burden. The bigger problem with a capacitive-sensing rod is that it would need to be approximately the length of the container. But not all kegs are created equal -- there are half a dozen different manufacturers of these things, and all of them vary in shape. The only consistent thing is the lid. (shrug) So I ended up going down a path where I put sensors in a modified lid, and use that to tell me the range to the fluid surface, which I use to derive volume. Keg shapes are accounted for with math, which is easier than accounting for different heights with a different rod.

One of the other interesting ideas was to use the resonant frequency of the head space. So when the keg is full, there's only a little air in the top, and the resonant frequency will be high. When the keg is near empty, there's a lot of space, and it'll resonate better at lower frequencies. I thought it was a very cool idea, but had no idea how to even start down that path, so I abandoned it. Haha.

So how accurate does the reading need to be? Is the application in a bar where they need to know when to order more? Or in a bar in a private club? Or somebodies house? Or is this a brewery? Is the level when almost full more important, or when it is almost empty?