Piezo transducer failure modes?

Thread Starter

Daniel McMath

Joined Dec 28, 2015
50
I have an application where I'm taking 40kHz piezoelectric transducers and installing them in keg lids to report how much beer is left. The environment is admittedly hostile -- the tank is cold (3-10C), pressurized (~10psi), and full of a wet, sugary liquid. I'd expect a certain amount of component failure, after a few years. Instead, I'm seeing a significant number of cases where transducers are failing within a few months.

The transducers in question are the PUI UTR-1440K-TT-R. According to the datasheet, they're rated for 120Vp-p at 40kHz at a very broad temperature range. (Link to Digikey for more info: https://www.digikey.com/product-detail/en/pui-audio-inc/UTR-1440K-TT-R/668-1550-ND/6071962). I'm using a 24VDC power supply, with no transformer, so I wouldn't expect voltage alone to be causing the failures. Aside from the aforementioned hostilities, there's no other engineering challenges -- no extreme heat, no cold below freezing, no vibration or mechanical shock, and so on.

First question: My test for failure is, initially, to look at the signal on the oscilloscope. Returns begin to get weaker, noise begins to increase relative to signal. Eventually, signal goes away entirely. When I pull transducers out of systems that have failed and apply a volt-ohm-meter, I find that there is relatively low resistance across failed transducers. While a known good transducer is typically showing infinite resistance, a failed transducer shows much less, sometimes as little as 2kOhm. Is resistance a good test for failure? Is there a smarter test that I should be using to confirm failure?

Second question: What would cause piezoelectric transducers to fail in this environment? Is there some combination of factors that's causing a problem? Does my test for failure indicate anything about the mode of failure? Is it possible that soldering at initial install is causing premature failure?

Thanks in advance for the help.

Dan
 
It's possible the manufacturer changed something to make manufacturing cheaper. I've had it happen. They changed from spot welding to 60/40 solder. No more high temperature for me.

Worked on a compromise even though the order was tiny. About 10 pieces at $20.00 USD ea.

They would do the soldering with Sn96, Upon receipt, I'd clean in boiling baking soda and electroless gold plate. They never did proper flux removal.
 

SamR

Joined Mar 19, 2019
5,031
The first idea off the top of my head was condensation but according to the datasheet it is IP65 compliant??? Is there any way moisture/humidity can get into the wiring connection side of the device. Is the face only IP65 compliant??? The combination of cold temp and any humidity (even during manufacture before sealing unless desiccant is used) is going to yield water... Have you contacted the sensor Vendor's Rep or the Manufacturer? Are any other sensor users having the problem?
 

SamR

Joined Mar 19, 2019
5,031
I looked at the datasheet again... The sensor is NOT pressure rated. The IP65 standard is for unpressurized water and this is being used in a wet pressurized environment. It may be that you need to go to a device that meets the 10PSI minimum pressure. Even IP67 is only good for 1-meter submersion, nowhere near 10PSI.
 

Thread Starter

Daniel McMath

Joined Dec 28, 2015
50
I looked at the datasheet again... The sensor is NOT pressure rated. The IP65 standard is for unpressurized water and this is being used in a wet pressurized environment. It may be that you need to go to a device that meets the 10PSI minimum pressure. Even IP67 is only good for 1-meter submersion, nowhere near 10PSI.
Huh. Wow. Ok, that's a great point!

So the face of the sensor is just aluminum, with aluminum sides, kind of a "can," if you will. The piezo element is encased in the aluminum "can" with some kind of epoxy paste. My assumption is that the IP65 rating applies to the unit as a whole, e.g. if you just clip some leads on to the exposed wires and pressurize the whole thing to 10psi, you'll get some water intrusion through the epoxy-paste.

In my installation, I've actually encased the whole thing in rubber, several layers -- see the attached image. I have to keep keg pressure inside the keg. As a pleasant side effect, I keep pressure off the back-side of the sensor. But now that you mention it, it might be possible to get moisture intrusion up the sides of the sensor, past the base, and in through the epoxy paste while still maintaining keg pressure. I seal everything with (food safe) RTV, so leaks are extremely rare, but it's entirely possible that if I'm not really careful about silicon around the full circumference of the sensor can, there's a path for moisture up through there.

Question: Would moisture intrusion be indicated by reduction in resistance across the element? Or in other words, is the failure indication that I'm getting consistent with the proposed failure mode?

So I guess Path #1 for me should be to make sure I'm getting a good seal around the sensor. I'll try that and report back in 6-9 months. :)

Path #2: Do you happen to know of any IP67-rated 40khz sensors? In my searching, these PUI units were about the best thing available.

Thanks for the great idea!

Dan
 

Attachments

SamR

Joined Mar 19, 2019
5,031
Dan,

I'm a retired US Process Control Engineer and the first thing to look at when choosing a sensor is; does it meet the environmental conditions for where I am going to place it. IP65 & IP67 are IEC (European) waterproofing standards, not pressure. Similar to our NEMA (National Electrical Manufacturer Association) dealing with water and dielectric capabilities and not pressure (other than washdown exposure). No idea who the IEC vendors are, but you need to look at pressure. Water will cause corrosion and that was my first thought. Pressure will cause deformation and may affect the sensor element. That MAY be what you are experiencing. It seems you are wrapping up the sensor? I would have nothing between the sensor face and the environment it is measuring in. Which means it needs to be food grade stainless (?), water, and pressure resistant to the limits of the environment it is used in. It seems that the back side is epoxied so it should be IP65 compliant and not leaking since it is not exposed to pressure. Not sure how you are connecting to it, but epoxy is heat sensitive if you are soldering it and may damage it's sealing. Most of my level detecting was for 10K to 1M gal. tanks with bubbler pressure gauging. So look for Sonic, Stainless (Teflon faced), Pressure proof to 10PSI (plus 10-20% safety margin). Find out exactly what the keg pressure is. 10PSI sounds high?

Good Luck, Sam
 

MisterBill2

Joined Jan 23, 2018
18,176
If the resistance has dropped a whole lot then either the sense element itself has changed or else the connections to the piezo device have become beer coated, or contaminated, or possibly a chemical reaction has occurred. Have you been able to open any of the failed transducers to see what may have changed inside?My guess is that there is some leakage and as a result some corrosion of the connections, and also a conductive coating forming. If there is any additional electronics inside the transducer housing that could be where the damage is happening.
One possible option, which I offer having no clue about the sensor construction, is to pressurize the inside of the housing to more than the 10 PSI, which should keep all the beer out.
 

Thread Starter

Daniel McMath

Joined Dec 28, 2015
50
If the resistance has dropped a whole lot then either the sense element itself has changed or else the connections to the piezo device have become beer coated, or contaminated, or possibly a chemical reaction has occurred. Have you been able to open any of the failed transducers to see what may have changed inside?My guess is that there is some leakage and as a result some corrosion of the connections, and also a conductive coating forming. If there is any additional electronics inside the transducer housing that could be where the damage is happening.
One possible option, which I offer having no clue about the sensor construction, is to pressurize the inside of the housing to more than the 10 PSI, which should keep all the beer out.
Good thinking. I've tried some dismantling in the past, with pretty limited results. The epoxy seems to be pretty ... rugged? Taking apart the epoxy to get to the sensor has historically destroyed the sensor in the process. I'll make another attempt at tearing down some sensors, though. Getting good visual confirmation on corrosion, moisture, or the like, would definitely confirm the suspicion.

I'll report back on that.
 

Thread Starter

Daniel McMath

Joined Dec 28, 2015
50
Reporting back ... I am not capable, apparently, of non-destructively opening these transducers. I've pulled a couple of them apart, but ... there's not much left when I'm done, so it's tough to tell if there's any corrosion. :-/
 

SamR

Joined Mar 19, 2019
5,031
Something else came to mind... Some of those (~40%) are designed to be used for ultrasonic cleaning and are required to be submersed and not in free air. Do you have the proper device?
 

MisterBill2

Joined Jan 23, 2018
18,176
Reporting back ... I am not capable, apparently, of non-destructively opening these transducers. I've pulled a couple of them apart, but ... there's not much left when I'm done, so it's tough to tell if there's any corrosion. :-/
It may make sense to investigate an industrial model of sensor, one made for pressure applications. And it might possibly be worth considering the use of an industrial float switch, except that it seems you are wanting an analog readout of the beer level.
 

Thread Starter

Daniel McMath

Joined Dec 28, 2015
50
It may make sense to investigate an industrial model of sensor, one made for pressure applications.
So ... "industrial" models with actual pressure ratings are *way* more expensive. Even these, at $9/ea are on the short list of "most expensive components in the assembly."

My goal is to build a low cost method of figuring out how much beer is left. With all of the other stuff that's required to make ultrasonic sensing work, it's already way more expensive than it should be. I'm actually shockingly close to a solid solution -- see the attached screen shot. I've spent a bunch of the last 3-4 months learning about what makes RF connectivity work, and how I could possibly have broken wifi connections so badly. Before that, I spent entirely too much time lurking on this forum and occasionally asking Really Dumb Questions (tm) trying to understand how to make an ultrasonic amplifier work correctly.

Ultimately, I think the big issue with sensor failure is most likely to be moisture intrusion from behind. I could redesign the mount assembly to provide more protection to the rear of the sensor, or to make it harder for moisture to get back there. But I think it comes down to an assembly process on my end. I've had some sensors in service for literally years, still providing good signal, but some sensors fail much faster than that. Assuming PUI is giving me consistent sensors from the factory, it almost has to be something I'm doing wrong in the assembly process here. Sam's moisture intrusion theory is the best answer I've seen, and it's really corroborated (imho) by the low resistance readings. And if I can fix that by just making sure that the sensor boot is well-filled with food-safe silicon when I assemble it, it solves a hard problem without forcing a complete redesign.

So I'll chalk this up as a solid path forward. :)
 

Attachments

MisterBill2

Joined Jan 23, 2018
18,176
So ... "industrial" models with actual pressure ratings are *way* more expensive. Even these, at $9/ea are on the short list of "most expensive components in the assembly."

My goal is to build a low cost method of figuring out how much beer is left. With all of the other stuff that's required to make ultrasonic sensing work, it's already way more expensive than it should be. I'm actually shockingly close to a solid solution -- see the attached screen shot. I've spent a bunch of the last 3-4 months learning about what makes RF connectivity work, and how I could possibly have broken wifi connections so badly. Before that, I spent entirely too much time lurking on this forum and occasionally asking Really Dumb Questions (tm) trying to understand how to make an ultrasonic amplifier work correctly.

Ultimately, I think the big issue with sensor failure is most likely to be moisture intrusion from behind. I could redesign the mount assembly to provide more protection to the rear of the sensor, or to make it harder for moisture to get back there. But I think it comes down to an assembly process on my end. I've had some sensors in service for literally years, still providing good signal, but some sensors fail much faster than that. Assuming PUI is giving me consistent sensors from the factory, it almost has to be something I'm doing wrong in the assembly process here. Sam's moisture intrusion theory is the best answer I've seen, and it's really corroborated (imho) by the low resistance readings. And if I can fix that by just making sure that the sensor boot is well-filled with food-safe silicon when I assemble it, it solves a hard problem without forcing a complete redesign.

So I'll chalk this up as a solid path forward. :)
OK, and it seems that "really expensive" is a relative term. If none of the sensors were lasting then that would be one thing, but if some last and some don't last then it may indeed be "a process variable issue," as they say in industry. It may be that external load cells would have been a more reliable choice, but with frequent replacement of the containers, and the variability of the connecting hose arrangement, that would have had other problems.
 

SamR

Joined Mar 19, 2019
5,031
The ultrasonic cleaning use transducers have to be submerged. They can withstand some cavitation but will degrade and fail if not submerged. It seems the ones you are using are for that purpose. They are apparently epoxy encapsulated which should rule out any moisture problems especially since they are washdown rated.
 

Thread Starter

Daniel McMath

Joined Dec 28, 2015
50
The ultrasonic cleaning use transducers have to be submerged. They can withstand some cavitation but will degrade and fail if not submerged. It seems the ones you are using are for that purpose. They are apparently epoxy encapsulated which should rule out any moisture problems especially since they are washdown rated.
I'm not entirely certain I agree with your ultimate conclusion. The below link lists this particular unit as a "transmitter / receiver," rather than as a "bender." The videos that PUI puts out associated with their ultrasonic products say that benders are the ones that are designed for ultrasonic cleaning sinks.
- https://www.digikey.com/en/product-...ultrasonic-receivers-transmitters-and-benders

That said, I think you were right up front when you pointed out that they're rated for wet, but not for pressurized-wet.

Also, if they *need* to be immersed ... I'm kind of putting them in cold damp places, and even the ones that sit out at my desk don't ever have any kind of heat problems. So I'm not sold on the idea that this particular unit *must* be immersed in order to work correctly.

I've ordered another batch of sensors. My plan is to actually attempt to make them fail in this way, specifically -- I'll build a set of lids, then put some water in some kegs, pressurize them, and turn them upside down. If it's a moisture impingement problem, that really ought to force it to occur, sooner rather than later. I'll test them periodically to see what kind of resistances I'm getting, and if they're working correctly. With any luck, they'll fail quickly. Then, I'll build a new set of sensor-lids, except I'll be really careful to seal around the edge, to make sure there's no easy path for moisture to get to that epoxy surface. Those should last indefinitely, or something like 20 years until the rubber breaks down.

I'll let you guys know how it goes. Thanks for all the help.
 

MisterBill2

Joined Jan 23, 2018
18,176
The ultrasonic cleaning use transducers have to be submerged. They can withstand some cavitation but will degrade and fail if not submerged. It seems the ones you are using are for that purpose. They are apparently epoxy encapsulated which should rule out any moisture problems especially since they are washdown rated.
Washdown is a WHOLE LOT Different from submersion under pressure. For starters, wash down is intermittent, and that constant pressure is persistent. Epoxy encapsulation depends on adequate bonding to the penetration materials to work, if that bond is not perfect then there may be leakage.
 

Thread Starter

Daniel McMath

Joined Dec 28, 2015
50
Washdown is a WHOLE LOT Different from submersion under pressure. For starters, wash down is intermittent, and that constant pressure is persistent. Epoxy encapsulation depends on adequate bonding to the penetration materials to work, if that bond is not perfect then there may be leakage.
That makes plenty of sense. In a perfect world, these are installed in the TOP of the keg, so even the front of the sensor shouldn't see moisture unless you splash stuff around. So if it's washdown rated, I'm adding a big rubber boot (see the picture I uploaded a few posts earlier), and I just need to make sure that I'm sealing around that, and it *should* be more than adequate to prevent the epoxy end from seeing any moisture.

I'm planning the pressurized immersion test to (a) determine if that's actually what's failing, (b) test to see if I can alter my manufacturing process to make it not fail that way, and (c) make it fail much faster, so I can do (a) and (b) without waiting a year. :)
 

SamR

Joined Mar 19, 2019
5,031
I'm not entirely certain I agree with your ultimate conclusion
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.
 
Top