Atomiser circuit - Changing resonent frequency

[eetaylog]

Im building a driver (see attached) for a water atomiser whose resonant frequency is 1.67MHz (standard for this part afaik).

The problem is that the user wants to fire a laser through the atomiser at the water that its humidifying, but in drilling the 5mm hole for the laser i have altered the resonant frequency of the ceramic, so the driver circuit no longer works on it.

I have done a frequency sweep on the atomiser with the hole and have found that its resonant frequency (highest atainable peak-to-peak voltage on the oscilloscope reading) is now in the region of 1.56MHz.

Im a bit confused as to which are the resonant components in the circuit however, so im struggling to match this new frequency. I was under the impression that L1 and C6 in the diagram are the tuned part, so i set about finding new values which would both maintain a matched impedance value for L and C, and combine to create the new desired 1.56MHz.

Using an I chose the new values:

L = 2.2uH
C = 4.7nF

These new values dont seem to work in the circuit however, so im a bit stuck. Can anybody give me any pointers?

TIA

 

[duffy]

Those aren't the resonant components, that's just a filter. If you plug L1 and C6 into your calculator, you will see it has a resonance of some 918khz.

The resonant component in this circuit is the piezo itself. These kinds of circuits are usually built around them because you want it to hit its natural resonant frequency, and piezo's are such good resonators they are routinely used for that exact purpose.

 

[duffy]

Here - this is an ordinary radio or computer crystal with the can removed. See that disc? Same idea as your ultrasonic resonator.




This is a slightly different version of your oscillator circuit. This one's an RF oscillator for a radio transmitter -



That transistor is your Q2.

Q1 in your circuit acts as a conductive switch, it provides the bias provided by the 47k resistor in my circuit. This bias voltage is necessary for Q2 to oscillate.

That filter I mentioned before is called an "RFC" - Radio Frequency Choke, same as the one coil in this circuit. It allows the collector voltage to bounce up and down like a spring. Dumb analogy, but it works.


You say you "drilled a hole" in the ultrasonic resonator?

 

[BrownOut]

eetaylog: Is there a link for that schematic. I'm interested in more information ( I didn't try to answer your question cause duffy already did )

 

[eetaylog]

Hi, thanks for the replies:

This is a slightly different version of your oscillator circuit. This one's an RF oscillator for a radio transmitter -

View attachment 58320

That transistor is your Q2.

Q1 in your circuit acts as a conductive switch, it provides the bias provided by the 47k resistor in my circuit. This bias voltage is necessary for Q2 to oscillate.

That filter I mentioned before is called an "RFC" - Radio Frequency Choke, same as the one coil in this circuit. It allows the collector voltage to bounce up and down like a spring. Dumb analogy, but it works.

Yeah, the first half of the circuit (Q1) is used to switch the oscillator part on and off. The base contacts (J3 and J4) of Q1 are mounted into a chamber so that the presence of water switches Q1, and hence Q2, on.

Like i say, the one part that i was having difficulty with was identifying the resonant components so that i could alter them to work with the 'holed' transducer and its new 1.56MHz resonant frequency.

So, are you saying that plugging the 'holed' atomiser into my original circuit should see it working correctly because the resonant frequency of the circuit is inherent on the load/transducer?

You say you "drilled a hole" in the ultrasonic resonator?

Yes, quite literally i drilled a hole through the middle of the atomiser using a mill and diamond tip drill-bit, and sealed it back up with a sample of glass so that a laser can be fired through it at the atomised water. A frequency sweep on it shows that it still resonates, but its properties (flexibility, capacitance...whatever) have altered very slightly so that the new res. frequency is now 1.56MHz.

eetaylog: Is there a link for that schematic. I'm interested in more information ( I didn't try to answer your question cause duffy already did )

The circuit was sent to me by the manufacturer of the atomiser.(prowave) You could try contacting them for more info if you like. There is a pdf attached to the website in the link with more info on the transducer itself as well.

 

[duffy]

A frequency sweep on it shows that it still resonates, but its properties (flexibility, capacitance...whatever) have altered very slightly so that the new res. frequency is now 1.56MHz.

When you did the sweep, did you look for harmonics? Do you have an undrilled transducer to compare? Harmonics will look like peaks in areas that are flat on an undrilled unit.

Reason I ask is that I cracked a piezo once, and remember that was how it behaved. Then I drove it with an oscillator, and it shattered.

Speaking of which - try hooking a signal generator to J4. Give it a square wave - try a lower frequency, say around 100 khz, 12V peak. Look at the waveform at the collector of Q2. You ought to at least see some "ringing" (sine wave with exponential decay). Check the frequency of that ringing, see if it really is 1.56MHz.

 

[duffy]

i drilled a hole through the middle of the atomiser using a mill and diamond tip drill-bit, and sealed it back up with a sample of glass so that a laser

Was it a diamond tip twist bit, or a burr?
How did you "seal it up" with glass - did you use a refractory process, or is it held in place with cement?
Is the water on the other side of the container under pressure?

 

[eetaylog]

When you did the sweep, did you look for harmonics? Do you have an undrilled transducer to compare? Harmonics will look like peaks in areas that are flat on an undrilled unit.

Reason I ask is that I cracked a piezo once, and remember that was how it behaved. Then I drove it with an oscillator, and it shattered.

Speaking of which - try hooking a signal generator to J4. Give it a square wave - try a lower frequency, say around 100 khz, 12V peak. Look at the waveform at the collector of Q2. You ought to at least see some "ringing" (sine wave with exponential decay). Check the frequency of that ringing, see if it really is 1.56MHz.

I used this circuit to find the resonant frequency of the holed piezo. The oscillation looked pretty clean to be honest.

Ill try using a sig gen. at the base of the switching half to see what i get back at the collector of Q2. In theory though should i be able to plug the holed piezo into the circuit and it will just resonate at its own inherent frequency?

 

[eetaylog]

Was it a diamond tip twist bit, or a burr?
How did you "seal it up" with glass - did you use a refractory process, or is it held in place with cement?
Is the water on the other side of the container under pressure?

Im not sure what type of drill bit it was, our mechanical guys did it for me. It looks a very clean cut though. The glass has been sealed with some epoxy, i can only imagine that the restrictions that the glass is now causing on the flexing of the ceramic is causing the change in RF.

and no, the water isnt pressurised, just 45mm under weight of gravity in a plastic chamber.

 

[duffy]

In theory though should i be able to plug the holed piezo into the circuit and it will just resonate at its own inherent frequency?

I THINK so, but I've never drilled a hole in a piezo before.

I know it's difficult to drill through ceramics. The barium titanate layer in the piezo is a type of ceramic. If there's a crack in it, it will never work right again - that one I have experience with, I worked on a miniature ultrasonic plastics welder and cracked plenty of piezo's.

Speaking of which, using epoxy to plug the hole might be the problem. Plastics absorb ultrasonic vibrations, that epoxy plug might be working as a damper. How big is it? Might be worth a try dissolving the plug (or drilling another transducer) and see if it will work without it.

 

[eetaylog]

Speaking of which, using epoxy to plug the hole might be the problem. Plastics absorb ultrasonic vibrations, that epoxy plug might be working as a damper. How big is it? Might be worth a try dissolving the plug (or drilling another transducer) and see if it will work without it.

Its around 7x7mm i would say (just enough to cover the 5mm hole). Even if the epoxy is dampening the vibrations the frequency sweep showed that the piezo was resonating quite cleanly regardless, just at a different frequency. Ill give it a try in my circuit anyway, thanks for all the help.

Just an after thought... the circuit is on strip board at the moment while its in test phase, probably not the best medium for high frequency propogation. Maybe this is part of the problem?

 

[duffy]

Only if you have a couple of feet of test leads to the transducer or something. Are you using a benchtop supply? It would be a smart idea to put a 1000µf filter cap right on the board itself, across the power supply leads.

 

[eetaylog]

Only if you have a couple of feet of test leads to the transducer or something. Are you using a benchtop supply? It would be a smart idea to put a 1000µf filter cap right on the board itself, across the power supply leads.

Yes im using a benchtop supply, and i only have maybe 8 inches of wiring from board to transducer. The one thing that confuses me about all this is a thread that i was reading on here:

https://www.electro-tech-online.com/threads/tuned-lcr-circuit-for-humidifier.102543/

...where the poster (4th post down on that page) says that changing Lr and Cr in his schematic will chage the operating frequency. I take it from your replies on my thread that this isnt the case, but it did throw me for a while.

 

[alec_t]

Adding inductance/capacitance to an oscillating piezo circuit will 'pull' the frequency slightly. This is what's done to trim the timing in a quartz clock/watch.

 

[eetaylog]

Adding inductance/capacitance to an oscillating piezo circuit will 'pull' the frequency slightly. This is what's done to trim the timing in a quartz clock/watch.

What are you referring to?

 

[alec_t]

That was a general reference to piezo oscillator circuits. For your particular driver circuit the values of C3, C4 and (to a lesser extent) C5 may have a slight effect on the operating frequency.

 

[eetaylog]

That was a general reference to piezo oscillator circuits. For your particular driver circuit the values of C3, C4 and (to a lesser extent) C5 may have a slight effect on the operating frequency.

and possibly the oscilloscope probe accross the piezo?

 

[duffy]

Have you tried driving that input with a signal generator and observing the piezo response at the collector of Q2 like I suggested? This will help you move forward on the project. The technique is like tapping a bell with a hammer and listening to how it rings. The waveforms will convey a good deal of information that can help us understand what the problem is.

 

[eetaylog]

Have you tried driving that input with a signal generator and observing the piezo response at the collector of Q2 like I suggested? This will help you move forward on the project. The technique is like tapping a bell with a hammer and listening to how it rings. The waveforms will convey a good deal of information that can help us understand what the problem is.

Hi, yes ive just this second tried it but i get nothing from the collector of Q2. Interestingly, if i replace the holed piezo with a normal one it rings just fine (both using a square wave sig. gen and just shorting J3 and J4 together as per normal), so the circuit must be working properly.

Its almost as if Q1 isnt being switched on and off by the sig. generator when i hook up the holed piezo, and yet if i plug the holed one straight into a signal generator i can make it resonate when pulsing 1.56MHz, so i know its not broken.

 

[duffy]

"Nothing" isn't a reading an oscilloscope makes. What does it show?