I have been tasked with designing a piezoelectric actuator driver at its
resonant frequency. This is to replace an arrangement where the sine
wave source is a function generator chip (the obsolete XR-2206) that
must be individually tweaked on a unit, by unit basis, to match the
crystal’s resonant frequency.

The crystal is used as an actuator to vibrate a mirror used in a laser
system. We have had systems come back from the field because the
function generator’s sinewave frequency became too different for the
system to work.

My idea is to use the crystal itself as part of an oscillator circuit so
that frequency is always exactly the crystal’s resonant frequency. I
have prepared an abbreviated concept schematic that can be viewed here:
https://i.imgur.com/4eBFQuD.png

This is an op amp version of a Colpitts Oscillator.

RV2 adds energy to the crystal to sustain its oscillations. I do not
know right now its optimum value.

U1 controls the oscillation amplitude by controlling the voltage on the
op amp’s power supply lines. The circuit is designed to have a single
supply so two regulators need not be ganged together.

I am aware there should be more capacitors in the U! adjustable
regulator circuit, and ops amp circuits. U1 also needs protection diodes
to prevent it being back driven in by capacitors during power down. All
this has been removed for clarity.

U1 might have to be replaced with an adjustable DC DC Converter. I do
not know right now how much current the crystal will require.

The cable to the crystal might be as long as five feet.

My questions are whether this should work in principle, and whether
there are any bettor ways to do it.

On a sunny day (Sun, 2 Jan 2022 23:03:26 -0800) it happened Artist
<sepflanze@sj.gmail.com> wrote in <squ740$eh0$1@dont-email.me>:

>I have been tasked with designing a piezoelectric actuator driver at its
>resonant frequency. This is to replace an arrangement where the sine
>wave source is a function generator chip (the obsolete XR-2206) that
>must be individually tweaked on a unit, by unit basis, to match the
>crystal’s resonant frequency.
>
>The crystal is used as an actuator to vibrate a mirror used in a laser
>system. We have had systems come back from the field because the
>function generator’s sinewave frequency became too different for the
>system to work.
>
>My idea is to use the crystal itself as part of an oscillator circuit so
>that frequency is always exactly the crystal’s resonant frequency. I
>have prepared an abbreviated concept schematic that can be viewed here:
> https://i.imgur.com/4eBFQuD.png
>
>This is an op amp version of a Colpitts Oscillator.
>
>RV2 adds energy to the crystal to sustain its oscillations. I do not
>know right now its optimum value.
>
>U1 controls the oscillation amplitude by controlling the voltage on the
>op amp’s power supply lines. The circuit is designed to have a single
>supply so two regulators need not be ganged together.
>
>I am aware there should be more capacitors in the U! adjustable
>regulator circuit, and ops amp circuits. U1 also needs protection diodes
>to prevent it being back driven in by capacitors during power down. All
>this has been removed for clarity.
>
>U1 might have to be replaced with an adjustable DC DC Converter. I do
>not know right now how much current the crystal will require.
>
>The cable to the crystal might be as long as five feet.
>
>My questions are whether this should work in principle, and whether
>there are any bettor ways to do it.

If you can find a crystal that matches the pieze frequency...
Or have one made.

I would personally go for a PIC with for example a 10 MHz crystal as clock
use the PWM outputs for the much lower frquency piezo drive, it will always be locked
to that crystal..
http://panteltje.com/pub/ultrasonic-anti_fouling_circuit_diagram_IMG_5132.JPG
Or use the PIC's internal oscillator if accurate enough (as I do).
Add a nice OLED didplay
http://panteltje.com/pub/ultrasonic_anti_fouling_status_on_display_IMG_5124.JPG

Have it do whatevver you want automatically (avi file)
http://panteltje.com/pub/ultrasonic_anti_fouling_OLED_running_5120.avi

and the part count is even less...
BUT.. you need to know how to program a micro (PIC 18F14K22 here), and [1] wind transformers
http://panteltje.com/pub/ultrasonic_antifouling_bigger_transformer_IMG_5179.JPG

Wonder if I should publish it in my new book 'The Fart Of Electronics'
Maybe I should keep it *secret* now that Russia and China know all about Dr Hobbs lidar
they may soon invade and take over, oh wait, already happening?

[1[] calcujaluate or something

On Monday, January 3, 2022 at 6:03:34 PM UTC+11, Artist wrote:
> I have been tasked with designing a piezoelectric actuator driver at its
> resonant frequency. This is to replace an arrangement where the sine
> wave source is a function generator chip (the obsolete XR-2206) that
> must be individually tweaked on a unit, by unit basis, to match the
> crystal’s resonant frequency.
>
> The crystal is used as an actuator to vibrate a mirror used in a laser
> system. We have had systems come back from the field because the
> function generator’s sinewave frequency became too different for the
> system to work.
>
> My idea is to use the crystal itself as part of an oscillator circuit so
> that frequency is always exactly the crystal’s resonant frequency.. I
> have prepared an abbreviated concept schematic that can be viewed here:
> https://i.imgur.com/4eBFQuD.png
>
> This is an op amp version of a Colpitts Oscillator.

<snip>

I suspect that what you want is a phase-locked loop.

I did it once, but forced the sustaining current to be exactly in-phase with the voltage across the device, which didn't give the maximum excursion. If I'd been able adjust the phase match to hit the actual resonant frequency of the actual device (and my circuit did offer a straight=forward way of doing it, but I was on holiday in Australia when the hardware came together) it probably would have gone in to production, but production device ended up having a remarkably consistent resonant frequency, so we didn't need it.

--
Bill Sloman, Sydney

On Sunday, January 2, 2022 at 11:03:34 PM UTC-8, Artist wrote:
> I have been tasked with designing a piezoelectric actuator driver at its
> resonant frequency
> My idea is to use the crystal itself as part of an oscillator circuit so
> that frequency is always exactly the crystal’s resonant frequency.. I
> have prepared an abbreviated concept schematic that can be viewed here:
> https://i.imgur.com/4eBFQuD.png
>
> This is an op amp version of a Colpitts Oscillator.

That's not clearly an oscillator; the offset voltage of the U3 op amp
allows lockup, for instance. A negative resistance oscillator is possible,
but the piezo might have more than one resonance (i.e. it wouldn't
be a stable sinewave drive), and there's startup issues to consider.

A triangle wave oscillator (like the XR206 core) can easilly be made adjustable,
and a microprocessor can be made to scan a range of frequencies and look
at the response, so could self-calibrate and park at or near a resonant frequency.

Or, you could twiddle knobs and find the resonance, then glue a frequency-determining
component (resistor?) to the piezo device, and run a four-wire cable to the control
center. Two wires to the resistor to determine the frequency, and two to drive the piezo.
That way, all the piezo targets and control modules are interchangeable.

Artist wrote:
=========
** Who are you ? What is you job title?

Why is the " BS" in front of your fake handle missing ??

> I have been tasked with designing a piezoelectric actuator driver at its
> resonant frequency.

** Piezo actuators are not " crystals " so have no special frequency.

>This is to replace an arrangement where the sine
> wave source is a function generator chip (the obsolete XR-2206) that
> must be individually tweaked on a unit, by unit basis, to match the
> crystal’s resonant frequency.

** Piezo actuators are not " crystals " so have no special frequency.

> The crystal is used as an actuator to vibrate a mirror used in a laser
> system.

** So much like a piezo disk driving a high frequency horn speaker ?

AKA a piezo tweeter.

> We have had systems come back from the field because the
> function generator’s sinewave frequency became too different for the
> system to work.

** So tweak it.

> My idea is to use the crystal itself as part of an oscillator circuit so
> that frequency is always exactly the crystal’s resonant frequency..

** Errr - what frequency is that exactly ?

The mechanical resonance of the whole mirror and wot not kaboodle ?
Surely you can reveal that here.

> U1 might have to be replaced with an adjustable DC DC Converter. I do
> not know right now how much current the crystal will require.

** You seem to know SFA - pal .

> The cable to the crystal might be as long as five feet.

** Shit eh ?

> My questions are whether this should work in principle, and whether
> there are any bettor ways to do it.

** Bettor find an engineer - cos YOU are not one

...... Phil

On Sun, 2 Jan 2022 23:03:26 -0800, Artist <sepflanze@sj.gmail.com>
wrote:

>I have been tasked with designing a piezoelectric actuator driver at its
>resonant frequency. This is to replace an arrangement where the sine
>wave source is a function generator chip (the obsolete XR-2206) that
>must be individually tweaked on a unit, by unit basis, to match the
>crystal’s resonant frequency.
>
>The crystal is used as an actuator to vibrate a mirror used in a laser
>system. We have had systems come back from the field because the
>function generator’s sinewave frequency became too different for the
>system to work.
>
>My idea is to use the crystal itself as part of an oscillator circuit so
>that frequency is always exactly the crystal’s resonant frequency. I
>have prepared an abbreviated concept schematic that can be viewed here:
> https://i.imgur.com/4eBFQuD.png
>
>This is an op amp version of a Colpitts Oscillator.
>
>RV2 adds energy to the crystal to sustain its oscillations. I do not
>know right now its optimum value.
>
>U1 controls the oscillation amplitude by controlling the voltage on the
>op amp’s power supply lines. The circuit is designed to have a single
>supply so two regulators need not be ganged together.
>
>I am aware there should be more capacitors in the U! adjustable
>regulator circuit, and ops amp circuits. U1 also needs protection diodes
>to prevent it being back driven in by capacitors during power down. All
>this has been removed for clarity.
>
>U1 might have to be replaced with an adjustable DC DC Converter. I do
>not know right now how much current the crystal will require.
>
>The cable to the crystal might be as long as five feet.
>
>My questions are whether this should work in principle, and whether
>there are any better ways to do it.

This kind of self-excited resonator circuit is very common. Look into
crystal oscillator circuits (low power, but frequency stable), and/or
ultrasonic transducer drivers (high power, frequency stability not
expected). It may be simpler to use discrete bipolar transistors or
FETs in this circuit.

The actual frequency will likely vary with local temperature and DC
power voltage.

Is a clean sinewave necessary, or can it be distorted or truly ragged?
What power level is needed?

Also think of how the resulting mirror-wobbled laser beam will be
used. Is synchronous detection in the picture? If so, you will need
to pick off and forward a copy of the actual drive signal to the
detectors function.

And so on.

Joe Gwinn

On Sun, 2 Jan 2022 23:03:26 -0800, Artist <sepflanze@sj.gmail.com>
wrote:

>I have been tasked with designing a piezoelectric actuator driver at its
>resonant frequency. This is to replace an arrangement where the sine
>wave source is a function generator chip (the obsolete XR-2206) that
>must be individually tweaked on a unit, by unit basis, to match the
>crystal’s resonant frequency.
>
>The crystal is used as an actuator to vibrate a mirror used in a laser
>system. We have had systems come back from the field because the
>function generator’s sinewave frequency became too different for the
>system to work.
>
>My idea is to use the crystal itself as part of an oscillator circuit so
>that frequency is always exactly the crystal’s resonant frequency. I
>have prepared an abbreviated concept schematic that can be viewed here:
> https://i.imgur.com/4eBFQuD.png
>
>This is an op amp version of a Colpitts Oscillator.
>
>RV2 adds energy to the crystal to sustain its oscillations. I do not
>know right now its optimum value.
>
>U1 controls the oscillation amplitude by controlling the voltage on the
>op amp’s power supply lines. The circuit is designed to have a single
>supply so two regulators need not be ganged together.
>
>I am aware there should be more capacitors in the U! adjustable
>regulator circuit, and ops amp circuits. U1 also needs protection diodes
>to prevent it being back driven in by capacitors during power down. All
>this has been removed for clarity.
>
>U1 might have to be replaced with an adjustable DC DC Converter. I do
>not know right now how much current the crystal will require.
>
>The cable to the crystal might be as long as five feet.
>
>My questions are whether this should work in principle, and whether
>there are any bettor ways to do it.

That circuit has problems, but I suggest that you understand the
crystal first.

Get a sinewave sig gen, a dual-channel scope, and a resistor. Drive
the piezo through the resistor and measure the relative phase and
amplitudes of the scope traces vs frequency. Then hack a Spice model
that behaves the same.

You can also connect the gen directly to the piezo and figure out how
much voltage it needs to do your optical thing.

Now you can Spice an actual oscillator. We can do that here.

Artists sometimes need help.

--

I yam what I yam - Popeye

Artist wrote:
> I have been tasked with designing a piezoelectric actuator driver at
> its resonant frequency. This is to replace an arrangement where the
> sine wave source is a function generator chip (the obsolete XR-2206)
> that must be individually tweaked on a unit, by unit basis, to match
> the crystal’s resonant frequency.
>
> The crystal is used as an actuator to vibrate a mirror used in a
> laser system. We have had systems come back from the field because
> the function generator’s sinewave frequency became too different for
> the system to work.
>
> My idea is to use the crystal itself as part of an oscillator circuit
> so that frequency is always exactly the crystal’s resonant frequency.
>
That's a good approach in general.

> I have prepared an abbreviated concept schematic that can be viewed
> here: https://i.imgur.com/4eBFQuD.png
>
> This is an op amp version of a Colpitts Oscillator.
>
> RV2 adds energy to the crystal to sustain its oscillations. I do not
> know right now its optimum value.

You're going to want a better oscillator circuit than that, for sure.

I doubt it would actually oscillate, for one thing--unless the resonator
has a lot of delay (e.g. a SAW device) it'll just look like an inductor
or a capacitor depending on frequency. There may be a higher-order
resonance that has enough delay.

> U1 controls the oscillation amplitude by controlling the voltage on
> the op amp’s power supply lines. The circuit is designed to have a
> single supply so two regulators need not be ganged together.

Most of the piezo actuators I've used have resonated in the 30 kHz
region, with Qs around 30. That's low enough that the nonsinusoidal
waveform of a self-limited oscillator can cause all sorts of uglies in
the motion--when the Q is low, the mechanical overtones often aren't far
enough from the electrical harmonics to get good rejection.

The other issue is that the electrical resonance won't be the same as
the mechanical resonance in general, so the vibration amplitude will be
tuning-sensitive. The mechanical resonance is what it is, and generally
coincides closely with the electrical series resonance.

A Colpitts has to run at a frequency where the piezo looks inductive.
Getting it to run exactly at its series resonance requires a series
inductor of just the right size.

<snip>

> The cable to the crystal might be as long as five feet.

Not a giant problem for a kilohertzy oscillator.

>
> My questions are whether this should work in principle, and whether
> there are any bettor ways to do it.

If you want to avoid tuning issues, I'd suggest using an external
sinusoidal VCO and driving the piezo via a resistor. Servo the VCO
frequency to where the phase shift across the resistor is zero. That
should nearly coincide with the amplitude minimum.

You might need an amplitude adjustment to take out unit-to-unit
variations in the piezoelectric sensitivity, but at least it'll be
running at its mechanical resonance.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

On Tuesday, January 4, 2022 at 8:41:36 AM UTC+11, Phil Hobbs wrote:
> Artist wrote:
> > I have been tasked with designing a piezoelectric actuator driver at
> > its resonant frequency. This is to replace an arrangement where the
> > sine wave source is a function generator chip (the obsolete XR-2206)
> > that must be individually tweaked on a unit, by unit basis, to match
> > the crystal’s resonant frequency.
> >
> > The crystal is used as an actuator to vibrate a mirror used in a
> > laser system. We have had systems come back from the field because
> > the function generator’s sinewave frequency became too different for
> > the system to work.
> >
> > My idea is to use the crystal itself as part of an oscillator circuit
> > so that frequency is always exactly the crystal’s resonant frequency.
> >
> That's a good approach in general.
> > I have prepared an abbreviated concept schematic that can be viewed
> > here: https://i.imgur.com/4eBFQuD.png
> >
> > This is an op amp version of a Colpitts Oscillator.
> >
> > RV2 adds energy to the crystal to sustain its oscillations. I do not
> > know right now its optimum value.
> You're going to want a better oscillator circuit than that, for sure.
>
> I doubt it would actually oscillate, for one thing--unless the resonator
> has a lot of delay (e.g. a SAW device) it'll just look like an inductor
> or a capacitor depending on frequency. There may be a higher-order
> resonance that has enough delay.
> > U1 controls the oscillation amplitude by controlling the voltage on
> > the op amp’s power supply lines. The circuit is designed to have a
> > single supply so two regulators need not be ganged together.
> Most of the piezo actuators I've used have resonated in the 30 kHz
> region, with Qs around 30. That's low enough that the nonsinusoidal
> waveform of a self-limited oscillator can cause all sorts of uglies in
> the motion--when the Q is low, the mechanical overtones often aren't far
> enough from the electrical harmonics to get good rejection.
>
> The other issue is that the electrical resonance won't be the same as
> the mechanical resonance in general, so the vibration amplitude will be
> tuning-sensitive. The mechanical resonance is what it is, and generally
> coincides closely with the electrical series resonance.
>
> A Colpitts has to run at a frequency where the piezo looks inductive.
> Getting it to run exactly at its series resonance requires a series
> inductor of just the right size.
>
> <snip>
> > The cable to the crystal might be as long as five feet.
> Not a giant problem for a kilohertzy oscillator.
> >
> > My questions are whether this should work in principle, and whether
> > there are any bettor ways to do it.
> If you want to avoid tuning issues, I'd suggest using an external
> sinusoidal VCO and driving the piezo via a resistor. Servo the VCO
> frequency to where the phase shift across the resistor is zero. That
> should nearly coincide with the amplitude minimum.
>
> You might need an amplitude adjustment to take out unit-to-unit
> variations in the piezoelectric sensitivity, but at least it'll be
> running at its mechanical resonance.

As usual Phil has got it pretty right. My problem was with a micro-stirrer which didn't need to produce a well-controlled excursion, but we did want to run it at the mechanical resonance , which was pretty sharp.

My problem was that maximum amplitude didn't happen when the phase shift across the resistor was zero.

The argument is that the phase shift goes from leading below resonance to lagging about resonance, but the reality is that it isn't exactly zero at resonance.

If you want to define resonance as maximum excursion, you can't use amplitude monitoring to get it exactly, because the amplitude doesn't change with frequency at the peak. Happily, phase does change quite rapidly at that frequency.

An approach that can work is to measure the phase when the amplitude is 50% down from the peak below resonance , and 50% down above resonance - both give well defined frequencies and thus phases, and set the system to lock the phase half-way between.

There's no guarantee that this will give the absolute maximum amplitude, but it will get you close, and close enough to let you work out almost the exact phase shift that will give you the maximum amplitude at the time.

The over-kill approach would be to use a direct digital synthesis (DDS) chip to generate the sine wave drive, and waveforms to drive in-phase and quadrature phase detectors. You'd have to digitise the outputs of both phase detectors and turn them into amplitude and phase signal and let a microprocessor do the tweaking.

There are simpler, cruder approaches. You can make a crude approximation to a sine wave as a "modified sine wave" - which is a switched drive which is off from 0 degrees to 30 degrees, high from 30 degrees to 150 degrees, off again from 150 degrees to 210 degrees, low from 210 degrees to 330 degrees and off again from 330 degrees to 360 degress.

That can be done with voltage controlled oscillator (VCO) in 4046 running at twelve times the desired frequency, and a bunch of digital dividers to spit out the drive waveforms and the outputs to drive the in-phase and quadrature phase detectors.

If you run the VCO faster you can push out waveforms at the operating frequency that can drive the "in-phase" detector to pick up the output amplitude at the phase you get when the system is operating at resonance and the "phase detector" so that its output goes through zero as the system goes through resonance, but you do have to know roughly where these points are. That was what I was trying to do back in 1992, but I hadn't spelled it out to my colleagues in enough detail.

--
Bill Sloman, Sydney

Artist wrote:
> I have been tasked with designing a piezoelectric actuator driver at its
> resonant frequency. This is to replace an arrangement where the sine
> wave source is a function generator chip (the obsolete XR-2206) that
> must be individually tweaked on a unit, by unit basis, to match the
> crystal’s resonant frequency.
>
> The crystal is used as an actuator to vibrate a mirror used in a laser
> system. We have had systems come back from the field because the
> function generator’s sinewave frequency became too different for the
> system to work.
>
> My idea is to use the crystal itself as part of an oscillator circuit so
> that frequency is always exactly the crystal’s resonant frequency.
Then *use* the crystal that way, instead of fiddling around with any
of the other driving schemes proposed?
BTW, your circuit as-is puts an adjustable (RV2) filtered (C1) DC
voltage on the crystal.

> I
> have prepared an abbreviated concept schematic that can be viewed here:
>  https://i.imgur.com/4eBFQuD.png
>
> This is an op amp version of a Colpitts Oscillator.
>
> RV2 adds energy to the crystal to sustain its oscillations.  I do not
> know right now its optimum value.
>
> U1 controls the oscillation amplitude by controlling the voltage on the
> op amp’s power supply lines. The circuit is designed to have a single
> supply so two regulators need not be ganged together.
>
> I am aware there should be more capacitors in the U! adjustable
> regulator circuit, and ops amp circuits. U1 also needs protection diodes
> to prevent it being back driven in by capacitors during power down. All
> this has been removed for clarity.
>
> U1 might have to be replaced with an adjustable DC DC Converter. I do
> not know right now how much current the crystal will require.
>
> The cable to the crystal might be as long as five feet.
>
> My questions are whether this should work in principle, and whether
> there are any bettor ways to do it.

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

I have some more information:

It is a piezoelectric element that is not crystalline. I was mistaken
when I described it as a crystal.

The resonance is totally mechanical, not electrical.

The piezoelectric actuating element drives a diaphragm. Here is where I
suspect the problem is. I suspect the elastic properties of the
diaphragm change with usage, and that changes the resonance frequency.

--
To email me directly remove sj. from my email address's domain name.
This is a spam jammer.

On Tuesday, January 4, 2022 at 7:55:23 PM UTC+11, Artist wrote:
> I have some more information:
>
> It is a piezoelectric element that is not crystalline. I was mistaken
> when I described it as a crystal.
>
> The resonance is totally mechanical, not electrical.
>
> The piezoelectric actuating element drives a diaphragm. Here is where I
> suspect the problem is. I suspect the elastic properties of the
> diaphragm change with usage, and that changes the resonance frequency.

You should be able to monitor the changing properties of the resonant assembly by looking at relationship between the current driving the oscillation and the displacement produced (which should show up in the voltage appearing across the piezoelectric element).

You've got to tell us more about the system you are driving if you want more specific help. You can certainly e-mail me at my ieee address, it you want to do it privately.

--
Bill Sloman, Sydney

Artist wrote:
> I have some more information:
>
> It is a piezoelectric element that is not crystalline. I was mistaken
> when I described it as a crystal.
>
> The resonance is totally mechanical, not electrical.
>
> The piezoelectric actuating element drives a diaphragm. Here is where I
> suspect the problem is. I suspect the elastic properties of the
> diaphragm change with usage, and that changes the resonance frequency.
>

Quartz crystals are mechanical resonators too, but their electrical and
mechanical characteristics are coupled via the piezoelectric effect.

There's nothing magic about quartz--lots of lower-performance
oscillators use ceramic resonators. Shifting the mechanical resonance
shifts the electrical one as well, because the piezoelectric effect is
intrinsically wideband--it's precisely the mechanical motion that's
responsible for the electrical frequency selectivity.

Resonator parameters such as peak amplitude, zero phase shift, and
minimum impedance appear close together at high Q values--with 'high'
meaning 'above about 3'. So your piezo should be fine.

For a fixed drive amplitude and a nonresonant diaphragm, tuning for the
maximum _in-phase_ current will get you the maximum mechanical
amplitude, because there's nowhere for the power to go except into the
mechanical motion.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

BS Artist wrote:
=============
> I have some more information:
>
** We had none before.

> It is a piezoelectric element that is not crystalline. I was mistaken
> when I described it as a crystal.
>
> The resonance is totally mechanical, not electrical.
>

** Something I picked up on immediately while other are still baffled by.

> The piezoelectric actuating element drives a diaphragm.

** So just like a piezo tweeter.
Have you never seen one?

The load presented by such a piezo transducer is purely capacitance.
About 100 to 200nF.

> Here is where I
> suspect the problem is. I suspect the elastic properties of the
> diaphragm change with usage, and that changes the resonance frequency.

** Exactly. So you tweak it.

...... Phil

On 5/1/22 1:20 am, Phil Hobbs wrote:
> Artist wrote:
>> I have some more information:
>>
>> It is a piezoelectric element that is not crystalline. I was mistaken
>> when I described it as a crystal.
>>
>> The resonance is totally mechanical, not electrical.
>>
>> The piezoelectric actuating element drives a diaphragm. Here is where
>> I suspect the problem is. I suspect the elastic properties of the
>> diaphragm change with usage, and that changes the resonance frequency.
>>
>
> Quartz crystals are mechanical resonators too, but their electrical and
> mechanical characteristics are coupled via the piezoelectric effect.

As they are with piezo-ceramics. But in P-Cs the resonance of the
ceramic isn't dominant, but the resonance of the substrate, often brass
or some other thin metal, like in birthday-card sounders.

Clifford Heath

Clifford Heath wrote:
> On 5/1/22 1:20 am, Phil Hobbs wrote:
>> Artist wrote:
>>> I have some more information:
>>>
>>> It is a piezoelectric element that is not crystalline. I was mistaken
>>> when I described it as a crystal.
>>>
>>> The resonance is totally mechanical, not electrical.
>>>
>>> The piezoelectric actuating element drives a diaphragm. Here is where
>>> I suspect the problem is. I suspect the elastic properties of the
>>> diaphragm change with usage, and that changes the resonance frequency.
>>>
>>
>> Quartz crystals are mechanical resonators too, but their electrical
>> and mechanical characteristics are coupled via the piezoelectric effect.
>
>
> As they are with piezo-ceramics. But in P-Cs the resonance of the
> ceramic isn't dominant, but the resonance of the substrate, often brass
> or some other thin metal, like in birthday-card sounders.
>
> Clifford Heath

Depends on the situation. The OP was talking about an optical scanner,
which needs to be stiff (i.e. nonresonant in the band of interest) in
order for the optical wavefront not to get screwed up by the mirror
bending.

Greeting cards may well present a more complicated picture. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

"the concepts "male" and "female" are essentially social constructions" (Bill Sloman)

"the Mueller investigation was about Trump only because Trump made it so" (Bozo paraphrased)

Bozo Bill Sloman is a chronic liar who cannot be reasoned with...

--
Anthony William Sloman <bill.sloman@ieee.org> wrote:

> X-Received: by 2002:ac8:6d35:: with SMTP id r21mr40185544qtu.9.1641196068408; Sun, 02 Jan 2022 23:47:48 -0800 (PST)
> X-Received: by 2002:a25:4e03:: with SMTP id c3mr39283382ybb.266.1641196068150; Sun, 02 Jan 2022 23:47:48 -0800 (PST)
> Path: eternal-september.org!reader02.eternal-september.org!news.misty.com!border2.nntp.dca1.giganews.com!nntp.giganews.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
> Newsgroups: sci.electronics.design
> Date: Sun, 2 Jan 2022 23:47:47 -0800 (PST)
> In-Reply-To: <squ740$eh0$1@dont-email.me>
> Injection-Info: google-groups.googlegroups.com; posting-host=123.243.66.234; posting-account=SJ46pgoAAABuUDuHc5uDiXN30ATE-zi-
> NNTP-Posting-Host: 123.243.66.234
> References: <squ740$eh0$1@dont-email.me>
> User-Agent: G2/1.0
> MIME-Version: 1.0
> Message-ID: <2329404f-444c-4d7c-bb9f-5f946d9f96c6n@googlegroups.com>
> Subject: Re: Crystal Controlled Oscillator
> From: Anthony William Sloman <bill.sloman@ieee.org>
> Injection-Date: Mon, 03 Jan 2022 07:47:48 +0000
> Content-Type: text/plain; charset="UTF-8"
> Content-Transfer-Encoding: quoted-printable
> Lines: 38
> Xref: reader02.eternal-september.org sci.electronics.design:656582
>
> On Monday, January 3, 2022 at 6:03:34 PM UTC+11, Artist wrote:
>> I have been tasked with designing a piezoelectric actuator driver at its
>
>> resonant frequency. This is to replace an arrangement where the sine
>> wave source is a function generator chip (the obsolete XR-2206) that
>> must be individually tweaked on a unit, by unit basis, to match the
>> crystal� Ts resonant frequency.
>>
>> The crystal is used as an actuator to vibrate a mirror used in a laser
>> system. We have had systems come back from the field because the
>> function generator� Ts sinewave frequency became too different for
> the
>> system to work.
>>
>> My idea is to use the crystal itself as part of an oscillator circuit so
>
>> that frequency is always exactly the crystal� Ts resonant frequency
> . I
>> have prepared an abbreviated concept schematic that can be viewed here:
>
>> https://i.imgur.com/4eBFQuD.png
>>
>> This is an op amp version of a Colpitts Oscillator.
>
> <snip>
>
> I suspect that what you want is a phase-locked loop.
>
> I did it once, but forced the sustaining current to be exactly in-phase with the voltage across the device, which didn't give the maximum excursion. If I'd been able adjust the phase match to hit the actual resonant frequency of the actual device (and my circuit did offer a straight=forward way of doing it, but I was on holiday in Australia when the hardware came together) it probably would have gone in to production, but production device ended up having a remarkably consistent resonant frequency, so we didn't need it.
>
> --
> Bill Sloman, Sydney
>
>

"the concepts "male" and "female" are essentially social constructions" (Bill Sloman)

"the Mueller investigation was about Trump only because Trump made it so" (Bozo paraphrased)

Bozo Bill Sloman is a chronic liar who cannot be reasoned with...

--
Anthony William Sloman <bill.sloman@ieee.org> wrote:

> X-Received: by 2002:a05:620a:9c3:: with SMTP id y3mr32730941qky.367.1641261509488; Mon, 03 Jan 2022 17:58:29 -0800 (PST)
> X-Received: by 2002:a25:ae13:: with SMTP id a19mr42789756ybj.327.1641261508993; Mon, 03 Jan 2022 17:58:28 -0800 (PST)
> Path: eternal-september.org!reader02.eternal-september.org!news.misty.com!border2.nntp.dca1.giganews.com!nntp.giganews.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
> Newsgroups: sci.electronics.design
> Date: Mon, 3 Jan 2022 17:58:28 -0800 (PST)
> In-Reply-To: <fdb23c12-119c-b0a3-cdec-da53833c28aa@electrooptical.net>
> Injection-Info: google-groups.googlegroups.com; posting-host=123.243.66.234; posting-account=SJ46pgoAAABuUDuHc5uDiXN30ATE-zi-
> NNTP-Posting-Host: 123.243.66.234
> References: <squ740$eh0$1@dont-email.me> <fdb23c12-119c-b0a3-cdec-da53833c28aa@electrooptical.net>
> User-Agent: G2/1.0
> MIME-Version: 1.0
> Message-ID: <44df5508-706a-4819-a188-49ad9e93ee34n@googlegroups.com>
> Subject: Re: Crystal Controlled Oscillator
> From: Anthony William Sloman <bill.sloman@ieee.org>
> Injection-Date: Tue, 04 Jan 2022 01:58:29 +0000
> Content-Type: text/plain; charset="UTF-8"
> Content-Transfer-Encoding: quoted-printable
> Lines: 123
> Xref: reader02.eternal-september.org sci.electronics.design:656633
>
> On Tuesday, January 4, 2022 at 8:41:36 AM UTC+11, Phil Hobbs wrote:
>> Artist wrote:
>> > I have been tasked with designing a piezoelectric actuator driver at
>> > its resonant frequency. This is to replace an arrangement where the
>> > sine wave source is a function generator chip (the obsolete XR-2206)
>> > that must be individually tweaked on a unit, by unit basis, to match
>> > the crystal� Ts resonant frequency.
>> >
>> > The crystal is used as an actuator to vibrate a mirror used in a
>> > laser system. We have had systems come back from the field because
>> > the function generator� Ts sinewave frequency became too differen
> t for
>> > the system to work.
>> >
>> > My idea is to use the crystal itself as part of an oscillator circuit
>
>> > so that frequency is always exactly the crystal� Ts resonant freq
> uency.
>> >
>> That's a good approach in general.
>> > I have prepared an abbreviated concept schematic that can be viewed
>> > here: https://i.imgur.com/4eBFQuD.png
>> >
>> > This is an op amp version of a Colpitts Oscillator.
>> >
>> > RV2 adds energy to the crystal to sustain its oscillations. I do not
>> > know right now its optimum value.
>> You're going to want a better oscillator circuit than that, for sure.
>>
>> I doubt it would actually oscillate, for one thing--unless the resonator
>
>> has a lot of delay (e.g. a SAW device) it'll just look like an inductor
>
>> or a capacitor depending on frequency. There may be a higher-order
>> resonance that has enough delay.
>> > U1 controls the oscillation amplitude by controlling the voltage on
>> > the op amp� Ts power supply lines. The circuit is designed to hav
> e a
>> > single supply so two regulators need not be ganged together.
>> Most of the piezo actuators I've used have resonated in the 30 kHz
>> region, with Qs around 30. That's low enough that the nonsinusoidal
>> waveform of a self-limited oscillator can cause all sorts of uglies in
>> the motion--when the Q is low, the mechanical overtones often aren't far
>
>> enough from the electrical harmonics to get good rejection.
>>
>> The other issue is that the electrical resonance won't be the same as
>> the mechanical resonance in general, so the vibration amplitude will be
>
>> tuning-sensitive. The mechanical resonance is what it is, and generally
>
>> coincides closely with the electrical series resonance.
>>
>> A Colpitts has to run at a frequency where the piezo looks inductive.
>> Getting it to run exactly at its series resonance requires a series
>> inductor of just the right size.
>>
>> <snip>
>> > The cable to the crystal might be as long as five feet.
>> Not a giant problem for a kilohertzy oscillator.
>> >
>> > My questions are whether this should work in principle, and whether
>> > there are any bettor ways to do it.
>> If you want to avoid tuning issues, I'd suggest using an external
>> sinusoidal VCO and driving the piezo via a resistor. Servo the VCO
>> frequency to where the phase shift across the resistor is zero. That
>> should nearly coincide with the amplitude minimum.
>>
>> You might need an amplitude adjustment to take out unit-to-unit
>> variations in the piezoelectric sensitivity, but at least it'll be
>> running at its mechanical resonance.
>
> As usual Phil has got it pretty right. My problem was with a micro-stirrer which didn't need to produce a well-controlled excursion, but we did want to run it at the mechanical resonance , which was pretty sharp.
>
> My problem was that maximum amplitude didn't happen when the phase shift across the resistor was zero.
>
> The argument is that the phase shift goes from leading below resonance to lagging about resonance, but the reality is that it isn't exactly zero at resonance.
>
> If you want to define resonance as maximum excursion, you can't use amplitude monitoring to get it exactly, because the amplitude doesn't change with frequency at the peak. Happily, phase does change quite rapidly at that frequency.
>
> An approach that can work is to measure the phase when the amplitude is 50% down from the peak below resonance , and 50% down above resonance - both give well defined frequencies and thus phases, and set the system to lock the phase half-way between.
>
> There's no guarantee that this will give the absolute maximum amplitude, but it will get you close, and close enough to let you work out almost the exact phase shift that will give you the maximum amplitude at the time.
>
> The over-kill approach would be to use a direct digital synthesis (DDS) chip to generate the sine wave drive, and waveforms to drive in-phase and quadrature phase detectors. You'd have to digitise the outputs of both phase detectors and turn them into amplitude and phase signal and let a microprocessor do the tweaking.
>
> There are simpler, cruder approaches. You can make a crude approximation to a sine wave as a "modified sine wave" - which is a switched drive which is off from 0 degrees to 30 degrees, high from 30 degrees to 150 degrees, off again from 150 degrees to 210 degrees, low from 210 degrees to 330 degrees and off again from 330 degrees to 360 degress.
>
> That can be done with voltage controlled oscillator (VCO) in 4046 running at twelve times the desired frequency, and a bunch of digital dividers to spit out the drive waveforms and the outputs to drive the in-phase and quadrature phase detectors.
>
> If you run the VCO faster you can push out waveforms at the operating frequency that can drive the "in-phase" detector to pick up the output amplitude at the phase you get when the system is operating at resonance and the "phase detector" so that its output goes through zero as the system goes through resonance, but you do have to know roughly where these points are. That was what I was trying to do back in 1992, but I hadn't spelled it out to my colleagues in enough detail.
>
> --
> Bill Sloman, Sydney
>
>

The John Doe troll stated the following in message-id
<sdhn7c$pkp$4@dont-email.me>:

> The troll doesn't even know how to format a USENET post...

And the John Doe troll stated the following in message-id
<sg3kr7$qt5$1@dont-email.me>:

> The reason Bozo cannot figure out how to get Google to keep from
> breaking its lines in inappropriate places is because Bozo is
> CLUELESS...

And yet, the clueless John Doe troll has itself posted yet another
incorrectly formatted USENET posting on Wed, 5 Jan 2022 02:07:45 -0000
(UTC) in message-id <sr2uhh$jeh$6@dont-email.me>.

nxPG9WsoBGo5

The John Doe troll stated the following in message-id
<sdhn7c$pkp$4@dont-email.me>:

> The troll doesn't even know how to format a USENET post...

And the John Doe troll stated the following in message-id
<sg3kr7$qt5$1@dont-email.me>:

> The reason Bozo cannot figure out how to get Google to keep from
> breaking its lines in inappropriate places is because Bozo is
> CLUELESS...

And yet, the clueless John Doe troll has itself posted yet another
incorrectly formatted USENET posting on Wed, 5 Jan 2022 02:09:07 -0000
(UTC) in message-id <sr2uk2$jeh$7@dont-email.me>.

MDqmsKWPklxN

The driven waveform must be sinusoidal. Purity is not critical. It can
be somewhat ragged.

A way to detect phase I am considering is to use comparators to convert
sinewave to square wave, and then detect phase difference in the similar
way the CD4046 does. The logic would have to be more complex than just
an exclusive OR gate though, because the XOR adjusts phase to 90 degrees.

I have trouble finding a successor to the EOL XR-2206. Suggestions would
be appreciated. If I do not find one I may have to try self excitation.

--
To email me directly remove sj. from my email address's domain name.
This is a spam jammer.

On 5/1/22 5:15 pm, Artist wrote:
> The driven waveform must be sinusoidal. Purity is not critical. It can
> be somewhat ragged.
>
> A way to detect phase I am considering is to use comparators to convert
> sinewave to square wave, and then detect phase difference in the similar
> way the CD4046 does. The logic would have to be more complex than just
> an exclusive OR gate though, because the XOR adjusts phase to 90 degrees.

Tried reading the datasheet? The CD4046 has two different phase
comparators. You'll find this document useful:

<https://www.ti.com/lit/an/scha002a/scha002a.pdf>

CH

On Wednesday, January 5, 2022 at 5:15:37 PM UTC+11, Artist wrote:
> The driven waveform must be sinusoidal. Purity is not critical. It can
> be somewhat ragged.
>
> A way to detect phase I am considering is to use comparators to convert
> sinewave to square wave, and then detect phase difference in the similar
> way the CD4046 does. The logic would have to be more complex than just
> an exclusive OR gate though, because the XOR adjusts phase to 90 degrees.

As Clifford Heath has pointed out, there are a variety of ways of detecting phase.

The classical approach is to use a four quadrant multiplier to gerenate the product of the sine wave being looked at and the your reference sinewave.

Trigometry lets you see that the product waveform has a DC content the corresponds to the phase difference between the two sine waves. You do have to filter out the alternating components.

A switching detector effectively multiplies the sine wave you are looking by a square wave of the same frequency. It doesn't give quite the same result, but it close enough for most applications.

Turning both sine waves into square waves before you multiply them together is even cruder, but can be good enough.
> I have trouble finding a successor to the EOL XR-2206. Suggestions would be appreciated. If I do not find one I may have to try self excitation.

Since we haven't got a clue what kind of voltage, current and frequency you have to deal with , we can't really give you any useful advice.
There are a lot of sine wave oscillators around. Some of them can be tuned to run over a range of frequencies.

I have suggested using a "modfied sine wave" to drive your transducer - it's just a three level approximation to a sine wave. It's virtue is that it doesn't have any third harmonic content, but it does have the higher odd harmonics you'd get from a square wave drive. if less of the fifth and a bit less of the seventh.

--
Bill Sloman, Sydney

Artist wrote:
> The driven waveform must be sinusoidal. Purity is not critical. It can
> be somewhat ragged.
>
> A way to detect phase I am considering is to use comparators to convert
> sinewave to square wave, and then detect phase difference in the similar
> way the CD4046 does. The logic would have to be more complex than just
> an exclusive OR gate though, because the XOR adjusts phase to 90 degrees.
>
> I have trouble finding a successor to the EOL XR-2206. Suggestions would
> be appreciated. If I do not find one I may have to try self excitation.
>
A filtered square wave, or (for fun) a 74HC4017 and some weighted
resistors. We had a thread on that awhile back--2011, it was. The OP
was George Herold. (George? You still out there, man?)

<https://groups.google.com/g/sci.electronics.design/c/sZWMknd6yjw/m/8HdJPZFUriAJ>

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

> Artist wrote:
>> The driven waveform must be sinusoidal. Purity is not critical. It
>> can be somewhat ragged.
>>
>> A way to detect phase I am considering is to use comparators to
>> convert sinewave to square wave, and then detect phase difference
>> in the similar way the CD4046 does. The logic would have to be more
>> complex than just an exclusive OR gate though, because the XOR
>> adjusts phase to 90 degrees.
[...]

Why does that matter?

Jeroen Belleman