Piezo Driver

I am looking for an amplifier capable of driving a 30nF load w/ supply rails +/- 40V (or as close as i can get). Not sure what class/type of amplifier would be best for this. Any suggestions?

  • Hi Kyle,

    A piezoelectric load often requires a driver that operate from relatively high supply voltages, can deliver high current (typically 100's of milliamperes, or amperes), dissipate the associated power, have moderately fast slew rate and remain stable when driving a large capacitive load. All of that must be carefully considered when deciding upon an amplifier. You've already established your capacitive load, so next you need to determine the highest frequency (large-signal) the amplifier must pass. Doing so the slew rate requirements may then be determined and maximum current the amplifier must supply.

    Power operational-amplifiers are often employed as piezoelectric drivers. They are linear amplifiers with output stages that operate class-AB. Adding negative feedback reduces the distortion to very low levels. Your supply requirements are just about at the top end of TI's power operational-amplifier supply rating capabilities. The OPA454 HV operational-amplifier is capable of operating with +/-50V supplies, but the 25mA output current may be a bit low for the application - that will depend on the maximum frequency it must amplify. Moving into our power operational-amplifiers the OPA544 is usable with +/35V supplies and can deliver 4A of output current. Other power operational-amplifiers to consider are the OPA547/8/9. All are usable with supplies to +/-30V. All of these operational-amplifiers have slew rates in the range of 6 to 10V/us which must be considered for the application.

    Most operational-amplifiers cannotdrive a 30nF capacitive load without compensation being applied to the circuit. Therefore, any of the operational-amplifiers I have mentioned will require the circuit be analyzed and compensated to assure good stability and optimum transient performance.

    Regards, Thomas

    PA - Linear Applications Engineering

  • In reply to Thomas Kuehl:

    Not so fast! Most op amps cannot drive heavy cap loads under certain conditions. One such condition is positive unity gain (worst case) but what if you had a high gain (noise gain that is)? Shouldn't that make it easier? Sure it would. Also not all piezos require a high voltage op amp, think of HDD, does that require a high voltage for the piezo interface? No, it doesn't. In this case, I'd suggest the OPA547, very robust and will get you the results you want.

     

  • In reply to zidane:

    Hi Zidane,

    Not so fast! Most op amps cannot drive heavy cap loads under certain conditions. I agree, but most applications can be externally compensated such that they are stable when driving a specific capacitive, or complex load (see Tim Green's series on amplifier stability at EN-Genius.net).

    One such condition is positive unity gain (worst case) but what if you had a high gain (noise gain that is)? Shouldn't that make it easier? Sure it would. Yes.

    Also not all piezos require a high voltage op amp, think of HDD, does that require a high voltage for the piezo interface? I answered the inquiry in general terms. Once I have more information about the application then a more specific solution can be recommended.

    No, it doesn't. In this case, I'd suggest the OPA547, very robust and will get you the results you want. I suspect the OPA547 does have the necessary current drive capability (0.5A), but let's see the complete application requirements.

    I appreciate your inputs!

    Regards, Thomas

     

     

  • In reply to Thomas Kuehl:

    I am currently using the OPA454 to drive the piezo. I am using 30V rails. The input signal that we are using to drive the amplifier is a sine sweep from 500Hz to 8kz. We are using the piezo to excite a metal surface. In the past this test has always been done using a Crown rackmount amplifier. Something very robust but not practical for field testing. The signal audibly sounds right when put through the OPA454, but after reviewing the output signal on an oscilloscope we have noticed spikes in our sine wave when sweeping about 1k. This is not a constant problem, it occurs and then clears up as the sweep continues. Happens in about 250Hz increments. After continuous use of the OPA454 to drive this piezo (couple hours a day for a month) the amplifier exhibits crossover distortion at every frequency. The piezo is an American Piezo Disc Bender, 3kHz resonance, 300 Ohm max impedance, and 25,000pF. Sorry that is the only info that i have on the disc. No frequency response charts or datasheets available online. Sorry i do not currently know how much current the OPA454 is drawing on its own driving piezo. The entire system (micro controller, preamp, DC/DC converters, LCD, filter, etc) only pulls 150mA. So I dont think we are reach the current limit of the OPA454. I apologize for my ignorance on this issue. I designed this system straight out of school and never considered something like capacitive load driving. Too much ideal op amp theory :)

  • In reply to Thomas Kuehl:

    Hi Thomas, you're great guy. I hope I didn't come across as rude, I'm really sorry if i did. I know you know your stuff and I just meant to shed some light on perhaps what could have been perceived as an easy problem.

     

  • In reply to Kyle Aven:

    Hi Kyle,

    Thanks for all the additional information about your application; that helps. The behavior that you are observing with the OPA454 may be due to the power demanded by the piezoelectric load, or the amplifier's current limiting and/ or thermal protections when driving the load. The highest current demand will occur when the piezoelectric load is driven at the highest frequency, 8kHz. The capacitive reactance, Xc, is -j796-Ohms for a 25nF load at 8kHz. If the OPA454 output swings to a peak voltage (Vpk) of 35V, then the peak current (Ipk) will be about 45mA - which is within the OPA454's output current capability but is close to the limit (50mA). The power delivered to the load is roughly 1.3W for a sine wave output. You didn't mention how you are heat-sinking the OPA454. If the OPA454 gets hot, then the internal protections may be at the onset of activation distorting the waveform. Also, note that the OPA454 open-loop gain degrades with Ipk and temperature which worsens distortion; including cross-over distortion (see Fig's 6 and 7, in the data-sheet typical characteristic curves). This is the worst-case scenario and the situation should improve as you lower the drive frequency.

    One way to test if there is a thermal issue at work is to cool an OPA454 in operation and exhibiting the cross-over distortion. Get some freeze spray and give the OPA454 a shot for a few seconds while observing the output waveform. Not only will this lower the junction temperature (Tj) but also improve the Ipk capability of the OPA454 (Fig. 11 curves).

    Let me know what you find.

    Regards, Thomas

    PA - Linear Applications Engineering

  • In reply to zidane:

    Hi Zidane,

    No problem! Since I am pretty much a  "jack of all trades, master of none," I can always use the help when it comes to specialized applications.

    Thanks, Thomas

    PA - Linear Applications