6
\$\begingroup\$

Project background: I'm building a plasma speaker. There are several examples online using a TL494 power supply IC and feeding audio into its feedback pin to get sound out. A sample schematic: enter image description here

I'm wondering if there is a better way to do this. I'm a hobbyist, so I'm not terribly interested in the absolute cheapest method since it's a one-off, and I'll also be tolerant of potentially dangerous or sketchy things since I don't have to get certification for anything.

I have access to a few micros, but little experience with them. I did get a Launchpad, and my understanding is one of the MSP430s that came with it has a built-in ADC. While going from analog to PWM seems a little underwhelming for a microcontroller, I do have an additional requirement that I must limit my upper duty cycle to remain below about 95% to prevent saturation of the flyback transformer. Operating frequency is 120-150khz.

To summarize my question:

Are there more straightforward ways to go from a high-impedance 1V P-P input signal to a 150khz output signal with proportional PWM, capped at a maximum 95% duty cycle? Unless built natively into the device, I will be building my own output stage to drive the power MOSFET.

\$\endgroup\$

3 Answers 3

10
\$\begingroup\$

Amusingly enough, I was literally just looking at single-chip PWM generators for my job.

If interest to you might be the Linear LTC6992 - TimerBlox: Voltage-Controlled Pulse Width Modulator (PWM)

You give it a 0-1V input, and you get a PWM pulsetrain out. It can also internally limit the duty cycle to 5%-95%, or 0-100%.

enter image description here


You would need a op-amp or two to scale your input to the chip's input range, but you would need to do that anyways, because very few ADCs are bi-polar (assuming you were going the microprocessor route).

Looking at the reference schematic you provided, I would also recommend using a MOSFET driver of some sort, to improve switching times, if nothing else.
It would also likely provide level-shifting for the gate-voltage.

\$\endgroup\$
3
  • \$\begingroup\$ The TL494 is designed specifically for SMPS and includes a low-side MOSFET driver. However, LTC6992 looks amazing, assuming I can slow it down to 150kHz. \$\endgroup\$
    – Bryan B
    Commented Aug 3, 2012 at 18:24
  • \$\begingroup\$ Holy poopscicle, that thing is perfect. \$\endgroup\$
    – Bryan B
    Commented Aug 3, 2012 at 20:48
  • \$\begingroup\$ Frequency Range: 3.81 **Hz** to 1MHz - So I would imagine that it won't be hard to get it to run 150kHz. \$\endgroup\$ Commented Aug 3, 2012 at 20:52
3
\$\begingroup\$

I had a signal processing project where I used an old cheap PIC to sample an incoming sine wave and output a PWM signal. The PWM configuration was all pretty much taken care of by the micro. Capping the output duty cycle would just be a matter of signal filtering, which you can do in the micro.

I would say IMO that using a micro to do this is pretty straightforward on its own. You would need to regulate the power for the micro, and maybe add an opamp input stage to protect the ADC on your micro.

After quick glance at the user manual for your MSP430, it should do PWM on its own, but having never dealt with TI micros, I cant tell you specifically how to do it.

That being said, what you have seems simple enough. I haven't worked with a PWM control chip before, but reading the data sheet it should do the job.

By using a micro, your swapping external components for internal programming. The upshot is you gain better control over the output PWM if you are willing to work for it (program).

\$\endgroup\$
2
\$\begingroup\$

You might want to read some articles about Class D Amplifiers. See for instance: http://www.maximintegrated.com/app-notes/index.mvp/id/3977

These amplifiers (the ones used for amplifying analog signals) usually have a first stage where the input signal is fed into a comparator, whose second input is fed with a sawtooth/triangle wave. This effectively converts the analog input to a PWM signal.

Creating a triangle wave is also an easy task.

\$\endgroup\$
1
  • \$\begingroup\$ I've used the triangle-wave + comparator route before on other projects, but with the parts I had on hand I couldn't get the frequency high enough. I can try again with a lower voltage so the opamps aren't swinging as far. \$\endgroup\$
    – Bryan B
    Commented Oct 30, 2013 at 15:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.