LF347 High Frequency Triangle Wave Generator

Post deleted.

Eliminating the second op amp in the feedback path cuts the problem in half allowing for a triange at 250KHZ. It won't scale with the caps of course because of the slew rate.

Hey guys I'm sorry for the late reply

I'm currently studying for the test this week still got 3 to go :S

@ronv Im sorry we cant use only just one opamp for this project.
@MrAl Sorry we'r not allow to use any component other then resistor and capacitors. (For PWM part)
@Roff

You are to design, build, test, and report on, a simple open-loop Class-D stereo audio power
amplifier that takes the audio output from a portable device and drives a pair of 8 Ω nominal
impedance loudspeakers. The amplifier needs to be able to provide a continuous output of
12W peak per channel into 8 Ω nominal impedance loudspeakers. A full set of specifications
are listed below:
• Input impedance >500 Ω.
• Amplifier gain set to match input device chosen (maximum voltage amplitude input
signal gives maximum voltage amplitude output signal).
• Minimum bandwidth from 50 Hz to 18 kHz (preferably 10 Hz to 22 kHz).
• Voltage source of up to ±10 V. For the project hardware, a dual bench-top power
supply set to ±10 V is to be used.
• Able to continuously produce output power of 12 Wpeak per channel (into 8 Ω
nominal impedance loudspeakers).
• As high efficiency as possible.
• As low noise and low distortion as possible.


---- update
I've talked to my lecture about slew rate problem, and indeed it is the slew rate problem for the square wave which cant keep put to maintain the waveform and disturbed the triangle output, From what he said is that at 300Khz the output from the top right opamp (refer to the attach at the page 1) should be triangle and the waveform for VF1 will maintain triangle but to do this we need to replace one of the component to other.

My guess is swap one of the resistor to a ceramics capacitor, but due to the test situation, I havnt had time to do the simulation yet might do it after Friday test.

Hi Grumpy,

Well then just design an oscillator and use the RC part of the oscillator as the triangle wave. It wont be a perfect triangle but it doesnt have to be and will still work for your PWM.

MrAl said:

Hi Grumpy,

Well then just design an oscillator and use the RC part of the oscillator as the triangle wave. It wont be a perfect triangle but it doesnt have to be and will still work for your PWM.

Click to expand...

Yeah, I ran a simulation of his circuit, but with different values. It can make pretty good looking 250kHz triangle wave.

MrAl said:

Hi Grumpy,

Well then just design an oscillator and use the RC part of the oscillator as the triangle wave. It wont be a perfect triangle but it doesnt have to be and will still work for your PWM.

Click to expand...

But imperfections in the triangle will create PWM distortion if used to generate an audio signal.

Hello Carl,

The kind of imperfection i was talking about would be in the slight curve of the triangle, nothing more. That shouldnt bother anything.

MrAl said:

Hello Carl,

The kind of imperfection i was talking about would be in the slight curve of the triangle, nothing more. That shouldnt bother anything.

Click to expand...

It may not be a problem for a PWM power supply, but for audio that translates into distortion in the recovered speaker audio signal. Whether it's a bother depends upon how much negative feedback the modulator circuit has and how much distortion you can tolerate.

Hi Carl,

I agree. For audio a good triangle is better but it depends as you noticed. The amplitude vs pulse width relationship is not linear for a curved triangle, but it is for a straight sided triangle. The degree of nonlinearity depends on various factors too.

Did anyone try ronv's circuit yet?

Hey guys sorry for the late reply, I've work out how to manage to get around 270 Khz which is fine for our project.

The key point is to use a capacitor between the output of 1st Opamp to the input of the 2nd opamp to maintain the voltage for the square wave form at high frequency.

Thanks for all your help