Hi All,

This is part of my thesis. I need to generate a 0V-5V 10kHz triangle wave as a carrier wave for my SPWM and as much as possible, the system is only single-supplied.

Any advice is appreciated.

Cheers,
Jay

 

Hi All,

This is part of my thesis. I need to generate a 0V-5V 10kHz triangle wave as a carrier wave for my SPWM and as much as possible, the system is only single-supplied.

Any advice is appreciated.

Cheers,
Jay

You probably need the op-amp version of the "inverter" oscillator - the output will be a square wave, but there's a sawtoot on the timing capacitor. Its probably worth using a JFET source follower so the next stage doesn't mess stuff up.

 

So you want a true triangle-wave (equal rise and fall times) and not a sawtooth?
Edit: Sawtooths are generally easier to generate.

 

So you want a true triangle-wave (equal rise and fall times) and not a sawtooth?

The PUT does sawtooth - nearly linear if you replace the charging resistor with a current source.

 

Since this is for *your* thesis, here is a plan: search the innergoogle for triangle wave circuits, study them, determine which properties are important to you, draw a first-pass schematic of the circuit you think will work, and post it here for comments. We don't solve homework problems, only advise.

ak

 

MOD NOTE: Moved to Homework Help where it can receive the kind of attention appropriate to an academic situation.

 

If you integrate a square wave you will get a triangular wave.

 

Myriad of ways -

1) Drive a C with a current source/sink
2) Stored wavetable approach with DAC
3) Harmonic summation
4) Transformation, integratrion, like previously mentioned

And I am sure many more....


Regards, Dana.

 

If you integrate a square wave you will get a triangular wave.

This is one of our labs for undergrads.

1. Build a Schmidt trigger with an op-amp
2. Use the output to charge an RC circuit, feed the capacitor voltage back to the Schmidt trigger input to create a relaxation oscillator
3. Build an integrator circuit to produce a triangle from the square wave.

Requires two op-amps, a handful of resistors and two capacitors. It's quick and easy.

 

You probably need the op-amp version of the "inverter" oscillator - the output will be a square wave, but there's a sawtoot on the timing capacitor. Its probably worth using a JFET source follower so the next stage doesn't mess stuff up.

So you want a true triangle-wave (equal rise and fall times) and not a sawtooth?
Edit: Sawtooths are generally easier to generate.

Yes, a true triangle waveform.

Something similar to the attached image below.

 

So I would follow TechWise's suggestion in post #9.
You may need a high-speed op amp for the integrator if you want a good fidelity (sharp corner) triangle-wave at 10kHz.

 

The attached image is one of my design references for obtaining this kind of waveform.

 

The attached image is one of my design references for obtaining this kind of waveform.

That's similar to what TechWise was referring to.
So, if you adjust it for the desired frequency, is that adequate for your purposes?

 

So you want a true triangle-wave (equal rise and fall times) and not a sawtooth?
Edit: Sawtooths are generally easier to generate.

Hi,

Yeah for SPWM if by that he means sine generation then yes it has to be a true triangle with equal rise and fall times. That's so that it produces a pulse that is centered on the average sine voltage wave for that period.
It is a little interesting that this is a somewhat common choice for a grad thesis project, yet not the most modern way to do it anymore.