Triangular waveform generation

I need to generate three triangle waveforms which are phase shifted from each other by 33%. The frequency is 100 KHz. Basically I need to generate three PWM waveforms that are phase shifted from each otherby 33% so I plan to compare these three triangle waveforms to generate my PWMs. AndI can only use basic analog chips. No microcontrollers are allowed.
Anybody know how to go about doing this?

TIA

At 100kHz, if you want a triangle waveform, how about the xr2206? (It's a chip that's been around for some time and is quite analog.) You'll want three, and you want them running from three oscillators that are phase shifted by 33%. The oscillators could be done via one fundamental oscillator and logic with three outputs.

Another possibility is three oscillators (as described above) feed three up-down counters which feed three DACs which feed a three LP filters. Bear in mind that the oscillators would be running fast so as to produce the ramps that went from 0 to Vmax for the first half cycle of the triangle output.

You could make a capacitor charge linearly and then discharge linearly or switch between two, just taking the charge part, inverting one and then summing to create a triangle. Then, of course, do this three times.

Ok, there are three ideas.

you can try rc phase shift oscillators with 3 rc sections. the output of this will be a sine. give it to a schmitt trigger to make it a square. then send it through a low-pass section to get a triangle.

hope this would help.

AMRITH.S.

usernam said:

I need to generate three triangle waveforms which are phase shifted from each other by 33%. The frequency is 100 KHz. Basically I need to generate three PWM waveforms that are phase shifted from each otherby 33% so I plan to compare these three triangle waveforms to generate my PWMs. AndI can only use basic analog chips. No microcontrollers are allowed.
Anybody know how to go about doing this?

TIA

Click to expand...

amriths04 said:

you can try rc phase shift oscillators with 3 rc sections. the output of this will be a sine. give it to a schmitt trigger to make it a square. then send it through a low-pass section to get a triangle.

hope this would help.

AMRITH.S.

Click to expand...

Yes, that, the filter part, is a much better idea. First get the square (odd harmonics) and then LPF to the triangle (odd harmonics also.)

Can a butterworth do that?

ofcourse it can... but the order should not be too high,else, the output would be somewhere between a triangle and a sine. i think a first order butter worth would suffice...

AMRITH.S.

Euler's Identity said:

amriths04 said:

you can try rc phase shift oscillators with 3 rc sections. the output of this will be a sine. give it to a schmitt trigger to make it a square. then send it through a low-pass section to get a triangle.

hope this would help.

AMRITH.S.

Click to expand...

Yes, that, the filter part, is a much better idea. First get the square (odd harmonics) and then LPF to the triangle (odd harmonics also.)

Can a butterworth do that?

Click to expand...

Try with 565/566 pll/vco with good design values

This would server your purpose