PWM amplifier 1 - The PWM modulator design III.

On my previous analysis, I didn't pleased with the quality of triangle signal, but a little modification made better result. In the second article, the triangle (and the PWM) signal is better, but not perfect. This is the reason why I simulated the modulator design with another oscillator circuit.



The design is very simple, simplest then the previous signal generator version. This new contains simple CMOS oscillator. The design is small, simple, and the output signal is perfect. The duty-cycle is always 50%, the amplitude of square signal is 5...15V if required. The frequency setting is very easy, on 10 MHz the square signal is really nice.

The red is the square output of CMOS circuit, the green is the triangle signal.

The triangle and the square signal is much better compared to the analog oscillator circuit.

This is the schematic, the difference between R1 and R2 is 10x, the capacitor can be very small value, the frequency is very stable.

Unfortunately, the upper ideal stage will be wrong if I continues the design with comparator circuit on the simulation. (See the schematic, I did not continued the triangle output to the comparator input). If the triangle signal after CMOS (or TTL) circuit wired to the comparator, the signal will be non-usable:

The square signal stay good, but the triangle signal going wrong on the simulation.

 The usefulness triangle signal.

I did nothing else, I just continued the signal with comparator.

The upper problem can be solved with a small modification, see the images and the power source of the comparator on the schematic:

On the circuit, I using independent DC power source for comparator chip only. After I getting back the nice signals, but with this design the comparator doesn't made PWM signal from sources. If I used common ground, the comparator started to work, but the triangle (and the PWM) signal is going wrong again:

Transient analysis result when the CMOS oscillator and the comparator have common ground:

At this moment I don't know how to solve the problem on the simulator. I wonder, (and not sure) about the result of real circuit. The simulation is usefulness, but the oscillator design is better and simpler before using comparator. So at this moment, the analog oscillator design is much better than this new one.

PWM amplifier 1 - The PWM modulator design II.

I have successfully cleared the last design example of PWM modulator circuit. This PWM modulator will be used before TAS H-bridge chip. In this article, I don't really like the signal at the end, and I made mistakes about power sources (with Zener diodes) on the last simulation schematic. Not too much, but I modified the design, and I got better result. I deleted the dead-time logic, because for the TAS chip not required, the soft start only is easier. At the end of this newer design, I got nice signal from both TTL and CMOS outputs and comparator without dead-time logic. See images.

In this image the green is the triangle signal, the red is 10 kHz sinus, a yellow is the PWM output on the CMOS 4009 inverter.

I know, that not really important, but how I know about the quality of modulated PWM signal? I have to convert back to the original with RLC network. This is the blue line.

I changed the CMOS 4009 (15V power for comparator and CMOS circuit) to TTL 7414 (5V power for comparator and TTL circuit). The TTL output is the yellow line.

 The decoded result at the end of TTL circuit is the blue signal.

I made small modifications on the circuit around the comparator. I changed Z diodes DC power sources, the output of the comprator continued with only one TTL or CMOS gate (inverter). I analyzed the PWM signal after the TTL/CMOS chip, but the output of comparator is really good too with this schematic: