# Modified Sine Wave with Stable Voltage Output by NE555 (Bipolar)

#### Wendy

Joined Mar 24, 2008
23,134
But they work great in Class D amps, full circle.

#### Wendy

Joined Mar 24, 2008
23,134
Either the schematic isn't complete or I'm missing something. Where is the PWM coming from?

#### Audioguru

Joined Dec 20, 2007
11,248
Pin 13 of the CD4047 is a 120Hz square-wave and pin 10 and pin 11 are 60Hz square-waves with opposite phases.

The integrating opamps convert the square-waves into triangle-waves.

The NOR gates will not work with a diode in series with one input. The diodes are not needed anyway.

There is no PWM and the output voltage is not regulated.

This is how i designed the modified sine-wave driver a couple of years ago:

#### Attachments

• 8 KB Views: 92

#### JUAN DELA CRUZ

Joined May 27, 2008
121
Here is my Exp. Modified sine-wave ckt. It uses CD4047 Square wave Osc. w/ a value of R= 39K, C=0.1uF for 60Hz output.

The Op-amps vary the pulse-width correspond to Battery(source) voltage variation.

Diodes rectify the PWM sine wave before going to the NOR gates to obtain the desired Modified-Sine waveform.

oK...Here is my revised ckt.

Last edited:

#### Audioguru

Joined Dec 20, 2007
11,248
Juan,
It won't work.
1) The first opamp's output is a square-wave, not a triangle-wave.
2) The second opamp is an integrator so its output is a triangle-wave, not a sine-wave.
3) Logic gates do not need a diode for them to gate a signal, and don't work if a diode is in series with an input.
4) The logic gate is a logic switch. If it has a slow input like a sine-wave then its output is a sudden abrupt switching rectangular waveform, not a smooth sine-wave.
5) If you feed a smooth sine-wave to power Mosfets then they will waste a lot of power and will get extremely hot.
6) You do not have a circuit to adjust the voltage gain and threshold voltage of each Mosfet. Some will begin to turn on when the input voltage is 2.0V. others need 4.0V.
Some have low gain and some have high gain.
7) The gain of the Mosfet depends on its load resistance. Therefore its output voltage will change when its load current changes.

#### Wendy

Joined Mar 24, 2008
23,134
The circuit I drew in my second post (though I would merge it with the updated triangle wave generator show in post 10) would work with MOSFET, because it is fundamentally a digital circuit (a Class D amp). Class D amps are the only way you are going to feed a sine wave through MOSFET and not get them hot. A plain unmodulated sine wave is analog, any analog signal generates heat because it is operating out of the modes described in post 14 that dissipate minimum wattage. As is stated in post 14, this is true of all transistors, not just MOSFETs.

Last edited:

#### JUAN DELA CRUZ

Joined May 27, 2008
121
Class D amps are the only way you are going to feed a sine wave through MOSFET and not get them hot. A plain unmodulated sine wave is analog, any analog signal generates heat because it is operating out of the modes described in post 14 that dissipate minimum wattage. As is stated in post 14, this is true of all transistors, not just MOSFETs.
Thanks Bill...

........the Op-amp must be a High Speed one right?..what did you use?
What will be the value of the Resistor & capacitor if I need an output of 60Hz?
...I will use this formula( f = 0.7/ RC )?

The circuit I drew in my second post (though I would merge it with the updated triangle wave generator show in post 10) would work with MOSFET, because it is fundamentally a digital circuit (a Class D amp).
You mean like this one?

#### Attachments

• 3.6 KB Views: 110
Last edited:

#### JUAN DELA CRUZ

Joined May 27, 2008
121
1) The first opamp's output is a square-wave, not a triangle-wave
Thanks Audioguru

I think...
..... I can now manage to Stabilize the output voltage of
My old Modified Sine Wave Inverter by varying the square wave first using op-amps before going to the NOR gate to make a M.S-wave.
Do you think it could work?

Anyway, I'm interested in making a Class D Amp. first, to make a Sine wave Inverter.

Last edited:

#### Wendy

Joined Mar 24, 2008
23,134
I don't think op amp speed is critical, but if I'm wrong use a Schmitt Trigger (shown) to square up the edges.

I redrew the schematic a bit...

Looks a lot like what you drew...

Last edited:

#### The Electrician

Joined Oct 9, 2007
2,916
I notice that in some of your schematics, you show what I assume is an unused opamp (1 of a quad) with a resistive divider from Vcc to ground, and with the + input connected to that divider. You also show the output of the opamp connected to the - input. But, why do you have the output of the opamp grounded? Won't this cause a lot of unnecessary dissipation in the opamp?

#### Wendy

Joined Mar 24, 2008
23,134
I notice that in some of your schematics, you show what I assume is an unused opamp (1 of a quad) with a resistive divider from Vcc to ground, and with the + input connected to that divider. You also show the output of the opamp connected to the - input. But, why do you have the output of the opamp grounded? Won't this cause a lot of unnecessary dissipation in the opamp?
This is your most fundimental power supply, a voltage follower. It allows me to use a single power supply, but makes the remaining op amps think it is a dual tracking supply. The output impedance of the op amp voltage follower is quite low, a few millohms at best. You'll note I'm not actually using a ground symbol, but something else. That something else is a virtual ground.

You are right about the disappation factor, so you don't load it down. Some loading is inevitable, but that is true of the other op amps too.

The main reason for it is it brings the other circuits to extreme predictablility. If I used just a voltage divider for the Vcc/2 the resistors would become ensnarled in the gain equations and worse, this way you can think of it as a virtual ground and treat it as such. For things like active filters and other designs this is pretty important.

Last edited:

#### Wendy

Joined Mar 24, 2008
23,134
It would be more correct to say I don't think it is. This is an opinion, subject to change if the facts don't support it. Even the worst op amp (which the 741 qualifies even though I defend it occasionally) has a .7V/us slew rate, which means every microsecond the op amp can move a max of .7V. It's cousins, the LM324, a pretty decent quad, isn't much better. If you want to upgrade the op amp a LM353 would be a good route, we discussed a lot of these ideas in [url="http://forum.allaboutcircuits.com/showthread.php?t=11847]this[/url] thread, which I have bookmarked for future reference.

My basic concept is if you run into trouble you can always add a 555 or some other digitial schmitt trigger (CMOS for example) later. This is why breadboarding a design is desirable, so you can make changes is good enough isn't.

#### Audioguru

Joined Dec 20, 2007
11,248
I notice that in some of your schematics, you show what I assume is an unused opamp (1 of a quad) with a resistive divider from Vcc to ground, and with the + input connected to that divider. You also show the output of the opamp connected to the - input. But, why do you have the output of the opamp grounded? Won't this cause a lot of unnecessary dissipation in the opamp?
The extra opamp buffers the 'half the supply bias voltage" that biases the comparator. The opamp is not needed since the input of the comparator has an extremely low DC current.

#### Audioguru

Joined Dec 20, 2007
11,248
Hi Juan,
Will the high frequency PWM signal go through your low frequency mains transformer??
A class-D amplifier does not have an output transformer.

So far the circuit does not provide for some dead-time to prevent both output Mosfets from shorting the supply (shoot-through current) when one is turning off and the other is turning on.

A PWM controller IC has everything that is needed. The TL494 PWM controller IC costs $.52US each at Digikey today. They have thousands in stock. Class-D amplifier ICs are also available. #### Attachments • 20.6 KB Views: 95 #### Wendy Joined Mar 24, 2008 23,134 The extra opamp buffers the 'half the supply bias voltage" that biases the comparator. The opamp is not needed since the input of the comparator has an extremely low DC current. I have to disagree with that assessment. Two things, it allows the setting of the input impedance of the audio in for a fixed value, it isn't just for the comparator, and it also sets up the integrator for the RC oscillator for a fixed mid level range and prevents the Audio In from getting involved with this oscillator. It's either two resistor dividers or that op amp circuit. The op amp allows for maximum linearity, if two separate resistor divider bridges were used they would not quite match exactly. Might not matter much, but it is also there if any other biasing needs arise. I have seen Class D amps go through transformers, this is a design decision. The switching delay could be a major problem which might be addressed with a combination of Schmitt Triggers and logic. Last edited: #### Wendy Joined Mar 24, 2008 23,134 I had a thought that might address the dead time concerns. I modified the schematic as shown... With this drawing it would turn on a little slower and turn off a little faster so there were never be a condition where both outputs would be on at the same time, plus there would be the inverted signal as a bonus. The 47Ω resistors values were arbitrary, and might need tweaking. Better? Last edited: #### Audioguru Joined Dec 20, 2007 11,248 Keep adding parts and you will make a discrete TL494 at 100 times its price and 100 times its space. Now add feedback for a regulated output voltage. The TL494 has it. Thread Starter #### JUAN DELA CRUZ Joined May 27, 2008 121 A PWM controller IC has everything that is needed. The TL494 PWM controller IC costs$.52US each at Digikey today. They have thousands in stock.
Class-D amplifier ICs are also available.
Thanks Audioguru...
Can you give me further idea/ LINK in making a Sine Wave Inverter using TL494 IC?
.....using a Class D amp IC is more efficient??