## how do pure sinewave invertors work?

Expand Messages
• I am curious about the how they create pure sine wave output on these invertors? I understand that the modified sine wave use pulse width modulation to
Message 1 of 30 , Jun 20, 2014
I am curious about the "how" they create "pure sine wave" output on these
invertors?

I understand that the modified sine wave use pulse width modulation to
roughly approximate a sine wave.

Do the "pure sine wave" invertors do the same thing except the use a much
finer PWM with better feedback to maintain a specific quality of the "sine
wave shape" under varying load conditions?

Or, instead of very closely approximating with PWM, do pure sine wave
inverters use a linear component and literally change the voltage output to
make the pure sine wave without using PWM?

I know if you smoothly vary the voltage to a transistor Base, the "output" (
collector-emitter voltage ) will vary smoothly (within limits).

But, for an inverter carrying heavy loads, wouldn't most transistors waste a
lot power through excessive heat?

Is there another electronic component that allows this process to be done
efficiently?

Always trying to learn,

Dan Nicoson
• Dan, It s a very good question. Here in hawaii, they have a LOT of roof top solar panels. And they feed the grid during the daytime and then the house use (get
Message 2 of 30 , Jun 20, 2014
Dan,

It's a very good question. Here in hawaii, they have a LOT of roof top solar panels. And they feed the grid during the daytime and then the house use (get fed) by the grid when there no sun or weak sun.

I always wondered how they (a) create sine wave to feed the power into the grid (b) sync the phase.

Also, for the current to flow backwards into the grid, the voltage has to be higher than the grid voltage, but if everybody in town has a panel, then the average grid voltage will be higher again and get no current (or power) flowing into the grid.

What up ????

=====

But first, the original OP's question.

Jong

On Jun 20, 2014, at 7:17 PM, "'A6intruder@...' A6intruder@... [Electronics_101]" <Electronics_101@yahoogroups.com> wrote:

> I am curious about the "how" they create "pure sine wave" output on these
> invertors?
>
> I understand that the modified sine wave use pulse width modulation to
> roughly approximate a sine wave.
>
> Do the "pure sine wave" invertors do the same thing except the use a much
> finer PWM with better feedback to maintain a specific quality of the "sine
> wave shape" under varying load conditions?
>
> Or, instead of very closely approximating with PWM, do pure sine wave
> inverters use a linear component and literally change the voltage output to
> make the pure sine wave without using PWM?
>
> I know if you smoothly vary the voltage to a transistor Base, the "output" (
> collector-emitter voltage ) will vary smoothly (within limits).
>
> But, for an inverter carrying heavy loads, wouldn't most transistors waste a
> lot power through excessive heat?
>
> Is there another electronic component that allows this process to be done
> efficiently?
>
> Always trying to learn,
>
> Dan Nicoson
>
>
>
> ------------------------------------
> Posted by: "A6intruder@..." <a6intruder@...>
> ------------------------------------
>
> ------------------------------------
>
>
>
>
• I don t know what you consider pure . In all large UPS systems, the waveform is generated as a kind of staircase by adding certain harmonics of the
Message 3 of 30 , Jun 21, 2014

I don't know what you consider 'pure'.  In all large UPS systems, the waveform is generated as a kind of staircase by adding certain harmonics of the fundamental frequency.  It doesn't take very many harmonics to get within a few percent of 'pure'.  It isn't perfect but it's nothing that a low pass filter can't straighten out.  And who cares?  There's no award for the lowest THD on utility power.  Most devices connected to utility power don't really care what the THD is and would be quite content with a square wave.  Motors don't care at all, PCs and other devices with switching supplies probably don't care.  And so on...

As a practical matter, no individual solar array is going to be able to drive the grid.  Even the small transformers supplying power to a few homes are pretty stiff compared to a solar setup.

I don't know if the solar arrays source power by leading in phase or by having a somewhat higher voltage.  It might work either way but utilities buy and sell power by adjusting their generators to slightly lead or lag the phase of the grid.  Synching to the utility is no problem.  There's no way in the world to get ahead of the utility phase angle!  Just drive as much current as the device is capable of providing.  It will go somewhere...

Richard

• The following is a link to a pure sine wave inverter that uses a 555.
Message 4 of 30 , Jun 21, 2014
The following is a link to a pure sine wave inverter that uses a 555.

For now it looks like black magic in how they get a pure sine wave from the PWM pulses.  When  I get more time in the future I will use LTspice to simulate the circuit to examine the waveforms.

The Home Made Circuits web site shows several other pure sine wave inverter circuits.

The other Howard

On 6/21/2014 12:17 AM, 'A6intruder@...' A6intruder@... [Electronics_101] wrote:

I am curious about the "how" they create "pure sine wave" output on these
invertors?

I understand that the modified sine wave use pulse width modulation to
roughly approximate a sine wave.

Do the "pure sine wave" invertors do the same thing except the use a much
finer PWM with better feedback to maintain a specific quality of the "sine
wave shape" under varying load conditions?

Or, instead of very closely approximating with PWM, do pure sine wave
inverters use a linear component and literally change the voltage output to
make the pure sine wave without using PWM?

I know if you smoothly vary the voltage to a transistor Base, the "output" (
collector-emitter voltage ) will vary smoothly (within limits).

But, for an inverter carrying heavy loads, wouldn't most transistors waste a
lot power through excessive heat?

Is there another electronic component that allows this process to be done
efficiently?

Always trying to learn,

Dan Nicoson

• I am guessing that pure sine wave may be a relative term. I ve seen the (filtered) waveform some inverters generate and it is neither sinusoidal nor square;
Message 5 of 30 , Jun 21, 2014
I am guessing that "pure sine wave" may be a relative term.  I've seen the (filtered) waveform some inverters generate and it is neither sinusoidal nor square; something like a badly distorted sine or triangle wave ... a first-order attempt to crudely approximate a sine wave without trying very hard.  So perhaps "pure sine wave" means using the same PWM technique but doing a better job of approximating the sinewave.

It is technically possible to generate a nice sine wave either way: by using PWM with the pulsewidth control so that its filtered waveform is a good sinewave, or by actually generating and amplifying a sinewave with linear techniques.  If you think of the inverter output as a power amplifier (which it essentially is), recall that amplifiers have types like Class A, Class B, Class D, etc.  A Class B (or AB) amp has excellent linearity so it could drive a clean sinewave if you give it one at its input, but it has poor efficiency, so it wastes a LOT of power as heat.  You wouldn't do that if you needed to drive a heavy load.  A Class D amp wastes less power but its output is binary and it requires low-pass (or bandpass) filtering to look right.

I remember from a few decades ago a device that generated a 120V 60Hz waveform, which actually used a linear audio power amplifier module, driven by a sine wave source.  The output drove a transformer to step up to 120 V AC.  (This device was designed to synchronize an Ampex tape recorder to something else.)

If you are feeding power into the grid, it should adjust the phase to match the grid (to make the power factor zero), then adjust the driven output amplitude so that its voltage is ever slightly greater than the grid's voltage, whatever that is at the moment.  The easier way to do the latter is to drive a current rather than a voltage waveform.  A phase lead or lag does not mean power flowing to vs. from the grid.  It just means power factor, with no net power transfer.

Regards,
Andy

• I was working for a government supply company. Came up request for a 1 KW 24 Vdc to 120 Vac sine wave power supply. I did a little innovation and using
Message 6 of 30 , Jun 22, 2014

I was working for a government supply company.   Came up request for a 1 KW 24 Vdc to 120 Vac sine wave power supply.  I did a little innovation and using several IR2110s and a PIC uP made a very nice sine wave power supply.   The PIC had a quarter cycle look-up table for the sine wave PWM data.  Then I would go up or down the quarter wave and invert the quarter wave to get the opposite cycle.   I had some 256 PWM points that made a very nice smooth 60 Hz sine wave.

Inside the DC to DC converter was a voltage regulator that kept the output level constant as the battery ran down in voltage.  The circuit design had very good efficiency as it was all a switching power supply.    The whole circuit was patentable as it was way off the usual power supply design.   But the company only wanted the paperwork on the power supply, so only the prototype was built to prove the design.

Yeah, it can be done!

Larry

From: Electronics_101@yahoogroups.com [mailto:Electronics_101@yahoogroups.com]
Sent: Saturday, June 21, 2014 10:21 PM
To: Electronics_101@yahoogroups.com
Subject: Re: [Electronics_101] how do pure sinewave invertors work?

I am guessing that "pure sine wave" may be a relative term.  I've seen the (filtered) waveform some inverters generate and it is neither sinusoidal nor square; something like a badly distorted sine or triangle wave ... a first-order attempt to crudely approximate a sine wave without trying very hard.  So perhaps "pure sine wave" means using the same PWM technique but doing a better job of approximating the sinewave.

It is technically possible to generate a nice sine wave either way: by using PWM with the pulsewidth control so that its filtered waveform is a good sinewave, or by actually generating and amplifying a sinewave with linear techniques.  If you think of the inverter output as a power amplifier (which it essentially is), recall that amplifiers have types like Class A, Class B, Class D, etc.  A Class B (or AB) amp has excellent linearity so it could drive a clean sinewave if you give it one at its input, but it has poor efficiency, so it wastes a LOT of power as heat.  You wouldn't do that if you needed to drive a heavy load.  A Class D amp wastes less power but its output is binary and it requires low-pass (or bandpass) filtering to look right.

I remember from a few decades ago a device that generated a 120V 60Hz waveform, which actually used a linear audio power amplifier module, driven by a sine wave source.  The output drove a transformer to step up to 120 V AC.  (This device was designed to synchronize an Ampex tape recorder to something else.)

If you are feeding power into the grid, it should adjust the phase to match the grid (to make the power factor zero), then adjust the driven output amplitude so that its voltage is ever slightly greater than the grid's voltage, whatever that is at the moment.  The easier way to do the latter is to drive a current rather than a voltage waveform.  A phase lead or lag does not mean power flowing to vs. from the grid.  It just means power factor, with no net power transfer.

Regards,

Andy

• I simulated this in LTspice and don t get the PWM output that the circuit claims. Pretty much 50% square waves, which is what I would expect given that the
Message 7 of 30 , Jul 6, 2014
I simulated this in LTspice and don't get the PWM output that the circuit claims. Pretty much 50% square waves, which is what I would expect given that the square wave generator (driving the reset on the 555) is synchronised with the triangle-wave generator. Anyone else get a different result?

Donald.

----- Original Message -----
> From: "Howard Hansen hrhan@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Saturday, June 21, 2014 4:32:44 PM
> Subject: Re: [Electronics_101] how do pure sinewave invertors work?
>
> The following is a link to a pure sine wave inverter that uses a 555.
>
> For now it looks like black magic in how they get a pure sine wave from
> the PWM pulses. When I get more time in the future I will use LTspice
> to simulate the circuit to examine the waveforms.
>
> The Home Made Circuits web site shows several other pure sine wave
> inverter circuits.
>
> The other Howard
>
[snip]
• I got the same disappointing results as you did when I simulated the circuit. . And like you I don t see ow the circuit could possibly work when the
Message 8 of 30 , Jul 7, 2014
I got the same disappointing results as you did when I simulated the circuit. .   And like you I don't see ow the circuit could possibly work when the modulating frequency is the same frequency as the PWM frequency.

The other Howard

On 7/6/2014 7:35 PM, Donald H Locker dhlocker@... [Electronics_101] wrote:

I simulated this in LTspice and don't get the PWM output that the circuit claims. Pretty much 50% square waves, which is what I would expect given that the square wave generator (driving the reset on the 555) is synchronised with the triangle-wave generator. Anyone else get a different result?

Donald.

----- Original Message -----
> From: "Howard Hansen hrhan@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Saturday, June 21, 2014 4:32:44 PM
> Subject: Re: [Electronics_101] how do pure sinewave invertors work?
>
> The following is a link to a pure sine wave inverter that uses a 555.
>
> For now it looks like black magic in how they get a pure sine wave from
> the PWM pulses. When I get more time in the future I will use LTspice
> to simulate the circuit to examine the waveforms.
>
> The Home Made Circuits web site shows several other pure sine wave
> inverter circuits.
>
> The other Howard
>
[snip]

• I m happy to share my .asc file (LTspice .asc file, using only standard LTspice components) for anyone else who wishes to try. I keep thinking perhaps I m
Message 9 of 30 , Jul 8, 2014
I'm happy to share my .asc file (LTspice .asc file, using only standard LTspice components) for anyone else who wishes to try. I keep thinking perhaps I'm missing something simple or fundamental or obvious.

Donald.

----- Original Message -----
> From: "Howard Hansen hrhan@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Monday, July 7, 2014 8:05:23 PM
> Subject: Re: [Electronics_101] how do pure sinewave invertors work?
>
> I got the same disappointing results as you did when I simulated the
> circuit. . And like you I don't see ow the circuit could possibly work
> when the modulating frequency is the same frequency as the PWM frequency.
>
> The other Howard
>
>
> On 7/6/2014 7:35 PM, Donald H Locker dhlocker@...
> [Electronics_101] wrote:
> >
> > I simulated this in LTspice and don't get the PWM output that the
> > circuit claims. Pretty much 50% square waves, which is what I would
> > expect given that the square wave generator (driving the reset on the
> > 555) is synchronised with the triangle-wave generator. Anyone else get
> > a different result?
> >
> > Donald.
> >
> > ----- Original Message -----
> > > From: "Howard Hansen hrhan@... [Electronics_101]"
> > <Electronics_101@yahoogroups.com>
> > > To: "Electronics 101" <Electronics_101@yahoogroups.com>
> > > Sent: Saturday, June 21, 2014 4:32:44 PM
> > > Subject: Re: [Electronics_101] how do pure sinewave invertors work?
> > >
> > > The following is a link to a pure sine wave inverter that uses a 555.
> > >
> > >
> > > For now it looks like black magic in how they get a pure sine wave from
> > > the PWM pulses. When I get more time in the future I will use LTspice
> > > to simulate the circuit to examine the waveforms.
> > >
> > > The Home Made Circuits web site shows several other pure sine wave
> > > inverter circuits.
> > >
> > >
> > > The other Howard
> > >
> > [snip]
> >
> >
>
>
• Please post your .asc file in the Temp folder and let the group know. I am an absolute beginner with LTSpice but I would like to see what you have. The thing
Message 10 of 30 , Jul 8, 2014

Please post your .asc file in the Temp folder and let the group know.  I am an absolute beginner with LTSpice but I would like to see what you have.  The thing is, without a lot of study, it seems to me that using a triangular waveform to change the width and a square wave to trigger the timer, well, it ought to work.  Apparently, under LTSpice, it doesn't.

There are 3 possibilities I can think of:  The simulation is not correct (the .asc is wrong), the whole idea is nonsense or LTSpice just can't simulate that circuit even if the .asc is correct.

But it seems right!  The idea of modulating the pulse width should work.  I suppose the easy way to test out the circuit is to actually build it.  I just don't have a satisfactory way of powering up a breadboard with multiple power supplies.  What I need is a 3 output bench type power supply.  Alas, I don't have one...

Richard

• LTspice gets it exactly right. The problems with the design as presented are: 1 - the square wave applied to the 555 Reset force its output to inactive for 50%
Message 11 of 30 , Jul 8, 2014
LTspice gets it exactly right. The problems with the design as presented are:
1 - the square wave applied to the 555 Reset force its output to inactive for 50% of the time.
2 - the triangle wave is generated from the same square wave as resets the 555, so it is always in phase and the timing will always be identical. Unless there is a modulation source other than those shown, the "PWM" output must always be at a fixed duty cycle. (The sequence I see is: square wave low holds off 555; square wave transitions high, releasing 555 and starting triangle ramp up; triangle starts ramping toward threshold; square wave resets 555 and starts triangle ramping down.)

I haven't experimented with a faster (or slower) triangle wave or modulating the triangle wave, but those are not included in the design as it is.

I'll post this evening after work.
Donald.

----- Original Message -----
> From: "rstofer@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Tuesday, July 8, 2014 10:39:08 AM
> Subject: Re: [Electronics_101] how do pure sinewave invertors work?
>
> Please post your .asc file in the Temp folder and let the group know. I am
> an absolute beginner with LTSpice but I would like to see what you have.
> The thing is, without a lot of study, it seems to me that using a
> triangular waveform to change the width and a square wave to trigger the
> timer, well, it ought to work. Apparently, under LTSpice, it doesn't.
>
> There are 3 possibilities I can think of: The simulation is not correct
> (the .asc is wrong), the whole idea is nonsense or LTSpice just can't
> simulate that circuit even if the .asc is correct.
>
> But it seems right! The idea of modulating the pulse width should work. I
> suppose the easy way to test out the circuit is to actually build it. I
> just don't have a satisfactory way of powering up a breadboard with
> multiple power supplies. What I need is a 3 output bench type power
> supply. Alas, I don't have one...
>
> Richard
>
>
>
• The following link shows a 555 PWM. The 555 PWM in the
Message 12 of 30 , Jul 8, 2014
The following link shows a 555 PWM.

The 555 PWM in the above link uses the same method to to generate a PWM output as the 555 PWM in the pure sine wave inverter circuit.  However, the circuits differ in the frequency of the waveforms applied to the  555 IC. in the pure sine wave inverter circuit the PWM frequency and the modulating frequency are the same frequency.  Whereas in the above link the the PWM frequency is a higher frequency than the modulating frequency.  Both Donald and I say the PWM frequency must be higher than the modulating frequency when using a 555 as PWM.

Another thing that is troubling with the pure sine wave inverter circuit is the very low amplitude of the triangle waveform.  With such a low amplitude it is hard to see how this low amplitude signal could vary the pulse width.

For reference here is a link to the pure sine wave inverter circuit.

The other Howard

On 7/8/2014 9:39 AM, rstofer@... [Electronics_101] wrote:

Please post your .asc file in the Temp folder and let the group know.  I am an absolute beginner with LTSpice but I would like to see what you have.  The thing is, without a lot of study, it seems to me that using a triangular waveform to change the width and a square wave to trigger the timer, well, it ought to work.  Apparently, under LTSpice, it doesn't.

There are 3 possibilities I can think of:  The simulation is not correct (the .asc is wrong), the whole idea is nonsense or LTSpice just can't simulate that circuit even if the .asc is correct.

But it seems right!  The idea of modulating the pulse width should work.  I suppose the easy way to test out the circuit is to actually build it.  I just don't have a satisfactory way of powering up a breadboard with multiple power supplies.  What I need is a 3 output bench type power supply.  Alas, I don't have one...

Richard

• It is not black magic, it is called filtering. In the 555 circuit, the output stage transformer seems to be the primary filtering element. When a sudden
Message 13 of 30 , Jul 8, 2014
It is not black magic, it is called filtering. In the 555 circuit, the output stage transformer seems to be the primary filtering element. When a sudden Voltage pulse is applied to an inductance, it takes time for the current to build up. Since the secondary/output winding only "sees" the current on the primary, it ramps up instead of the instantaneous Voltage change of the primary.

Perhaps the Spice simulations are not properly modeling this inductance or the load attached to the secondary? Again, I suspect that the actual waveform will change with different loads/current levels: low or high current, inductive, capacitive, etc.

That being said, I do not see any way of using a square wave and a triangular one to provide the PWM needed to create a pure sine wave. I strongly suspect they are generating something more like a compromise between a triangular wave and a real sine wave and calling it a "pure" sine wave.

---In Electronics_101@yahoogroups.com, <hrhan@...> wrote :

The following is a link to a pure sine wave inverter that uses a 555.

For now it looks like black magic in how they get a pure sine wave from the PWM pulses.  When  I get more time in the future I will use LTspice to simulate the circuit to examine the waveforms.

The Home Made Circuits web site shows several other pure sine wave inverter circuits.

The other Howard

On 6/21/2014 12:17 AM, 'A6intruder@...' A6intruder@... [Electronics_101] wrote:

I am curious about the "how" they create "pure sine wave" output on these
invertors?

I understand that the modified sine wave use pulse width modulation to
roughly approximate a sine wave.

Do the "pure sine wave" invertors do the same thing except the use a much
finer PWM with better feedback to maintain a specific quality of the "sine
wave shape" under varying load conditions?

Or, instead of very closely approximating with PWM, do pure sine wave
inverters use a linear component and literally change the voltage output to
make the pure sine wave without using PWM?

I know if you smoothly vary the voltage to a transistor Base, the "output" (
collector-emitter voltage ) will vary smoothly (within limits).

But, for an inverter carrying heavy loads, wouldn't most transistors waste a
lot power through excessive heat?

Is there another electronic component that allows this process to be done
efficiently?

Always trying to learn,

Dan Nicoson

• On 07/08/2014 03:11 PM, palciatore@gt.rr.com ... (snip) There is something not quite right, there. The instantaneous voltage applied to the primary is the
Message 14 of 30 , Jul 8, 2014
On 07/08/2014 03:11 PM, palciatore@...
[Electronics_101] wrote:
>
>
> It is not black magic, it is called filtering. In the 555
> circuit, the output stage transformer seems to be the
> primary filtering element. When a sudden Voltage pulse is
> applied to an inductance, it takes time for the current to
> build up. Since the secondary/output winding only "sees" the
> current on the primary, it ramps up instead of the
> instantaneous Voltage change of the primary.
(snip)

There is something not quite right, there. The
instantaneous voltage applied to the primary is the same as
the instantaneous voltage produced by the secondary, (in
volts per turn), assuming the windings are tightly coupled.

So a normal, well coupled transformer does not make much of
a low pass filter.

--
Regards,

John Popelish
• If the PWM duty cycle doesn t vary, the output voltage won t either. It has nothing to do with filtering. Donald. ... [snip]
Message 15 of 30 , Jul 8, 2014
If the PWM duty cycle doesn't vary, the output voltage won't either. It has nothing to do with filtering.

Donald.

----- Original Message -----
> From: "palciatore@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Tuesday, July 8, 2014 3:11:59 PM
> Subject: [Electronics_101] Re: how do pure sinewave invertors work?
>
> It is not black magic, it is called filtering. In the 555 circuit, the output
> stage transformer seems to be the primary filtering element. When a sudden
> Voltage pulse is applied to an inductance, it takes time for the current to
> build up. Since the secondary/output winding only "sees" the current on the
> primary, it ramps up instead of the instantaneous Voltage change of the
> primary.
>
> Perhaps the Spice simulations are not properly modeling this inductance or
> the load attached to the secondary? Again, I suspect that the actual
> waveform will change with different loads/current levels: low or high
> current, inductive, capacitive, etc.
>
> That being said, I do not see any way of using a square wave and a triangular
> one to provide the PWM needed to create a pure sine wave. I strongly suspect
> they are generating something more like a compromise between a triangular
> wave and a real sine wave and calling it a "pure" sine wave.
>
>
>
> ---In Electronics_101@yahoogroups.com, <hrhan@...> wrote :
>
> The following is a link to a pure sine wave inverter that uses a 555.
>
> For now it looks like black magic in how they get a pure sine wave from the
> PWM pulses. When I get more time in the future I will use LTspice to
> simulate the circuit to examine the waveforms.
>
> The Home Made Circuits web site shows several other pure sine wave inverter
> circuits.
>
> The other Howard
>

[snip]
• has the model. There is only one file there SinInverter.asc
Message 16 of 30 , Jul 8, 2014
<https://groups.yahoo.com/neo/groups/Electronics_101/files/Temp/ClaimedSineWaveInverter-555/> has the model. There is only one file there "SinInverter.asc"

Enjoy.
Donald.
----- Original Message -----
> From: "rstofer@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Tuesday, July 8, 2014 10:39:08 AM
> Subject: Re: [Electronics_101] how do pure sinewave invertors work?
>
> Please post your .asc file in the Temp folder and let the group know. I am
> an absolute beginner with LTSpice but I would like to see what you have.
> The thing is, without a lot of study, it seems to me that using a
> triangular waveform to change the width and a square wave to trigger the
> timer, well, it ought to work. Apparently, under LTSpice, it doesn't.
>
> There are 3 possibilities I can think of: The simulation is not correct
> (the .asc is wrong), the whole idea is nonsense or LTSpice just can't
> simulate that circuit even if the .asc is correct.
>
> But it seems right! The idea of modulating the pulse width should work. I
> suppose the easy way to test out the circuit is to actually build it. I
> just don't have a satisfactory way of powering up a breadboard with
> multiple power supplies. What I need is a 3 output bench type power
> supply. Alas, I don't have one...
>
> Richard
>
>
>
• Sorry for the misdirection. The square wave is applied to the trigger input (not reset). The result is just inverted from what I was thinking. Same 50%
Message 17 of 30 , Jul 8, 2014
Sorry for the misdirection. The square wave is applied to the trigger input (not reset). The result is just inverted from what I was thinking. Same 50% waveform at 1.8+/-kHz

Donald.

----- Original Message -----
> From: "Locker, Donald" <dhlocker@...>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Tuesday, July 8, 2014 11:52:02 AM
> Subject: Re: [Electronics_101] how do pure sinewave invertors work?
>
> LTspice gets it exactly right. The problems with the design as presented are:
> 1 - the square wave applied to the 555 Reset force its output to inactive
> for 50% of the time.
> 2 - the triangle wave is generated from the same square wave as resets the
> 555, so it is always in phase and the timing will always be identical.
> Unless there is a modulation source other than those shown, the "PWM"
> output must always be at a fixed duty cycle. (The sequence I see is:
> square wave low holds off 555; square wave transitions high, releasing 555
> and starting triangle ramp up; triangle starts ramping toward threshold;
> square wave resets 555 and starts triangle ramping down.)
>
> I haven't experimented with a faster (or slower) triangle wave or modulating
> the triangle wave, but those are not included in the design as it is.
>
> I'll post this evening after work.
> Donald.
>
> ----- Original Message -----
> > From: "rstofer@... [Electronics_101]"
> > <Electronics_101@yahoogroups.com>
> > To: "Electronics 101" <Electronics_101@yahoogroups.com>
> > Sent: Tuesday, July 8, 2014 10:39:08 AM
> > Subject: Re: [Electronics_101] how do pure sinewave invertors work?
> >
> > Please post your .asc file in the Temp folder and let the group know. I am
> > an absolute beginner with LTSpice but I would like to see what you have.
> > The thing is, without a lot of study, it seems to me that using a
> > triangular waveform to change the width and a square wave to trigger the
> > timer, well, it ought to work. Apparently, under LTSpice, it doesn't.
> >
> > There are 3 possibilities I can think of: The simulation is not correct
> > (the .asc is wrong), the whole idea is nonsense or LTSpice just can't
> > simulate that circuit even if the .asc is correct.
> >
> > But it seems right! The idea of modulating the pulse width should work.
> > I
> > suppose the easy way to test out the circuit is to actually build it. I
> > just don't have a satisfactory way of powering up a breadboard with
> > multiple power supplies. What I need is a 3 output bench type power
> > supply. Alas, I don't have one...
> >
> > Richard
> >
> >
> >
>
• Naught to do with filtering. A 50% duty cycle will remain 50% no matter how well filtered. The output will be more or less noisy, but will remain at 50% of the
Message 18 of 30 , Jul 8, 2014
Naught to do with filtering. A 50% duty cycle will remain 50% no matter how well filtered. The output will be more or less noisy, but will remain at 50% of the input voltage.

Donald.

----- Original Message -----
> From: "palciatore@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Tuesday, July 8, 2014 3:11:59 PM
> Subject: [Electronics_101] Re: how do pure sinewave invertors work?
>
> It is not black magic, it is called filtering. In the 555 circuit, the output
> stage transformer seems to be the primary filtering element. When a sudden
> Voltage pulse is applied to an inductance, it takes time for the current to
> build up. Since the secondary/output winding only "sees" the current on the
> primary, it ramps up instead of the instantaneous Voltage change of the
> primary.
>
> Perhaps the Spice simulations are not properly modeling this inductance or
> the load attached to the secondary? Again, I suspect that the actual
> waveform will change with different loads/current levels: low or high
> current, inductive, capacitive, etc.
>
> That being said, I do not see any way of using a square wave and a triangular
> one to provide the PWM needed to create a pure sine wave. I strongly suspect
> they are generating something more like a compromise between a triangular
> wave and a real sine wave and calling it a "pure" sine wave.
>
>

[snip]
• Err, excuse me, but my physics professors must have been wrong. I was taught that it is the current in the windings that produces the magnetic field in a
Message 19 of 30 , Jul 9, 2014
Err, excuse me, but my physics professors must have been wrong. I was taught that it is the current in the windings that produces the magnetic field in a transformer and it is the CHANGE in that magnetic field that induces a Voltage/current in the secondary. When a sine wave is applied to the primary winding the current in that winding lags 90 degrees behind the instantaneous applied Voltage. This is the primary difference between an inductor and a resistor where Voltage and current are always in phase. In the inductor/transformer, this produces a corresponding lag in the Voltage induced in the secondary winding.

Now, with a single frequency sine wave, this is of little importance because you still get a sine wave output on the secondary, just 90 degrees later. Works well for the power grid: nobody cares if the power is delayed by 1/240th of a second by each transformer on the route from the power house to your home. And, according to both mathematical theory and experimental fact, the output will still be a sine wave. All in all, nobody ever notices any difference.

BUT, if you apply a square wave to the primary, that square wave, according to Fourier Analysis, is actually a series of sine waves of different, harmonically related frequencies, superimposed over each other (added together). And they each behave exactly like a sine wave of it's own, individual frequency. So those 90 degree phase shifts in an inductor or transformer become different amounts of actual time shifted for each one of them. The lower frequencies will be shifted by larger amounts of time and the higher frequencies will be shifted by shorter amounts. This has the effect of changing the shape or form of the wave. The sum total effect will be a rounding off of that sharp rise time of the square wave. It will no longer have an instantaneous rise time, but instead it will take some time to reach the maximum value. If the circuit values are adjusted to bring that rise time to about one half of 1/60 of a second (power line frequency / 2) then this rising edge will start to look a lot like the rising half of a sine wave. The same thing occurs on the trailing edge of the square wave. So, with the proper circuit values, square wave in and sine wave (or at least an approximation of one) out.

FILTERING!

This effect is one of the reasons why the frequency of a digital signal is limited over a transmission line, weather it be twisted pair, coaxial, waveguide or even fiber optic. All of these lines exhibit both inductive and capacitive effects and these, in turn induce different delays for the different frequencies present in a pulse (square wave) that represents a digit. Thus, over the length of the line the leading and trailing edges of these pulses will be stretched out until the trailing edge of one pulse reaches the leading edge of the following one and the two pulses merge into a single one and the data is lost.

Paul A.

---In Electronics_101@yahoogroups.com, <jpopelish@...> wrote :

On 07/08/2014 03:11 PM, palciatore@...
[Electronics_101] wrote:
>
>
> It is not black magic, it is called filtering. In the 555
> circuit, the output stage transformer seems to be the
> primary filtering element. When a sudden Voltage pulse is
> applied to an inductance, it takes time for the current to
> build up. Since the secondary/output winding only "sees" the
> current on the primary, it ramps up instead of the
> instantaneous Voltage change of the primary.
(snip)

There is something not quite right, there. The
instantaneous voltage applied to the primary is the same as
the instantaneous voltage produced by the secondary, (in
volts per turn), assuming the windings are tightly coupled.

So a normal, well coupled transformer does not make much of
a low pass filter.

--
Regards,

John Popelish
• This is all true, modulo physical reality and parasitics, but the effect that we are looking at is from a relatively high-frequency square wave, unmodulated
Message 20 of 30 , Jul 9, 2014
This is all true, modulo physical reality and parasitics, but the effect that we are looking at is from a relatively high-frequency square wave, unmodulated (despite the triangle wave applied to the 555 control pin). All output sine waves based on this square wave will be odd harmonics of the fundamental frequency. If the transformer succeeds at filtering all higher harmonics, the output will be a high-frequency sine wave, not at all synthesised by PWM, quite contrary to what the article claims the circuit was designed to provide.

Donald.

----- Original Message -----
> From: "palciatore@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Wednesday, July 9, 2014 3:41:11 PM
> Subject: Re: [Electronics_101] Re: how do pure sinewave invertors work?
>
> Err, excuse me, but my physics professors must have been wrong. I was taught
> that it is the current in the windings that produces the magnetic field in a
> transformer and it is the CHANGE in that magnetic field that induces a
> Voltage/current in the secondary. When a sine wave is applied to the primary
> winding the current in that winding lags 90 degrees behind the instantaneous
> applied Voltage. This is the primary difference between an inductor and a
> resistor where Voltage and current are always in phase. In the
> inductor/transformer, this produces a corresponding lag in the Voltage
> induced in the secondary winding.
>
> Now, with a single frequency sine wave, this is of little importance because
> you still get a sine wave output on the secondary, just 90 degrees later.
> Works well for the power grid: nobody cares if the power is delayed by
> 1/240th of a second by each transformer on the route from the power house to
> your home. And, according to both mathematical theory and experimental fact,
> the output will still be a sine wave. All in all, nobody ever notices any
> difference.
>
> BUT, if you apply a square wave to the primary, that square wave, according
> to Fourier Analysis, is actually a series of sine waves of different,
> harmonically related frequencies, superimposed over each other (added
> together). And they each behave exactly like a sine wave of it's own,
> individual frequency. So those 90 degree phase shifts in an inductor or
> transformer become different amounts of actual time shifted for each one of
> them. The lower frequencies will be shifted by larger amounts of time and
> the higher frequencies will be shifted by shorter amounts. This has the
> effect of changing the shape or form of the wave. The sum total effect will
> be a rounding off of that sharp rise time of the square wave. It will no
> longer have an instantaneous rise time, but instead it will take some time
> to reach the maximum value. If the circuit values are adjusted to bring that
> rise time to about one half of 1/60 of a second (power line frequency / 2)
> then this rising edge will start to look a lot like the rising half of a
> sine wave. The same thing occurs on the trailing edge of the square wave.
> So, with the proper circuit values, square wave in and sine wave (or at
> least an approximation of one) out.
>
> FILTERING!
>
> This effect is one of the reasons why the frequency of a digital signal is
> limited over a transmission line, weather it be twisted pair, coaxial,
> waveguide or even fiber optic. All of these lines exhibit both inductive and
> capacitive effects and these, in turn induce different delays for the
> different frequencies present in a pulse (square wave) that represents a
> digit. Thus, over the length of the line the leading and trailing edges of
> these pulses will be stretched out until the trailing edge of one pulse
> reaches the leading edge of the following one and the two pulses merge into
> a single one and the data is lost.
>
> Paul A.
>
>
> ---In Electronics_101@yahoogroups.com, <jpopelish@...> wrote :
>
> On 07/08/2014 03:11 PM, palciatore@... mailto:palciatore@...
> [Electronics_101] wrote:
> >
> >
> > It is not black magic, it is called filtering. In the 555
> > circuit, the output stage transformer seems to be the
> > primary filtering element. When a sudden Voltage pulse is
> > applied to an inductance, it takes time for the current to
> > build up. Since the secondary/output winding only "sees" the
> > current on the primary, it ramps up instead of the
> > instantaneous Voltage change of the primary.
> (snip)
>
> There is something not quite right, there. The
> instantaneous voltage applied to the primary is the same as
> the instantaneous voltage produced by the secondary, (in
> volts per turn), assuming the windings are tightly coupled.
>
> So a normal, well coupled transformer does not make much of
> a low pass filter.
>
> --
> Regards,
>
> John Popelish
>
>
• OK, I am not doing a SPICE simulation. No time for that. I am working on code for another project. But they do show a pulse width modulated waveform. This is
Message 21 of 30 , Jul 9, 2014
OK, I am not doing a SPICE simulation. No time for that. I am working on code for another project. But they do show a pulse width modulated waveform. This is probably from the triangular wave input at the modulation input (pin 5). So, it would seem that the shorter pulses would only rise to a smaller Voltage value before they end and the longer ones would rise to a higher Voltage value. I suspect that the drawing is not accurate as to the time duration BETWEEN these pulses because the 555 is being triggered by the square wave on pin 2, the TRIGGER input. So they will be at regular intervals. Therefore the start times of the pulses will be at regular intervals and the duration of the pulses will vary. So the off times will be varying also, not constant as their drawing shows.

So the transformer only has to FILTER at the higher frequency of the actual pulses and not at the power line frequency. Or some combination of the two. But my argument, that it is the filter, still holds. And, as I also said, the secondary winding in conjunction with the load resistance/impedance will only add to this filtering effect. There is virtually no way that you will see that PWM square wave at the load. And it is the Voltage/current at the load that counts.

If you are doing a SPICE simulation, please include all of the circuit; all the way to the load. And get and use the actual characteristics of the transformer actually shown. Oh, I am sorry, they are not very specific. I doubt that they tested the circuit with more than one transformer so your results may vary.

Now, since we are using a triangular wave instead of a sine wave, the output will not be an actual sine wave. But it will not be a square wave either. Something between sine and triangular.

Or, has anybody actually verified that the second waveform is triangular and not sine?

Paul A.
• On 07/09/2014 03:41 PM, palciatore@gt.rr.com ... Almost. It is the instantaneous sum of amp turns in all windings of the transformer that produces the core
Message 22 of 30 , Jul 9, 2014
On 07/09/2014 03:41 PM, palciatore@...
[Electronics_101] wrote:
>
>
> Err, excuse me, but my physics professors must have been
> wrong. I was taught that it is the current in the windings
> that produces the magnetic field in a transformer and it is
> the CHANGE in that magnetic field that induces a
> Voltage/current in the secondary.

Almost. It is the instantaneous sum of amp turns in all
windings of the transformer that produces the core magnetic
flux. It is the instantaneous rate of change of that flux
that induces voltage across every turn on the transformer
(both primary and secondary). Even as the flux passes
through zero, as it changes directions, it has a rate of
change that produces voltage across all turns. For a sine
wave, the highest induced voltage takes place, just as the
flux passes through zero, because that is where the highest
rate of change for a sine wave of flux is.

> When a sine wave is
> applied to the primary winding the current in that winding
> lags 90 degrees behind the instantaneous applied Voltage.

If the secondary is unloaded, this is true, because the
primary is just an inductor.

> This is the primary difference between an inductor and a
> resistor where Voltage and current are always in phase. In
> the inductor/transformer, this produces a corresponding lag
> in the Voltage induced in the secondary winding.

There is a lag in magnetization current, not in secondary
voltage.

> Now, with a single frequency sine wave, this is of little
> importance because you still get a sine wave output on the
> secondary, just 90 degrees later.

This is wrong. The unloaded primary produces a current and
flux delayed by 90 degrees, but the rate of change of that
flux leads the flux by 90 degrees, so the secondary voltage
is in phase with the primary.

> Works well for the power
> grid: nobody cares if the power is delayed by 1/240th of a
> second by each transformer on the route from the power house
> to your home. And, according to both mathematical theory and
> experimental fact, the output will still be a sine wave. All
> in all, nobody ever notices any difference.

Except that this difference is not as you say.

> BUT, if you apply a square wave to the primary, that square
> wave, according to Fourier Analysis, is actually a series of
> sine waves of different, harmonically related frequencies,
> superimposed over each other (added together). And they each
> behave exactly like a sine wave of it's own, individual
> frequency. So those 90 degree phase shifts in an inductor or
> transformer become different amounts of actual time shifted
> for each one of them. The lower frequencies will be shifted
> by larger amounts of time and the higher frequencies will be
> shifted by shorter amounts. This has the effect of changing
> the shape or form of the wave. The sum total effect will be
> a rounding off of that sharp rise time of the square wave.
> It will no longer have an instantaneous rise time, but
> instead it will take some time to reach the maximum value.
> If the circuit values are adjusted to bring that rise time
> to about one half of 1/60 of a second (power line frequency
> / 2) then this rising edge will start to look a lot like the
> rising half of a sine wave. The same thing occurs on the
> trailing edge of the square wave. So, with the proper
> circuit values, square wave in and sine wave (or at least an
> approximation of one) out.

This is all just wrong. A perfect transformer uses a little
current to magnetize the core, and that current lags the
primary voltage, but the voltage across the secondary is in
phase with the voltage of the primary.

If you don't believe me, take a look at any small
transformer, primary and secondary voltages, with a two
trace scope. Under no load (where leakage inductance
doesn't make a difference). You are about to learn something.

As the secondary load current increases, that current also
magnetizes the core, but inversely to what the primary is
doing. This lowers the rate of change of the flux,
slightly, decreasing the voltage induced across each turn
(which has been buckling the applied voltage), slightly.
This drop in induced voltage across the primary increases
the current passed through the primary. Effectively, the
its current in in phase with its voltage) gets algebraically
added to the magnetization current of the primary. But
since that load current is in phase with the voltage, the
primary sees a less and less lagged current, as that current
increases. The magnetization component is almost fixed
(actually decreases a little) as the in phase load current
component rises.

> FILTERING!
>
> This effect is one of the reasons why the frequency of a
> digital signal is limited over a transmission line, weather
> it be twisted pair, coaxial, waveguide or even fiber optic.
> All of these lines exhibit both inductive and capacitive
> effects and these, in turn induce different delays for the
> different frequencies present in a pulse (square wave) that
> represents a digit. Thus, over the length of the line the
> leading and trailing edges of these pulses will be stretched
> out until the trailing edge of one pulse reaches the leading
> edge of the following one and the two pulses merge into a
> single one and the data is lost.

Since the effect you describe is not happening, I think you
might want to go back and rethink this paragraph, also.

--
Regards,

John Popelish
• The problem is that the op amps that attempt to modify the control voltage to produce the variable width pulses don t actually do that. If you look at the
Message 23 of 30 , Jul 9, 2014

The problem is that the op amps that attempt to modify the control voltage to produce the variable width pulses don't actually do that.  If you look at the change in control voltage, it is a very tiny signal.  All it manages to do is flip the state of the 'comparator' op amp which, in turn, controls the direction of the slope of the triangle.

The triangle waveform varies from +7.4V to +7.6V and since it is biased at 1/2 of 15V in my case, it is really 7.5V +- 0,1V.  This is simply insignificant when applied to the control input of the 555.

So, all we get is a string of symmetric equal width square waves from the 555 at about 2 kHz.  The photos in the article that show PWM aren't related to how this circuit is working.  I didn't stumble across the solution because it just isn't interesting enough to spend the time.

Richard

• Well, you ve already spent more time writing emails than it would have taken to download and install LTspice and the circuit I posted and to run it. I stand by
Message 24 of 30 , Jul 9, 2014
Well, you've already spent more time writing emails than it would have taken to download and install LTspice and the circuit I posted and to run it.

I stand by my analysis and the simulation. There is no modulation of the pulse width because the [very small] triangle waveform is synchronous with the trigger events.

Donald.

----- Original Message -----
> From: "palciatore@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Wednesday, July 9, 2014 5:01:30 PM
> Subject: [Electronics_101] Re: how do pure sinewave invertors work?
>
> OK, I am not doing a SPICE simulation. No time for that. I am working on code
> for another project. But they do show a pulse width modulated waveform. This
> is probably from the triangular wave input at the modulation input (pin 5).
> So, it would seem that the shorter pulses would only rise to a smaller
> Voltage value before they end and the longer ones would rise to a higher
> Voltage value. I suspect that the drawing is not accurate as to the time
> duration BETWEEN these pulses because the 555 is being triggered by the
> square wave on pin 2, the TRIGGER input. So they will be at regular
> intervals. Therefore the start times of the pulses will be at regular
> intervals and the duration of the pulses will vary. So the off times will be
> varying also, not constant as their drawing shows.
>
> So the transformer only has to FILTER at the higher frequency of the actual
> pulses and not at the power line frequency. Or some combination of the two.
> But my argument, that it is the filter, still holds. And, as I also said,
> the secondary winding in conjunction with the load resistance/impedance will
> only add to this filtering effect. There is virtually no way that you will
> see that PWM square wave at the load. And it is the Voltage/current at the
>
> If you are doing a SPICE simulation, please include all of the circuit; all
> the way to the load. And get and use the actual characteristics of the
> transformer actually shown. Oh, I am sorry, they are not very specific. I
> doubt that they tested the circuit with more than one transformer so your
> results may vary.
>
> Now, since we are using a triangular wave instead of a sine wave, the output
> will not be an actual sine wave. But it will not be a square wave either.
> Something between sine and triangular.
>
> Or, has anybody actually verified that the second waveform is triangular and
> not sine?
>
> Paul A.
• Well, then the circuit diagram is wrong. The author must have made a mistake somewhere because he/she clearly intended for there to be that modulation. I guess
Message 25 of 30 , Jul 9, 2014
Well, then the circuit diagram is wrong. The author must have made a mistake somewhere because he/she clearly intended for there to be that modulation. I guess I looked at it too quickly.

This does not mean that the proper circuit would not work.

You can't believe everything you see on the web.

Paul A.

---In Electronics_101@yahoogroups.com, <dhlocker@...> wrote :

Well, you've already spent more time writing emails than it would have taken to download and install LTspice and the circuit I posted and to run it.

I stand by my analysis and the simulation. There is no modulation of the pulse width because the [very small] triangle waveform is synchronous with the trigger events.

Donald.

----- Original Message -----
> From: "palciatore@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Wednesday, July 9, 2014 5:01:30 PM
> Subject: [Electronics_101] Re: how do pure sinewave invertors work?
>
> OK, I am not doing a SPICE simulation. No time for that. I am working on code
> for another project. But they do show a pulse width modulated waveform. This
> is probably from the triangular wave input at the modulation input (pin 5).
> So, it would seem that the shorter pulses would only rise to a smaller
> Voltage value before they end and the longer ones would rise to a higher
> Voltage value. I suspect that the drawing is not accurate as to the time
> duration BETWEEN these pulses because the 555 is being triggered by the
> square wave on pin 2, the TRIGGER input. So they will be at regular
> intervals. Therefore the start times of the pulses will be at regular
> intervals and the duration of the pulses will vary. So the off times will be
> varying also, not constant as their drawing shows.
>
> So the transformer only has to FILTER at the higher frequency of the actual
> pulses and not at the power line frequency. Or some combination of the two.
> But my argument, that it is the filter, still holds. And, as I also said,
> the secondary winding in conjunction with the load resistance/impedance will
> only add to this filtering effect. There is virtually no way that you will
> see that PWM square wave at the load. And it is the Voltage/current at the
>
> If you are doing a SPICE simulation, please include all of the circuit; all
> the way to the load. And get and use the actual characteristics of the
> transformer actually shown. Oh, I am sorry, they are not very specific. I
> doubt that they tested the circuit with more than one transformer so your
> results may vary.
>
> Now, since we are using a triangular wave instead of a sine wave, the output
> will not be an actual sine wave. But it will not be a square wave either.
> Something between sine and triangular.
>
> Or, has anybody actually verified that the second waveform is triangular and
> not sine?
>
> Paul A.
• The circuit at http://homemadecircuitsandschematics.blogspot.com/2011/12/how-to-build-pure-sine-wave-inverter.html is SERIOUSLY broken. I doubt that anyone
Message 26 of 30 , Jul 9, 2014
The circuit at http://homemadecircuitsandschematics.blogspot.com/2011/12/how-to-build-pure-sine-wave-inverter.html is SERIOUSLY broken. I doubt that anyone actually built it let alone got it to work.

For one thing, the PWM signal should be a much higher frequency than the triangle wave. If you want a 60Hz "sine wave", the triangle wave should be 120 Hz, and the PWM frequency generated by the 555 should be somewhere above 20 KHz.

The Square_Wave signal can't be connected to TRIG. (I'm referring to Locker's LTSpice circuit at https://groups.yahoo.com/neo/groups/Electronics_101/files/Temp/ClaimedSineWaveInverter-555/. Thanks, Don!) TRIG should be connected to a 20 KHz oscillator instead.

R4 and R6 need to be adjusted to get the proper frequency and amplitude of the triangle wave. Of course that's a joke, because the frequency will drift with temperature. It will also be tricky to get the proper voltage swing on CV.

There needs to be a flip-flop between the Square_Wave output and the power section of the circuit. I don't feel like explaining in detail, but it has to do with getting a (more) reasonable output wave shape.

Even if the PWM signal can be filtered (which might happen if the PWM frequency is high enough and there is enough leakage inductance in the transformer and you throw some capacitance on the transformer output...), the output won't be very close to a sine wave. It will be approximately a triangle wave, if you can tune it properly.

A practical, robust circuit also should have some overload protection. This circuit will go from "Sine" wave generator to smoke generator if the output is shorted.

• I spent as little time as possible on this. First look said it can t do what it claimed but I had to do a quick simulation just to verify my expectations. It
Message 27 of 30 , Jul 10, 2014
I spent as little time as possible on this. First look said "it can't do what it claimed" but I had to do a quick simulation just to verify my expectations. It could be made to do something useful, but I have better things to do with my time.

The text accompanying the diagram was reasonably correct, but does not actually describe the behaviour of the circuit as drawn.

Donald.

----- Original Message -----
> From: "Jan Kok jan.kok.5y@... [Electronics_101]" <Electronics_101@yahoogroups.com>
> To: "Electronics 101" <Electronics_101@yahoogroups.com>
> Sent: Thursday, July 10, 2014 2:21:23 AM
> Subject: Re: [Electronics_101] Re: how do pure sinewave invertors work?
>
> The circuit at
> is SERIOUSLY broken. I doubt that anyone actually built it let alone got it
> to work.
>
> For one thing, the PWM signal should be a much higher frequency than the
> triangle wave. If you want a 60Hz "sine wave", the triangle wave should be
> 120 Hz, and the PWM frequency generated by the 555 should be somewhere
> above 20 KHz.
>
> The Square_Wave signal can't be connected to TRIG. (I'm referring to
> Locker's LTSpice circuit at
> https://groups.yahoo.com/neo/groups/Electronics_101/files/Temp/ClaimedSineWaveInverter-555/.
> Thanks, Don!) TRIG should be connected to a 20 KHz oscillator instead.
>
> R4 and R6 need to be adjusted to get the proper frequency and amplitude of
> the triangle wave. Of course that's a joke, because the frequency will
> drift with temperature. It will also be tricky to get the proper voltage
> swing on CV.
>
> There needs to be a flip-flop between the Square_Wave output and the power
> section of the circuit. I don't feel like explaining in detail, but it has
> to do with getting a (more) reasonable output wave shape.
>
> Even if the PWM signal can be filtered (which might happen if the PWM
> frequency is high enough and there is enough leakage inductance in the
> transformer and you throw some capacitance on the transformer output...),
> the output won't be very close to a sine wave. It will be approximately a
> triangle wave, if you can tune it properly.
>
> A practical, robust circuit also should have some overload protection. This
> circuit will go from "Sine" wave generator to smoke generator if the output
> is shorted.
>
• Donald s results exactly match mine whether I use his simulation circuit or mine. His is more faithful to the published design, I rolled the fixed resistors
Message 28 of 30 , Jul 10, 2014

Donald's results exactly match mine whether I use his simulation circuit or mine.  His is more faithful to the published design, I rolled the fixed resistors into the potentiometers (and then used a fixed value) since they were only there to protect from zero ohms on the pot and their fixed value was swamped by the pot value.

When I printed the schematic from the web page, I got 26 pages of comments.  I didn't read through them but there might be a correction buried in there somewhere.

The Internet is full of examples of using a 555 to generate PWM.  Some of them may actually work.

Richard

• OK, I admit that I did not think this out properly. Yes, as you say, the secondary Voltage is in phase with the primary. I am going to cut that part of your
Message 29 of 30 , Jul 11, 2014
OK, I admit that I did not think this out properly. Yes, as you say, the secondary Voltage is in phase with the primary. I am going to cut that part of your argument out below as you are right and we do not need to debate it.

---In Electronics_101@yahoogroups.com, <jpopelish@...> wrote :

On 07/09/2014 03:41 PM, palciatore@...
[Electronics_101] wrote:
>>
>>
>> Err, excuse me,

...<snip>...

>> BUT, if you apply a square wave to the primary, that square
>> wave, according to Fourier Analysis, is actually a series of
>> sine waves of different, harmonically related frequencies,
>> superimposed over each other (added together). And they each
>> behave exactly like a sine wave of it's own, individual
>> frequency. So those 90 degree phase shifts in an inductor or
>> transformer become different amounts of actual time shifted
>> for each one of them. The lower frequencies will be shifted
>> by larger amounts of time and the higher frequencies will be
>> shifted by shorter amounts. This has the effect of changing
>> the shape or form of the wave. The sum total effect will be
>> a rounding off of that sharp rise time of the square wave.
>> It will no longer have an instantaneous rise time, but
>> instead it will take some time to reach the maximum value.
>> If the circuit values are adjusted to bring that rise time
>> to about one half of 1/60 of a second (power line frequency
>> / 2) then this rising edge will start to look a lot like the
>> rising half of a sine wave. The same thing occurs on the
>> trailing edge of the square wave. So, with the proper
>> circuit values, square wave in and sine wave (or at least an
>> approximation of one) out.

>This is all just wrong. A perfect transformer uses a little
>current to magnetize the core, and that current lags the
>primary voltage, but the voltage across the secondary is in
>phase with the voltage of the primary.
>
>If you don't believe me, take a look at any small
>transformer, primary and secondary voltages, with a two
>trace scope. Under no load (where leakage inductance
>doesn't make a difference). You are about to learn something.
>
>As the secondary load current increases, that current also
>magnetizes the core, but inversely to what the primary is
>doing. This lowers the rate of change of the flux,
>slightly, decreasing the voltage induced across each turn
>(which has been buckling the applied voltage), slightly.
>This drop in induced voltage across the primary increases
>the current passed through the primary. Effectively, the
>its current in in phase with its voltage) gets algebraically
>added to the magnetization current of the primary. But
>since that load current is in phase with the voltage, the
>primary sees a less and less lagged current, as that current
>increases. The magnetization component is almost fixed
>(actually decreases a little) as the in phase load current
>component rises.

I believe you are thinking in terms of a single frequency sine wave. As I was trying to say, a square wave applied to an inductor is going to act a lot different. A sine wave rises and drops over time, over a half cycle for each of these. But a square wave jumps from a low Voltage to a high one almost instantaneously. So, the current will try to jump instantaneously also. But it can not. It will take time to increase and that means that the core flux will take time to build up. And that, in turn, means that the secondary Voltage will not mimic the primary Voltage as it does with a sine wave. The leading edge of the square wave will be stretched out over time. Once the secondary reaches the full Voltage level it is going to reach, the primary is at a constant Voltage level at the top (or bottom) of the square wave, so the primary current is not changing and the secondary Voltage will no longer be generated. So, after the secondary reaches the maximum level, that is, if the circuit parameters allow it to reach that level, it will start to drop, much like a capacitor that is discharging. The inductor/transformer in series will essentially act somewhat like a capacitor to ground: that is it will act as a filter.

In a classical Pi style filter, there is a series inductor with capacitors to ground on each end of it.

I have tried to use different language to explain this. The Fourier Analysis should say much the same thing.
>> FILTERING!
>>
>> This effect is one of the reasons why the frequency of a
>> digital signal is limited over a transmission line, weather
>> it be twisted pair, coaxial, waveguide or even fiber optic.
>> All of these lines exhibit both inductive and capacitive
>> effects and these, in turn induce different delays for the
>> different frequencies present in a pulse (square wave) that
>> represents a digit. Thus, over the length of the line the
>> leading and trailing edges of these pulses will be stretched
>> out until the trailing edge of one pulse reaches the leading
>> edge of the following one and the two pulses merge into a
>> single one and the data is lost.

>Since the effect you describe is not happening, I think you
>might want to go back and rethink this paragraph, also.

No, I think that paragraph is also correct. From the point of view of Fourier Analysis you find that the various frequencies that are traveling down a transmission line are actually traveling at different velocities. This causes a square pulse to spread out into something resembling a single cycle of a sine wave. Nature abhors sharp transitions.

>--
>Regards,

>John Popelish

• On 07/11/2014 10:02 AM, palciatore@gt.rr.com [Electronics_101] wrote: (snip) ... You are still stuck in the mental concept that it takes flux to produce
Message 30 of 30 , Jul 11, 2014
On 07/11/2014 10:02 AM, palciatore@...
[Electronics_101] wrote:
(snip)
> I believe you are thinking in terms of a single frequency
> sine wave. As I was trying to say, a square wave applied
> to an inductor is going to act a lot different. A sine
> wave rises and drops over time, over a half cycle for
> each of these. But a square wave jumps from a low Voltage
> to a high one almost instantaneously. So, the current
> will try to jump instantaneously also. But it can not. It
> will take time to increase and that means that the core
> flux will take time to build up. And that, in turn, means
> that the secondary Voltage will not mimic the primary
> Voltage as it does with a sine wave. The leading edge of
> the square wave will be stretched out over time. Once the
> secondary reaches the full Voltage level it is going to
> reach, the primary is at a constant Voltage level at the
> top (or bottom) of the square wave, so the primary
> current is not changing and the secondary Voltage will no
> longer be generated. So, after the secondary reaches the
> maximum level, that is, if the circuit parameters allow
> it to reach that level, it will start to drop, much like
> a capacitor that is discharging. The inductor/transformer
> in series will essentially act somewhat like a capacitor
> to ground: that is it will act as a filter.

You are still stuck in the mental concept that it takes flux
to produce voltage. But it takes only rate of change of
flux to produce voltage. The magnitude of flux that rate of
change is changing, is immaterial, except for core
saturation concerns.

Look at it this way. The primary is a short circuit (a
piece of wire) across the applied square wave (of the PWM
circuit under consideration, for instance), with one
exception. And that exception is the instantaneous voltage
that the rate of change of core flux produces across the
primary winding. That rate of change of flux produces an
instantaneous step of voltage across the primary, to buck
the instantaneous applied step of voltage (as any inductor
would).

But the secondary is also just turns around that same rate
of change of flux, so it also produces a similar step of
voltage. (Each turn of wire has no idea if it is part of a
primary or secondary. winding.) All that is required for a
winding to produce an instantaneous voltage step is for the
core flux to have an instantaneous change of *slope*, not
magnitude.

There is a limit to the maximum frequency a transformer can
transform, primary to secondary, but it does not involve
this classic transformer action. The limits involve things
like eddy current circulating in the core laminations, core
hysteresis and inter winding capacitance. But we should
discuss those limits separately.

> In a classical Pi style filter, there is a series
> inductor with capacitors to ground on each end of it.
(snip)

And, as I just mentioned, those pi filter (and other)
effects do exist in real transformers and do limit the
frequencies that can pass, efficiently, through actual
physical transformers, but they are completely separate
effects than ideal transformer operation, and transformer
designers have some control over those effects by the way
they wind transformers and what core materials they use.

we don't want any basic misconceptions to get propagated
about fundamental electronic effects or components, here.

If you want to discuss further, perhaps we should start a