Combining many oscillators?

[audioguru]

shouldn't this circuit oscillate...?

An oscillator uses noise in the circuit to get it going. But a simulation program has no noise, you must "kick" the circuit with a pulse to get it going.

Your circuit has a few things wrong, I fixed it and explained why:

 

[synox89]

Hey, thanks! I just tried it but it doesnt seem fully to work... In Yenka i have the ability to analyze the current/voltage etc from several parts.
And it doesnt seem to oscillate yet. What do you mean that i shouldnt phase the transormator? Do you mean the coil direction? Because it doesnt look like there is an option to change that.. surprisingly..

And to the amplifier... actually if i plug it to a sinus generator, it releases a correct tone, what is exactly the reason i could not plug in the way i had?

Alex

 

[cowboybob]

...What do you mean that i shouldnt phase the transormator? Do you mean the coil direction? Because it doesnt look like there is an option to change that.. surprisingly.. ...

Some SIMs don't allow (or have the means for) adjusting the polarity of transformers. All that's necessary is to wire them to acheive the polarities needed, as below:

(Credit to AG for the schematic)

 

[audioguru]

What do you mean that i shouldnt phase the transormator? Do you mean the coil direction? Because it doesnt look like there is an option to change that.. surprisingly..

The transformer provides feedback so that the circuit can oscillate. If the coil directions (the big dots on my schematic) of the primary and secondary windings of the transformer are not correct then the phasing will be wrong and the circuit will not oscillate.
Either Yenka needs to go to electronics school or you need to go to Yenka school to change it.

And to the amplifier... actually if i plug it to a sinus generator, it releases a correct tone, what is exactly the reason i could not plug in the way i had?

Your circuit had a few problems that I discussed on my schematic.

 

[alec_t]

shouldn't this circuit oscillate...?

That would depend on transistor gain, the phasing of the transformer windings, the winding resistances, the loudspeaker inductance .....
What do the numbers 500:1000 represent? Turns ratio? Inductance values?
BTW the variable resistor does nothing useful in that circuit.

 

[synox89]

Well, I tried to recreate this circuit... Its given as an example for an feeder circuit...

 

[audioguru]

Well, I tried to recreate this circuit... Its given as an example for an feeder circuit...

I doubt the circuit will work:
1) The potentiometer is in series with the primary of the transformer.
2) The speaker will get an extremely small signal if the oscillator works.
3) The phasing of the transformer is not shown.

 

[cowboybob]

AG, I guess you noticed the text is in German (which I certainly cannot read).

Any idea what the "500W" and '1000W" values for the transformer windings mean? (Surely not Watts!?)

At least with my SIM, no matter what transformer phase manipulations I try, the circuit will not oscillate...

 

[MrAl]

Hello there,

Not sure if you are interested or not, but the modern way to do this would be to use DSP using a microcontroller or similar.

The instrument sound could be sampled and stored, then it's just a matter of generating that same waveshape with different frequencies and amplitudes. For close to real sound modeling you'd have to sample several takes on different music to catch the different ways the flute can be played by a real life human being (pure tone or 'breathy' tone for example).

Alternately you might find information on the web about the flute itself with specifications paying more attention to the first three or four harmonics. Look up "Timbre" for more information.

 

[synox899]

Hey, 1000W means windings of course.
MrAl: I would like to make it the vintage way. In fact, i just need to put the frequencies of several oscillators. (That shouldnt be so hard, should it?)
And then they next keypoint is to program how loud every frequency should be, because they are not static and change over time (adsr envelope: attack, decay, sustain and release). How could this be programmed?

 

[audioguru]

I would like to make it the vintage way. In fact, i just need to put the frequencies of several oscillators.

Then you will simply have a bunch of noises.
In the first post you wanted to demonstrate the sounds from different acoustical musical instruments. But a musical instrument does not simply produce several frequencies. Instead most of the frequencies are harmonics of the fundamental frequency and they are all synchronized (except not for a piano). The fundamental and harmonic frequencies must all be sinewaves, not digital squarewaves. That is difficult to do without a computer or a recording doing it.

 

[Nigel Goodwin]

Hey, 1000W means windings of course.
MrAl: I would like to make it the vintage way. In fact, i just need to put the frequencies of several oscillators. (That shouldnt be so hard, should it?)
And then they next keypoint is to program how loud every frequency should be, because they are not static and change over time (adsr envelope: attack, decay, sustain and release). How could this be programmed?

Refer back to my post #8, which is what you appear to be wanting to do, which is an electronic organ - as you mentioned 'programming' are you aware there's a simple project on the net using a PIC to create a polyphonic electronic piano?.

http://www.pic24.ru/doku.php/en/osa/articles/pk2_osa_piano

 

[audioguru]

The polyphonic electronic piano sounds like a simple electronic organ without the overtone harmonics and attack, decay, sustain, and release of a real musical instrument.
Its tuning sounded interesting because some chords produce a slow beat frequency (wha, wha, wha), I expected NO beat frequency.

I have a Casio portable electronic organ that is about 25 or 30 years old that I bought for my kids to learn music. It produced real-sounding musical instruments but it don't work no more no more.

 

[MrAl]

Hey, 1000W means windings of course.
MrAl: I would like to make it the vintage way. In fact, i just need to put the frequencies of several oscillators. (That shouldnt be so hard, should it?)
And then they next keypoint is to program how loud every frequency should be, because they are not static and change over time (adsr envelope: attack, decay, sustain and release). How could this be programmed?

Hi,

I'd go with Nigel's suggestion. He stated that you would create first the 12 top octave generators and then divide them down to get all the lower octaves. After that you need to filter each one of them to create sine waves from the square waves. This is going to take some work.

The easy part is creating the generators and divide counters. With 12 generators (square wave oscillators) you get the highest notes, then divide them down with digital counters to get the lower frequencies. To get all the way down to 40Hz you'd need counters that can divide by 2^10, unless you really only need the flute tones and then maybe less (check actual instrument range).

Once you do that the hard part comes next. You need one filter for each frequency output. With 12 generators and 8 bit counters, that's 96 filters required. You can probably get away with 2nd order bandpass filters, maybe using op amps. Using quad op amps that's a total of 24 op amps, 48 capacitors, maybe 72 resistors.

Next you need to decide if you really only want one note at a time. If so that makes it easier.

Then you need to use gain controlled amplifiers to generate the timbre and choose and set the harmonic content.
To program the timbre you could control the gain controlled amps with a charging and discharging capacitor.

You'll also want to generate a sequence of tones to create a full musical passage. This will require some kind of memory and some counters to step through the memory picking out the notes one by one. Another oscillator will act as the note timing generator and that drives the memory address counter.

As each note is fetched it selects which gain controlled amps get turned on and triggers the timbre generator that actually controls the gain of the gain controlled amps.

So you've got your work cut out for you.

I'd be glad to help as long as you promise to record at least one Bach piece with flute solo and post it here as i would love to hear this thing when it was completed

 

[audioguru]

I searched in You Tube for Bach flute music but couldn't find one that I know.

 

[MrAl]

Hi audioguru,

Oh ok, well im sure there are a lot of them. The one i was thinking of is a very beautiful piece but i'd have to remember the title. It's been a long long time since i heard it now. It's not just a solo but the solo is part of the whole piece.

 

[alec_t]

Is this it?
https://www.google.co.uk/url?sa=t&r...uOIz14Oaxye4njKI9gGK-pA&bvm=bv.64125504,d.ZG4

 

[synox89]

Hey, I wrote that flute is actually not that I wanted to have. A theremin, just like the violin dont have separated note steps, they slide between them. Isnt that going to be easier?

 

[audioguru]

I have a Casio portable electronic organ that is about 25 or 30 years old that I bought for my kids to learn music. It produced real-sounding musical instruments but it don't work no more no more.

Goodee! I got it to work when powered from a wall-wart. The battery contacts are probably corroded.
I tried to play the very complex chord in the PIC demo but I ran out of phonics (it has a max of only 4) and I nearly ran out of fingers. My organ (the electronic one, hee hee) is perfectly in tune with the demo.

 

[synox89]

Isnt that going to be easier?