Sine wave oscillator

...dual integrator approach...four phases...generate a signal based on the DC average of the four phases
Here you go (LTSpice simulation only). It works, but needs seconds to settle down, and even with a JFET servo loop to control gain, I get lower sinewave distortion and quicker start-up by shaping a triangle into a sine. The triangle-to-sine approach also requires considerably less circuit complexity.

The usual caveats apply - this is only a simulation, and the real thing might not behave exactly like this.

My personal guess is that the heterodyne approach will also prove a very tough nut to crack if you want to generate beat frequencies as low as 0.5 Hz, because the two oscillators will want to lock up and generate a zero-hertz signal instead. If I were a betting man, I would give you long odds on this actually working.

IMO a relaxation oscillator of some kind is the way to go for this application, whether made with solid-state, or hollow-state devices.

I guess time will tell. :)

-Gnobuddy
 

Attachments

  • 3RC_4-phase_rectifier_Sine_Osc.png
    3RC_4-phase_rectifier_Sine_Osc.png
    71.5 KB · Views: 328
Rectifying and filtering the output of the oscillator to make a (nearly) DC control voltage is/was routinely used with the LED/LDR and JFET stabilization methods. But adding the extra poles in the servo feedback loop tends to make settling times long, and/or lead to outright servo oscillation, with the LFO producing amplitude-modulated output, or short bursts of sinewaves with dead silence in between!

One of the approaches suggested in this thread was to use a 3RC oscillator with a six-phase rectifier. With so many phases, what comes out of the rectifier requires only light filtering to become almost pure DC. So it is still capable of the fairly rapid changes needed to make a stable servo loop, so the LFO output has a reasonably fast settling time.

Personally, I think Rod Elliott's "miniosc" is an excellent solution for the purposes of this thread. It's relatively simple, distortion is far lower than necessary for this application, and Mr. Elliott has been nice enough to provide a proven schematic that only needs to be tweaked slightly to lower the frequency range: Miniature Audio Oscillator

OP didn't want to go digital, otherwise there are inexpensive DDS chips and modules that will generate a decent sine for you. The cheapest modules come from Hong Kong vendors, but there are $30 versions marketed by North American vendors, such as this Sparkfun module: SparkFun MiniGen - Pro Mini Signal Generator Shield - BOB-11420 - SparkFun Electronics


-Gnobuddy
this was Rod Elliott's old website (dead in the meantime). go to
Miniature Audio Oscillator
are the residual distortion (showed in fig. 4) from the OP-Amps ?
If yes, maybe other op-amps provide lower THD-level or more ear-friendly THD-character ? - go also to
Choosing of best sounding OP AMPs for the lowest possible THD+N -really the best Way?
 
I have found that with a 6 phase (full wave rectified three phase oscillator) that this only needs a single filter capacitor to yield a fairly respectable low distortion (typically 0.003% was about as low as I could measure). The filter capacitor however has to change with the frequency range and is approximately 10 times the value of the timing capacitor(s). The other main pole is caused by a lag in the LDR so adding more filtering generally caused squegging, or motor boating as used to be said, as Gno' indicated because of the additional poles.
So I'd recommend a multiphase oscillator. Similar designs are around and there was a four-phase one in Wireless World sometime in 1984 or 1985. I used OPA2604's but the newer 1611's might give even lower distortion. I don't think I'll be building another though.
 
Member
Joined 2004
Paid Member
I designed a very sophisticated motor driver for the Rockport turntable. The oscillator was pretty straightforward (6 speeds since it also needed to do 1/2 speed transfers). 2 phase motor drive with current feedback and transformers was a little more challenging since it needed a start drive ands then an operating drive at a lower level. I think maybe 5 were built.
You can sense the harmonics in the motor with an accelerometer. However the current drive helped since the internal magnetic structure is not perfect so you get harmonics. The current drive minimized the harmonics at the motor. I suspect a more ideal drive waveform could be found somehow.
 
@1audio
I am in need for a sine three phase star driver for 220v ac motors as well, not synth sine though, no pwm allowed.Would you give some help?

What follows is recall from my youth--- about 60 years ago:

My dad was a farmer, a self-taught electrician, and a wizard with relay logic; he did a lot of mechanization/automation of grain handling equipment for fellow farmers.

He would sometimes buy multiple-horsepower three-phase motors because they were plentiful and cheap on the used market. He would make these three-phase motors work on farms that had only single-phase power (120-neutral-120) by using a "phase-convertor." This was a large free-running (no mechanical load) 3-phase motor; as I recall, he applied 220VAC power across one winding of this motor and there there were large AC caps tying the powered-winding to the other two motor phases to "balance the phases." The converter would be started with no load; once running, the three phases were the power source for other multi-horse motors. I'm sorry that I don't remember any details about Y versus delta power connections.

I know he didn't invent this technique and I imagine there's technical literature to be found.

Goodluck!
 
Last edited:
Account Closed
Joined 2010
@BSST

I know that technique wich i employed several times, it's calculus is actually found in many electrical books, not a big deal, but in practice is simply impossible to obtain equal 120 degrees phase lag between all three phases and the voltage of each phase will be different, thus you get mechanical imbalance in the motor creating huge stres on the bearings and vibrations.The technique is only helped by large loads that significantly lower the speed and make the imbalance lower, but other than that is not your best pick for a turntable motor .Most probably the cleanest and simplest conversion technique is to use a 3 phase generator , namely a 3 phase motor(generator in reverse action) attached to a one phase motor...
 
Last edited:
If you search for signal generators on Amazon, you will find XR2206 modules for $10, but XR2206 is a rounded triangle wave, not very clean. But they also have a DDS box for $30, which will be more than you need. Of course you can use Audacity etc. on your computer or a RPi to generate sine waves etc.
But maybe you just want to tinker, in which case just make an amplifier with AGC / compression and add a phase shift feedback. Such is the essence of analog sine wave generators. ~70 years ago, HP started with a tube circuit that was one such, and used an incandescent lamp for automatic gain control. HP Virtual Museum: Model 200A audio oscillator, 1939 HistoryP
 
Last edited:
I designed a very sophisticated motor driver for the Rockport turntable. The oscillator was pretty straightforward (6 speeds since it also needed to do 1/2 speed transfers). 2 phase motor drive with current feedback and transformers was a little more challenging since it needed a start drive ands then an operating drive at a lower level. I think maybe 5 were built.
You can sense the harmonics in the motor with an accelerometer. However the current drive helped since the internal magnetic structure is not perfect so you get harmonics. The current drive minimized the harmonics at the motor. I suspect a more ideal drive waveform could be found somehow.
How can I realize a current control mode (current drive) in retrospect?
Maybe by replace R18/R24 by the motor winding coils (and reducing R26/R19 in the right manner) ? - go to PDF in post #16 under
Homebrew Motor Control Linn LP12 - Circuit Description wanted for Sine Wave Osc.
 
yes:
this 110VAC Motor:
9904 111 31813 McLennan Servo Supplies | McLennan 9904 Reversible Synchronous AC Motor, 1.8 W, 1 Phase, 24 Pole, 110 V, Flange Mount Mounting | 441-0423 | RS Components
this power amp
https://www.st.com/resource/en/datasheet/tda7294.pdf
and 2 pcs. of this 24 VAC transformer
VC 10/1/24 | Printtrafo, Trafo 24V ac/230V ac Durchsteckmontage 1-Ausgang 10VA | RS Components
Actually the amp was suited only for 24VAC motors - therefore the transformer nearly 1:10 ratio.

The schematic is in the attachment.
Currently is connected at each output of TDA7294 the secondary 24VAC winding of the transformers. the 230VAC primary winding is connected with the above mentioned motor

The main question is, how I must calculate the NFB resistors R3/25, that determine the voltage gain factor together with the 50Hz impedance of the transformer incl. connected motor.
 

Attachments

  • Dr Fuß 50 Hz Linn LP12 Motorcontrol clone (1).pdf
    274.9 KB · Views: 60
Last edited:
Member
Joined 2004
Paid Member
Attached is the motor controller I designed many years ago. There are some elements that need to be figured out. - There was two gang pots for setting the speed and this supported 1/2 speed playback for transferring vinyl to digital with better fidelity. And there were interlocks to protect the air bearing (which would be destroyed if operated without air). The platter was something like 30 pounds. It has a start mode to get the platter moving and a run mode with reduced torque for lower noise.

I hope this helps understand some of the issues around a good driver for hysteresis motors. DC motors and hall effect motors are a different story.
 

Attachments

  • Sony Motor Controller.pdf
    39 KB · Views: 56