advertisement


Turntable Power Supply Amplifier

J

John Channing

fruit box forever
I am looking to build a variable frequency power supply for 2-phase AC synchronous motors (e.g Airpax) and need some advice on the amplifier stage. I guess the requirements are somewhat different to the type of amplifier you would use to drive a typical speaker but ideally, I want an off the shelf solution I can drop into this project. What can you recommend?
 
Nope, no different at all. A Pair of |Lm3886 will do the job fine, one for each phase.
 
Depends on the power and voltages. For Airpax motors you only need a few Watts. Generate 90V using a small step up mains transformer. Add a few back emf catching diodes to the rails for the inductive kick back when you switch off
 
You can use really low cost amplifier like the TDA2050. No need for the LM3886.
A single supply design with output capacitor is easier, avoiding dc offset problems into the transformer winding.
 
davidsrsb said:
You can use really low cost amplifier like the TDA2050. No need for the LM3886.
A single supply design with output capacitor is easier, avoiding dc offset problems into the transformer winding.
Click to expand...

Are you suggesting using a capacitor for the phase shift? If so that is not what I am looking to do as I want precise control and will do that in the signal generator.
 
You have three options, I think.

1. Build a high voltage, low power amp.
2. Transformer couple a standard audio amp.
3. Change the motor to a lower voltage similar type.

Something like the TDA7293 can run on +-50v rails, so a pair bridged would give you +-100v swing, which might be enough. No idea on cost.

Transformer coupling works, and allows you to get something running using completely off the shelf parts, or an amp you might have spare. Changing the motor gets you the same advantages, but you've changed the motor and have to fit pulleys etc.

Somewhere I have a small program I wrote that creates quadrature sines on a PC with complete control of the frequency and phase. I used it in conjunction with a spare power amp and a pair of small toroids to persuade an LP12 to run at 78rpm.

The capacitor coupling David is talking about would be at the output of power amp.

Paul
 
Paul R said:
You have three options, I think.

1. Build a high voltage, low power amp.
2. Transformer couple a standard audio amp.
3. Change the motor to a lower voltage similar type.
...l
Click to expand...

Paul,
Many thanks, that is exactly the sort of answer I was looking for.

I guess I could look at the Rega 24V motors as despite what their website claims, I believe they are a stock Premotec item. It's then just a case of popping the pulley off the current motor or getting a new one machined (which obviously I can do, but at a cost).
John
 
Paul R said:
Somewhere I have a small program I wrote that creates quadrature sines on a PC with complete control of the frequency and phase. I used it in conjunction with a spare power amp and a pair of small toroids to persuade an LP12 to run at 78rpm.
Click to expand...

There are lots of options there, both off the shelf boards or software based. What I am interested in investigating is whether I can reduce/eliminate torque ripple through tweaking all of the available parameters (voltage, phase, waveform shape, etc.). I have been looking through the theory and it certainly looks possible.
John
 
I haven't pursued it (like a lot of other stuff...) but the idea was software signal generator on laptop->spare power amp->transformers->standard LP12 motor. Then all parameters can be varied and the results evaluated before constructing anything more permanent. The total cost was a pair of small mains transformers and it worked well enough on an old LP12 to do some transcribing at 78rpm.

Obviously resurrected software is now crashing, I'll tidy it for my own sense of order.

Paul
 
Paul R said:
I haven't pursued it (like a lot of other stuff...) but the idea was software signal generator on laptop->spare power amp->transformers->standard LP12 motor. Then all parameters can be varied and the results evaluated before constructing anything more permanent.
Click to expand...

Yes, that is the kind of thing I have in mind.
 
How about using programmable boards such as Arduino and write a short code to generate sine wave throigh 2 outputs that is perfectly out of phase by 180 degrees and then have it amplified?
 
screenshotmotorsteuerun.jpg


Stufe 1 and Stufe 2 are the two turntable speeds, 33.3 and 45, you can set the frequency to two decimal places. Phasenversch is the offset of the second phase, in degrees, adjustable to one decimal place and beschleunigung is the start up accelertation of the platter, lower is a smoother start up.

I think there's actually more controls shown in a later image.
 
I remember seeing a simple wave form generator using a 1 of 16 line divider 74250 I think and pots so you could crudely make a wave of any shape.

Might be fun to see if a sine wave is the best shape for minimum coging.

Pete
 
Pete, it is the best shape, more important than the absolute accuracy of the waveform is that you have some load on the motor. It's light platters and free running bearings that interact with the motor to give most cogging.

There's a piece on one of the Mark Kelly threads where someone tests various waveforms and the conclusion was the one you get out of the back of the motor when you turn it is the one you should put into it.
 
John Channing said:
Are you suggesting using a capacitor for the phase shift? If so that is not what I am looking to do as I want precise control and will do that in the signal generator.
Click to expand...
No, the capacitor is just to prevent the amplifier seeing the dc short of the transformer winding. It should be large enough not to affect phase. As it has about half the dc rail biasing it, it can be electrolytic
 
sq225917 said:
Pete, it is the best shape,
Click to expand...

No, a sine wave will not be the best shape, you are confusing simplified theories and real world electric motors. A single loop of wire spinning in a uniform magnetic field produces a sinusoidal current, but real motors are not made of single coils and do not have uniform magnetic fields. If it was that simple, we would not need motors with 24 poles to produce approximately smooth torque.


sq225917 said:
more important than the absolute accuracy of the waveform is that you have some load on the motor. It's light platters and free running bearings that interact with the motor to give most cogging.
Click to expand...

If you had a motor with perfectly constant torque, fly-wheel and viscous drag effects are largely irrelevant. I also think what you are describing is a mechanical form of changing the phase relationship and waveform. They are also techniques for trying to make a platter spin at a constant angular velocity driven by a motor with varying torque which is always going to be impossible. It is much simpler to make the motor spin at a perfectly constant speed and those things become less relevant (or irrelevant).

sq225917 said:
There's a piece on one of the Mark Kelly threads where someone tests various waveforms and the conclusion was the one you get out of the back of the motor when you turn it is the one you should put into it.
Click to expand...

That is correct. If you could make the motor spin at a perfectly constant angular velocity, the waveform you got out would be exactly what you need to make it spin at a constant speed with no torque ripple. The problem is, how do you do that?

One idea I had was to take two identical motors (i.e. same type and batch etc, I realise nothing is truly the same). Drive the first motor with a sine wave and connect it to a fly wheel. Then use the fly wheel to drive the second motor at what should be an approximately constant speed. The difference between the waveform driving the first motor and the one produced by being driven by the flywheel can then be added back into the drive signal to produce a more constant speed.
 
You must log in or register to reply here.


advertisement


Back
Top