Audio Home

Motors for Turntables.

This is intended as an overview of available motor technology for turntables and as background for the sections on the drives for the various motors:

I have separate pages for electronic drives:

AC synchronous Drives
EC (3 phase) Drive
DC brushed Drives
DC drive Part 2

Plus a separate section on the mechanics of belt, idler wheel and direct drive systems which quantifies some of the limits to performance

Turntable Mechanics

Motors

Motors for turntables fall into the following categories:

AC induction motors (single phase)

AC synchronous motors (single phase)

DC brushed motors

EC (Brushless) motors, Iron stator and Ironless.

Fig.1. Speed / torque curves for typical motors, taken at similar rated torque levels. Note that an AC shaded pole motor will be much larger than an AC synch motor for a given rated torque and will require more input power. A DC motor will be smaller again, and require less input power.

 

AC Induction (usually Shaded Pole).

The original motor in the Garrard 301 (and 401) falls into this category. The motor itself is a two pole single phase AC shaded pole induction motor rated at 16 watts. The efficiency of shaded pole induction motors is very low (around 25%) so the motor produces around 4watts. As an aside, AC synchronous shaded pole motors do exist but are rarely seen, partly because they are even less efficient. Some AC eddy current motors have been used in turntables, most famously the Papst "flywheel motor" in Empire and some Thorens machines. These set the record for inefficiency - I have measured one empire motor at about 6%. Because of their construction most AC induction motors show some degree of flywheel effect.

Induction motors are called that because the rotor magnetisation is induced by relative motion between the rotor and the stator field. This relative motion shows up as non-synchronous speed, slower than the supply frequency by an amount known as “slip”. Because the current in the rotor (and thus the total torque) is dependent on the degree of difference, the higher the torque the higher the slip. This allows speed control by load variation which is the principle behind eddy current brakes. The degree of slip for a given load is affected by supply voltage and the interaction between these three variables gives a torque/speed curve which is not linear, as can be seen in the graph above. Shaded pole motors also give some cogging torque although this can be minimised by skewing the rotor. Technically eddy current motors are not induction motors but I have included them here as they fit nowhere else.

Motors such as the shaded pole in the 301 are designed to run at a torque which corresponds to 5-10% slip, which is in a reasonably linear part of their speed / torque curve. Other types of induction motors generally run at lower slip levels. Note that the motor will produce 200% of rated torque at lower speed, so it will start a heavy platter quite quickly. The motor can be very noisy, both electrically and mechanically. The AC supply also acts as a source of noise.

AC synchronous

The original motor in Linn Sondek (and most other TTs of that ilk) falls into this category. Because the rotor magnetisation is permanent, there is no slip induced between the rotor and the stator field so the rotor moves in synchrony with the stator field, hence the name. Generally the motors found in TTs are two phase multi-pole synchronous motors with iron stators. The run speed equals 120 x fsupply / pole number so a 24 pole motor gives 250RPM on a 50Hz supply and 300 RPM on a 60Hz supply. A synchronous motor must run at this speed until its pull out torque is reached where it abruptly stops, giving the "squared off" speed / torque characteristic shown in the graph. Loading techniques like eddy current brakes therefore have no effect. Changing the voltage supply only changes the pull out torque so cannot be used for speed control.

To vary speed an AC synchronous motor needs a variable frequency drive. These can be either digital or analogue, with digital being far more common as they are cheaper and easier to design to a given specification. Electronic supplies can be designed to give two output waves "in quadrature" - 90 degrees apart. On a single phase supply (like the mains) the motor uses a phasing capacitor to generate the cosine equivalent of the sine wave supply.

In theory the torque cogging produced by each of the sine and cosine phases cancels. This is because the torque produced by each phase is proportional to the square of the current and sin squared + cos squared = 1. In practice this only applies to the actual drive currents and then only if the two currents are exactly equal. In any real motor there will be excess currents due to the motor running at less than pull out torque and to winding resistance. The motor responds to torque requirement by changing the angle between the stator field and the rotor field. At zero load the two fields are aligned, at pull out torque they are at (90 degrees / pole pairs) apart. 90 degrees is the point of maximal torque conversion efficiency, so any greater torque than that which gives 90 degree separation will cause the motor to lose synchrony and stop.

Often the starting load is the highest load a motor will see, so in normal running the motor is at some fraction of its pull out torque. Lesser loads mean that the motor runs at a lower angle and thus a lower torque conversion efficiency. The lower efficiency lowers the inductance of the windings, changing the ratio of reactance (drive) to resistance (heating). The excess current therefore simply heats the motor windings and it appears that these heating currents do not follow the sine squared form of the drive currents so they create cogging torque, but I have been unable to confirm this. Lowering the drive voltage reduces the excess and thus the vibration but to eliminate it completely the voltage would have to be lowered to the point where the motor was at pull out torque at which point any disruption would stop the motor. An AC synchronous motor can usually be held in the stopped position for long periods without harm as long as the rated motor inout power is not exceeded.

There is also some cogging due to the attraction / repulsion between the rotor magnets and the stator pole pieces in iron stator motors. These motors are quite sensitive to their power supply which can be another source of noise.

DC (brushed).

Many modern machines use these and they are offered as upgrades for existing AC TTs. The current required by a DC motor is directly proportional to torque produced so the voltage required to drive this across the windings is also proportional to the torque and the winding resistance. The motor also produces a back emf (voltage) which is directly proportional to its speed. Taken together these mean a DC motor will always accelerate to the point where the sum of the voltage required for the given torque and the back emf exactly equals the drive voltage.

Together these mean that the motor has a linear speed torque curve and is always operating at the balance point of its torque requirement / forward voltage so it has no "excess current" as does an AC motor. This makes a DC motor potentially much quieter than an AC one, but also makes speed control much more difficult. Although some headway can be made by using current compensation (also known as I x R compensation) a true precision DC drive requires some method of sensing the motor speed and compensating for it, such as an add on DC tachometer or an external encoder loop. This adds complexity and can lead to problems with the loop "hunting" either side of its set speed.

The linear characteristic means that DC motors can put out many times their rated torque at slower speeds. The maximal torque possible is the stall torque for the motor (eg locked rotor) which can be from 4 to 20 times the rated torque. Care must be taken to ensure that the motor cannot draw more current than its rated torque equivalent for any period of time as this will burn out the motor, but as long as this is observed the motor will pull even the heaviest of loads up to speed, making it ideal for the massively overweight and underdesigned tables currently fashionable.

The fact that there is no drive at all in the “space” when the brushes are changing polarity can make them prone to torque cogging but this can be addressed by appropriate design. The brushes create mechanical noise as they run on the commutation rotor, and the commutation gaps create arcs at changeover giving rise to commutation noise on the supply. The use of precious metal brushes and capacitive loading can reduce this to miniscule levels.

 

EC (Electronically Commutated, aka Brushless DC)

The only current turntables using these are the SMEs and the Caliburn. The application of EC motors is complex and is therefore often poorly understood. The most common EC motors are actually 3 phase AC motors with permanent magnet rotors, so the motor will run at synchronous speed (governed by the supply frequency according to the number of poles).The ideal supply to these motors is therefore a three phase AC supply with three precisely sinusoidal waveforms of equal amplitude and equal phase delay (120 degrees). In this condition the speed torque curve looks like that for an AC synchronous motor except that the EC motor will be much more powerful for a given size.

The three sinusoidal phases sum to zero so the net current into the motor does not fluctuate. The sum of the torques produced by the three phases is also unity at all points so they do not produce torque cogging. The absence of brushes means that mechanical noise and lifetime are limited only by the bearings. Ball bearings give the longest life at high speeds so are the usual choice in EC motors. The ball bearings may be replaced with sleeve bearings which are much quieter but require frequent lubrication which is difficult for the lower motor bearing due to its position inside the motor.

Many EC motors use stator poles to increase the sensitivity of the magnetic circuit and this gives rise to attraction / repulsion forces between stator and rotor, causing cogging. Ironless stator motors do not suffer from this but must use expensive rare earth magnets to support the air gaps in the magnetic circuit. High speed and high torque (due to permanent magnet construction) means high power ratings for their size but this is offset by the fact that they are many times the cost of the other motors.

The complications come with driving them from a DC source. They cannot be driven directly on DC as the supply needs to be fed to each of the three windings in sequence. For this reason these motors are usually fed from a specialised supply which can synthesise the required waveforms. The motor speed is controlled by the voltage supplied and the speed for commutation is fed back from the motor itself, usually from hall effect sensors mounted on the “back” of the motor

In the simplest case, known as “block commutation” the three waveforms are simply approximated by turning the positive and negative polarity DC supplies on and off at intervals determined by the output from the sensors. The output to the motor thus consists of a series of pulses on three channels resulting in a great deal of electrical noise. These pulses can have their leading and trailing edges ramped to reduce this, known as trapezoidal commutation. A slightly more complex form uses multiple voltage steps to better approximate a sinusoidal supply. All these schemes suffer from torque cogging due to the deviation from the sinusoidal wave.

To reduce this some drives use a digital waveform generator to synthesise sinusoidal waveforms, generally using pulse width modulation. Like all switching systems this creates a great deal of high frequency noise. The inductance of the motor windings is often used as part of the smoothing filter for these designs. In addition the bit count of the PWM generator limits the fidelity attainable.

If we do this with analogue waveform generation and linear power amplifiers, we can eliminate digital noise and power line artefacts and get a synchronous motor with far superior performance to the single phase type. The great hurdle to this has always been the difficulty of making a three phase variable frequency analogue supply. If a single phase waveform generator such as a Wien bridge is used so as to maximise waveform purity, it is difficult to make the other phases track correctly because analogue phase shift is always a function of frequency. It can be done:

3 phase EC motor installed in Garrard 301 chassis

But in truth I don't think it is worth the trouble and expense.

Stepper Motors

Stepper motors are a specialised type of synchronous motor which are constructed to allow accurate positioning of the rotor WRT the stator. Jones on stepper motors is an authoritative source of information on the subject. The interesting features to note are that to achieve micropositioning accuracy (the design goal of the manufacturers) these motors use teeth on the rotor and stator. This will inevitably result in some torque cogging in use, estimated as 5% in this app. note by Oriental Motor . They have significant benefits in terms of the ease with which they can be driven to achieve very accurate rotational speeds, and could form the basis of a direct drive system.

SECTION 2

Motor vibration analysis.

The following are phots of swept spectra of vibration for several turntable motors. The start frequency is 0Hz, swept at 50Hz per division so end frequency is 500Hz. Bandwidth was 10Hz so anything under 10Hz is the carryover from the DC response (full scale deflection). There were no significant data points in the first few Hz when using 1Hz bandwidth and the 10Hz BW sweep is faster.

The data were collected using a microphone stethoscope taped to the front of a guitar body. The motors were mounted on an aluminium plate which was in turn attached to the guitar using blu-tack. and stand offs screwed into the aluminium (not the guitar!) The idea here was to create a reasonably uniformly resonant structure to elevate the motor vibration well above the noise floor. These data are relative only - I know of no absolute reference. Also please note that my digital camera sometimes won't focus correctly in these low light conditions.

First up a baseline sweep with the vibration rig set up but no motor running. This gives an ambient noise baseline which is basically at -70dB once past the DC response edge, with the exception of some mains noise at 50Hz at -40dB, probably from the compact fluorescent lighting.

Next the motor from the Garrard 301, with and without the eddy current brake in action. The peak levels are similar in each case, about -3dB and are centered at 100Hz (twice mains frequency) . Note that the brake does not contribute any vibration, in fact it effectively damps some of the higher frequncy vibration.

No brake

 

 

With brake

And now the 3 phase Maxon motor running at 1200rpm (20 Hz supply) Peak level is -41dB, mostly at the third harmonic (60Hz) as is to be expected with three phase supply.

As a reference hee are three standard commercial motors: first an AC synchronous motor from Landis and Gyr

 

 

And one from Thorens:

 

 

My thanks to Brian Kearns for lending me these motors for my experiments.

Note that both of these motors have peaks at about -20dB but have quite differenct harmonic spectra.

 

Lastly a brushed DC motor also from Maxon.

 

This peaks at about -29dB.

Lest you think these are insignificant differences remember that these are logarithmic power relationships so the vibrational energy relative to the lowest energy (the 3 phase motor) are:

Motor Peak Level Energy ratio
3 Phase
-41dB
x1
DC
-29dB
x16
Thorens
-22dB
x80
Landis Gyr
-18dB
x 200
Garrard (w brake)
-3dB
x6000
Garrard
-3dB
x6000