under the term direct-drive we understand a turntable where the platter is connected directly to the rotor of a motor, or even forms a integral part of the rotor.
The rotational speed of the platter equals the rotational speed of the motor.
The direct drive is the pinnacle of all turntable drives.
At the same You find in the surrounding field comments like plastic-bomber, cheap-drive, and inferior-to-belt-drives.
Now there´s no such thing as the direct drive, as there is no the belt drive.
Reason enough to spread some light into the dark and to possibly correct some crude views.
Direct-drives can be built with different motor types and control systems. I´d like to concentrate on the brushless DC motor and its associated control systems, as they were used by the majority of companies.
AC motors were iirc only built in considerable number of units by Denon.
Dual, Technics, Sony, Pioneer, Kenwood et al resorted to DC motors.
To get a true grasp on why the direct drive in its most refined incarnation using a PLL-servo loop is truely the superior drive, one needs to get rid of some widespread but totally wrong
thinking about the functioning of such a control loop.
Obviously many user think that a direct drive servo loop works ´quasi digital´, meaning that it accelerates the platter with multiple discrete pushes with the full motor torque when the platter runs too slow, and shutting temporarily down when running too fast.
Things don´t work that way! The idea that the platter is moved by single discrete pushes and it requires the platter´s inertia to smoothen the movement to a somehow constant rotation is fundamentally wrong. Instead the motor current is regulated in a continuing, soft way and of a value that is just sufficient to overcome the drag of the bearing, the needle and possibly a brush. The amount of motor torque is supplied as demanded at the time. Due to this fine regulation action and the fast reacting of the servo loop it is factually impossible that the platter ever runs too fast.
For home useage the lower torque drives like from the Technics SL5210 or SL-Q2 are equally well suited as the high torque drives like Technics SL-1200MK2 et al, that even features the same control chipset.
The DJ-turntables require the higher torque to cope with the higher load from Scratching and to be able to keep the speed stable. Said in other words, the speed keeps constant over a larger load range, no more, no less.
If You throw an eye into the service manuals you will notice that the difference between the turntables is rather only the raised motor supply voltage of 21V instead of 12V.
The same motor is used, the same bearing, the same control chipset and -apart from a few- even nearly identical parts values.
Historically the first and simplest control loop of a brushless DC-directdrive-motor is the so called E-servo. Drive current flows in only one of the commonly used three phases of the motor. Meanwhile the two other current-free phases generate voltages proportional to the turning speed due to the electromotive forces (hence E-servo), we know from Dynamo action. A network of Diodes sums up and rectifies these voltages. A following filter smooths out the ripple. Then a voltage comparator stage compares the speed-dependent voltage with a reference voltage, that can be switched between 33,3 and 45Turns/min and that is pitchable with potentiometers.
The output signal of the voltage comparator controls the motor drive current and as such it controls the platter speed.
The smoothing filter adds considerable delay due to their necessarily large time constants.
This results in less regulation precision and load regulation, and is prone to drift over time and temperature, and makes the other control systems seemingly superior. The effort for the electronics is the lowest though and no dedicated speed sensor is required.
Typically You find these control system in combination with heavier platters from ~2kg up that also have higher inertia.
Examples of this kind are the DUAL 701 (featuring the EDS1000 motor, probabely the best that has ever been built), Technics SP-10MK1, SL-1800, SL-1900, etc. till into the early 80s.
The F-servo presents a improvement and refinement. Here all three phases are conducting current. A speed sensor is connected to the motor or platter which generates a speed-related pulse signal. Mostly electro-magnetic sensors are used which work in that a rotating magnet induces pulsed voltages of 50Hz-200Hz in a meander-shaped coil. Sony, Denon and Fisher used a different principle similar to a tape machine with a magnet tape located at the inside rim of the platter and a fixed tapehead.
It allows for higher pulse-frequencies of 500Hz-1000Hz. Rather seldom one can find optical sensors with up to 4000pulses/turn.
The sensor signal is converted from a train of pulses into a speed -proportional dc-voltage with a Frequency-to-Voltage- converter (F/V) and a smoothing filter.
More refined systems apply a Sample-and-Hold-circuit insted of the classical F/V-converter. The advantage is that the S/H-stage omits with the smoothing Filter alltogether.
It ´reacts´ faster and with increased regulation resolution.
Like in the E-servo, the output voltage of the S/H-stage is then compared to a reference voltage and the motor current and platter speed can be controlled this way.
The detection of platter speed via a train of pulses reduces the number of possible points of error. As such it allows for a more precise control which is free of drift. The effort is higher though than for the E-servo. An example for this kind of servo is the Technics SL-D2.
From the mid-70s on the F-servo eventually replaced the E-servo.
The best results are achieved by implementing a PLL-servo control at the cost of the highest effort.
One notices that a oscillator and a Phase-to-Voltage converter stage (P/V) is added to the F-servo schematic.
The oscillator is either built from a RC-oscillator or from a Quartz-oscillator. Due to the easy frequency tuning of the RC-oscillator its often used when the speed is pitched, while a Quartz-oscillator is activated for fixed speed (Quartz-Lock). The SL-1200mk2 makes use of a additional specialized IC that allows Quartz-locked pitch-mode also.
The oscillator provides the P/V-converter with a highly stable clock signal as reference. The F/V-converter provides for a second speed-related signal. The P/V-converter compares the timing relationship of the signal flanks and converts the difference into a dc-voltage which is then compared to a reference voltage and that is used to control the motor speed. Because the P/V, resp. the PLL control mechanism relates to the timing of the signal flanks, resp. their phase relationship, it ´locks-in´ precisely onto the reference clock from the oscillator. The F/V and the P/V control loops are acting kind of parallel, whereby the F/V loop serves as a Preset loop and the P/V loop provides for the precision and fine detail.
The difference between the control loops can be viewed on the position of the stroboscope markings. If You slow down the platter manually the strobe marking will shift its position.
When releasing the platter again the strobe marking will remain at that new position with a F-servo, but will run back to its initial position with a PLL.
Prominent examples of (Quartz-) PLL-servos are the top range models of most japanese manufacturers, like the Technics SP-10mk2 and its successors.
Since the second half of the 80s, F-servo turntables concentrated on the lowend entrance model range.
Comparing the three control systems, one esily recognizes the PLL-servo as the most precise, utmost stable and finest resolving control loop.
That the PLL and direct drives in general earned a bad reputation among many HighEnders is not founded on its principal capabilities, but due to inferior implementations and clever propaganda of belt-driven turntable agents and enthusiasts.
The improved servo controls allowed the manufacturer to use ever less weighing platters and weaker motors, because for speed constancy neither high platter inertia nor high torque motors were required.
Slotless, coreless motors with no cogging and reduced number of phases/poles, that also built more compact, could be used and were cheaply manufactured as subassemblies, containing the bearing and the electronics at the same.
The manufacturers rather put their eye on more comfort functionality and features.
Overall the build quality, haptics and optics suffered. The term ´plastic bomb´ speaks for itself.
The development of belt-drives turned in a different direction towards ever more impressive and more massive devices. The term ´oil-rig´ says it all.
The often performed overkill-designing into the wrong or irrelevant points or subassemblies is glossed over by the sheer massiveness or altar-like presence of these machines. The Bluff is accompanied by marketing speech that creates the impression that only this manufacturer has found the final and sole solution to a herculian task, that dwarfs the moon landing to a simple finger exercise.
Since the associated press has mostly given up fully on their technical expertise a important measurement standard is lost. Manufacturers and distributors could fake specs and data after their wish without anybody noticing.
That the all to often off of physical and electronics known trueths designed heavyweight reacts hyper-sensitive to any variation, must be accepted, because in the short periods in between where it eventually hits the correct 33.3Turns/min, it may sound spectacular.
Such bugs -sorry, features of course- open up a whole new toy world of tweaks and tunings to the addicted HighEnder, who spends time on his hobby with brave indulgance, in his lasting search for the holy grail.
He´ll be using a thread when the sun shines, a tape when the weather is indifferent and a belt when it rains. Finely balanced of course regarding the mechanical tension and inspected every second hour, otherwise the timing could be off.
What a shock it must be when the DD-oldtimer from those good old days won´t give in an inch, presenting the long sustained piano note with stupendous pitch stability, not waggling around for ages.
Without question are there many superb belt-driven turntables. But whom wants a manufacturer possibly fool, who isn´t even capable of designing a decent control system for a motor, but needs to source a measly sine-oscillator for a banal synchronous motor from a OEM?
Unfortunately turntable direct drive motors, servo-electronics included, are hardly produceable economically nowadays. As DIYer though one can easily source used donor drives, that can be tuned and tweaked into great turntables.
My own projects base on Technics drives featuring the control chipset used in the SL-1200mk2, SL-Q2, SL-Q3, Grundig PS-4500 and others.
Some servicing, some tuning of the electronics, a decent plinth and a good tonearm is all that is needed to lift these reliable stressless drives to a niveau that stands proud in any comparison regarding haptics, optics and acoustics.