The Korf Blog

In our previous posts, we have talked about the turntable belt drive and explored its function as a lowpass filter, rejecting the motor's torque ripple. Then, we went through the key problems associated with the belt drive—and some of the solutions.

Then we shifted our attention to the idler drive, and went through its typical troubles.

We have saved the best for last. This is the first post about direct drive, how it works, what is good about it and what isn't.

While the inventors of belt and idler drives are unknown, Matsushita (Panasonic/Technics) claims the invention of direct drive turntable with their 1970 SP-10. This isn't completely preposterous. But, strictly speaking, neither is it true. To understand why, let's first define what a direct drive turntable is.

The most logical definition I found states that in a direct drive turntable, the motor A and the platter B share the same shaft C and rotate at the same angular speed.

This is a very straightforward configuration, and it would be really strange if nobody tried it before 1970. And, of course, during the prewar shellac/78 RPM era, everybody and his dog did, both in turntables and in cutting lathes. After the Paillard buyout in 1963, Thorens mentioned in their brochures a 1928 Paillard patent that supposedly covers a direct drive motor for an electric grammophone. Whether this patent is relevant or not, by 1930s direct drive was quite common.

Then what did Matsushita invent? And why can't we find any mention of postwar direct drive turntables until Technics came along?

The answer is, those early direct drive motors were unusable with slower (45 or 33 RPM) speeds. They depended on mechanical commutation—carbon brushes supplying current to the rotating copper collector. With a suitable flywheel, it didn't create too much trouble at 78 RPM. But any slower, the jerkiness of mechanical commutation became unbearable.

With belt and idler drives, the mechanical diagrams were pretty much self explanatory. These are not very helpful in understanding direct drives. All the cool things are in how the motor itself is designed and driven. Shall we have a closer look at one?

For illustration purposes, we will use a high quality first generation motor: Dual EDS 900. Just like with digital audio a decade later, the caliber of engineers working on early direct drives was often a lot higher than during the technology's supposed "maturity". If you don't believe me, compare the sonic quality of Donald Fagen's fully digital "The Nightfly" from 1982 with anything recorded in the 1990s.

If you are not familiar with the way brushless motors work, I can recommed the following tutorial. Otherwise understanding of what follows might be a bit challenging.

How does this all comes together and rotates the platter at the desired 33 1/3 or 45 RPM? Here's a block diagram of EDS 900. It's fairly typical of the 1970s DD motors.

The frequency generator (in EDS 900's case, a coil under the soft iron gear) generates impulses with a frequency that is proportional to the motor speed.

This frequency is then converted to voltage: the higher the RPM the higher the voltage. The lowpass filter discards the high frequency fluctuations and provides the damping, preventing "hunting" or runaway oscillations of the whole feedback loop.

The resulting voltage is compared to a reference. This is how the speed selection (33/45/78 RPM) is usually done—by changing the reference voltage.

Comparison gives us the error voltage. This voltage is then driving the motor controller itself, telling it to increase or decrease the current through the coils.

If the motor rotates too fast, the reference will be lower than the FG voltage. The comparator will thus "tell" the controller to put less current through the coils, slowing the motor down. If the motor is too slow, the error voltage will increase, "telling" the controller to increase the coil current.

Precision voltage references are quite hard to implement. Instead, later DD motors often used a a quartz stabilized frequency reference inside of a phase locked loop. On paper, this is an amazing advantage: a common inexpensive quartz is stable to 50 ppm (0.005%), while a very good modern voltage reference can be stable to 50 ppm per degree of change in the ambient temperature!

Some turntables, like Micro Seiki's DQX-1000, actually gave the user a choice of two control strategies.

If you read any advertising copy or magazine reviews for direct drive turntables, you are likely to see the following statements:

1. The turntable's wow and flutter are below the measurement floor. It plays music with perfect timing and metronomic precision: the speed error is less than a second over the whole LP side!
2. There's something called cogging that affects our competitor's turntables, but is compeltely eliminated in ours

As is typical in advertising, these statements are obfuscating the very nature of the challenge while being technically true. In our next post, we will try to accurately and impartially describe the complexity of direct drive's speed behaviour.

To whet your appetite, here's a little chart. It shows speeds distribution—how many times a given momentary speed was measured over a given period of time. The speed was taken 5000 times a second with a proprietary contactless Korf Audio system, while a pop music LP was playing. Thin vertical grey line is 33 1/3 RPM.

Brown is a top quality late 1970s Japanese DD turntable. Blue is a similarly priced belt driven turntable of equally high standard. Both were tuned to exactly the same average speed (a little slow).

DD's momentary speed wanders over a much wider range, having at least 2 distinct peaks. The belt driven turntable has the correct speed almost twice as often—despite its much worse wow and flutter specification.

Why is it so?

Please subscribe to receive blog updates in your inbox!