Rega tonearm on Technics SL 1210 mk2
- Thread starter Andrei
- Start date
JP
7480 74111110101115
How did you come to that gem?
For those who are wondering:
The basis of Quartz Lock is a quartz crystal that vibrates at a very precise, constant frequency when electricity is applied.
Additionally, actual motor speed is monitored by a sensor, which generates a voltage based on its current rotational speed.
A phase-locked loop (PLL) circuit compares the voltage supplied by the motor sensor to the constant (reference) voltage supplied 'by' the crystal.
If motor speed drops or increases, the circuit immediately adjusts the voltage supplied to the motor to bring motor speed and thus platter speed back to reference.
/Mansplaining.
Make up your own mind if it matters.
The basis of Quartz Lock is a quartz crystal that vibrates at a very precise, constant frequency when electricity is applied.
Additionally, actual motor speed is monitored by a sensor, which generates a voltage based on its current rotational speed.
A phase-locked loop (PLL) circuit compares the voltage supplied by the motor sensor to the constant (reference) voltage supplied 'by' the crystal.
If motor speed drops or increases, the circuit immediately adjusts the voltage supplied to the motor to bring motor speed and thus platter speed back to reference.
/Mansplaining.
Make up your own mind if it matters.
JP
7480 74111110101115
No, that's an F/V servo. PLL compares the phase of the reference to the phase of a frequency generator from the motor rotation.
flavio81
Turntable technician
Hi,
My claim: "Furthermore, for most turntables, the way the circuit works, it doesn't even have major influence on wow and flutter either."
I will explain.
Wow is speed variations on frequencies around 0.55Hz.
Flutter is speed variations upwards of that frequency, with more importance (weighting) given to approx the 10Hz-100Hz range.
Quartz lock or not, a direct drive turntable uses a servo mechanism to keep good speed.
Let's set a goal of wow or flutter less than 0.04%, for this exercise.
The servo has basically two signals:
a. speed input from the turntable rotation, typically obtained through a tacometer device that produces an AC or pulse signal or voltage.
b. speed reference, which can either be an AC/pulse signal (from an oscillator) or a fixed voltage.
(b) speed reference, it is easy on the analog domain to produce a constant voltage that stays stable with less than 0.01% variation for a dozen or more rotations of the turntable, that is, very small fractions of a Hz, well out of the "wow" frequency region.
it is also easy on the analog domain to produce an AC pulse signal whose frequency stays stable with less than 0.01% variation for a dozen or more several rotations of the turntable.
The servo needs to compare (a) and make it match (b). It can be trying to match to voltages to one another, like on the oldest servos, or it can use a Phase Locked Loop to try to match the two frequencies against each other. Both method work. We assume the servo filter has been well designed, of course.
Thus, to achieve wow and flutter of less than 0.04, what will dominate the real wow and flutter is:
- the mechanics (quality of bearing, centering of platter, balancing of platter, available inertia)
- the electromechanics (precision and centering of the tacho generator)
- and of course, interactions between stylus friction and the turntable
Why quartz lock? One thing is to generate a frequency of x Hz with less than 0.01% variation for a few minutes. Another is to keep the same frequency constant for several hours. Such analog circuits have drift over time, for example due to temperature fluctuations. Here is where quartz lock is useful, to keep the absolute (average) operational speed constant for long amounts of time. Here is where quartz lock works. It won't have a major influence on W&F for the typical turntable.
Thus, the benefit of Quartz lock is mainly in elliminating drift, which is also the same reason quartz oscillators were introduced in watches: so the correct time does not drift over weeks, days or months.
Last edited:
Basically, I like Quartz Lock because I don't have to worry about adjusting the (+/-) speed dial for different record weights. Or even flipping between 33 and 45.
I wish my older Sony 2251 had the Quartz technology, though the sound quality is every bit as good as my newer Technics.
flavio81
Turntable technician
Yes, for everyday use it is practical.
Maybe we can get a jeweller make us a quartz record weight that can be placed on top of the record to "lock it" in place, then we'd have a "quartz locked 2251"
JP
7480 74111110101115
You're conflating two independent aspects of servo design:
- Reference stability over time: affects drift
- Servo topology: F/V (voltage comparison) vs. PLL (phase comparison) — affects W&F rejection
The difference is not about reference stability. Both topologies can have excellent reference stability — a precision voltage reference for F/V, a crystal oscillator for PLL. The question is what the error detector sees.
F/V is a rate servo, PLL is effectively a position servo. What you actually care about for pitch stability is angular position over time, not averaged velocity. PLL controls that directly. F/V throws away phase information and hopes averaged velocity is close enough.
An F/V servo converts the tach signal to a DC voltage proportional to average speed over some integration window. That averaging is required to reject ripple from the discrete tach pulses. Any speed perturbation shorter than the integration window is invisible to the loop until it has persisted long enough to move the averaged value.
The F/V topology also has a structural problem at wow frequencies specifically. The F/V time constant must be long enough to reject ripple from the discrete tach pulses — but that same time constant makes the loop sluggish to slow speed variations. Wow is slow, exactly where the F/V filtering is fighting you.
A PLL compares phase directly. Every tach edge is compared to where it should be relative to the reference. A velocity error doesn't need to persist and average out — it immediately begins accumulating phase error, which the detector sees on the very next edge. PLL has no time constant tradeoff — accumulated drift is visible regardless of how slowly it occurred.
This gives PLL an additional integrator in the loop. Phase is the integral of velocity, so a phase servo has one more order of integration than a velocity servo — better rejection of low-frequency disturbances, which is the definition of wow.
When you write that mechanics dominate W&F, you're describing the disturbance spectrum. The question is how well the servo rejects those disturbances. A PLL rejects them better because velocity error accumulates into visible phase error immediately, rather than being averaged away. This is easily demonstrable by looking at the instantaneous frequency vs. time of a well designed PLL implementation - subsequent revolutions will nearly perfectly overlay each other.
Thanks, but WOW! Way over my head! 
JP
7480 74111110101115
Thanks, but WOW! Way over my head!
Punchline - PLL (Quartz Lock) is fundamentally different in very real and very material ways. Absolutely makes a difference.
flavio81
Turntable technician
That's a big strawman argument you have assembled there, and I'm really puzzled since I am not sure it is due to you intentionally wanting to do a strawman argument, or if it is due to ignorance of turntable control methods.
You have presented a false dichotomy: "F/V (voltage comparison) versus crystal PLL".
It was never that way. Turntable servos use either F/V or PLL and this is independent of using quartz oscillators or not. You can do a servo using a PLL without the need for any quartz oscillator. PLL servo control has been used for servos for long before the appearance of quartz oscillators small enough to fit in a compact equipment. They are used when we need to match two signals in phase -- more about that below.
Again, conflating two orthogonal, independent concepts.
For example, the top of the line JVC TT-101 "quartz locked" turntable, does not use a PLL.
Furthermore, i'm a bit dissapointed that your reply implies a lack of fully understanding (or, perhaps, fully considering the details around) these control mechanisms. A PLL (phase locked loop) also has a comparison voltage output, just as the F/V control method also compares two voltages and gives an output. Moreover, you can generate a very stable voltage reference out of a very stable frequency reference and viceversa.
The only practical difference between a F/V comparison and a PLL comparison is that, while both can make sure the frequency of the signal matches exactly to our reference, the PLL can match them in phase, not just frequency. But phase matching is not necessary to our "low W&F goal" since wow and flutter are depending only on the frequency of rotation.
Last edited: