Again: "FG servo" what is it?

[BillWojo]

Well I have had both FG and QL Victor made turntables. On my FG turntable, a Victor TT 61 motor drive the speed would drift for a few minutes as evidenced by the built in strobe. Once it settled down it would remain perfectly stable for days. On the QL 71 motor drive it was stable as soon as I turned it on and stayed that way.
I could be happy with either version but was offered a rebuilt TT 71 motor drive so I sold the TT 61 off. Both are very beautifully built drive motors and Victor (JVC) are known for their steadfast speed control.

BillWojo

 

[Wolfie62]

I agree, QL is better. Specs are better. FG Servo, at least in my samples, don’t have as good of specs; but the differences get into largely inaudible differences.

 

[Montmorency]

I found an additional fairly good comparative explanation of DC servo and FG servo systems in Kenwood KD-500 series service manual (http://www.tru-burn.com/ProductDoc/Kenwood_kd-500_550_service.pdf).

Combining what is said there with the existing answers here I can currently sum it up as follows:

DC Servo

A feedback-based system. The amplitude, the voltage of the feedback signal is the feedback parameter. The feedback is generated by a dedicated coil, which is typically wound inside the motor itself. In fact the whole speed control system is often implemented inside the motor.

The feedback signal from the aforementioned coil is converted to quazi-DC voltage. The feedback voltage and stable reference voltage is then compared by voltage comparator, which produces the corrective signal for the motor.

The feedback voltage can be thought of as DC voltage, but in reality it doesn't have to be ideally flat DC. Some ripple is acceptable.

FG Servo

A feedback-based system. The frequency of the feedback signal is the feedback parameter. The feedback is generated by either dedicated coil, a photo-coupled pair or a similar pulse-producing sensor.

The feedback signal passes through a pulse shaper (differentiator+mutivibrator etc.) and the result converted to quazi-DC feedback voltage. After that the system is essentially equivalent to DC Servo. I.e. the resultant feedback voltage and stable reference voltage is fed into voltage comparator, which produces the corrective signal for the motor. Note that in this system the signal processing stages ensure that the amplitude of the feedback signal does not matter, only the frequency does.

Again, after the pulse shaper and integrator the feedback voltage can be thought of as DC voltage, but in reality it doesn't have to be ideally flat DC. Some ripple is acceptable.

Phase-locked loop (PLL)

A feedback-based system. The phase of the feedback signal is the feedback parameter. The feedback signal is usually generated by a dedicated coil.

The feedback signal (apparently passed through a pulse shaper) is directly fed into phase comparator. The comparator immediately compares the phase of the feedback signal with the phase of the stable reference signal. Note that now the amplitude of the reference signal does not matter, only the frequency does. Any misalignment of signal phases detected by the comparator is converted to the corrective signal for the motor.

PLL systems can (and normally will) use quartz-based reference signal generators, in which case the speed control system is referred to as quartz PLL or quartz-locked system.

 

[egellings]

I have a ReVox A77 recorder, and the capstan outer-rotor motor has grooves cut into it around its periphery, and a head is placed near the rotor where the grooves move past under it, and that is used to set the speed of the capstan motor. Would that qualify as FG Servo?

 

[SaSi]

Both systems, PLL and FG Servo, compare feedback from the motor to a reference value.
With FG servo systems, the frequency received from the motor is converted to a dc voltage proportional to the frequency. That voltage is compared to a reference voltage and the "servo circuit" of the controller adjusts the power sent to the motor to either speed it up or slow it down, so that the error of the comparison is minimized.
This approach works great but has two points that can be improved.
a. The tolerances of various components, like resistors and comparator gains, affect how the error is derived and so the error calculation has it's own errors hence the speed regulation is not extremely precise.
b. The adjustment is as good as the reference value. A zener is a typical reference used in electronics. There are better reference components, like the LM285. It has an error of +/- 1,5% which is not great. The MAX6126 has an error of +/- 0,02% which is far better.
The error is not only affected by component tolerances but also drifts with temperature.

PLL systems again use a pulse train generated by the motor but utilize a frequency as a reference and compare the phase of the feedback signal to the phase of the reference one. Doing that, the amplitudes of reference signal and feedback signal are irrelevant and the only important aspect is the accuracy of the reference oscillator.
Now, it's really feasible to make a reference signal using, say, a 555 timer or even make an oscillator with a couple of transistors. But that would defeat the purpose as these oscilators are not accurate enough.
This is where quartz oscillators come into play. Typically, a crystal oscillating at 1-2MHz is utilized and the native frequency is divided down to a couple of kHz (1000 times or so). Bringing down the reference frequency 1000 times brings down the error of the crystal 1000 as well. And the crystals are pretty accurate and have much lower drift with temperature.
So, quartz PLL systems are delivering a more precise result that is much less affected by tolerances, temperature and age.

To answer the question about the Revox A77 (and B77 and PR99), since there is no crystal on the capstan speed control, it most certainly isn't quartz controlled. The board utilizes a 555 timer chip, so could this be a PLL? No it isn't because the feedback pulse train is converted to an average DC voltage that is compared against a reference (that can be modified to achieve vari-speed) to drive the motor to a constant regulated speed.

 

[Mike in SC]

Sorry to bump this year old post, but Question: are FG Servo motors interchangable? between say SL20 & 23 with SL B series?

Example, my SL20 (bought new, still working) FG Servo MHX-5P2RDA: motor dated 1976 ; red, white, blue, green wires
SLB5 FG Servo MMX-3H2RPA, dated 1980, red, white, yellow, black wires.
they are both 12VDC FG Servo's: I bet the green and black wires are the ground, presume red and white stayed the same: so use yellow where blue wire was on old motor.

I do notice there is an adjustment on the back of newer ones, so they might have a different feedback loop, not sure...

Has anyone tried to exchange different FG Servo motors with success?

wish you could upload pics: IMHO the tonearm on the SL20 was it's weak link (pivots were not adjustable); I installed the tonearm from SL B3, which is same design, but adjustable needle bearings, but with longer tonearm wand, so i carefully shortened it up to proper length to match the SL20 table setup. It aligned nicely and plays great.

The SL20 specs rival much more expensive tables sold today. I've had to replace cheap pots, and i think about 4 old electrolytic caps just to be safe, and seems to run fine. I have experienced speed drops, not sure if it's the motor itself, thats the last thing I havent been able to disassemble to see...