As I opined in my piece on the Garrard 301 the best way to control the speed of any mains frequency AC motor turntable is to vary the frequency of the supply. But how are you going to do this? ringing up the local utility and asking them to change their frequency isn't going to work, so you're going to have to do it for yourself. Lots of the winery equipment with which I work runs on three phase variable frequency inverters which convert the supply to DC and then cut it up to give the required frequency, but these are a bit large and expensive for the job, besides which they put out huge amounts of RFI and run in class D so they would probably sound terrible. What we need is a variable frequency local oscillator able to supply enough power to run the motor on the 301 16 watts max. The oscillator should be variable through about +/- 10%. The oscillator part is easy since we don't want to change the frequency very often we can use a high purity sine wave oscillator such as a Wien bridge. I built a straight forward Wien bridge oscillator based on an LM833 IC chip using tungsten filament PTC stabilisation and made it variable between 40 and 60 Hz (for the 50 Hz pulley). This is much more variability than the original brake gave but I figured that the extra might come in handy for 78s. For functional reasons I split the Wien bridge in half and connected the variable resistor to the speed control knob on the Garrard, giving the circuit in fig 1.
Click pic for full size version
Any sensible person would then use one of the cheap IC power amplifiers like the LM1875T combined with some form of step-up transformer, to get a circuit like that shown in fig 2. This is simple, cheap, reliable, performs extremely and boring (but see AC synchronous drives). In the spirit of thermionic overkill, I decided to use a push-pull V/T amplifier. The motor is a very reactive load so a triode output is the best choice, but with the cost of power triodes being the way it is I decided to use triode strapped EL34s or KT88s instead. The motor draws 16 watts max which is just within the capability of a pair of EL34s or KT88s in ClassA with cathode bias but the measured performance turned out better when I converted it to fixed bias Class AB. There's a fair amount of feedback to keep the damping on the motor as high as possible. In order to keep this lot stable the open loop gain is deliberately rolled at 1kHz and a phase compensation capacitor is connected across the motor winding. The output transformer does not need to perform to audio grade I used a power transformer from an old console radio, using the HT windings for the plate connections, the mains winding as the output and the heater winding for feedback.
This transformer was designed to give 500V centre tapped from 240V supply so running it backwards and connecting it to the 120 V terminals on the TT motor means that I need 2 x 125 V RMS from the plates of the output V/Ts, which is well within capability with a 375V supply. The final version uses KT88s running at 60mA per plate, peak plate current is around 200mA so the total dissipation both plates is approximately 375 * (.200 * .637 + .050/2) 16W giving about 41W so they aren't being burnt up too much. The driver stage uses a Schmitt phase splitter, and I decided to use a diff amp front end to make feedback mixing easier. The swings at the grids of the KT88s need to be about 15V RMS so with 2.75 volts from the feedback winding and 20 dB of feedback we need forward gain of times 55 (35 dB). This is easily attainable using trusty 6SN7s in both the diff amp and the Schmitt phase splitter, giving the circuit in Fig 3.
Click pic for full size version
The power supply is similarly straightforward although it turns out to be more important to keep it quiet than I first thought excessive mains level hum will intermodulate with the output signal to create small regular speed variation which might well be audible. The closer the two frequencies are the worse this will be it might be worthwhile using the 50Hz pulley and a 50 Hz oscillator in 60Hz countries and vice versa. It is also very important to decouple the input stage from the output the first version of this suffered terribly from motorboating due to inadequate decoupling. I increased the decoupling to the present values and everything calmed down very nicely. The power supply is shown in figure 4. Note that the relay for the HT centre tap is operated from a 12V line which runs via the power switch on the 301 more on this later. The strange looking circuit attached to the 6.3V heater supply is a single phase line frequency LED strobe light for speed checking.
Click pic for full size version
I packaged this lot in a plain old budbox chassis, so it's not the prettiest of my creations.
Stop Press:
Ipod drives Garrard