Brushed DC motor speed stabiliser

Here I describe the design of a simple brushed DC motor speed controller and stabiliser used in gramophone players and cassette players. The schematic shown above is for a 2-speed (33/45 RPM) belt-driven phonograph player, but can be adapted to 3-speed as well as single speed (for cassette players).

I have an old Kenwood belt-drive record player whose original motor was damaged and I had a motor scavenged from a from a wrecked Panasonic micro-system which I repurposed. Now the original motor had a resistance of about 40Ω, while the replacement is about 12Ω and ran way fast when connected the existing speed control circuit of the gramophone player. Also, nowadays motors with built-in speed control are available tailored to cassette or gramophone use. But nothing compares to DIY! So I decided to build a circuit to suit this motor instead of buying a new one 😄.

Working principle:

For speed control and stabilisation of a brushed DC motor, the preferred method is a tachometer or some kind of sensor (hall or auxiliary winding) based feedback system which will continuously monitor the motor's RPM and adjust the motor voltage (and speed) according to load, in effect maintaining constant motor speed under varying load conditions. This is what is required for a phonograph player - the total platter mass will increase when a record is placed over it and the motor should compensate for this.

The other way is to indirectly determine the load on the motor using its back-emf. With no mechanical load, when the rotor of a brushed DC spins as a result of a voltage being fed to it, an EMF is induced into its windings, which is apparent as if the motor is drawing less current (i.e. higher resistance). This EMF is known as the motor's back-emf. When a mechanical load is applied to the motor's shaft (i.e. rotor),  its speed will decrease causing in the back-emf to fall. When this happens the motor draws more current, and the apparent terminal voltage of the motor reduces.

In the present design we use the op-amp to detect the change in the motor's terminal voltage as a function of its back-emf. When the motor operates with a fixed mechanical load, its back-emf is fixed and the reference voltages that we measure in the following design process (2.45V/1.85V) are the ones with just the platter (but no record). When a record is placed on it, the mass increases and therefore the RPM will drop. If the motor were to be driven by just a regulated voltage source it would simply draw more current, but not increase its speed (the regulated voltage source will supply more current but not more voltage). However in this design, we use the motor in the feedback network of the op-amp. The op-amp will try ensure that whatever voltage is present at its non-inverting terminal, also appears across its inverting terminal - which is also the motor's terminal voltage. As the motor's RPM drops due to load, the back-emf drops - this appears to  the op-amp as a voltage reduction at its inverting pin. The inverting & non-inverting pins are now unbalanced, and the op-amp increases its o/p voltage make both pins equal. Note here that the op-amp's o/p voltage will increase in proportion to the motor's speed reduction, causing the motor's speed to be corrected. Also, the collector of Q1 is connected to the regulator's output (+9V) instead of the +15V supply rail. This is done to reduce the dissipation in Q1, though there is no harm in connecting it to the +15V rail except that Q1 will need a larger heat-sink.

Note:

1. There are many  ways of deriving a constant voltage to get the  reference voltages - for example by using +/-15V or +/-12V regulated rails and doing away with the 7809 altogether. However this design uses one regulator since the +/-15V rails are unregulated. Also the regulated voltage (whatever it be, as long as it can be used to get up to 5V by the voltage divider) will have to be used to calculate the resistor values.

2. This circuit has been simulated as well by modelling the motor as a series combination of the motor resistance and a (varying) voltage source of 200mV to 500mV.

3. The circuit uses dual-rails because the op-amp o/p can swing only to within 2V of each rail. In 33.33 RPM mode the output has to swing to a bit less than 2V which the 741 cannot do with just a single rail - this I found after building it with a single rail (therefore changed to dual rail).  The circuit can be used with a single 15V supply if an op-amp capable of rail-rail o/p and single-supply operation is used - in this case the op-amp's negative supply pin would be tied to ground. However, the 7809 regulator will still be needed (if the supply itself is not regulated) as it provides the constant voltage needed to derive the reference voltages. For higher resistance motors (40Ω for example) the motor voltage for 33.33 RPM is greater than 2V and a single rail can be used. For 16 RPM even with a higher resistance motor a dual rail might be needed - this has not been considered in this design.

Design process:

This process can be used to suit any DC motor of resistance 10Ω to 40Ω (even 50Ω) rated for 2 to 6V operation, BUT all values have to be recalculated. The only things that need not change are the op-amp, the +/-15V supplies, the 7809 regulator, the diode D1 (1N4007), the transistor Q1 (BD139), mounted on a small heat-sink) and the pot VR1 which serves as the pitch control. The supply rails need not be regulated but must be capable of about 1A.

Tools needed:

• Variable voltage power supply capable of at least 1A.

• Stroboscope disc or any mobile app to measure turntable RPM.

• A multi-meter.

Steps to design:

• Measure the motor's winding resistance for different rotor positions and take the average value. This is Rm. In the motor I used it is about 12Ω.

• With the new motor mounted and pulley attached to it, put the belt around the turntable and the pulley.

• Set the power supply voltage to zero before beginning. Connect it to the motor observing correct polarity for clockwise rotation. 

• Place the stroboscope disc (or mobile if using the RPM app) on the platter. Advance the voltage slowly to determine the voltage for every speed of interest. For example, in the present case, the voltages were about 1.85V for 33.33 RPM and 2.45V for 45 RPM. These values need not be exact (try to get as close as possible) and the final speed can anyways be adjusted using the potentiometer VR1 which serves as the pitch control.

• Select R6 to be the same as or closest to the motor's resistance (Rm) - a 10Ω to 12Ω resistor rated for 1W will work fine. This will also set the op-amp's gain to about 2. Note that with a higher resistance motor, the resistor will have to be increased - if it exceeds 18Ω, use one rated for 2W, if it exceeds 27Ω use 5W. The power dissipation is found by simulation (about 0.5-0.6W for 12Ω) and the wattage specified is double that for reliability.

• Start with the 45 RPM mode (SW1 closed). From the schematic we can see that the op-amp is operating with a gain of around 2 since R6 is the feedback resistor and M1 (Rm) is the gain resistor. G = 1 + (R6/Rm) = 2. This gain is not particularly critical, however, the voltage at Q1's emitter will be limited to about 8.4V if the op-amp's output exceeds 9V because Q1's collector is at 9V. For greater output connect Q1's collector to the +15V rail keeping in mind the increased dissipation in Q1.

• For 2.45V at the op-amp's inverting pin , we need 2.45V at the op-amp's non-inverting pin (op-amp principle which says that it will always try to maintain zero volts between its inverting & non-inverting pins). Choose VR1 to be 1k (it can be 2k or even 5k, but 1k seems best in terms of "aggressiveness" for pitch control). Assuming  VR1 is centred, the upper part is Rx = (R2 + 500)Ω (SW1 is closed so R1 is not in the picture) and the lower part is Ry = (R3 + 500)Ω. R4 is not considered since it is much greater than Ry. Rx & Ry form a voltage divider. Voltage across Ry = Vry =  2.45V = 9Ry / (Rx + Ry). Assume a value of 4k7 for R3, which means Ry is 5200Ω.  Upon calculation Rx is about 13.9k, therefore, R2 = 13.9k - 500Ω = 13.4k. Use a standard value of 12k. However for further calculations we still use the calculated value of 13.4k.

• Find R1 for 33.33 RPM. Since R2, VR1 and R3 are known, this time Rx becomes (R1 + 13.4k + 500Ω) = (R1 + 13.9k). For 33.33 RPM voltage across Ry (the lower part) is 1.85V. Recalculating, we get Vry = 1.85V = 9Ry / (Rx + Ry). But Ry is still R3 + 500Ω or 5200Ω. Upon calculation, Rx is about 20.1k, Since Rx = (R1 + R2 + 500Ω) = 20.1k, R2 = (20.1k - R1 - 500Ω) = (20.1k - 13.4k - 500Ω) = 6.2k. Therefore, R2 = 6k2. Use a standard value of 5k6, since 6k8 might be a bit too large. In practice it is found that 5k6 works well such that once VR1 is set for correct 45 RPM operation, switching to 33.33 RPM  mode gives a measured speed of about 33.5 RPM which is quite close.

This design deliberately uses fixed resistors for R1 & R2, and VR1 to control the pitch. If needed R1 & R2 can be replaced with trim pots adjusted to the exact calculated values - the only downside is that with a standard trim-pot, dust or age can make the trim-pot wiper contacts unreliable and maintaining the speed will become problematic. A multi-turn trim-pot would be much reliable though. The diode D1 is used to suppress any fly-back voltage the motor may generate when power is removed.