The Sony minidisc recorder/player can be made to record and play at speeds other than the speed it was designed for. Although not tested to the limits, a variation of greater than ±20% should not be attempted because other factors come into play such as the data being presented to the decoder will be too fast or too slow for the decoder to function properly. This is not a problem for the design of a variable speed unit for callers and cuers since a change of between ±10% to ±15% is more than adequate to meet our needs. In terms or record RPM equivalent, a ±10 change would provide a speed range from 40.5 rpm to 49.5 rpm which will fit 99.999% of our needs.


1. Speed control of approximately ±10%.

2. A visual speed indicator.

3. Easy speed adjustment.

4. An integrated mechanical design that is both small and effective.

5. The ability to easily restore operation to the MZ-R3 internal crystal oscillator.

6. Reasonable cost

7. As little modification as possible to the MZ-R3.


A requirement that is dictated by the design is the supply voltage. It must be regulated and be 5 volts in order to reduce speed fluctuation as much as possible. Since the battery of the MZ-R3 cannot supply 5 volts and tapping the internally generated 5 volts was not considered as a viable option and the included charger does not provide sufficient voltage to permit proper operation, a new power pack must be provided.

The controller was built into the external battery case (Sony EBP-MZ3) supplied with the Sony MZ-R3. Nobody will use this case anyway because the lithium battery to fit it and the charger to charge that battery cost almost $200. Sonys other options include a rechargeable pack with 2 AA nickel metal hydride (Sony BP-DM20) batteries in a special holder (the holder is essential if you want the batteries to be charged) and regular AA batteries. I would recommend the nickel metal hydride battery pack (about $30) because it is a buffer between the external power pack and the MZ-R3 electronics. Getting the battery case ready for installing the control circuitry consists of clipping off a little plastic pin on the side of the case that goes into a hole in the MZ-R3. The purpose of that pin is to tell the MZ-R3 that an external battery is connected and to disable the internal battery. You need to fool the system because you are not installing an external battery. A lot of the excess plastic inside the case was removed with a dremel tool to provide as much open space as possible. You'll make that choice when you see what your final mechanical design looks like. (Note: the case comes apart - remove 2 screws that are on the sides of the large thumb wheel and now all that is left holding the halves together is the plastic catches that unlock when one edge is pressed in or the other is plied out. Takes a little playing the first time until you see the way it is designed.)

The design of the control consists of three main components, the voltage regulator, the display section and the voltage controlled oscillator.

The easiest to design was the voltage regulator. A Linear Technology LT1121CZ-5 or LT1121Z-5 micropower low dropout regulator was chosen because it can supply 150ma of current, has reverse battery protection, a low 0.4 volt dropout voltage and was designed for this type of application. The only additional components needed are capacitors on the input and output that both stabilize the regulator and bypass the RF from the voltage controlled oscillator. A 1N4001 rectifier diode can also be used to reduce the voltage supplied to the MZ-R3. It will provide a 0.7 volt drop. My design did not use this diode but it is probably a good idea to keep the supply to the MZ-R3 as close to that supplied by the original power pack. The MZ-R3 does have internal regulation that will handle the 6 volts but personally I would rather keep the voltage as low as practical. The connection between the power pack and the speed control circuit was accomplished using a male dc plug from radio shack and a female dc socket from Sony. The Sony socket is a replacement part for the MZ-R3's built in power pack socket and is p/n 1-691-099-51. The power pack I used is Radio Shack p/n 273-1664 that provides 2 regulated voltages - 4.5 and 6 volts and supplies 550ma. The yellow tipped adapter plug fits the sockets of the Sony. You can use the 4.5 volt setting for operation without the speed control and 6 volts when you are using the speed control. There was 3 capacitors used on the input and three used on the output. Since we are working at 45.158 MHz, bypassing can get to be fun and using several TYPES of capacitors reduces the chances of resonance in the bypass network. A combination of electrolytic, monolythic and ceramic are a good choice. The electrolytic should be at least 10uf, the monolythic in 0.47uf range and ceramic in the 0.001uf range. Keep leads as short as possible to reduce inductance. More about capacitors as we go along since in this application actual value in not as critical as placement and other attributes. In a sense, I'll give you ideas where to place capacitors and some values to breadboard but your final choice will be done through experimentation.

The next circuit is the display. We have little room for complex circuitry and using the principle of KIS (keep it simple) the following design was chosen. The design is based upon a National Semiconductor Dot/Bar Display Driver LM3914 and 10 LED's. The LED's were 9 red and 1 blue. You can save about $5.00 if you use some other color than blue. I thought about using one of those prepackaged bar graph led's with 10 bars in it but I rejected it because the center was not easily discernible and the form factor was poor (fairly large 20 pin dual inline style). My final choice was individual T-1 3/4 style units such as Digi-Key p/n P390-ND which is a Panasonic Blue LED shown on the back cover of Catalog 974Q. Your choice. Pick what you want. Mechanical as well as appearance are the most important factors. The design does not require individual current limiting resistors for the LEDs. The LM3914 acts as a current source and 1 resistor selects the brightness you want for all LEDs (A 1K resistor from pin 7 to ground provides about 12 ma to each LED). The LM3914 is operated in dot display mode (Pin 9 open circuit). The only additional parts are bypassing capacitors and a voltage divider which will be discussed further on in the design. For experimenting to see how the circuit functions you can use a 1.8K resistor from +5 volts to one end of a 1K potentiometer and pin 6 on the LM3914. The other end of the potentiometer goes to pin 4 on the LM3914 and a resistor of 3.3K to ground. The center of the potentiometer goes to pin 5 on the LM3914. Pin 7 of the LM3914 has a resistor valued at 1K to ground. An important design constraint is that the voltage at pin 6 must be about 1.5 below the supply voltage. This, as you will see, will cause us a little indigestion.

So, up to this point you can breadboard the regulator and display an try it out.

Now for the good part. Using a Motorola MC1658 Voltage Controlled Multivibrator you can make a simple 45.158MHz oscillator. A 10pf to 15pf capacitor is connected between pins 11 and 14. Pins 1 and 5 go to +5 volts. Pin 8 goes to ground. A few bypass capacitors go between +5 and ground close to the package. Pin 2 goes to a voltage divider. Pin 6 goes to a 510 ohm resistor to ground and a 0.001 uf capacitor. The other end of the capacitor is the RF output. Try around 3.5 volts into pin 2 as a starting point, measure the frequency at the output. Adjust the voltage to get 40.141Mhz. Record the voltage. Change the voltage to get 45.158MHz. Record the voltage. Finally, adjust the voltage for 49.172 MHz and again record the voltage. Looking at the waveform on a 100Mhz or better scope, play with bypassing to get as clean of a signal as possible. Bypass capacitors between Pin 2 and ground will probably be needed. Use ate least 10uf and a .47uf. My scope had a built in frequency counter so my job was easy.

Now for the fun parts. There is a conflict between the voltage required by the MC1658 to give you the frequency needed and the LM3914. It seems that you need too high a voltage for the MC1658 than will properly function on the LM3914. I resolved this as follows:

The MC1658 capacitor was lowered as much as possible so that the voltage required would be as low as possible as still remain out of a very nasty nonlinear region. In fact, you may need to go to 12pf or even 15pf if you are having terrible problems. Symptoms would be bad oscillation, frequency jumping (slight change in voltage - great change in frequency), inability to get the right frequency. This will leave you about a ½ volt too high to work properly with the LM3914. How to get rid of at least a ½ volt. I used transistors but I recommend darlington transistors prefabricated that way. A transistor set up in an emitter follower mode will give about .5 volt drop. A darlington set up the same way will give about a 1 volt drop - an advantage that would really give extra room to adjust the oscillator to get further away from that nasty nonlinear region. The downside of the approach is that we are trying to divide up a small voltage change into 10 parts and selection of the transistors and associated emitter resistors is necessary to give the same drop across all three reference points - the high end, the variable and the low end. In fact, thru proper selection, you can get some interesting display characteristics.

At this point you should have a fully breadboarded system. An option I thought about and even purchased some parts to try was eliminating the display and using a pot with a detent at the center point. By proper selection of the resistors in the divider you can make the detent 45.158Mhz and the ends the other frequencies. I didn't go this way because I wanted the visual display.

A few comments on adjusting the components. First measure the voltages needed for the oscillator at the end points and the center (45.158Mhz). The difference between the end point voltages will determine the values in the voltage divider as follows (assuming currents to pin 2 on the MC1658 and to the display are negligible):

At 40.141 MHz the voltage is = 3.68 volts = Vl

At 45.158 MHz the voltage is = 3.77 volts = Vc

At 49.172 MHz the voltage is = 3.84 volts = Vh

Making a voltage divider network consisting of 3 resistors in series connected between ground and +5 volts we have:

Potentiometer= 1K = Rp

R1 goes to ground.

R2 to +5 volts.

Vr is the voltage difference between the high frequency voltage and the low frequency voltage. It is the range that will be provided by the potentiometer and will provide for the speed adjustment of 10%:

Vr = Vh - Vl.......... Vr = 3.84 - 3.68 ..........Vr = 0.16 volts

Current thru the potentiometer (Ip) is then:

Ip = Vr / Rp.......... Ip = 0.16 / 1000.......... Ip = 0.16 ma

Calculating for R1:

I = 0.16 ma.......... Vl = 3.68

R1=I/Vl..........R1 = 3.68 / 0.16 ma.......... R1 = 23K ohms

Calculating for R2:

I = 0.16 ma..........Voltage = 5.0 - 3.84 = 1.16 volts

R2=V/I..........R2 = 1.16 / 0.16 ma.......... R2 = 7.25K ohms

I would use a 50K 10 turn pot for R1 and a 20K 10 turn pot for R2

To reduce experimenting, I would measure the currents needed by the display and the oscillator control line and take those into consideration. As a first analysis, the current needed by the display using emitter followers is approximately the voltage applied to the base of the emitter follower transistors divided by Beta of the transistor times the emitter resistor. If the emitter resistor is 100K. The voltage is 3.75 volts and the Beta is 100 than the current will be:

I = 3.75 / (100 * 100K)........... I = 0.375ua

This may look small but it really does have a large effect on the frequency.

A more serious impact is the current required by the oscillator. Unfortunately the spec. Sheet I have is not much help but I would assume it could be 25 ua or bigger which is really significant.

Bypassing the voltage divided is important. At least it was in mine.

Construction: If this was 10 years ago I would have done it with a printed circuit card. Unfortunately, I no longer have the facilities available to me so I did mine using perf board - bad, bad, bad for 45 MHz. You may be able to avoid a lot of problems I had.

If I was going to make pc boards I would proceed as follows:

Board #1 would be the display with the emitter followers and probably the regulator. I would put the LEDs down the center of foil side and the LM3914 on the other side. The spacing of the LEDs would match holes drilled along the top (from front to back of the external battery case) so that when placed inside the case the LEDs fit into the holes and hold the board in place. (A well placed drop of super glue on one of the LEDs after testing will make sure the board doesn't pop loose.)

Board #2 would be the oscillator and voltage divider network. I definitely would use plenty of ground plane. Good RF practices are highly recommended. The 1K pot would be on one end of the board facing the front so it will go into a hole in the case front and hold the board in place. Next would be the divider circuitry, then the oscillator and finally toward the back would be the regulator. I would make the divider pots accessible from the edge of the board that faces away from the MZ-R3. This way, tiny holes could be made in the case if needed for final adjustment instead of opening the case each time.

I mounted the power pack socket to the case near the battery cover and cut the cover to fit around the socket when closed. Ground from the socket goes to the added circuitry and to the connector that feeds the MZ-R3. + voltage from the socket goes thru a 1N4001 diode to the MZ-R3 connector. It also goes to the regulator input. I would put a capacitor or two at the socket and connector if possible to keep RF from leaving the case and also prevent RF from a remote mic. from entering since some work in that frequency range.

The fun part now is to get the RF to the MZ-R3. My solution was to open the MZ-R3 back cover by removing 6 small phillip head screws. What you will see is the main circuit board. With the battery compartment to your left and the MZ-R3 face down the IC in the middle (IC601 p/n CXD2536R) is the one containing the oscillator circuit. The crystal is directly above and is X601, RVR528, 45.158 MHz. Carefully remove the crystal. Attach 2 wires, one to the left pad on the pc board and one to the side of the crystal that was attached to this pad. MAKE SURE THE PAD ON THE BOARD IS THE ONE CLOSEST TO THE BATTERY COMPARTMENT SIDE OF THE MZ-R3. Connecting to the wrong pad WILL KILL THE IC WHEN YOU FEED IN THE EXTERNAL SIGNAL. Use as little solder as possible and put a small piece of tape over the connection on the pc board. I then replaced the crystal into it's original location (left edge is lifted up slightly) and reattached the crystal to the remaining pad on the card. If you didn't see the wires you would think the crystal was back into it's original position except one end is tilted upwards.

These wires then went to a subminiature mono 3/32" socket so that when the plug is removed the crystal is back in the circuit. The pad connection is to the tip contact and the crystal wire goes to the shorting tab. It has a shorting bar on the center conductor. This way when you remove the plug the MZ-R3 works as it was designed. For the wires I tried coax, shielded audio single and double conductor and thin wire wrap wire. The one that worked the best was the thin wire wrap wire. The coax and shielded stuff caused great problems with load on the oscillator causing instability.

Placement of the 3/32" socket. In an attempt to keep RF lines as short as possible and being half blind, mostly deaf and utterly unfamiliar with the internal working of the MZ-R3 I found what I thought was a location that was PERFECT in every way. The area under the circuit board in the upper right hand corner appeared to have plenty of space beneath it. It is also perfectly lined up so a plug could be mounted into the battery case along the same edge as the power plug that goes into the MZ-R3 but at the other end. That way, when you remove the battery case, plug comes out with it and lo and behold the MZ-R3 works as it always did. One small problem. It would no longer record. Matter of a fact, if you tried to record you would erase your disc. After removing whatever hair I had left, taking the thing apart many times with great fear, I discovered the problem. The socket I installed interfered with a mechanical arm that would try to move into position from a carefully hidden spot during record mode. So much for that. I believe I can file a little off the arm and maybe file a little off the socket and make it work but I haven't tried yet since I really don't care if it records. I may even try and disable the record switch so I don't erase a disc (I did that to one already). There are other spots to put the socket although they are not the PERFECT spot I thought I found. Maybe you can come up with a smaller socket to fit into my perfect spot. Or you may opt for making the speed control non removable or some other solution.

I have tried to outline as best as I could the speed controller and I hope you find it understandable.

This page was last revised on June 3, 1998

Webmaster -