MPX96 FM Stereo Transmitter

The North Country Radio MPX96 FM stereo low-power broadcast transmittter kit was a bit of an ugly duckling when it was introduced in the mid 1990s. Consisting of a circuit board, a bag of parts and several xeroxed instruction sheets, the kit was not suited for the novice kit builder. Nor was it suited to someone without an oscilloscope and a firm knowledge of the principles of the circuit's operation. However, for someone with the requisite building skills, knowledge and test equipment, the MPX96 could be turned into a swan.

But alas! Even properly built and adjusted, the completed original transmitter was flawed in a number of ways. The flaws are listed here in approximate order of importance (most important first). Some or all of these problems may have been corrected by the designers in subsequent versions of the MPX96:

1. The audio inputs were insufficiently low-pass filtered. This allowed high frequency artifacts in the audio signal, produced by the D/A converter of a CD player, for example, or the bias signal from a tape player, to beat with the 19 kHz stereo pilot. This was heard in the receiver as low level sputtering or chirping sounds.
   
2. The PLL was poorly filtered, allowing a 10 kHz artifact from the crystal reference to appear in the unit's modulated output. This was heard in the receiver as a faint high-pitched squeal or whine in the background.
3. The phase-locked loop (PLL) controlling the unit's carrier frequency was insufficiently damped, causing the PLL to "fight" the modulating audio at low frequencies. Because of time delays in the circuit, the result was an exagerated bass response. I suspect the low-frequency roll-off mentioned below was an attempt to compensate for this.
4. The unit's low frequency audio response was poor.
5. There was no modulation limiter.
6. The pre-emphasis networks were not accurate.
7. The unit had only so-so stereo separation performance (about 30 dB).
8. The unit had no convenient way to switch to mono mode.

The modifications outlined here range from easy on-board component changes and additions, to fairly involved external circuits added to the audio signal path. Listed below are the modifications, this time in order of increasing difficulty, sort of.

Easy:

1. The problem of poor low-frequency audio response was cured by changing the value of C11 to 10 uF (from the original 1.0 uF), changing C8 to 1.0 uF (from 0.1 uF), changing C5 to 1.5 uF (from .47 uF) and eliminating entirely capacitors C3 and C4 from the audio signal path..
2. The problem of the poorly filtered PLL was cured with the addition of a .001 uF capacitor from pin 3 of IC6 to ground.
3. The PLL damping problem was fixed by changing the values of C19 and C20 to 10 uF and 1.0 uF respectively (from their original values of 0.1 uF and .01 uF).
4. Stereo separation performance was improved almost 20 dB by substituting 1% (22.1k) resistors for R5, R6, R7 and R8 in the circuit and by installing a .1 uF capacitor at C7 in place of the original .01 uF.

Slightly more difficult:

5. A mono/stereo mode switch was added as follows: 1) Pin 11 of IC4 was isolated from ground by cutting the foil trace between IC4 pins 11 and 8, and between IC4 pin 11 and IC3 pins 1 and 2; 2) a jumper was added from IC3 pins 1 and 2 to ground (restores the ground connection to those pins); 3) a 100k resistor was added from pin 11 to ground; and 4) a wire from pin 11 leading to an off-board SPST switch was added. The switch, when closed, applies +12 volts DC to pin 11. This forces the unit into mono mode.

More hassle, but well worth the effort:

6. An external audio filter/pre-emphasis network board was added. This circuit solves two problems by providing both audio input low pass filtering and accurate pre-emphasis. The schematic for the filter/pre-emphasis board is shown in Figure 2.

Figure 2



Note that, for direct connection to the transmitter input (limiter circuit below not used), capacitors C3 and C4 are removed from the MPX96 board and the audio input connections are made directly to the junctions of R5 and R7 (left channel) and R6 and R8 (right channel).

Lots more hassle:

7. A relatively complex (for this project) modulation limiter circuit is shown in Figure 3. It responds slowly to average input levels and more quickly to short term peaks, but does not respond instantaneously (because its operation depends on an error signal derived downstream from the gain control portion of the circuit). The limiting action is soft. The limiter will not necessarily prevent overmodulation in the legal sense (75 kHz peak deviation) but does provide leveling and protection from distortion at moderate to high audio input levels. The circuit required quite a bit of fiddling to get right but resulted in a very good sounding transmitter.

   Figure 3

   
   
   The source for the 'BB in' signal input is TP4 on the MPX96. This is the composite 'baseband' signal from summing amp IC6 which directly modulates the transmitter's voltage-controlled carrier oscillator. By sampling the transmitter at this point the entire modulation signal envelope is monitored: the main audio channels (L+R); the sub-carrier signal (L-R); and the 19 kHz pilot. The baseband signal is converted to DC in a high-speed precision rectifier (LM6171BIN), then peak-detected in the following stage. The voltage from the peak detector is fed to LEDs in close proximity to two light-dependent resistors (LDRs) which make up part of a variable attenuator circuit in the audio path preceeding the transmitter input. The LDR's resistance varies inversly with the LED's light intensity, controling the loss through the attenuators. As the signal level, and thus the LED's intensity, increases, the loss through the attenuators increase.
   
   The LM3914 is the 'monitor' portion of the circuit. It monitors the voltage output from the peak detector and drives three LED's. No LEDs lit means no audio signal is being applied; the amber LED lit means an audio signal is being applied but it is of insufficient level to fully modulate the transmitter; the green LED lit means the transmitter is being fully modulated; the red LED lights when the audio input level is too high and the circuit is into limiting mode.

Not necessary, but neat:

8. A "lock" indicator was added. See Figure 4.

   Figure 4

   

   This circuit monitors the voltage at TP3 on the MPX96. This is the output from pin 28 of the unit's PLL chip. When the voltage at this point rises above a set point the LED lights, indicating a lock condition. The LED is mounted on the front panal of the exciter/stereo generator and serves as its "power on" indicator.

Two nit pick items, hardly worth mentioning:

9. A 82 pF capacitor was installed across R38; a .01 uF capacitor was installed across R28. (Refer back to Figure 1.) These capacitors reduce harmonics of the 19 kHz pilot signal.
10. An LF353 was installed in place of the LM1458 for IC1. The LF353 is a slightly better choice for this application.

Finally, an item useful for testing:

11. A direct external connection to the modulator was added. This allows a test tone from an external modulation source to be applied directly to the modulator. This would be useful, for example, in a test setup in which the Bessel null method is used to calibrate a monitor receiver. This mod consists of an added external BNC jack connected to IC6 pin 2 through a .1 uF capacitor and a 100k resistor. (Refer back to Figure 1.)

The MPX96 is still available from North Country Radio. I would recommend all of the above modifications except number 7, the modulation limiter, to anyone building one of these transmitters. You might consider tackling #7 after all of the others are in place and the unit is functioning satisfactorily.

Here are the Figures on this page in printable (PDF) format.