It may be interesting to make a stereo encoder on an experimental basis, for example to broadcast in another room of the home, using a small FM transmitter, the sound of a program from a TV or NICAM a stereo VCR.
Principle of transmission in stereo
The aim is to transmit on the same carrier frequency two audio signals (left channel and right channel, or G and D), with a rate of crosstalk as low as possible (maximum separation between the channels), while ensuring compatibility monophonic. This is achieved by passing the sum signal L + R (mono track) and their difference GD (stereo channel). It follows that the amplitude of the monaural channel signal is maximum when the signals applied to channels L and R are exactly the same amplitude and phase. The stereo channel signal is then theoretically zero. Conversely, the amplitude of the channel signal reaches its maximum when stereo signals L and R are of equal amplitude but opposite phase. The monaural channel signal is zero. At the reception, the decoder recreates the stereo left and right channels by performing the relations G = 1 / 2 ((L + R) + (GD)) and D = 1 / 2 ((L + R) - (GD)) . The stereo channel signal is transmitted by amplitude modulation of a carrier at 38 kHz. This carrier is removed so that it does not interfere with the signal. A signal at 19 kHz, said pilot frequency, and synchronous with the carrier at 38 kHz, allows the receiver to recreate the carrier to extract the signed content in its sidebands. A transmission channel FM stereo thus consists of a frequency band from 30 Hz to 15 kHz corresponding to the mono channel, the frequency of 19 kHz pilot, and the 38 kHz sidebands, ranging from 23 to 53 kHz corresponding to the stereo channel. In this case, the left and right channels are multiplexed at a frequency of 38 kHz, each channel being transmitted sequentially over a period of about 13 microseconds. The top of the channel can be used for various data transmissions, such as RDS (Radio Data System). Figure 1 shows the spectrum of such a FM channel.
The circuit diagram
There are now integrated circuits performing the coding function stereo (used for example in some stereo headphones without HF son), they, as well as some components they implement, are poorly distributed in France . The realization that we propose uses conventional components in an assembly with some certainty on the spectral space, with the settings and levels to adapt to most situations. The diagram of Figure 2 summarizes the operation of the stereo encoder, and Figure 3 shows the electrical diagram.
The generation of 38 kHz and 19 kHz is obtained from a 456 kHz oscillator. Indeed, rather than using a quartz frequency "exotic", we use here in a ceramic resonator 455 kHz. Low cost, this component is cut by the manufacturer to operate at its characteristic frequency of 455 kHz, when loaded with a capacity of about 40 pF, but willingly accepts to run at 456 kHz with a series capacitance , consisting in this case, by C3 in parallel with CV1, in the assembly formed by the trigger IC 1a.
The output signal 3 of it is buffered by IC 1b, in addition to switching to mono before being applied to the clock input of IC 2. This is a 4018, here in wired divider 6. The 76 kHz signal available at the output is directed then to the input clock of D flip-flop IC 3a, in which we have output of 38 kHz, edited by IC IC 1c and 1d, for carriage on control of the multiplexer IC 9. This is a classic case of 4053, one third is dedicated to the multiplex channels L and R, which reached its came into pin 12 and 13. The input is connected to VEE-8V, in order to exploit the maximum possible momentum. In monophonic mode, force the door 1c IC switching the right channel.
The second D flip-flop contained in IC 3 is responsible for the divide by 2 signal of 38 kHz. The square wave signal at 19 kHz resulting undergoes integration R3/R4/P1/C5 the network before being applied to the input of the operational amplifier, mounted in an open loop, consisting of IC 4a. The signal available at the output of the latter, adjustable phase from P1, is applied to the input of low pass filter IC 4b. The oscillating circuit formed by L1/C8/C9 selects the fundamental frequency of 19 kHz, while the C10 capacity weakens harmonics. There is finally the output of IC 7 4b, a sinusoidal signal at 19 kHz, the amplitude is adjustable by the potentiometer P2.
The input circuits of the left and right channels are identical, with two differences: the left channel input has a set of two riders who can reverse the polarity of the signal for this adjustment, as we shall see further. The resistance values related to the floor IC5 are such that the voltage gain is the same if the amplifier is operating in reverse. The right channel has an offset adjustment, operated by P5, of IC 8, to cancel the residual voltage to 38 kHz, the output of IC 9.
The output of the IC first stage 5 (or IC 7 on the right track) is a rejector circuit tuned to 19 kHz, consisting of L2 (or L3), and its associated capabilities. This notch is designed to weaken the signals to 19 kHz, or close to this frequency, possibly at the entrance. Indeed, because of the following pre-emphasis, the signal to reach an amplitude detrimental to the pilot sub-carrier.
Then there is the cell pre-emphasis to 50 microseconds, formed by C15 (or C21) and R14/R15 (or R22/R23). Recall that the pre-emphasis is intended to increase the amplitude of high frequency signals to keep the signal to noise ratio. Indeed, the amplitude of the signal is inherently much lower than its frequency is high. Preemphasis is expressed usually by the time constant of the cell making it. Its standard value is 50 microseconds in Europe, and which is 75 microseconds in the USA. At the reception, a de-emphasis unit performs the inverse operation to restore the original amplitude of the signal.
It should be noted in passing that the transmitter which will connect the stereo encoder, should not behave itself pre-emphasis circuit. If one has, the latter must be disabled. Of course, this transmitter will also be able to "pass" the band takes the signal multiplex.
The multiplex signal output 14 of IC 9 is buffered by IC 10 before being applied to low pass filter, low impedance, which follows. This filter is designed to eliminate the harmonics caused by multiplexing at 38 kHz and its amplitude-frequency response theory is reproduced in Figure 4.
Decoupling supply:
An active filter, and features, based on operational amplifiers, would have required a greater number of precision components. The potentiometer P6 can fine-tune the high band of the filter. It is important that this band is as "flat" as possible in order not to introduce crosstalk between channels L and R The following amplifier compensates insertion loss, and applies the signal to the adder stage. Output of the latter, the gain is adjustable using P7, there are many of the multiplex signal including the 19 kHz pilot.
Except for input and output, all are direct connections between floors. This is made possible by the use of a balanced diet.
Food
Power is carried on a separate circuit, and requires no special comment. The low mounting allows to simply 2X9V 3VA transformer. Regulators are also used without radiators. The zener D1 allows the lighting of the LED if both power sources are present. The wiring diagram is shown in Figure 5.
Realization
There is no difficulty in reporting on the achievement of the two circuits. We will start preferably by obtaining the components. In case of physical differences (not many), it will still be possible to modify the layout of the circuit board. It is against the imperative to respect more closely the values of the components, especially those related to low-pass filter. If one is fortunate to have lots of components, and a multimeter features on selfmètre and capacitance, can be advantageously select the closest values.
As usual, it is recommended to solder first the straps. The designs of printed circuit boards are shown respectively in Figure 6 and 7, while the implementation is shown in Figures 8 and 9.
The knobs are positioned according to plan. Once wired, the cards will be mounted in a housing preferably made of metal, taking care to have the encoder card so that its inputs L and R are as much as possible away from the power transformer.
The connections between input and output multiplex card with the connector housing are preferably made with screened cable.
Startup and settings
After the usual controls, turning on and check the supply voltages. To complete the setting of this assembly, you must have at least a frequency counter, an oscilloscope and an audio generator (sinus). In the adjustment procedure that follows, we will assume that the nominal level at 1 kHz audio sources G and D, and the output level multiplex (or transmitter input) is 0 dBm, or 775mVeff/600 ohms.
- Connect the input of the frequency of TP1.
- Put the reverser position stereo, and adjust CV1 to get exactly 456 kHz on the frequency.
- Connect the frequency of TP2, and make sure you measure well 38 kHz.
- Check the oscilloscope (DC input) of TP2, the rectangular signal has an amplitude of about 8V and the fronts are steep.
- Connect the oscilloscope probe to TP3. One must observe a perfect sinusoidal signal whose amplitude, uncritical, should be around 8V C / C. Check the frequency that the frequency is 19 kHz.
- Remove from the socket IC4 Place the jumpers JP1 and JP2 set to "normal", ie in 1 and 2. Connect the generator audio channels simultaneously on G and D (strap inputs in parallel).
Connect the oscilloscope probe to TP4 (left channel) and verify that the DC component does not exceed a few millivolts. Set the generator to 1 kHz and 0 dBm, 2.2 V C / C (check this level to the oscilloscope) and adjust P3 for 2.2 V C / C.
- Reduce the level of the generator 10 dB (730 mV or C / C) and enn set the frequency to 10 kHz. Check that there are approximately 2.2 V C / C, the effect of pre-emphasis. Gradually increase the frequency of the generator. The level TP4 should continue to grow to around 15 kHz, then gradually decline to reach a minimum at 19 kHz, as a result of the rejection. This minimum must be at least 5 times lower than the level measured at 15 kHz (-14 dB).
- Make the same adjustment on the right track by taking the points 7 and 8 and replacing TP4 TP5 by P3 and P4 per.
- Disconnect the generator and put the G and D inputs shorted.
- Connect the oscilloscope probe to the output multiplex encoder. Adjust P5 for a minimum of residue 38 kHz (push the vertical sensitivity of the oscilloscope for easier observation).
- Remove the short circuit on the inputs and plug the generator to 1 kHz and 0 dBm. If the previous settings were done correctly, one must observe the output multiplex track perfectly fine. If this is not the case we can take a little adjustment of P3 or P4 (but not both!) So that the traces G and D are combined.
- Set P7 for the amplitude of 2.2 V C / C output multiplex.
- Disconnect the generator one of the inputs G or D, and short-circuit it.
- Synchronize the oscilloscope's input external synchronization by connecting it to the generator or, if one has a second probe, according to TP3 or TP4 input connected to the generator.
- Set the generator at around 5 kHz so you can see a stable image of the envelope of the output signal multiplex.
- Adjust P6 to obtain the basis of the signal as flat and horizontal as possible, as shown in Figure 10. The ratio of signal peak and that of the ripple signal to the base expresses the rate of crosstalk between the channels L and R
- Disconnect the generator set and input short circuit. Replace IC4 on its support. Leave the probe of the oscilloscope connected to the output multiplex board and the oscilloscope internal sync.
- Adjust P2 to obtain a sinusoidal signal at 19 kHz to about 200 mV C / C. Indeed, there must be a ratio of 10 between the trip due to the modulating signal and the total due to the driver alone.
- Place the jumpers JP1 and JP2 in position "setting", ie in 2 and 3. This provides two signals G and D of phase without using a generator with 2 independent outputs, rare in amateur.
- Reconnect the generator on the two inputs L and R, re-synchronize the oscilloscope and adjust the external generator around 1 kHz in order to visualize the points of the signal in a stable manner.
- Adjust the phase of the pilot with P1 to align the best points of the signal as shown in Figure 11 (possibly by pushing the vertical sensitivity of the oscilloscope for easier observation).
- Replace the jumpers set to "normal". The encoder is set.
Test
Connect the output of the encoder multiplexes the input of the transmitter. Turn on the encoder and transmitter and an FM receiver switch (stéréo!) tuned to the frequency of the transmitter. Strictly speaking it should now adjust the input level of the issuer for an excursion of 7.5 kHz without signal encoder inputs (pilot only). Should be to do this, have a receiver-deviation meter or at least a spectrum analyzer, which is not the lot of everyone! In practice, and inside the home with a low power transmitter, we can proceed as follows:
- No signal on the inputs, gradually increase the input level of the issuer or, if the output level by P7 the encoder until the light comes on the stereo receiver. Continue to rise slightly above the level in order to have a small margin, and be sure that the receiver decoder is positively locked.
- Connect the generator to 1 kHz, 0 dBm on the left input of the encoder.
Check that one "sort" good of the left speaker and receiver. If the receiver has lower left and right outputs, we can connect the oscilloscope to the output slightly right and back adjustment of P1 and P6 on the encoder to obtain the minimum signal.
- Disconnect the generator from the left lane to connect on the right track, and verify that this is the right speaker of the receptor that is sought.
- Switch the inverter mono encoder position and check that the two speakers and receiver are in action and seeing the stereo is off. Note that in mono position, only the right channel of the encoder is active.
- You can now connect audio sources dedicated to the left and right channels.
Remember that the show is prohibited without special permission and that these tests can not be and experimental using a low power transmitter.
Principle of transmission in stereo
The aim is to transmit on the same carrier frequency two audio signals (left channel and right channel, or G and D), with a rate of crosstalk as low as possible (maximum separation between the channels), while ensuring compatibility monophonic. This is achieved by passing the sum signal L + R (mono track) and their difference GD (stereo channel). It follows that the amplitude of the monaural channel signal is maximum when the signals applied to channels L and R are exactly the same amplitude and phase. The stereo channel signal is then theoretically zero. Conversely, the amplitude of the channel signal reaches its maximum when stereo signals L and R are of equal amplitude but opposite phase. The monaural channel signal is zero. At the reception, the decoder recreates the stereo left and right channels by performing the relations G = 1 / 2 ((L + R) + (GD)) and D = 1 / 2 ((L + R) - (GD)) . The stereo channel signal is transmitted by amplitude modulation of a carrier at 38 kHz. This carrier is removed so that it does not interfere with the signal. A signal at 19 kHz, said pilot frequency, and synchronous with the carrier at 38 kHz, allows the receiver to recreate the carrier to extract the signed content in its sidebands. A transmission channel FM stereo thus consists of a frequency band from 30 Hz to 15 kHz corresponding to the mono channel, the frequency of 19 kHz pilot, and the 38 kHz sidebands, ranging from 23 to 53 kHz corresponding to the stereo channel. In this case, the left and right channels are multiplexed at a frequency of 38 kHz, each channel being transmitted sequentially over a period of about 13 microseconds. The top of the channel can be used for various data transmissions, such as RDS (Radio Data System). Figure 1 shows the spectrum of such a FM channel.
The circuit diagram
There are now integrated circuits performing the coding function stereo (used for example in some stereo headphones without HF son), they, as well as some components they implement, are poorly distributed in France . The realization that we propose uses conventional components in an assembly with some certainty on the spectral space, with the settings and levels to adapt to most situations. The diagram of Figure 2 summarizes the operation of the stereo encoder, and Figure 3 shows the electrical diagram.
The generation of 38 kHz and 19 kHz is obtained from a 456 kHz oscillator. Indeed, rather than using a quartz frequency "exotic", we use here in a ceramic resonator 455 kHz. Low cost, this component is cut by the manufacturer to operate at its characteristic frequency of 455 kHz, when loaded with a capacity of about 40 pF, but willingly accepts to run at 456 kHz with a series capacitance , consisting in this case, by C3 in parallel with CV1, in the assembly formed by the trigger IC 1a.
The output signal 3 of it is buffered by IC 1b, in addition to switching to mono before being applied to the clock input of IC 2. This is a 4018, here in wired divider 6. The 76 kHz signal available at the output is directed then to the input clock of D flip-flop IC 3a, in which we have output of 38 kHz, edited by IC IC 1c and 1d, for carriage on control of the multiplexer IC 9. This is a classic case of 4053, one third is dedicated to the multiplex channels L and R, which reached its came into pin 12 and 13. The input is connected to VEE-8V, in order to exploit the maximum possible momentum. In monophonic mode, force the door 1c IC switching the right channel.
The second D flip-flop contained in IC 3 is responsible for the divide by 2 signal of 38 kHz. The square wave signal at 19 kHz resulting undergoes integration R3/R4/P1/C5 the network before being applied to the input of the operational amplifier, mounted in an open loop, consisting of IC 4a. The signal available at the output of the latter, adjustable phase from P1, is applied to the input of low pass filter IC 4b. The oscillating circuit formed by L1/C8/C9 selects the fundamental frequency of 19 kHz, while the C10 capacity weakens harmonics. There is finally the output of IC 7 4b, a sinusoidal signal at 19 kHz, the amplitude is adjustable by the potentiometer P2.
The output of the IC first stage 5 (or IC 7 on the right track) is a rejector circuit tuned to 19 kHz, consisting of L2 (or L3), and its associated capabilities. This notch is designed to weaken the signals to 19 kHz, or close to this frequency, possibly at the entrance. Indeed, because of the following pre-emphasis, the signal to reach an amplitude detrimental to the pilot sub-carrier.
Then there is the cell pre-emphasis to 50 microseconds, formed by C15 (or C21) and R14/R15 (or R22/R23). Recall that the pre-emphasis is intended to increase the amplitude of high frequency signals to keep the signal to noise ratio. Indeed, the amplitude of the signal is inherently much lower than its frequency is high. Preemphasis is expressed usually by the time constant of the cell making it. Its standard value is 50 microseconds in Europe, and which is 75 microseconds in the USA. At the reception, a de-emphasis unit performs the inverse operation to restore the original amplitude of the signal.
It should be noted in passing that the transmitter which will connect the stereo encoder, should not behave itself pre-emphasis circuit. If one has, the latter must be disabled. Of course, this transmitter will also be able to "pass" the band takes the signal multiplex.
The multiplex signal output 14 of IC 9 is buffered by IC 10 before being applied to low pass filter, low impedance, which follows. This filter is designed to eliminate the harmonics caused by multiplexing at 38 kHz and its amplitude-frequency response theory is reproduced in Figure 4.
Decoupling supply: An active filter, and features, based on operational amplifiers, would have required a greater number of precision components. The potentiometer P6 can fine-tune the high band of the filter. It is important that this band is as "flat" as possible in order not to introduce crosstalk between channels L and R The following amplifier compensates insertion loss, and applies the signal to the adder stage. Output of the latter, the gain is adjustable using P7, there are many of the multiplex signal including the 19 kHz pilot.
Except for input and output, all are direct connections between floors. This is made possible by the use of a balanced diet.
Food
Realization
There is no difficulty in reporting on the achievement of the two circuits. We will start preferably by obtaining the components. In case of physical differences (not many), it will still be possible to modify the layout of the circuit board. It is against the imperative to respect more closely the values of the components, especially those related to low-pass filter. If one is fortunate to have lots of components, and a multimeter features on selfmètre and capacitance, can be advantageously select the closest values.
As usual, it is recommended to solder first the straps. The designs of printed circuit boards are shown respectively in Figure 6 and 7, while the implementation is shown in Figures 8 and 9.
The knobs are positioned according to plan. Once wired, the cards will be mounted in a housing preferably made of metal, taking care to have the encoder card so that its inputs L and R are as much as possible away from the power transformer.
The connections between input and output multiplex card with the connector housing are preferably made with screened cable.
After the usual controls, turning on and check the supply voltages. To complete the setting of this assembly, you must have at least a frequency counter, an oscilloscope and an audio generator (sinus). In the adjustment procedure that follows, we will assume that the nominal level at 1 kHz audio sources G and D, and the output level multiplex (or transmitter input) is 0 dBm, or 775mVeff/600 ohms.
- Connect the input of the frequency of TP1.
- Put the reverser position stereo, and adjust CV1 to get exactly 456 kHz on the frequency.
- Connect the frequency of TP2, and make sure you measure well 38 kHz.
- Check the oscilloscope (DC input) of TP2, the rectangular signal has an amplitude of about 8V and the fronts are steep.
- Connect the oscilloscope probe to TP3. One must observe a perfect sinusoidal signal whose amplitude, uncritical, should be around 8V C / C. Check the frequency that the frequency is 19 kHz.
- Remove from the socket IC4 Place the jumpers JP1 and JP2 set to "normal", ie in 1 and 2. Connect the generator audio channels simultaneously on G and D (strap inputs in parallel).
Connect the oscilloscope probe to TP4 (left channel) and verify that the DC component does not exceed a few millivolts. Set the generator to 1 kHz and 0 dBm, 2.2 V C / C (check this level to the oscilloscope) and adjust P3 for 2.2 V C / C.
- Reduce the level of the generator 10 dB (730 mV or C / C) and enn set the frequency to 10 kHz. Check that there are approximately 2.2 V C / C, the effect of pre-emphasis. Gradually increase the frequency of the generator. The level TP4 should continue to grow to around 15 kHz, then gradually decline to reach a minimum at 19 kHz, as a result of the rejection. This minimum must be at least 5 times lower than the level measured at 15 kHz (-14 dB).
- Make the same adjustment on the right track by taking the points 7 and 8 and replacing TP4 TP5 by P3 and P4 per.
- Disconnect the generator and put the G and D inputs shorted.
- Connect the oscilloscope probe to the output multiplex encoder. Adjust P5 for a minimum of residue 38 kHz (push the vertical sensitivity of the oscilloscope for easier observation).
- Remove the short circuit on the inputs and plug the generator to 1 kHz and 0 dBm. If the previous settings were done correctly, one must observe the output multiplex track perfectly fine. If this is not the case we can take a little adjustment of P3 or P4 (but not both!) So that the traces G and D are combined.
- Set P7 for the amplitude of 2.2 V C / C output multiplex.
- Disconnect the generator one of the inputs G or D, and short-circuit it.
- Synchronize the oscilloscope's input external synchronization by connecting it to the generator or, if one has a second probe, according to TP3 or TP4 input connected to the generator.
- Set the generator at around 5 kHz so you can see a stable image of the envelope of the output signal multiplex.
- Adjust P6 to obtain the basis of the signal as flat and horizontal as possible, as shown in Figure 10. The ratio of signal peak and that of the ripple signal to the base expresses the rate of crosstalk between the channels L and R
- Disconnect the generator set and input short circuit. Replace IC4 on its support. Leave the probe of the oscilloscope connected to the output multiplex board and the oscilloscope internal sync.
- Adjust P2 to obtain a sinusoidal signal at 19 kHz to about 200 mV C / C. Indeed, there must be a ratio of 10 between the trip due to the modulating signal and the total due to the driver alone.
- Place the jumpers JP1 and JP2 in position "setting", ie in 2 and 3. This provides two signals G and D of phase without using a generator with 2 independent outputs, rare in amateur.
- Reconnect the generator on the two inputs L and R, re-synchronize the oscilloscope and adjust the external generator around 1 kHz in order to visualize the points of the signal in a stable manner.
- Adjust the phase of the pilot with P1 to align the best points of the signal as shown in Figure 11 (possibly by pushing the vertical sensitivity of the oscilloscope for easier observation).
- Replace the jumpers set to "normal". The encoder is set.
Test
Connect the output of the encoder multiplexes the input of the transmitter. Turn on the encoder and transmitter and an FM receiver switch (stéréo!) tuned to the frequency of the transmitter. Strictly speaking it should now adjust the input level of the issuer for an excursion of 7.5 kHz without signal encoder inputs (pilot only). Should be to do this, have a receiver-deviation meter or at least a spectrum analyzer, which is not the lot of everyone! In practice, and inside the home with a low power transmitter, we can proceed as follows:
- No signal on the inputs, gradually increase the input level of the issuer or, if the output level by P7 the encoder until the light comes on the stereo receiver. Continue to rise slightly above the level in order to have a small margin, and be sure that the receiver decoder is positively locked.
- Connect the generator to 1 kHz, 0 dBm on the left input of the encoder.
Check that one "sort" good of the left speaker and receiver. If the receiver has lower left and right outputs, we can connect the oscilloscope to the output slightly right and back adjustment of P1 and P6 on the encoder to obtain the minimum signal.
- Disconnect the generator from the left lane to connect on the right track, and verify that this is the right speaker of the receptor that is sought.
- Switch the inverter mono encoder position and check that the two speakers and receiver are in action and seeing the stereo is off. Note that in mono position, only the right channel of the encoder is active.
- You can now connect audio sources dedicated to the left and right channels.
Remember that the show is prohibited without special permission and that these tests can not be and experimental using a low power transmitter.
(From the article by B. Lebrun)
No comments:
Post a Comment