  |
In the first place, the easiest blocks for DIYers to install are
microphone amplifiers, tone amplifiers that color the sound, and speech
amplifiers. It is difficult to touch the inside of recent
manufacturer-made equipment, so many accessories are made by hand. Among
them, transmitters have an ALC function that limits the output to a
certain level. Speaking loudly into the microphone can cause splatter,
so I think radio waves with less distortion and no splatter sound
better. I hear the term Hi-FiSSB, but I'm not sure what the definition
of Hi-Fi is. I don't think there is any Hi-Fi in the 3KHz band SSB, but
I think they are probably calling it Hi-FiSSB to mean a sound that is
easier to hear (pleasant to the ears) within the 3KHz band. When I was
listening to QSOs around the time I started radio again (around 1987), I
often heard things like ``never let the ALC of the transmitter swing''
or ``lower the microphone gain because it will cause splatter.''
Certainly, with a 100W rated transmitter, it is desirable to operate it
at the rated output as much as possible, resulting in overdrive and
distorted radio waves. Therefore, I made a limiting amplifier so that no
matter how loud I shout, the radio waves will not be distorted. However,
the reason I made it myself was that I was listening to QSOs with
various stations and longing for a cockpit-like radio wave with a very
powerful signal and a loud background noise, and I wondered how to emit
such a radio wave. It was possible to do so by increasing the microphone
gain, but I wanted to generate a powerful signal without splatter or
distortion. Generally, it is a feedback type, which suppresses the gain of the previous stage when the output level exceeds the target level, and cannot be controlled unless the result exceeds the target level. The time from detecting the output level to controlling the gain is called the attack time, and the key is how to shorten this period. The feedforward type does not detect the target output level from the rear stage of the gain control circuit, but detects it from the front stage and predicts the output level and controls the gain, but even with this, it is difficult to make the attack period zero. The method is divided into two types, depending on whether the excessive input signal is detected from the front stage or the rear stage. Furthermore, there is a forward delay type, which is a method that completely eliminates distortion caused by the attack period, and this completely eliminates attack distortion by delaying the time (or more) until the detection signal for gain control is generated, thereby controlling the gain in advance. 【Feed-forward delayed】 This method delays the time (attack time) until the level of the detected signal exceeds the specified value and gain control is performed, and the level is controlled in advance by this time to eliminate attack distortion. The problem is whether there is a device with a good S/N ratio and low distortion to be used as a delay element. This can be easily done by digitizing the entire system. When digitizing the signal, the heart of the system is the analog to digital (A/D conversion) and digital to analog (D/A conversion) blocks, and especially when making it yourself, the layout and wiring process of the conversion section greatly affects the performance. There are three types of delay elements that can be realized: ① BBD (transfers electric charge from bucket to bucket) device, ② reverberation springs used in old audio echo systems, and ③ RAM read/write processing. ① BBD The signal is sampled and sent to the next stage, and the more stages there are, the longer the delay time can be secured, and the lower the sampling frequency, the longer the delay time can be secured. The theoretical playback frequency is 44.1KHz/2 = 22KHz, just like CD sampling (44.1KHz). However, considering sampling noise, it is desirable to have a frequency four times higher than the maximum playback frequency, so the sampling frequency used here is > 3KHz x 4 = 12KHz, and the number of stages for the required delay time value can be selected. However, this type of device was developed for karaoke equipment, and it is difficult to obtain distortion of more than 50dB.  The concept is the same as digital signal processing, where sampling discretizes continuous time (meaning it is not continuous but is discrete) and converts it into a numerical value at regular intervals, and this quantized data is called digital. ② Reverberation Spring In the old days of stereo sound equipment (vacuum tube era), this component was used to achieve an echo function. It converts an electrical signal into mechanical vibration with a transducer, passes it through a mechanical spring, and then converts it back into an electrical signal with a transducer. The length of the spring determines the delay time, and a feedback loop is created. The amount of feedback is varied with the VR to increase or decrease the echo function. I remember playing around with this when I was making vacuum tube amplifiers. This component does not have a sufficient bandwidth, so it is not suitable. Or rather, it is anachronistic. ③ RAM read/write processing (digital delay) The most likely method will be this, with sampling > 12KHz and the number of stages determined by the RAM capacity.  (Write address) - (Read address) = Delay time. If the ring memory ranges from 1 to 100, a deviation of 1 address will result in the minimum delay time, and a deviation of 99 addresses will result in the maximum delay time. Sampling = 1/12KHz = 0.083ms X99 address = 8.25ms is the maximum delay time. The number of bits is set according to the required D-range. This is the first device I built when I resumed wireless 30 years ago (1987). The only drawback is that the monitor sound is delayed, so the bone conduction sound generated by me and the delayed sound overlap, which can be a little strange. However, as you continue to use it, your ears will get used to it and the strange feeling will almost disappear.   【Feed-forward type】 The basic structure is the same as the delay type, but the delay element has been removed. Feed-forward is a method that detects whether the level is too high before controlling the gain and adjusts the level in advance, while this method has a fuzzy element, and judges that if it is adjusted to a certain level, it will probably be the same level and controls it accordingly. Therefore, if the input-output characteristics of the control signal and the gain control circuit do not match, constant control is not possible, and it is difficult to match the characteristics. Because it is a pre-control, attack distortion can be avoided, but if you allow for a margin, you will want to insert a delay element. 【Feedback type】 Since the output signal is directly detected and fed back, it can be extracted at a stable level by setting the loop gain. However, depending on the distribution setting of the loop gain, you may feel a sense of discomfort such as a hiccup. This is an old method that has not many devices that can be used for the gain control circuit, and in most cases, it was made by assembling FET elements and analog multiplication circuits discretely, but I remember that there was nothing that was satisfactory in terms of characteristics and it was not very good. The only drawback of the feedback type is that attack distortion occurs. Since no action can be taken unless the specified level is exceeded, it is natural that the head will be over-leveled (attack period). It is wrong to call it attack distortion, and the slight overshoot during the attack period is just because the level is high. As a transmitter, the distribution setting is set with a margin for the subsequent D-range, and the attack/release time constant setting is no problem in practice. Thirty years ago, when I started radio again, there was an electronic volume device for Hi-Fi audio equipment that controlled the level by DC voltage control, and it had very excellent characteristics. When I made a PSN transmitter, I needed a stereo (W circuit) circuit for audio circuit processing to set up an ISB mode, and the electronic volume was perfect with a single stereo chip built in. I would like to introduce a feedback type limiting amplifier using the M5283P.  Circuit Diagram Depending on where you put the limiter block, the output level will vary, and I often put it at the varimodule input and set the output level to 1Vpp. If you supply this to the microphone input terminal of a manufacturer's rig, for example, it will be at most a few hundred mV. Input a signal level about three times higher than the output amplitude you want to set. I will explain using the conditions I use, and there are only two adjustments when assembled. ① Input a 1KHz 3Vpp signal to the L-CH or R-CH input, and adjust VR2 so that the L-CH or R-CH output is 1Vpp. ② In this state, adjust VR1 so that the compression level meter shows two lights. Input-output characteristics for 2T signal  Level
shifting from zero input  【Limiting amplifier insertion position】 Ideally, it is desirable to control within the band components emitted from the antenna. For example, in a filter-type transmitter (300Hz to 3KHz), the signal input from the microphone has the strongest energy in the first formant (100Hz), and if this component is limited, the output band is 300Hz to 3KHz, and the level is limited by the audio signal, so the power does not go out from the antenna. Therefore, it is desirable to limit the components that are band-limited by audio (300Hz to 3KHz). When actually building a transmitter, it is better to pass the audio through the HP-F/LP-F and input the band-limited signal to the limiting amplifier, but if you try to respond flexibly to different filter types, it will be difficult in terms of the D-range. In reality, since the first formant (100Hz) is the problem, it would be sufficient to pass at least the HP-F and input it to the limiting amplifier. To limit other bands as well, you would need to install an ALC circuit in the RF stage. |