【Design
Specifications】
【Block diagram】  When the power is turned on, the command is set so that the 1PPS output of the NEO6M is 10KHz Duty = 50%, so that 10KHz can be used as a fixed reference clock for radios, and 10MHz can be used as a reference clock for other measuring instruments, etc. Also, the TXD/RXD of the NEO6M can be used as an interface for the CPU of this unit and a USB converter, and the 'U-CENTER' software can also be started from a PC. 【Circuit configuration】 The 10KHz output from the NEO6M is a 3.3V pulse, and the 10MHz-VCXO power supply is 3.3V, so the 10MHz output is also a 3.3V pulse. Therefore, it would be fine to adjust the power supply for the LCD/CPU/others to 3.3V, but since the LCD I have is for 5V and I wanted the 10KHz/10MHz output pulse to be 5V, I used a 5V power supply. Therefore, the 10KHz output from the NEO6M and the 10MHz output from the 10M-VCXO are converted to 5V pulses using part of the 74HCT86 used as a phase comparator. The 10MHz output of the 10M-VCXO is divided by 1/500 (20KHz) using a 12-bit binary counter (74HC4040). This IC can divide up to 1/4096, but when dividing with a general counter, it is possible to obtain a duty of 50% with 1/2^n division, but with other divisions, there will be a point where it is not possible to keep it at 50%. Therefore, if you want a duty of 50% for the target signal, first divide it up to the desired double frequency, and then divide it by 1/2 to get a duty of 50%. The phase of the 10KHz obtained in this way and the 10KHz from the NEO6M are compared using a 74HCT86 (X-OR). The phase comparator (74HC86) is an Xclusive-OR gate and operates as follows.  With 0-180 degree detection, the signal is output at twice the frequency of the input signal within the active region. When the output signal passes through a low-pass filter, it becomes a 0V to 5V control voltage, and this signal is used to control the VCXO to lock the phase. The feature of this phase comparator is that when one side of the input signal is not connected (solid L or solid H), the output signal is the other input signal output as is with a 50% duty cycle. When this is passed through a filter, the output voltage becomes 1/2 the voltage, minimizing the difference in frequency between the UN-Lock state when there is no reference signal or comparison signal and the VCXO in the normal Lock state. In other words, the lock-up time can also be shortened. For this reason, it is desirable to supply one side of the input signal in the UN-Lock state with a 50% duty cycle. The 10MHz signal output is only the VCXO signal, but there are two types of 10KHz: 10KHz from the NEO6M output and 10KHz divided from the VCXO. The 10KHz output signal is switched by a switch (relay). The TXD/RXD interface can be used for both USB and CPU, and a terminal is provided so that software can be started from a PC. Circuit Diagram Bill of Materials |
| 【NEO6M】 Previously, I made a GPS reference signal generator using a GPS unit (TU30/TU60). At that time, I had the impression that it was cheap because I had previously purchased GPS receivers costing tens of thousands of yen, or used expensive reference clocks such as OCXOs and rubidiums. However, this time, the model number is 'NEO6M', which is sold very cheaply (1680 yen) on Amazon and auction sites, and can be purchased online for 780 yen these days. Several stations I know have purchased multiple units, and I also purchased two. There was a reason for purchasing two units, so I purchased one and tested its operation.   This unit has an RS232C terminal, and if you connect it to a PC via a USB conversion unit and install U-Center on the PC, you can display the GPS positioning results and the link with the map on the monitor. Basically, without connecting it to a PC, if you apply power (+5V) and connect the antenna, it will link with several satellites in about 2 minutes and output a precise reference clock of 1PPS. The default output is 1PPS (1Hz). The output frequency can be set arbitrarily in the configuration settings, but the stable range is up to about 10KHz. If it is more than that, it will lose stability. In any case, to output 10KHz, it is necessary to connect it to a PC and configure it so that the output is 10KHz.   How to connect the conversion module to the NEO6M TX →RX RX →TX GND→GND When supplying power from the PC-USB terminal, connect VCC → VCC; however, as far as I could tell from my computer, the C/N ratio of the carrier output when supplying Vcc from the computer was terrible, so using computer power is a no-no. For us amateur radio operators, many devices use clocks of 15MHz to 30MHz as the reference clock frequency for the transmitters and receivers we use, and to create these reference clocks in a PLL configuration, all we need is a stable 10KHz as the PLL master clock.  The first unit I purchased was the blue unit. To output 10KHz, I connected it to a PC, started U-Center, and configured the output signal to be 10KHz with a duty of 50%, and it output 10KHz from the 1PPS output as specified. This unit is in a configured state, so if you click 'send' on the screen below and save it to the EEPROM, the settings should be reflected the next time you start it up.  There is no problem when the power is turned off and the backup is done, but when the backup is erased, it starts up in the default state and outputs 1Hz. Since it has an EEPROM, isn't this strange in terms of specifications? I thought so, and to check, I removed the battery (supercon) and found that it performed a save operation and returned to the default state when the power was turned OFF and ON in a short period of time. When I checked with various stations I know, the operation was different, with some (normal units) always starting up in the saved state even if the battery was gone once it was saved, and about half of them started up in the default state like my first unit. The EEPROM is connected with an I2C interface, and I checked the DATA-LINE, but it was not accessed at all when saving or when the power was turned ON. Even though an EEPROM is installed, it is not used at all. So I bought another one, this time a red unit with a different look. I immediately removed the batteries, connected it to a PC, set it to 10KHz output, and saved it. After that, it outputs 10KHz when I turn the power OFF then ON (either after a week or a month). I was able to confirm that the DATA-LINE was also accessed when saving and when the power was turned on. Even if it were possible to save and back up the data, connecting to a PC would be a hassle, and to use it as a master clock, all you need to do is connect the antenna, turn on the power, and leave it as is. To configure it as a standalone type, this unit unconditionally executes clock output = 10KHz (50%) when the power is turned on, and is configured to display only the minimum necessary information on the LCD, with a built-in 10MHz-VCXO that locks at 10KHz, and outputs 10MHz (2 systems) and 10KHz (2 systems) as output clocks. In addition, the 10KHz clock is fully SW-equipped with a function to switch between 10KHz directly output from the NEO6M and 10KHz divided by 10MHz. |
【Production】     Even if there is no backup function, the 1PPS output is unconditionally set to 10KHz (duty = 50%) when the power is turned on. The number of satellite links varies depending on the region and time of day as well as the antenna installation conditions. 【Power supply circuit】 When driving the LCD/relay, the current consumption is approximately 130mA. Do not use power from the computer's USB (+5V).  Completed (Panel painting failed)  【Production Results】 When connecting to a USB converter, it is possible to supply power (+5V) from the computer USB, but never use this power source. If you want to generate a pure system clock, supply it from an independent 5V power source. I checked it this time and it has a terrible C/N.   I don't have a high-precision oscillator like a rubidium oscillator, but I do have a ninth-power class OCXO, and I previously adjusted the frequency of the OCXO to 10 MHz locked to the color burst of a TV. This time, I adjusted the frequency of my OCXO to a 10 MHz-VCXO locked to GPS, and compared the 10 MHz output of each. The upper waveform is 10 MHz locked to GPS, and the lower waveform is 10 MHz output from the OCXO. Although there is some phase push-pull, the frequency accuracy is quite satisfactory. In the first place, it is not possible to measure frequency accuracy; it can be done by comparing the two and measuring the time it takes to complete one cycle, but although it is possible to measure the deviation of the OCXO against the GPS, when compared to something with much higher precision, I don't think it is possible to actually measure the one-cycle shift, even if it is possible to observe the push-pull of the phase. A 20-second video comparing GPS lock 10MHz (top) and OCXO-10MHz output (bottom)  In the past, in the analog TV era, we used a 9th power class OVCXO (oven controlled crystal oscillator) as a reference clock to lock the master clock of the radio. At that time, we adjusted the OVCXO by comparing it with 10MHz locked to the TV burst signal for adjusting it. After that, analog TV disappeared and there was no reference clock to check the adjustment. Even after moving to digital TV, analog composite video signals are still output, but the burst signal generated by this depends on the accuracy of the crystal used in the NTSC encoder inside the device. Therefore, it varies depending on the manufacturer and high accuracy cannot be expected. Therefore, nowadays, it is possible to generate a reference clock with an inexpensive and simple GPS unit, so we input the GPS-locked 10MHz of this unit into the comparator we created previously and took a video of how the OVCXO rise characteristics change. The OVCXO is aged sufficiently and the frequency is temporarily adjusted to the GPS 10MHz. After that, the OVCXO was turned off and a few minutes later, a video was taken after turning it back on; the frequencies of both were perfectly matched in just under three minutes. If the OVCXO has been turned off for an extended period of time, it will take a considerable amount of time for the frequencies to match, so this time the experiment was conducted with an off period of just a few minutes. The comparison frequency of both was set to 1 MHz by dividing 10 MHz each. This was created at 1 MHz so that the frequency discrepancy can be easily determined, and results are immediately obtained in terms of how many seconds it takes to complete one rotation, or how many rotations there are per second. The upper waveform is the GPS clock, and the lower waveform is the OVCXO clock. When the phases of the two signals stop (locked state), the LED will no longer light up or move.  When the OVCXO is powered on, the frequency will overshoot once, shift in the opposite direction, and then return over time. This can be used to adjust the OVCXO frequency to match the GPS clock frequency above (to a position where there is no left or right drift). When the frequency is matched, the LED will not move and will be fixed. |