Programming
Program the micro in the unit, and while the programmer is connected, set the EEPROM parameters for the exact crystal frequency to be used. Two calculations are involved - the baud rate, and the division ratio. The crystal frequency should be in the range 2 - 10 MHz, and MUST be a multiple of 2 kHz (or 2.5 kHz if you change the firmware so the software divider divides by 2500). In the examples below, the unit is to be operated on 3600.000 kHz.Baud Rate
Make the following calculation:UART = (Fcrystal / (BAUD x 16)) - 1
For example, using a 3600 kHz crystal and operating at 9600 bps, the calculation will be:
UART = (3,600,000 / (9600 x 16)) - 1 = 22.4375
Take the nearest integer (22) and convert to hexadecimal (16HEX or 0x16)
By reversing the calculation, work out what the actual baud rate will be using the integer value. For example, using 3600 kHz and an integer of 22, the baud rate is 9782, or an error of 1.9%. If the error is much >1%, consider using a lower baud rate or choose a different operating frequency. Convert the value to hexadecimal as described, (you can use the Windows calculator) and program this value into the fourth byte of the EEPROM (address 0x03).
Frequency Setting
You will need a high quality 30pF parallel mode crystal for the chosen frequency, which must be an exact multiple of 2kHz between about 2MHz and 10MHz. To discover the correct counter divisor, divide the crystal frequency in Hz by 2000, subtract one and convert the value to hexadecimal:DIVISOR = (Fcrystal / 2000) - 1
For example, for 3600.000kHz, the value is 1799, or 0x0707 (0707HEX). Program this value into the fifth and sixth bytes of the EEPROM (address 0x04 - 0x05), with the high byte first.
Oscillator Setting
Run the PC monitoring software, and connect up the RS232 cable to the unit. Two windows will appear, a very slow moving large one at the top, and a faster moving narrow one at the bottom. Concentrate on the lower trace. Set the switch SW1 in the FREE position.Adjust the trimmer C2 so that first the red line becomes straight, and then fine tune so that the blue line is also straight. This puts the oscillator at an exact multiple of 50 Hz. It is possible to choose the wrong frequency, but highly unlikely unless the crystal chosen has widely different properties to those specified. If the straight lines cannot be achieved with the trimmer C2 near the middle of the range, consider changing C3 slightly. You will be able to tell by which way the trace moves as the trimmer is adjusted, whether to add or remove capacitance. (The red line shows coarse phase, and the blue line fine phase).
With a frequency counter, you can also check the frequency range of the VCXO, which should be at least 10Hz either side of the set frequency. Set the switch SW1 in the HOLD position. Short the point labelled VCXO to 0V and measure the frequency after 10 seconds. Then short the VCXO point instead to +5V, and measure the frequency again. The unit should work fine without this check being made, but this would be a good check to make if there is trouble with frequency lock.
Checking Lock
With a well warmed-up and accurate HF communications receiver, dial up the intended frequency of the Super-Stable Exciter, and then tune down in frequency by EXACTLY 1kHz. Operate the receiver in USB. Switch SW3 to the ON position, SW2 to the OFF position, and arrange for the receiver to hear the Exciter by adding a metre or so of wire to the Exciter output connector.Connect the receiver audio output to the PC sound card Line Input, and operate a spectrogram program such as ARGO or SPECTRAN. Using a width of 50 Hz (ARGO 3 sec mode) note the audio tone from the Exciter. Now change switch SW1 to the LOCK position. On the spectrogram display there will be an abrupt change in the trace frequency, as it smoothly curves to the correct spot, overshoots slightly, and after a minute or two the trace will be absolutely stable, and it should be at exactly 1kHz. If it is not, suspect your receiver calibration!
If you can still watch the phase display at the same time (see picture), you will notice an abrupt change in the red line,
followed by the blue line eventually (after a minute or two) slowing down its pace, and eventually going along
in a straight line as well. (The picture is a detail of the PC software display, and shows the oscillator
locking up in under one minute, with slight but well damped overshoot.)
Check lock a few times by switching SW1 to FREE and back, or removing power from the unit. Check the ticks and ID transmission by changing SW2 to the TICK/ID position, and set SW3 to the KEY position. In this position of SW3 the carrier will not be transmitted until lock has been reached (just the ticks will be heard).
Note that you won't see the phase data during the CWID (the PC display will seem to freeze) with most versions of the firmware. This is to ensure that the Morse keying is not interrupted by the serial comms. Use the above procedure to check lock, rather than turning on and off the power. The ID message comes out as soon as power is applied, so wait until the message is over.