RSGB HF Committee 

HF Awards 
 

1999 HF & IOTA Convention 
 

LF bands index 

          Useful LF Links:  Email me! 

RSGB HF Committee 
Files which relate to LF operation

Files for download from G4JNT 

Attached is SLOWCW.ZIP which contains an updated version of SLOWCW.EXE. This contains a number of improvements and extra CW characters suggested by DJ8WL. The main operational change is the ability to change CW speeds and messages without exiting the programme completely, by pressing escape once. The original message / dot period can be retained by pressing [rtn] at the appropriate places. Common punctuation characters have been included and the CW pattern for 6 corrected. This uses the serial port to interface the PC to the Tx, earlier versions of SLOWCW may have used the parallel port. Also included in the .ZIP file is the circuit diagram, SLOWCWIF.PCX, for the hardware interface.  

xtaltx.zip - two useful circuits to provide a simple crystal controlled source at LF...... 

ssbgen.zip - Image Cancelling Upconverter: This is aimed at providing a simple SSB transmit driver that does not have to rely on an existing transverter. Audio phase shifting is performed by a pair of quad op amps arranged as two sets each of three all pass filters. The basic design for this network appeared in QST a few years ago, and the circuit with some very minor differences also appears in the Phillips RF Communications IC data book. Coupled with a pair of NE612 mixers and 74F74 quadrature signal generator the complete system achieved a WORST CASE sideband rejection of 26dB over the range 300 - 5000 Hz, voice bandwidth plus more, and in the 500 - 800 Hz region over 30dB was possible. Roll of in performance was quire fast below 300 Hz and much gentler above the upper figure. This was achieved over the entire RF range 0 to 30 MHz. Carrier rejection below 3.5MHz output was at least -40dB and worsed to -20dB at 30 MHz output. As this system is intended for upconverting signals generated by Codecs and A/D converters up to LF (and HF), and not voice, no additional audio filtering was included. The mixers used alone with SIN and COS signals generated by DSP at zero IF, (meaning the carrier leakage is at the centre frequency with the image sideband appearing as interference on top of the wanted signal) look to be capable of 40dB image rejection. This sort of performance should be more than adequate for amateur data comms and avoids having to use an HF transceiver plus Tx mixer to generate linear modes at LF.  

With a simple 2 or three section band pass filter to get rid of RF LO harmonics, there is enough signal to drive an LF audio type amplifier directly. For HF, broadband power amplifiers are available (Cirkit) for a few tens of watts. 

jntfile2.zip containing  

MAGLOOP.WK1 and MAGLOOP.XLS Spreadsheets for magnetic loop design. Both the same, for Lotus 123 or Excel  

73NOISE2.BMP Two day plot of 73kHz noise levels, 09/1997  

LFXTAL.EXE and .BAS Simple prog to give the division ratios required for output in the 73kHz band from any arbitrary crystal frequency. Assumes divide by 2 at the output to give a 1:1 duty cycle waveform.  

73DRIVER.BPM & 73SYNTH.TXT Circuit diagram and description of a 73kHz frequency synthesiser (G3WKL: users pse chack that my amendment to VDD and VSS is correct on the MC145151 - this is an amendment to the original cct diagram supplied by G4JNT) 

lfwebpg.zip containing  

TEEANT1.PCX Picture of the 'JNT Tee antenna  

73NOISE.BMP Seven days of noise measurement results in a 3 Hz bandwidth by G3PLX, made in July 1997  

73NOISEA.PCX Two days of noise measurement made by G4JNT in the same period  

73RX2.PCX Circuit diagram of portable 73kHz receiver. Very non optimum gain distribution so don't copy too closely !  

73DRIVER.PCX This has now been removed from this archive file, as it is on jntfile2.zip. Reference is retained for any who downloaded the file earlier. WARNING: Pins 2 and 3 of the MC145151 were incorrect on this circuit - please check with data sheet before powering up - pin 3 should go to the +5volt supply and pin 2 to ground. The correct version, 73DRIVER.bpm can be downloaded within the jntfile2.zip 

73RXLOOP.PCX Diagram of typical receiving loop for 73kHz  

32BPF.PCX Circuit of approx 1 Hz bandwidth crystal filter using watch crystals at 32765.5Hz 

adcsw.zip containing: 

DISPLAY SOFTWARE FOR PIC BASED DSP INTERFACE  

(PIC code references - SERADxxx and DSPIFxxx)  

The latest display software supplied includes options at start up for sampling rates of 2500 and 1000 Hz in addition to 10000/7142 Hz as originally shown in the RadCom article. PIC software has been updated to incorporate these additional sampling rates, but PCB modification (included) will be needed because of the need to then monitor two rate select inputs. The associated PIC software has the title SERADC06. A few alternative versions are floating around titled SERAD06x which are customised code designed to be programmed on top of earlier software versions in one-time-programmable devices. This is possible by overwriting the old code with NOP (no operation) codes which are 000 and appending the new code to the end. SERADC04 and SERADC05 can be updated on request, but see the notes at the end of this file.  

To enable the display software to find the COM port to which the interface is attached, the environment variable DSPCOM needs to be set. Before running the software type SET DSPCOM=2 (replacing 2 by whichever COM port is chosen). It will probably be convenient to incorporate this command into the AUTOEXEC.BAT file so the variable is set automatically at start up.  

On the waterfall plots, on screen help isn't very helpful! The various plotting functions may be changed whilst plotting, but the functions are slightly different for the 16 and 256 colour plots.  

WFALL16 C / [ctl]-C changes the colour scale T / [ctl]-T for the background threshold. These only affect the lines subsequently plotted  

the R key resets the display to top of screen.  

FFT size can be changed by the +/- keys  

For FFT sizes greater than 1024 the screen may be scrolled left to right by using the appropriate cursor keys. Or </>  

D / [ctl]-D controls the decimation of sampling rate, ie by averaging samples gives 10000/5000/3333 Hz etc  

F1 generates a .BMP file of the current display.  

WFALL256 +/- R D/[ctl]-D T/[ctl]-T F1 as above  

C/[ctl]-C changes the colour mapping, but now alters the whole display V / [ctl]-V changes the frequency averaging P changes the colour palette used for the waterfall display N resets the colour palette  

CAPTURE  

A capture to file utility that can be started instantly. The data is stored immediately, and a filename is requested when capture has finished. A unique default filename can be assigned by just pressing [rtn]  

TONEMON  

Monitors the amplitude of a single tone and plots its variation over time. Use for monitoring CW signals whose precise frequency is known  

The following keys can be used whilst the software is running.  

T / [ctl] T - Tone Frequency B / [ctl] B - Bandwidth V / [ctl] V - Averaging number L - Plot Colour (use to set a marker event) C - Clear and start new display F1 - Generate a .BMP file [esc] - Exit  

LEVELMON  

Continuously monitors, plots and records the level of the entire input signal to the interface. Used for monitoring noise levels or wider band signals that fill the filter bandwidth.  

Interacting keys :  

[ESC] Exit F1 to make a .BMP file C to clear the display and start a new plot.  

DAT2WAV  

Generates a .WAV file which enables recorded data to be replayed using a Soundblaster card and appropriate software.  

SLOWCW  

Not really part of the DSP software, but included anyway. Allows the PC to act as a keyer for very slow CW. For dot periods more than a few seconds long, depending on the length of the text, the message is sent at 10WPM at the end of every dot period to comply with identification requirements.  

PIC SOFTWARE DSPIFxxx  

This version of the PIC software is compatible with the display software detailed above for sampling rates of 1000/7142/2500 Hz selectable via the links / DIP switches. The 1000 Hz option (both pins low) has been replaced by a Complex sampling routine. This samples a signal centred on 1kHz and band limited to the range 500 to 1500 Hz in I/Q format. The effect is to give output samples that treat 1000Hz as if it were a frequency of zero. 500 - 999 Hz appears as negative frequency, and 1001 to 1500 Hz as positive frequency. The sign can be inferred by the relative phase of the I/Q samples. To allow a simple interface to measure in I/Q format, some liberties have been taken with the maths required to perform the digital downconversion of the signal from 1000 to 0 Hz centre frequency. The effect of this is to require the audio to be band limited to the range 500 to 1500 Hz. Any frequency components outside this range will be alliased back into this passband. A CW filter of bandwidth less than 500Hz is sufficient, but note that the CW tone has to be centred on 1kHz, not the usual 600 - 800 Hz normally used for listening. It is up to your ingenuity how to arrange this.  

The output data format may be changed between the 3 Real modes and the one Complex mode at any time whilst running, by altering the link /DIP switch settings. For a fuller description of the reason for and value of I/Q sampling read the article by G3PLX in November 1997 Radio Communication.  

The I/Q data output format is not compatible with the real mode display software listed above which is just repeated 8 bit data samples at the requested sampling rate. The I/Q data consists of pairs of 10 bit 2's complement I and Q values; the data format is made up of four data bytes per 1kHz sample as follows:  

1) The 2 Most significant bits of the In-phase sample in bit positions 0 and 1, with the six highest order bits set to zero - IHI.  

2) The eight Least significant bits of the I sample - ILO.  

3) The 2 Most significant bits of the Quadrature sample in bit positions 0 and 1, with the six highest order bits set to one as a flag marker for I/Q labelling - QHI.  

3) The eight Least significant bits of the Q sample - QLO.  

Programmers intending to read this data format need to note the following:  

Initially and periodically test for data slippage (missed data bytes) and incorrect framing, by looking at the 6 MSBs of the high order bytes which are 000000 for I data and 111111 for Q data. The chances of these patterns occuring in both low order bytes at the same time is low enough to ignore - 1 in 4096 or on average one every 4 seconds, and that is assuming two bytes have been lost in sychronisation in the first place!  

Mask off these bits, and the remaining data can then be gathered into IHI / ILO , QHI / QLO pairs of 10 bit two's complement form. This needs sign extension to fit into the normal 16 bit integer format of most software.  

Display software to support this sampling mode is under development. One programme supplied is PHASEPLT which monitors the phase of a carrier on and X/Y plot. Software tuning to zero the exact carrier frequency is posible usingthe cursor keys whilst displaying.  

As an experiment, and to check the accuracy of your receiver calibration, tune in to BBC Radio 4 from Droitwichon 198kHz using a CF IF filter that attenuates all signals outside a 1kHz bandwidth by 40dB or more, then arrange for this to give a beat note of EXACTLY 1kHz. This is easiest if a tunable BFO is available on CW reception mode. When properly tuned, a stationary trace will appear on the X/Y plot. If tuning is in error, the trace will rotate at the difference frequency; even at a few Hz error this will form a continuous circle too fast to observe! The linear plots at the top of the screen show the I and Q data separately and for a 1kHz tone these will be straight lines, becoming sinusoidal shaped for off tune signals. The ability to tune exactly to frequency is necessary for this sampling method, although software correction can be made this is best used for fine tuning between the (often 10Hz) steps of typical synthesized receivers.  

When Droitwich can be observed as a stationary signal, is becomes possible to make out the phase modulation of the carrier used for remote control purposes. Compare with MSF on 60kHz which has no phase modulation imposed but the on/off pulsing gives a differently shaped plot. The French longwave station on 162kHz also has phase modulation imposed on it, but of a differant format to Droitwich.  

For the real ultimate stability is is necessary to derive ALL frequencies used in the receiver, as well as the PIC clock oscillator, from one master source which is ideally locked to a primary standard such as MSF. When this is achieved, some very interesting measurements can be made of ionospheric phase shifts etc.  

A version of the PIC code , DSPIF090 is available which can be programmed into PICs containing SERADC04/05/06 BUT NOT IF THIS HAS ALREADY BEEN DONE with code SERAD06x. In this even contact me for a customised version, or details of how to do this yourself if a programmer is available.  

Thanks go to Peter martinez, G3PLX, who provided the idea and method of I/Q sampling incorporated into the PIC software. The algorithm used can be seen by examining the source code DSPIF.ASM which includes comments sufficient to work out the technique used.  

Andy Talbot G4JNT 15/10/97  

Last Updated: 11 April 1999