|VK5JST HOMEPAGE- PART 10|
|VK5JST AERIAL ANALYSER|
This project was published in the May 2005 issue of the Australian magazine "Amateur Radio", and has been designed using parts which are very readily available (see the circuit and parts list). There is little point in supplying a kit of parts to amateurs who have been around a while - most will be in your junk box. However, the printed circuit board (double sided and hot air solder reflow coated) is another matter, and is available at cost to hobbyists only, here.
BEFORE YOU SEND ME YOUR MONEY, check that boards are in stock by emailing me (Jim) at the following address, then click on the Paypal button.
PLEASE NOTE THAT ORDINARY AIRMAIL IS USED FOR SHIPMENT. THIS MEANS THAT ALL FORWARDING OCCURS AT YOUR RISK UNLESS YOU SPECIFICALLY REQUEST (AND PAY EXTRA FOR) FORWARDING BY REGISTERED MAIL.
MK2 PRINTED CIRCUIT BOARD.........$A12.80 plus post and packing
However, if you don't have a comprehensive junk box, then from December 7th, 2011, you can obtain an excellent and complete kit of parts (including the printed circuit board and pre-programmed Picaxe processor) from the Adelaide Hills Amateur Radio Society (AHARS) . This kit is brilliant value, and actually sells for less than the price you would pay for the individual parts. All profits go to the club too, and this link takes you there........ AHARS Kit...... Please also note that kits are no longer available from the South Coast Amateur Radio Club, which sold its last kit on November 29th, 2011.
Download the files for the "Aerial Analyser" here by right clicking on the appropriate link and using the "SAVE TARGET AS" Windows feature.
The code for the PICAXE 28X was prepared using the Revolution Education Programme Editor and is in BASIC format. This will program any PICAXE chip. Use the following link to get it........ Revolution Education
Note that since submission of the article for publication, a battery voltage monitoring modification has been included at the suggestion of Barry Williams VK5ZBQ. To add this feature to MK1 printed circuit boards, cut the PCB connection between AN3 (microprocessor pin 5) and ground. Then add a 16K resistor between the +12volt rail and AN3, and 3K9 between AN3 and ground. At switch on the battery voltage is displayed for 1.5 seconds.
To obtain the correct voltage indication, simply vary the division ratio in the line "w0=w0/100" in the "battery:" subroutine of the microprocessor code. The divider should lie in the range 94-106, and compensates for the 5% component tolerances. This modification ( and other minor layout and circuit mods ) have been included on the Mk2 version of Analyser detailed below
NEW PICAXE 28X1 CODE
As at the date of writing (June 23rd, 2007), Revolution Education has replaced the older Picaxe 28X chip (based on the PIC16F873) with the newer Picaxe 28X1 (based on the PIC16F886) and supplies of the 28X are drying up fast. The new chip is, as claimed by Revolution Education, pin for pin compatible with the old one, but is not directly software compatible because the processor is run at 16MHz in the Analyser, rather than the default 4MHz. To use the code above with the new chip, simply add the instruction SETFREQ EM16 to the code as the very first line. This tells the new Picaxe 28X1 to run at 16MHz and no other code alteration is required. Note that chips with firmware of version A.2 (version 2) or later must be used. Earlier versions of the chip firmware contained serious errors and the display routines simply will not work. It is also worth noting that the new chip is very much more powerful, and that enthusiasts can take advantage of this to extract greater accuracy from the Analyser. I used the Programming Editor 5.1.3 from the Revolution Education website to load the chips. Download the new code below.
February 3rd 2010. At long last someone has taken advantage of the extra power locked up in the 28X1 processor! The code below, prepared by Stan VA3SMM and friends, includes component mode by David Milne VK3DPM and eliminates display flicker. But you will have to do some simple programming to enter your own name and callsign. It is for the Picaxe 28X1 only, as the code is slightly too large to jam into the 28X chip.
NEW PICAXE 28X2 CODE
November 4th 2010 ***** NOW IN GLORIOUS TECHNICOLOUR***** No, only kidding. Stan Madore VA3SMM (with able assistance from Anthony Howes VK2AJH and others) has even further upgraded his code "with the lot". The code is now in the form of four files which have to be loaded into the "slots"in the powerful Picaxe 28X2 (built around the Microchip PIC18F25K22 chip). The only differences from a practical standpoint are that the 28X2 processor must be fitted with a new crystal, and that each time the function switch is toggled, the analyser switches to a new function (4 digit frequency display-fast gate, 5 digit frequency display- slow gate, component mode, angle mode). See Stans "Aerial Analyser Code Manual" as to how to install the code etc and... Get them here.:-)
TRY WRITING YOUR OWN CODE FOR THE AERIAL ANALYSER!! ADD YOUR OWN DESIRABLE FEATURES, AND HAVE THE CODE DISPLAYED HERE TOGETHER WITH YOUR NAME AND CALLSIGN.
e-mail me at
WARTS, PROBLEMS AND COMMENTS
***********December 22nd 2009- MINOR PCB ERROR- After producing thousands of pcbs, the artwork at the board supplier simply wore out and was replaced. Production of the new artwork involves converting a .pcb file to Gerber form. During this computer conversion, a couple of adjacent pads became minutely disconnected even though they are joined in the master pcb file. Why? Who knows..... . If you have one of these pcbs please use a solder bridge to connect the two pads. The artworks at the supplier have now been corrected. See diagram below for where to look.
***********May 20th 2009- A lot of grief is being caused by PN2222 transistors which although supplied by Fairchild, have a mirror image pinout to that shown on the component overlay (also for a Fairchild PN2222!). The central pin is always the base but the collector and emitter pins are reversed. Check your transistors with the current gain checking feature found on most DVMs. When placed into the test socket correctly, the transistor will typically exhibit a current gain in excess of 50. If the transistor collector and emitter terminals are reversed, a very low current gain will be seen.
On June 23rd 2007 when I reviewed this section, many thousands of units have been built, and apart from the odd dry joint, only 3 units have given serious trouble. For two of these units, the trouble was identical and involved the upper frequency range of the oscillator. Instead of producing a good sinusoidal output, the generator produces a near sawtooth waveform of about half normal amplitude.This is caused by the 1uH RF choke in the oscillator tank having much lower losses than normal,and consequently causing much higher than normal gain around the oscillator loop. The oscillator consequently "squeggs" producing a weird output, and to overcome the problem, simply reduce the tank Q by paralleling the 1uH RFC with a damping resistor. Both of the troublesome units produced good flat sine outputs when the 1uH choke was paralleled with a 3k3 resistor, but depending on your choke characteristics, it may be necessary to go as low as 1k5. Keep an eye on the waveform purity when you are experimenting.
The remaining unit gave trouble simply because it would not oscillate on the two highest frequency ranges, and had odd looking waveforms on the remaining ranges (flat topped sines etc). This was traced to the use of ordinary standard 100nF disc ceramic bypass capacitors, which being physically large single plate devices, have relatively high losses and low self resonant frequencies. There is no place in a wide band power circuit (which this instrument is) for such components. Use only miniature multilayer MONOLITHIC bypass (and coupling) capacitors in this circuit with lead lengths of as near to zero as possible, or you WILL end in trouble. The bypass capacitor at T3 base is particularly important.
Minimizing stray capacitance in the test circuit is important, as it appears in parallel with the load and can upset measurement accuracy at high frequencies and high SWRs. It can be minimised by mounting the following components away from the ground plane by 1-2mm. (a) D5 and the associated 100pF and 47K (b) D6 and the associated 100pF and 47K. (c) The 2 X 100 ohm test cct series resistors.
ADDING LOW FREQUENCY RANGES
If you wish to extend the frequency range of the analyser down to around 200KHz, modifications must be done to both the oscillator tank circuit and the time constants in the envelope detectors (to avoid AC being passed to the microprocessor inputs- with the result that the display readings vary randomly). Here are the mods. Add 330uH, 1mH, and 3.3mH to the existing oscillator tank circuit to get continuous coverage down to 200KHz. Change the filter capacitor values in each envelope detector low pass filter (now 100pF, 47K, and 1nF) from 100pF to 470pF and from 1nF to 0.01uF. Do all setting up at 2MHz.
Interested in adding a serial port to the analyser for data logging? Jeff Garrett ZL1BIV has done just this, and has written some clever code which allows the results of frequency sweeps on a load to be displayed in Excel. For more details, contact him at the following email address