2009 Project Description: Tri-Band Transceiver

This all began during a tour of eBay in December 2008 when I spotted a Crystal Filter used in the Heathkit HW/SB Series and my lucky bid was less than $10. I followed that up with a set of Carrier Oscillator and CW Crystals and that was about $13. Next were the Heterodyne Crystals. Why I bought them, I don't know? Then I remembered I had a PTO from a Ten Tec Triton Transceiver. A little time spent with some pencil and paper convinced me I could build a solid state tri-band transceiver. The design goal was to have a 5 Watt transceiver suitable for QRP or for driving an outboard Solid State Amp I have.

For the most part circuits are broad band using small ferrite FT-37-43's as matching devices and the selectivity is provided by several series of Band Pass Filters that were designed based on easily calculable formulas. The circuit boards were all made using my mini-milling machine that I purchased many years ago from Harbor Freight. Initially the boards were single sided copper board that I simply milled out small squares. Subsequently I actually laid out the circuit and only milled where required. Most of the RF boards were built using this latter method.

The radio was initially built on a bread board and tested for several weeks before any attempt was made to put it in a box. This process has some very beneficial results. It provides lots of room for experimentation and changes -- and there were a lot of performance improvement changes. The down side to this is that I got so focused on building block circuits that I gave no initial thought about how you put this all in a box. Well the box is large measuring 12X12X5 inches. No miniaturization here.

Let me not fool you, this design was based on a lot of some of my previous work and solid designs from others such as Ron Taylor, G4GXO, Ed Skelton, EI9GQ, Wes Haywayrd, W7ZOI, Zack Lau, W1VT and others. Not too much is original other than integrating all of the pieces into a working transceiver. I purposefully picked 40, 20 and 15M as the bands of choice as the heterodyne crystals abound for these bands. In actuality if I were to spend money having custom crystals manufactured I would have picked 17M versus the 15M Band.

The "heart" of the radio is the bi-lateral IF strip designed by G4GXO. To that I added some bi-directional amplifier stages and soon we had a working transceiver. One of my ideas that proved successful was to mix the Heterodyne Oscillator signal with the PTO in a subtractive mix which is then fed to the digital display that has offsets for USB and LSB. Thus the frequency displayed is the true transmit frequency. An SBL-1 is used as the mixing device for the Heterodyne Oscillator and the PTO. The output of the SBL-1 is fed to a series of Band Pass Filters that only recognize the difference frequency. One problem I encountered was the low signal level to the display on 15M. I solved that problem by using a small 10 DB gain utility amplifier between the Band Pass Filters and the Display.

Another key element is the use of relay switched band pass and low pass filters. One of my other design goals was to eliminate the use of multiple gang band switches. I was successful in using relay switching and in the case of the filters used in the display scheme diode steering was used. The Band Pass Filters were designed using equations developed by W7ZOI that are in the Appendix of the Solid State Design Manual. They worked perfectly. The Low Pass Filters in the output are a straight lift from the work of W3NQN. The miniature DIP relays were purchased from All Electronics and in quantity were about $1 each. (These are no longer available.)

To assure that I minimized signal leakage in other parts of the radio especially where I am mixing the Heterodyne Oscillator output with the PTO great thought was given to this issue. The first step was to sample the output from the Heterodyne oscillator and feed that to a separate amplifier. Two amplifiers follow the Heterodyne oscillator, one of which feeds the first mixer a TUF-1 and the second independent amp is used with the display. The use of the SBL-1 as a mixing device provides additional isolation of signals leaking through. So that has worked nicely.

A word about the first mixer stage. The mixing device is a TUF-1 from Mini-Circuits and that is followed by a Band Pass Filter centered on 8.645 MHz with a 500 KHz bandwidth. This Band Pass Filter was designed using the Hayward equations. Thus the first conversion for all bands is in the range of 8.395 MHz to 8.895 MHz. Following the Band Pass Filter is a bi-directional amplifier that uses a 3N211 dual gate mos fet. A pair of J310's in Cascode can be used as well. Based on work by W7ZOI, I biased the the Dual Gate MOSFet hot (30 Ohms in the Source to ground) and so a heat sink is required if you use the 3N211. The actual circuit used is a direct lift from Ron Taylor and is the same used in the IF strip for the amplifier.

The output of the first mixer is routed to the main board and there it passes through a second identical 8.395 to 8.895 MHz Band Pass Filter. Following the filter is another TUF-1 where the signal from the 5.0 to 5.5 MHz PTO is mixed to provide a constant output of 3.395 MHz which is the IF filter frequency. If you followed the mathematics, the PTO tunes backwards BUT is the same for all bands. Thus at 5.5 MHz it is tuning 7.0, 14.0 and 21.0 MHz and for 5.0 MHz it is 7.5, 14.5 and 21.5 MHz.

The "heart" of this radio is the bi-directional/bi-lateral Main Board. On this board the following circuit functions are included; 8.395-8.895 Band Pass Filter, Second Mixer (TUF-1), Relay switching scheme for apply AGC to the first IF Amplifier, (on transmit this is switched to a fixed gain), the Heathkit Model 404-238 Crystal Filter, Second IF Amplifier (operated at a fixed gain) and the Product Detector/Balanced Modulator. Later a small relay was added in the open space along the right hand edge. This was for tune up where a small value resistor is switched to ground thus unbalancing the balanced modulator providing a carrier signal for tune up purposes. Just below this relay pads are used to terminate the pick up coil for the signal that is fed to the AGC. Thus the AGC is now IF derived. Initially an audio derived AGC was used but the results were unsatisfactory. A small board just above the main board now contains the AGC circuitry.

The Product Detector/Balanced Modulator was built with discrete components and offers the advantage of low cost and the ability to adjust both the carrier and the phase balance. I have also used an SBL-1 as there are pins on the package where a small value pot can be introduced to provide a carrier balance as well switching in a small resistor for tune up. The discrete component version by far wins hands down. I used a small 15 Turn pot for the carrier balance control. Two version of the IF amplifiers were built one with 3N211 Dual Gate MOSFets and the second with J310's. The current "production version" uses the J310's. The following is a link to the G4GXO design of the Bi-lateral IF amp. His article appeared in the Autumn 2006 issue of the GQRP SPRAT, Issue Number 128. The input and output impedance of the Crystal Filter was 2K.

Initially I used a large 4PDT Relay to do the switching from receive to transmit, which included the switching of voltages to diode steer the Bi-lateral circuits and to provide voltages to those parts of the circuits as appropriate for receive only and for transmit only. Every time I switched the relay, the display would change and despite using diodes across the relay and an attempt at a soft start and soft off, the problem persisted. Then I remembered the sequenced switching circuit used by W1VT in his 7.0 MHz transciever that is desicribed in the ARRL Publication "QRP Power". I built that circuit "as was" and found that the current draw larger than the circuit could handle and so I upsized all of the sequenced voltages so that there is a darlington pair (2N3906 and TIP30) for the three sequenced voltages. In normal operation, voltage #1 is applied to all the receive cicrcuits. On transmit, voltage #1 goes to zero, voltage #2 comes on and that goes to the bi-lateral circuits that are used on transmit. Shortly after that voltage #3 comes on and that goes to the other transmit stages (mostly the RF stages, but also the microphone amplifier). Upon disengaging the Push To Talk cicruit the design reverses the order of the voltages being disengaged. Works pretty cool and no large spikes on the frequency display. Thanks Zack!