G6KSN describes a CB to 6m transverter

(from http://hamradio.lakki.iki.fi, with re-layout)

I offer the following information in 'the spirit of amateur radio'. I accept no responsibility whatsoever for any misfortune that may occur regardless of the use that the information is put to.

Firstly please note that the following is not intended to be a complete description down to the last nut and bolt but more as 'food for thought'.  It is open to individual adaptation to suit what is in your particular junkbox. Treat it as such and feel free to experiment. By doing so you will probably come up with something better (feedback please!!)

The six metre band really is a superb place for packet and FM voice operation. OK the band is a little noisy and suffers from anomalous propagation from time to time (don't they all) but I really don't understand why there is such a
complete lack of activity. Being a relatively low frequency it is quite easy to build equipment for and actually get it up and running without the need for thousands of pounds worth of test gear.

Packet radio channels on 2 metres in the Midlands are overflowing and yet 6 metres is totally devoid of activity in this area at least (and I suspect in most areas). There are lots of channels available for both packet, FM simplex and soon I believe FM voice repeaters. So why not take the plunge? Get on a new band and free up two metres from all that desense when the packet fires up!!

What I intend to describe to you is a cheap way on to the band but not one  that compromises on performance. My design brief is as follows..

(1) To produce a transverter that when used in conjunction with a cheap UK (27/81) FM CB set will generate 10 watts of spectrally clean RF power over a 400 KHz portion of the 6 metre amateur band.

(2) Receive performance must be better than 0.3 microvolts pd for 12dB SINAD. Overload performance and out of band rejection must be similar to that achieved by Japanese 'black boxes' (depends to some extent on the CB set
used).

(3) Alignment must be simple and must not require any expensive test equipment.

(4) The design must be cheap and tolerant of slight modification here and there.

There is a certain amount of conflict between design brief (4) and the other 3! Many of the components can be salvaged secondhand from ex PMR equipment, particularly the transistors in the PA strip. Probably the most expensive
component is the SBL1 double balanced mixer. These retail for about 8.00ukp but I have seen them for sale at rallies much cheaper than this. It was necessary to use an expensive mixer because twice 27 (the CB's operating frequency) is 54 and twice 23 (the local conversion oscillator) is 46. Both a little too close to 50-52 MHz for comfort!

TR1 and it's associated components form the conversion oscillator. This runs at a nominal 23 MHz. (23 MHz crystals can be found in PYE PF2 handhelds). This gives coverage from 51.60 to 51.99 MHz with a standard CB set (ideal for packet operation). The actual choice of crystal frequency is up to the individual and if you don't have one to hand I suppose you could always order one!

If you do intend to order a crystal then you should bear in mind the following formula.

6 Metre operating freq = (Xtal freq) + (CB freq)

The output from TR1 is buffered by TR2 and splits 2 ways. A feed to the receive converter is taken straight from the emitter of TR2 (via a short length of 50 ohm coaxial cable) and the transmit mixer is fed via T1 which is tuned to the crystal frequency. This is done to ensure spectral purity (remember 46 MHz hi hi!) and to provide a low impedance feed to the mixer (nominal impedance 50 ohms). IC1, the mixer (SBL1) requires around 5 milliwatts of drive (+7dBm).



The 27 MHz signal is fed first to a dummy load consisting of R1 to R4 in  parallel and then to IC1. With the values shown, 4 watts of drive will give the required 5 milliwatts of drive to IC1. If your set puts out more power then increase the value of R5 (we don't want to fry IC1 do we?).

Oh yes a word of caution about IC1, it has some very fine coils of wire between pins 1 and 2, 7 and 8. Even relatively small amounts of current can destroy them, not to mention 30 amps from the shacks psu as I found out to my cost.
Be very careful not to mistreat this device and certainly don't attempt to construct the transverter with power applied as I was doing when the unit was under development!

The output from IC1 contains sum and difference components ie 50 MHz and 4 MHz. I suppose you could always transvert to 80 metres but I doubt you would be too popular using narrow band FM (now there's a thought hi).

The wanted 50 MHz output is extracted by T2 and amplified by TR3. Further filtering/matching is effected by T3 and TR4 brings the RF level up to around 3-400 milliwatts. TR5 and TR6 are the ones that consume all the current and should be mounted on heatsinks. A clip on TO5 heatsink will suffice for TR5 but TR6 needs to see a substantial piece of aluminium. I built this bit of the circuit on a heatsink as used in the PA stage of a Pye Westminster. TR5/6 and associated components can be separated from the earlier circuitry by introducing some 50 ohm coax between C18 and R20.   This is how mine is built but it is up to you. Remember I am not telling you how to build the transverter by giving a blow by blow account. The design is quit flexible and open to your interpretation. It really does depend on what you have in the junkbox or can obtain cheaply.

The receive converter is designed around a couple of dual gate mosfets, this was done because a certain electronics company was selling them very cheaply at one time and I bought a load of them (wish I had bought more now because I
have run out). I haven't tried other types but I would imagine that 40673's or 3N201's etc would work just as well. The noise level is pretty high on six so we are not all that concerned with extracting the last half a dB out of the receiver. In fact when I plug the aerial in at this QTH I have around S3 of  noise (with the computers switched off) confirming that any improvement in sensitivity would be a complete waste of time.

TR8 is configured as a standard tetrode amplifier in common source. P1 allows the resting bias on G2 to be adjusted for best noise performance. L12 acts as a resonant auto transformer coupling RF into G1 at the hot end. L13 is bought
to resonance by VC11 and is coupled to L14. The amount of coupling between  these two circuits is not all that critical and with the centres of the two coils spaced around 12 millimetres seems to give good performance. L14 is
bought to resonance by VC12 and feeds the amplified 50 MHz signal to G1 of TR9, the mixer. 23 MHz from the injection oscillator is fed to G2 of TR9 and all manner of mixing products appear at the drain of the device. The wanted 27 MHz signal is extracted By T4 and C35. A low impedance feed from the secondary is coupled to P2, an attenuator. This is set to give good sensitivity withoutoverloading the CB set. It can be set to give 'sensible' S meter readings.

One final piece of circuitry is required to make the whole thing operate i.e transmit/receive switching. I have used relays to perform this function because very good ones can be obtained very cheaply from 'certain surplus outlets'. A voltage doubler is fed with a portion of the 27MHz carrier which then biases TR7 on, causing the relays to change state on transmit. The time for the relays to pull in does not seem to be a great problem for 1200 baud  packet. If you want something a little faster then you may wish to consider diode switching although this may compromise the dynamic range of the receiver somewhat.



A double tuned bandpass filter (L10, VC8, L11, VC9 and C28) tuned to 50 MHz is included in both the receive and transmit paths to help minimise spurii on transmit and to help prevent breakthrough from PMR and broadcast stations on
receive.

You may very well have your own ideas about how to construct the transverter.  My particular version was mainly built 'rats nest' style on the copper side of PCB with a track cut along one edge to carry 12 volts to all sections of  the transverter. If you wish you could build the circuits up on 'proper' PCB's but may I suggest that the earthplane idea is retained to help maintain stability throughout the unit.

A diode probe (I won't insult your intelligence by including details) or a wavemeter would be most useful for aligning the various tuned circuits. A receiver or scanner that can cover six would also be very useful. Lets start with the oscillator, connect your diode probe to the secondary side of T1 and tune the slug for maximum indication. Hopefully this will be 23 MHz and not 46! You can confirm this with a wavemeter but if you don't have one, a hf receiver tuned to 23 MHz will suffice. A strong carrier should be received when pins 3 and 4 of IC1 are touched with a finger! Tuning the slug either
way should result in the received signal strength weakening. Incidentally if you require the crystal to be bang on frequency you can substitute C1 for a 60pF trimmer and adjust using a known good frequency counter or synthesised HF set.

Connect the CB set and key up on channel 20. Ensure that the relays change over. If they do not then you must find out why! Assuming they do then connect the diode probe to the tap on T2 and tune for maximum. Transfer the probe to the secondary of T3 and adjust for max. If you don't yet have a dummy load and power meter attached to the output then now would be a good time!

Set VC8 and VC9 to their mid positions. If power is evident at this point then adjust VC8 and VC9 for maximum power as indicated on the meter.

Connect the diode probe to R20 (hot end hi) and adjust VC1 for maximum indication. Connect the probe to the base of TR5 and explore settings of VC2 and VC3. It is useful to have 2 trimming tools for this purpose, you must tune
for an absolute maximum. Resetting one capacitor will influence the setting of the other to bring the circuit to resonance.

Move the probe to the base of TR6 and adjust VC4 and VC5 in the same fashion.

Adjust VC6 and VC7 for maximum indicated power, readjust VC8 and VC9 for maximum indicated power. go back to VC4, VC5. Repeat until no improvement can be obtained. Power output should be 10 watts with the specified transistor. If you are a novice and require less power in order to comply with your license conditions then replace L7 for a 5 watt wirewound resistor.  You will have to experiment with the value until the power drops to 3 watts  (try around 22 ohms).

Now to the receiver. Do not readjust VC8 or VC9!!

set P1 midway, P2 to maximum, VC10, 11, 12 to mid position. You will need an off air signal or a generator. You might consider using the local oscillator of a scanner or the like, this makes a good sig gen hi.

Assuming you can hear something, adjust VC10, VC12, VC11 in that order for best signal, reduce the signal level until noisy and repeat. Now adjust P1 for best quieting (not maximum indicated signal). Finally reduce P2 to the point where
the signal to noise ratio just starts to degrade and then turn back up a fraction.

The design was not really intended for this purpose but a little experimentation may very well prove that it is possible. Firstly the time constant of the TX/RX switching circuitry will need to be lengthened by a substantial amount. Secondly TR5 and TR6 are biased in class C. This will need to be changed to class A in order to obtain some sort of linearity.
Further modifications to the circuitry around TR3 and TR4 may well be necessary. The receive section will work well on SSB without any modification.

If you intend to use your transverter to talk to a friend who is also using a CB set then you will have no problems. If on the other hand you want to talk to other stations it might be worthwhile changing the ceramic filter in the set for a CFW455F, the one fitted in CB sets is for 10 KHz channelling and 'proper' ham equipment will sound distorted. You could also turn up the sets deviation (if this hasn't already been done hi hi). Around 4 KHz is acceptable with 20 KHz channelling (which 6 metres is). Some sets can actually be modified so that the synthesiser steps in 20's rather than 10's. Examples of which are early versions of the Binatone Route 66, The Lake Manxman, Cheiza etc. Sets with the LC7137 will not unfortunately (most seem to be fitted with this one). Neither will the 9119 as used in Uniden equipment. One set that I did cone across that would was the 'Lancaster'. I have only ever seen one of these and would like to come across some more (any ideas?)

R1.....240 ohm 1 watt.
R2.....240 ohm 1 watt.
R3.....240 ohm 1 watt.
R4.....240 ohm 1 watt.
R5.....1K.
R6.....56 ohm.
R7.....47K.
R8.....22K.
R9.....330 ohm.
R10....220 ohm.
R11....560 ohm.
R12....75 ohm.
R13....100K.
R14....270 ohm.
R15....100 ohm.
R16....470 ohm.
R17....24K.
R18....15 ohm.
R19....15 ohm.
R20....75 ohm.
R21....15 ohm.
R22....15 ohm.
R23....1K.
R24....10 ohm.
R25....1K.
R26....220K.
R27....22K.
R28....1K.
R29....8.2K.
R30....2.2K.
R31....1K.
R32....330 ohm.
R33....47K.
R34....330 ohm.
R35....1K.

P1.....100K.
P2.....100 ohm.

C1.....33pF.
C2.....not used
C3.....56pF.
C4.....56pF.
C5.....10nF.
C6.....22nF.
C7.....100pF.
C8.....68pF.
C9.....22nF.
C10....22nF.
C11....10pF.
C12....10nF.
C13....6.8pF.
C14....22nF.
C15....22nF.
C16....22nF.
C17....22nF.
C18....22nF.
C19....47pF. S/M
C20....100nF.
C21....10nF.
C22....22nF.
C23....100nF.
C24....10nF.
C25....15pF.
C26....100nF.
C27....10nF.
C28....4.7pF S/M
C29....1.5nF.
C30....1.5nF.
C31....1.5nF.
C32....10nF.
C33....1.5nF.
C34....47nF.
C35....22pF.
C36....10nF.
C37....47nF.
C38....100pF.

VC1....60pF.
VC2....60pF.
VC3....60pF.
VC4....60pF.
VC5....60PF.
VC6....60pF. HV.
VC7....60pF. HV.
VC8....60pF. HV.
VC9....60pF. HV.
VC10...60pF.
VC11...60pF.
VC12...60pF.

T1.....KANK3335.
T2.....KANK3335.
T3.....KANK3335.
T4.....KANK3335.
L1.....8 1/2 turns 18 SWG 8mm o/d 18mm long tapped at 1 1/2 and 3 turns.
L2.....8 1/2 turns 18 SWG 8mm o/d 14mm long.
L3.....2 1/2 turns on VHF ferrite bead.
L4.....6 1/2 turns 18 SWG 8mm o/d 12mm long.
L5.....8 1/2 turns 18 SWG 8mm o/d 14mm long.
L6.....2 1/2 turns on VHF ferrite bead.
L7.....8 turns 22 SWG 5mm o/d 20mm long.
L8.....6 1/2 turns 18 SWG 8mm o/d 12mm long.
L9.....7 1/2 turns 18 SWG 8mm o/d 12mm long.
L10....8 1/2 turns 16 SWG 8mm o/d 20mm long tapped at 2 1/2 turns.
L11....8 1/2 turns 16 SWG 8mm o/d 20mm long tapped at 2 1/2 turns.
L12....8 1/2 turns 18 SWG 8mm o/d 20mm long tapped at 2 turns.
L13....8 1/2 turns 18 SWG 8mm o/d 20mm long.
L14....8 1/2 turns 18 SWG 8mm o/d 20mm long.

Note.. L13 and 14 are mutually coupled and are separated by 12 1/2mm between
centres.

RLA1...Magnetic devices RF changeover relay available from SENDZ components.
See the back pages of Television Magazine (as used in the pye Westminster).

TR1....BC182.
TR2....BC182.
TR3....BSX20.
TR4....2N3866.
TR5....2N4427.
TR6....PT4166. (10W, 100 MHZ, 15V, 10dB gain).
TR7....BC182.
TR8....3SK88.
TR9....3SK88.
D1.....1N4148.
D2.....1N4148.
D3.....1N4148.
LED1...Red 0.2" standard LED.
IC1....SBL1.

X1.....23 MHz. (gives 51.6-51.99 on a standard CB 27/81 set, you choose!).

Case to suit.
Power connector.
2 SO239 Sockets.
led bezel.
Nuts, bolts etc.
Heatsink for PA.
Clip on heatsink for driver.
Fibre glass or SRBP PCB.
About 1 metre of 50 ohm miniature coaxial cable for board interconnection.

I hope the foregoing will inspire one or two of you to have a go at building something and hopefully populating a presently very underused allocation. If you come up with any improvements to the design (I did think about using the SBL1 on receive as well) then please let me know.