The G4OO Homebrew Transmitter and Receiver.

(Updated March 17th 2002)

This project was born when, following the death of Dennis, G4OO in 2000, it was suggested that some of his gear should remain in use in the Spalding and District Amateur Radio Society clubroom, as a memorial to Dennis, who was one of the founders of the Society.

Also, the Society members hold a Sunday morning topband AM net, as no AM facility is available at the clubroom, and also as Dennis was a regular on the net, we thought that something to use on this mode would be appropriate.

Amongst Dennis' gear was this receiver and transmitter.

 

The receiver was a strange beast indeed, built on what appears to be a commercially manufactured chassis, it uses Command receiver components, including the tuning dial and gang, front end coils and IF transformers.  However, it uses miniature valves of the 6BA6, 6AM6 etc. variety, with an ECH81 mixer and an ECL80 audio valve.  It had been much modified over the years, but what remains of the original wiring is very neat indeed, I think that it was either commercially made, or a very well assembled kit, but for what purpose?  It tunes 3 - 6 Mc/s, as on the original command receiver.  The clue was that, on the right hand side there is an empty space, with two holes in the front panel to take Belling sockets, this suggests that converters were fitted, maybe for two and four metres, so it was designed as a tuneable IF unit.

Following appeals for information on the receiver, Robin Holderness G3XDA wrote:

I recognise that rx as one which came from the late Brian Loveday, Weldon (near) Corby. He died approx 3 or 4 years ago and I helped dispose of most of his vast collection of bits & pieces. He thoroughly enjoyed experimenting/modifying/constructing, with interests from vlf up to microwaves, and was a quiet but friendly person who having retired spent a lot of time enjoying his hobby.

I am sure it was originally built by him. It ended up in my possession and then I took it to a Spalding junk auction where it was knocked down to Dennis for (I think) £4!

It is quite likely that with his vhf/uhf interests that it was a tunable IF.

It has subsequently transpired that the receiver was in fact built by Doug Hill G8BEH.

The transmitter consists of two chassis, a power supply and the TX unit itself, very obviously homebrew.


Underside view of the PSU, as found.
The TX consisted of a Geloso VFO unit covering 80 - 10 metres driving a pair of 807's, the dial of the VFO being marked in Italian, the exact same unit was used in the KW Vanguard transmitter, but with English markings.  The power supply consisted of two transformers, one a 400-0-400 volt HT unit, with a 5R4 rectifier, the other a multi-voltage unit with several 6.3 volt windings and a higher voltage winding used with a selenium rectifier to supply bias and relay voltages. There was  also two B9A sockets wired for ECC83 type valves, I guess used as a mic amplifier for some sort of low level modulation system.  As Dennis was a CW man, this would have been the primary mode of the TX, the AM being a second thought.
The TX had a crude screening cage made from the perforated zinc that was once used for "meat safes".  It had been stored in a damp place for many years and was in a very sorry state indeed.

Restoration.

Where to start?  Well, the theory  "if you can't hear 'em, you can't work 'em" is a good one, so I started with the receiver.  There was a great big space which had obviously once been occupied by a mains transformer, so I had a hunt round and found a Parmeko transformer with 250 - 0 - 250 and several 6.3 volt windings that would fit in the space, and, as there was a B7G size hole nearby, with a 6X4 rectifier a simple PSU was connected up.  Soon, the heaters were all glowing nicely, and as the HT rose, noises, horrible noises, emerged from the on - board speaker.  Not for long though, a bang,  a lot of smoke, and an obnoxious smell filled the shack.  Why didn't I reform the electrolytics?  Well, I did actually, the problem was that the RX contained an inordinate number of 0.01µF red "eurocapacitors".  Needless to say these were almost without exception leaky, and one of these was the source of the explosion.  Most had to be replaced, and with this done, more promising noises emerged.  In fact,  tuning to 80 metres showed that the receiver was quite lively.


Underside of the Rx chassis, after some restoration.
The topband converter and muting relay are on the left hand chassis.

The receiver includes a product detector using an ECC81 (12AT7) and a separate BFO, and also a separate AVC amplifier, with two 0A81 type germanium diodes for AM  and AVC detection.  the latter has a switchable time constant, this combination provides absolutely superb SSB reception.  A Q - Multiplier provides variable selectivity in the IF, which incidentally is the standard Command 1,415 kc/s, and with careful adjustment can provide selectivity as good as modern day transceivers.  A few more mods and tweaks were required, these produced a very creditable receiver, with very little drift and very pleasant audio quality and which is very nice to use on the 80m band.

Topband.

The primary purpose of this project is to provide a 160m station; as found this Rx covers 3 - 6Mc/s, so what would be the best way around this?  I decided against modifying the front end coils as they were so nicely made it seemed a shame to butcher them, also 80m coverage was already there and would be a bonus.
The simplest way would be to convert topband signals to a frequency already covered by the receiver, and as it had originally been designed as a tuneable IF unit this seemed the best way forward.  A quick root through my stock of crystals produced one on 6,800kc/s, some simple maths revealed that using this, 1.8 - 2 Mc/s would appear at 5 - 4.8 Mc/s (reversed tuning), and as this area is free of broadcast stations, and is relatively quiet, this was the way forward.

A simple converter using a 6BE6 (EK90) and the 6,800 kc/s crystal was built up on the blank chassis, using the circuit shown here.  The bandpass input circuit was found necessary to prevent breakthrough of broadcast stations from the 25m broadcast band (6,800 + 4,800 = 11,600).  This simple circuit is not excessively noisy, and provides enough sensitivity without an RF stage.  A simple switching arrangement was added to allow the converter to be switched in or out as needed.  An added bonus is that the local MW broadcast stations can be tuned in as well, by tuning further up the IF band.

Addition of a muting relay, and a few other minor mods were completed by Christmas 2000 and the receiver was used as the main shack receiver for topband and 80 for several months for testing.  One further fault developed, AM reception became distorted on strong signals, this was traced to an open circuit 0A81 diode, the AVC detector. 

The Transmitter.

Turning back to the transmitter, closer examination showed that a 'restoration' would not be feasible for several reasons, so I began to think along the lines of a 'rebuild'. 

As mentioned above, the transmitter had been stored in a damp environment for some time, the zinc screening and aluminium chassis were both very corroded, as was the tuning gang of the Geloso unit.  The dial cover and pointer for this were missing as well.  The electrolytic capacitors were of course beyond redemption, and much of the wiring was in screened rubber insulated cable, the insulation crumbled away when touched.  Also as it was the unit didn't cover topband!

So the decision was made to strip the unit completely and rebuild it as a topband transmitter using as many of the original components as possible, together with some others from Dennis' collection.

The transmitter chassis, stripped of all components and cleaned ready for reassembly. Also visible are the diecast box for the rebuilt VFO, the meter, mod transformer and PA tuning capacitor.

The TX and PSU chassis were stripped of all components and cleaned.  The zinc screening was too brittle to reuse and was discarded, but most other components (except the electrolytic capacitors) were cleaned and tested, and found to be serviceable.

Now, Dennis by his own admission was not a craftsman in the art of either metalwork or soldering, as long as it worked well it was OK. Also, he was never satisfied with a project, always modifying, improving or pulling down and starting again, and, judging by the number of holes in the chassis, this rig had been no exception.  So I saw no reason why I shouldn't carry on the tradition!

Beginning with the power supply, the transformers both tested OK, so I refitted them, together with the large choke and the two paper 8µF 1000 volt capacitors.  This was all rewired, together with a 10Kohm dropper and a 0A2  for the 150 volt VFO supply, and a simple 12 volt supply for the relays.  All outputs were made available at a 12 way Jones socket on the rear panel.

The microphone amplifier was also rewired, using new components, its output was also routed via the Jones socket.

Transmitter chassis
I had decided not to use the Geloso unit, both because of its condition, and because  as it is it doesn't cover topband, so instead I built up a VFO unit into a diecast box that was amongst Dennis' components.  It too had been used for several previous projects! I used a 6J5 triode in a standard Vackar circuit, supplied from the 150 volt stabilised supply and tuned with a 65pF Jackson variable capacitor from the same source. It tunes 1.75 - 2.0 Mc/s, with a view to using it on Topband and 80m (by doubling).  An aluminium plate was fixed over the large hole in the TX chassis where the Geloso unit had been fitted, and the diecast box fitted on top of this.
The rebuilt power supply. The orange object is a 10Kohm resistor of some colossal wattage.
Modulation
At an early stage I had decided to retain the CW facilities, but as the intended use will mainly be AM, I planned to include a proper anode and screen modulation system.  There were the remains of part of a Pye F27 base station amongst the junk, this provided a nice compact 25 watt modulation transformer and the 6V6 valves to drive it. These components were fitted to the TX chassis, the mic amp on the PSU was modified to provide a push pull output to drive them. This tested OK into an audio dummy load, so on now to the RF section.

Transmitter chassis partly rebuilt, visible are the new VFO unit, modulation transformer and 807 PA valve in its screened holder. The 6V6 modulator valves are on the rear left of the chassis.

Inside the VFO unit. The relay is to control the standby/
netting facility.

 

RF section

With the VFO working, the next step was to get its output amplified enough to drive the 807.  I tried several different valves in the driver stage, finally settling on a 6SG7.  There were several reasons for this choice;
(1) I had one! (2) Being a "metal" valve, no additional screening would be needed, (3) It has a relatively high µ
so adequate gain was available to drive the 807 into class C without needing additional bias.

The two tuned circuits in the driver anode - one for topband and one for 80m.

I wound two coils for the anode circuit, one for topband and one for 80, the driver acting as a doubler for this band, an all-insulated toggle switch being used here to select the band, and another Jackson variable capacitor as a peaking control. This produced about 3mA of drive on 160 and 2mA on 80, just about enough.

On to the PA anode, a tapped coil was rewound onto the original ceramic former, rewinding being necessary as the transmitter didn't originally cover 160m. I also used a new RF choke and a new 2kV 1000pF coupling capacitor, not wanting to take chances with this component:- a breakdown here can result in the full HT voltage being applied to the aerial!  With the original tuning and loading capacitors I got 15 watts of RF into a dummy load, but the only about 30% modulation was obtainable before severe distortion set in.  Scope checks revealed that this was because the 6V6 output valves simply weren't producing enough audio power.  I experimented with various levels of fixed and auto bias, but no combination produced any great improvement.  I came to the conclusion that there was a mismatch somewhere around the modulation transformer, though it was originally used with 6V6's.  Thinking about it, the Pye F27 base station from where it originated never achieved full modulation either.  In the end I simply substituted the output valves for EL34's and adjusted the bias to suit those.  Full modulation was then achievable.

The transmitter rigged up ready for testing

Temporary addition of an aerial relay and and a pair of SO239 sockets got the TX ready for testing and on Sunday 8th April 2001 the TX was tuned up ready for the 11am net.  First contacts reported that the modulation was OK, but later reports revealed that there was some 100Hz hum present, and distortion on speech peaks.  Both of these are visible on my modulation monitor, and are now under investigation.

Further on-air testing revealed that the modulation was poor due to RF instability on some parts of the modulation cycle, also the hum was because the power supply simply couldn't cope with the additional load of the modulator output, HT the voltage fell on modulation peaks, causing the general nastiness.

 

 

 This is what my shack workbench looked like on the morning of April 8th 2001, during the first tests of the transmitter. The receiver is visible in the background, and was in use for the tests as well.

A rethink.....

In fact almost a redesign in true G4OO style! It crossed my mind that a single 6L6 valve should be able to modulate an 807 (which after all is only a 6L6 with a top cap) so I disconnected the push pull setup and lashed up a single output stage, using the mod transformer secondary as a choke in a Heising arrangement.  The modulation immediately looked better, so I decided to look round for a proper choke to use.  The F27 chassis came to the rescue again, its psu has a choke in it which is identical in appearance to the mod transformer, and fits in the same mounting holes.  Connecting this up, and a bit of fiddling with the bias on the 6L6 and hey presto, 25 watts output, almost fully modulated and no distortion, hum or instability.  The class A modulator imposes a constant load on the psu, and in fact  less of a load than the greedy EL34's did in class B (or somewhere near that!) So out came the EL34's, the phase splitter on the psu chassis  and the Pye mod transformer, in went the 6L6 and Pye choke.  The next on-air test with G3MMS on  April18th proved the results, and the rig was used on the net for the next couple of weeks with satisfactory results all round.  Meantime the front panels for the psu and transmitter were cleaned up, sprayed with matt grey car paint and fitted, together with a screen between the PA and VFO. A Muirhead vernier dial (also from Dennis' junk box) was fitted to the VFO tuning caopacitor.

The transmitter with the 6L6 and Pye choke fitted.

80 metres

Try as I might, I couldn't get more than a few watts of poorly modulated RF out on 80, because there was not enough drive to the PA.  After trying several different coupling arrangements, I decided that I was faced with either giving up on 80, or adding another stage. Having got this far, I opted for the latter, especially after listening to a particularly well attended Saturday morning AM net on 80.  The 6SG7 was moved to the hole vacated by an EL34 to form an aperiodic buffer amplifier, and a 6V6 added  as a doubler/driver.  Now I had 20 watts out on 80 and all sounding good too. The two units were bolted together ready for more air testing, and the panels were labelled using Dymo, as was all of Dennis' homebrew equipment.

Here with the front panel fitted.

 

Casing.

As can be seen from the first picture, the complete rig was sort of encased in thin perforated zinc. This would have been OK in a shack where no-one but the builder would ever go near the unit, and would make for quick repair and adjustment, but would hardly be suitable in the clubroom, bearing in mind that nasty voltages (mains and 500 volts DC!) are present.  I paid a visit to the local metalworking emporium in Boston, with a rough sketch of what I wanted, and came away with a nicely cut and folded casing made of 18swg perforated steel.  Here I confess to spending a little cash, but I could never have drilled all those holes myself, let alone folded the metal accurately!  The casing was treated to a couple of coats of matt black car paint and fitted to the transmitter.

Front and rear views of the finished transmitter.

Aerial Tuner Unit.

As it is planned to use the rig with a simple end fed longwire aerial,  all that is required is a series inductance to resonate the aerial, and some means of indicating when resonance is achieved. The series inductance needs to be variable, and the best form of this would be of the 'roller coaster' type. However, there was nothing like one of these in Dennis' collection of bits and pieces; the alternative is to use a tapped coil, and use a variable capacitor for fine tuning.

Again from Dennis' collection, I obtained a small aluminium box, a 10mA meter, a ceramic 4 way switch  and a 2x500 pF broadcast type tuning gang.

On a panel cut from a piece of aluminium, painted and lettered to match the transmitter, I mounted these components and a coil wound on a piece of plastic pipe.  The aerial current is measured by passing the RF along a wire through the centre of a small toroid on which is wound a few turns of wire.  This is connected to a bridge rectifier which feeds the meter via a series variable resistor, the idea being copied from the 19 set variometer and other military equipment.  The photos below should give the general idea.

 

  

 

To tune up, all that is necessary is to peak the tuning, loading and aerial capacitors for maximum indication on the aerial current meter.  Not too difficult, even for a novice!  It is sensible, however, to have a frequency meter handy, as it is possible to tune either the PA or aerial to double the operating frequency on initial tuneup!

 

This equipment is now installed in the new clubroom.