The 'Cyclone' AM MW / LW Wireless Receiver
Last modified: 18 Nov 2018
T
HE STORY IS THAT my neighbour had an 'Ekco AC-85' A.M. wireless receiver which, we think, was made about 1934 by E. K. Cole at the Southend-on-Sea factory in Essex. This model is now a museum piece, particularly as it features the world's first ever truly large-scale, single-piece, Bakelite moulding for the case. E.g. see the
British Wireless And Television Museum See if you can spot it on page 1 of 'The Tour'...
A white version which I think was later and is rarer:
In February 2001, my remit was to clean it up and get it working if possible. This mostly consisted of cleaning up some corrosion and replacing some rubber wires, but it needed 3 new valves (of 6 in all), very early Mullard types with 4V heaters and which have long been obsolete.
Thanks to Harry Rosenbaum of Chelmer Valve (now changed) for finding new I.F. and output pentodes, and to Gerry of
Crowthorne Tubes for supplying 2 versions of the detector diode. (The original output had a blown heater, the I.F.'s top connector had broken off and the detector diode had been destroyed through plugging it into the rectifier socket by mistake!)
Once it was all working, I liked it so much I wanted one too. Real ones are somewhat difficult to come by so I decided to make a 'copy' for myself (which acquired the working title 'AC-85 Clone'). All parts are new, currently available components where possible, and the valves are later alternatives as Octal 6.3V.
One idea was to use up some items I had kicking around and so make use of them, such as a 290-0-290V 'Parmco' mains transformer with 2 x 6.3V heaters (1 centre-tapped, 1 for a rectifier). The chassis is a 12" x 8" x 2" high aluminium case from
Maplin Electronics , and the o/p transformer is a prototype VT1087.
Hassles
Apart from finding suitable valves (Harry and Gerry again), there were a couple of other problems. The biggest was getting hold of the proper equivalent moving vane type, air-dielectric, variable tuning condenser. Manufacture of these has completely ceased (Maplin used to sell Jackson ones), and it's such a problem that the US based
Crystal And Shortwave Society have had to have a batch made specially for their own projects and kits. They have single and two section types which they also offer for sale on their web-site.
However I needed a 3-section ganged type in order to duplicate the original, since that uses a double-tuned aerial transformer for best selectivity. Thanks again to Gerry at
Crowthorne Tubes for finding a 2nd hand one, plus a mating concentric reduction drive. However he insisted it should be 500pF per section, whereas I figured it should be 465pF, but never mind. However the top mounted compression trimmers – for fine adjustment and alignment – were missing and had to be reconstructed from thin brass sheet and mica.
The second problem is that no suitable coil making hardware is available anymore, so all coils had to be handmade. This was achieved using invaluable information in Babani's
Coil Design And Construction Manual, but I still needed at least two goes at it to get it right.
Mainly this was because of errors in the number of turns (too few) due to too much capacitance while testing the tuned circuits; even my X10 oscilloscope probe had too much – a bit like quantum physics, just looking at it altered the behaviour! So I had to create a special very low capacitance oscilloscope probe, as follows:
Appendix
It occurs to me latterly that a simpler way might be make a small pick-up coil and connect the 'scope across that. (?)
Extra Notes
The AC-85 I.F. transformers are on Bakelite tubes – there are no ferrites whatsoever in the AC-85, presumably it was not possible to make such materials in the 1930's. So the I.F. coils are air-cored and each winding tuned by a compression trimmer capacitor. In the 'clone' this is duplicated by a 65pF trimmer in parallel with 150pF low K ceramic plate.
According to Babani's
Coil Design And Construction Manual, the 0V or 'earthy' ends of both the aerial transformer primary and secondary windings should be adjacent, the 'live' ends being furthest apart. This is to prevent
capacitive coupling between the 'live' ends.
The I.F. coils were duplicated by winding with 40 s.w.g. enamelled wire to produce narrow layers, each layer separated by a
1/
2" wide strip of masking tape. The tape emulates the spacing normally afforded by cloth covered wire. Narrow layers produces a large final diameter that is greater than the coil length, as originally. The spacing between primary and secondary must be right for correct coupling and requires some experimentation (ideally, one winding should be movable). Too close and the coils damp each other; too far apart and the coupling is poor with a low Q. The best compromise was 0.75" between the actual windings.
The screening cans are (believe it or not!) made from card and covered in self-adhesive aluminium tape. T5 (3rd I.F.) is varied slightly in that it actually functions as a double-tuned resonator for the AGC detector, and so is wired differently. All this duplicates the AC-85 as closely as possible.
Unusually, the I.F. is 120kHz, same as it is in the AC-85. This is actually helpful in this case, because ideally the oscillator section of the 3-gang tuning unit (VC1) should have a corresponding, slightly smaller value and reshaped vanes for best tracking of the oscillator frequency vs the R.F., but it doesn't, as all sections are the same value. So it was necessary to add a
padding capacitor, C1, in series with the oscillator section (VC1a), to lower its value. This works because the I.F. offset is relatively small at 120kHz. C1 is made up from several combined capacitors to achieve the desired value, that is, to set the lowest local oscillator frequency to the equivalent R.F. value +120kHz. The top end is then set by the VC1a top compression trimmer (not shown in detail in the circuit but implied).
A similar provision exists for the L.W. band, with additional padding capacitors C4 & C5 setting the bottom end limit of the L.W. oscillator range. This is necessary because the L.W. band is lower than M.W., and so the offsets are not the same. The AC-85 also has padding capacitors for the L.W. band for the same reason. C3 then sets the upper frequency limit.
Note that the M.W. range must be set with the VC1a top trimmer
first,
then C3 for L.W. afterward!
It's still a bit of a compromise but good enough. Babani reckons it's too difficult to achieve in practice, and recommends using a
separate tuning cap just for the local oscillator! Indeed this solution is not perfect, resulting in the latter addition of an extra trimmer to make sure that the receiver can be tuned for maximum signal strength at any position of the tuning scale. This is acheived by the 22pF compression trimmer between T2 M.W. secondary and ground. This parallels VC1c for the M.W. band – VC1c is set fairly loose, which works for the lower half of the tunable band, while the 22pF trimmer can be adjusted to make up the difference at the top end. This trimmer is accessible for screwdriver through a hole in the front panel and in fact it duplicates a similar feature found in the AC-85, which I didn't understand at the time!
The oscillator coils are wound on a plastic till roll tube, and enclosed in a screening can identical to the I.F. type. The anode coil winding must be small and of thin wire to minimise the capacitance between itself and the tuned circuit. Although coupling must be close to work, so it must be wound on top of the tuned winding, but is separated by 2 layers of masking tape.
Currently the fine-tuning trimmer (left) is adjacent to wave-change switch on front panel.
At first the oscillator's output was much more efficient at the top end than at the bottom, because of the reactance of the anode coil developing a larger A.C. voltage at the top end, even to overdriving the valve to clipping. Babani's
Coil Design And Construction Manual mentions that commercial practice was to make the anode coil from resistive wire – the added resistance pads out the reactance and so minimises its effect. It's impossible to obtain equivalent insulated resistance wire nowadays, but the same effect can be achieved by adding a 220R resistor, R2, in series with the anode coil.
Choke L1 (22µH) prevents a tendency to drop into parasitic oscillation at an alternative, much higher frequency, usually when L.W. is selected.
Once the receiver was properly aligned and working for the first time, there was revealed feedback oscillation and general instability caused by too much gain around V2. This was reduced by slightly loading T3 secondary with R8 (120k). The connection from T3 secondary to the top connector for V2 grid is made by a flying lead exiting directly out from the top of the T3 can by way of the shortest route, and again this is copied directly from the identical situation in the AC-85. Furthermore, this lead is wrapped with a spiral of thin wire by way of a screen, and that is also duplicated here, and earthed by a solder tag and screw to the top of the T3 can.
The audio detector time constant is determined by R11 & R12 and C14. C14 may seem a bit too small, but this produces the best tone and R11 & R12 were reduced from 130k to 100k to further increase the speed and improve treble. C21 is the I.F. filter and is mostly the capacitance of a screened lead for the volume control, but may be augmented with a 100pF capacitor. The output can be as much as 2V (depending on signal strength).
The separate A.G.C. detector operates from the anode signal of V2, via the double-tuned resonator T4, and again is copied directly from the AC-85. This makes for a very steep A.G.C. slope while tuning through stations and considerably sharpens the selectivity. The 'LOCAL – DISTANT' control of the AC-85, which worked like a squelch control, could not be duplicated exactly, because the valves are different. Initially I used it to provide the A.G. C. line with a negative DC bias offset down to –10V, as a manual control over I.F. gain. But latterly I successfully duplicated the squelch function by moving the 0V end of the pot to V2 cathode, R12 to the wiper and tying the 470k AGC line bias resistor to 0V (hitherto to the wiper of the pot). The continuously variable squelch allows weak stations, of whatever the chosen strength, to be muted so you only hear the stronger ones when searching the dial. Also it can help suppress interference from other broadcasts of near identical frequency, especially at night.
Alignment
The L.O. upper and lower limits of both M.W. and L.W. bands must be set first. Setting up requires e.g. an oscilloscope with calibrated timebase suitable for frequency measurement, and used with the very low capacitance probe.
Alignment of the receiver requires a special item of test gear, the 'wobbulator' (see below). This is a M.W. oscillator with F.M. capability. Here the tuned frequency is modified by the added capacitance of 4 reverse biased 1N4007 diodes, which in turn is altered by the varying reverse bias voltage applied from V2 anode.
V2 is driven from the timebase output of the oscilloscope and so, while using the 'scope to monitor either the audio or A.G.C. detector output, a negative going pulse is produced on screen as the wobbulator frequency sweeps through the fixed, tuned frequency of the receiver. It is then simply a question of peaking the R.F. and I.F. coils for maximum output and symetrical shape. However, due to the small capacitance of these diodes, it only really works with the wobbulator frequency at or near the top end of the range, and in any case the receiver's tuning unit trimmers only have any real effect at the top of the band.
The receiver was first working on Christmas Eve 2002, and has been in regular use ever since. It currently works through a small loudspeaker taken from an old music centre and so, just like the original, it has a good tone – nothing like any of the squeaky little 'portable trannies' of more recent decades!
The backlit tuning scale pointer is copied directly from that which the AC-85 uses. Reduction drive is a concentric friction type. Note EM84 voltage indicator tube at bottom left – functions as signal strength meter directly from the AGC line.
IF transformers were made from card, wooden dowel, plain matrix board and aluminium tape. Local oscillator coil (front 'left') is basically the same but the former is a plastic tube, formerly the centre of a till roll. Output transformer (back 'right') is a VT1087. Phono socket allows either recording off radio, or playback through the AF amplifier when the wave change switch is turned to 'off' (position 3). The 6J5 single triode valve (V4) is virtually identical to half a 6SN7 (same characteristics), but includes its own metal screen, as does the (very small) 6H6 to its left.
Note aerial coils (actually double-tuned transformers) on dowelling inside own screening. Input is by long wire aerial to wander socket on back panel.
Circuit Diagrams
Coil Winding Details
Aerial coil:
IF and oscillator coils: