Eddystone 770R Overhaul

 The Eddystone 770R is a VHF receiver that tunes from the bottom end of the short-wave band to the VHF marine band. In its day it was probably rarely surpassed as stability is very good, and its range of features pretty comprehensive... for the 50s.
To make the coil-pack easily manufacturable, and the tuning scale reliably accurate, the set uses a large turret; similar to the ones used in the wartime R206 and the post-war Murphy marine receivers.

 Above: General view of the 770R and below the RF chassis with covers removed, showing the top of the turret carrying its eighteen coil packs

From left to right, the rotary controls are Receive Mode switch, AF gain; Wavechange, Tuning, Aerial tuning, and RF gain.

Underneath these to the left are the headphone socket and the switch for the Noise Limiter, and to the right, Mains on/off, Muting and Standby.


 The three 6AK5 front-end RF valves are fitted into sockets soldered directly to the tuning capacitor.

At the top is the IF strip and to the right, the AF amplifier with push-pull output valves

 The power supply, with its very heavy mains transformer is on the left. HT rectifier, bottom left and voltage stabiliser to its right.

Along the right hand side are five IF transformers, and to their left the discriminator transformer for FM.

The BFO can be seen behind the front panel, above the mains transformer



 One of the three 6AK5s, the first RF amplifier with replacement disk capacitors fitted.

Underneath can just be seen some of the resistors, which like most of the components in these stages, are not easy to get at.


 Side view showing the underside of the IF strip. The screen grid resistors are just to be seen mounted vertically in the foreground.

The AF amplifier chassis can be seen along the right hand side, at the rear of the turret. This is virtually impossible to work on without first unsoldering the wiring harness and unbolting the sub-chassis from the main chassis.


 The Eddystone turret carries three coils per range. As it has six ranges the number of coil packs is 18. For some reason, best known to the Eddystone engineers, the coils can only be adjusted in-situ after first removing the adjacent set. This is not the only shortcoming. Another is the fact that the IF coil lower slugs can only be adjusted by threading the adjusting tool between close-packed components under the chassis. As some of the obstructions are at HT potential a metal stemmed adjusting tool must first be insulated.
What else can I moan about?
Well the audio section is very difficult to work on. Because of the long narrow shape of the unit, in order to change any deep-seated component, one has to unsolder all the cable connections to the chassis and remove it from the set. Once removed, all the work must be completed blind as the whole thing must be refitted and re-wired before one can judge the outcome of any remedial work.
Over many years of use, or for that matter storage, lots of problems will be apparent when looking at one of these receivers for the first time. There are lots of plastic-moulded paper capacitors, little dark brown things with coloured rings that lose their capacity, develop leaks and just generally deteriorate until they behave either like resistors or near open circuits.
The B7G valve sockets also deteriorate as their spring contacts lose their springiness resulting in intermittent loss of IF gain in particular.
The three 6AK5's in the front-end are mounted directly on the large 3-gang split-stator tuning capacitor. This makes replacement of the many brown capacitors used in these stages a trifle awkward.
Once I'd first got the set working, albeit with very low gain, I tried the various wavebands and discovered that the highest wasn't working at all. The calibration of the second highest was miles out and, try as I might, I just couldn't make the dial calibrations line up. After changing the padding capacitors in the oscillator things were slightly better but eventually I removed the oscillator coil can and squeezed the coil turns closer together. This certainly improved matters but the dial was still miles out. Suddenly the penny dropped. I looked at the dial pointer. It was tied to the dial cord with a piece of string and this had slipped allowing the pointer to move sideways by about an inch. After I'd reset the pointer the Range 2 tuning range had plenty of adjustment in hand. I reset the Range 2 oscillator coil to its original spacing and the band aligned perfectly.
IF alignment is not easy. Because there is so much potential gain in the four amplifier stages, and because of the different receive modes, such as AM, CW, narrow and wideband FM its possible to falsely align the strip. Without a scope and wobbulator, only by placing a 5.2MHz generator at the input to each IF stage, starting at the last is it possible to get an initial accurate alignment. Once the strip was tuned to 5.2MHz the gain increased dramatically. Note that most transformer tuning has two peaks. The peak nearest the top or bottom of the can is correct... except the first transformer, where one must use the inner peak.
The tail end of the IF strip is different for AM and FM and the tuning of the penultimate transformer must be done at AM as must be the last transformer, used in that mode. The FM discriminator transformer is awkward. I found that removing the top slug and tuning the lower slug from the top provided an easy way to get alignment. Once the lower slug was set for best gain the top slug could be set for best FM reception. This is done by centring the station tuning then using speech sound quality as a guide as the top slug is turned, finally adjusting the meter centre-zero pot for centre reading at best tuning.

The set was still relatively deaf. To get enough sensitivity to get this far I'd had to short-circuit the biasing of the first two IF stages. What next? I'd already changed all the brown capacitors in the front end. Once the capacitors in the mixer and oscillator had been replaced, I'd found the highest range was beginning to respond to signals. At least I could hear the signal generator, whereas previously I couldn't as I'd suspected the oscillator wasn't oscillating.

During the work I'd noticed that flexing the cans of the middle IF transformers had increased sensitivity quite dramatically. One reason was loose pins in the valve sockets, specifically the connection to the screen grid of the fourth stage. The third stage was also intermittent and this turned out to be a dry solder joint at the screen grid of the third IF valve.

Changing all the old tubular IF-strip decoupling capacitors, together with the screen grid feed resistor of the last IF stage improved matters yet again.

Next a cursory look at the other components in the IF strip... Although most resistors are not particularly critical in valve circuits, I found several that had drifted to almost double their marked values, reducing screen and anode voltages. To be on the safe side I also fitted a new diode detector, using a germanium type, together with a couple of other semiconductor diodes, a couple of resistors and some decoupling capacitors on the audio chassis.

The receiver was now producing quite acceptable results.
These old Eddystones were favourites with the listening posts of the Foreign Office, when to help fight the cold war, National Servicemen, with a rudimentary understanding of "technical" Russian spent many months tuning up and down the wavebands listening out for their opposite numbers in the armed forces across the east-west border. This particular set may have come from that source as it has certainly seen a lot more than average use. The tuning gears are badly worn and were only improved when a liberal amount of grease was applied to their teeth.

The worst mechanical fault is wear in the turret tuner selector arrangements. There is a large circular Bakelite, or phenolic, sheet with an ident for a bearing to precisely locate the drum's position. Over the years the hole mating with the bearing has worn so that the drum can be wobbled from side to side. This results in a lack of preciseness to wave changing and results in a degree of resetting error. The wear could be fixed but would necessitate the complete dismantling of the drum and either a repair to the phenolic plate or the fitting of a new one made from the pattern of the old one.

Next I looked at a few odds and ends. The RF gain control connects to its knob via a short length of Bowden cable. This had perished resulting in a very springy and imprecise feel in the control. I removed the fittings and made a brass coupler to replace the old type, passing the new quarter inch spindle through the outer part of the old panel fitting. Not perfect but considerably better than before.

Once reliable operation had been achieved I checked the accuracy of each of the tuning ranges. All were pretty good with scale markings quite acceptable once the trimmers in the coil packs had been adjusted. Fortunately it was not necessary to remove coil packs to gain access to the slugs in the coils. Possibly this was so because I'd fitted a new padding capacitor in the oscillator in my efforts to sort out the dial discrepancy.

There are a number of other areas needing to be checked out. The BFO did actually work during initial testing but stopped working when I got round to finishing off loose ends. The user manual is misleading as it gives the anode voltage of the BFO valve as something between 80 and 55 volts depending on the type of voltmeter one is using. Unfortunately what the quoted voltgae refers to is the tapping point between a 68k and a 22k resistor. The true anode voltage is nearer 15 volts.

The main components for the BFO are buried inside a metal can that needed to be removed from the chassis in order to carry out any checks. Several solder points are first dealt with, removing the valve heater supply, anode feed, ground and RF output. After wiring it up on the bench the unit stubbornly refused to oscillate. Initially I confirmed that the internal components where in order; three capacitors,and the 22k and 47k resistors. Next I tested the coil with an old grid dip meter (a real grid dip meter) and found it could be brought to resonance by twiddling the slug in the coil former. Each time I applied power however the thing refused to oscillate until I poked at the coil, whereupon it suddenly started up. Presumably the enamel had perished in some way and had allowed a short to develop?

When the BFO had been reassembled and soldered back into the circuit it worked perfectly. The manual suggests it is offset from the 5.2MHz IF frequency by 1000Hz and this was done by setting the drum between bands, injecting a 5.2MHz carrier via the aerial socket and setting the tuning slug for a suitable tone.

Before I'd had chance to drip wax into the tops and bottoms of the IF coils to prevent movement of the (very loose) slugs I discovered a tube of Radiospares patent compound which seems to me to be very similar to the sticky substance used in two-part Araldite. I secured the dial pointer with superglue to prevent it wandering again and finally checked the tuning of the IF transformers and the discriminator coil. The latter I achieved using a sweep generator.

A note here on IF alignment won't go amiss. It's imperative that the IF tuning is carried out using the CW position of the mode switch and not the other settings. AM provides AGC, which can upset the tuning and both the narrow and wideband FM settings introduce additional coils into the transformers. This is done so that the IF bandwidth can be broadened and any attempt to twiddle coils in the FM settings will ruin AM gain (excepting the discriminator of course which has to be set up in FM).

During the alignment procedure I'd noticed that there are three resistors that are switched into the bias circuits of the first couple of IF amplifiers. That for the AM and CW positions is marked in the component listing as "value to be determined". As the value of this resistor has a dramtic effect of overall IF gain I experimented with it and found that a 68ohm component produced a much higher gain than the 220ohm part that was fitted. Any less than 68ohms seemed to introduce excessive noise and some instability.

Overall performance is now quite good, especially on the lower wavebands. Band II is quite respectable for an old-timer but the highest band is pretty poor in comparison. I suspect that this is due to deterioration, and subsequent leakiness, of the various parts used in the front-end.

Now that I consider myself something of an expert on the 770R, perhaps I'll investigate my own Eddystone. At least mine has a silky smooth tuning mechanism, although I do expect to find precisely the same "stock" faults as I discovered in the customer's example. Oh, and I seem to have acquired another example of this receiver so now I have two in the queue. One of these I'll certainly be selling once it's been sorted out.

Click here to find out more about the specification

return to entrance