Eddystone 840A

[G6Tanuki]

The 840A has no voltage-regulation on either the LO or the BFO; unless your mains is very stable the tuning's always going to be a bit wandery. Even altering the RF Gain setting shifts the HT rail by several volts which shifts the tuning.

There are no screening cans on any of the 840A's valves. It could be worth swapping the BFO-valve with the audio-driver valve - they are both UAF42 - to see if this makes a difference. Swapping identical RF/IF valves round in TV sets to find the least-microphonic/drifty one was a sometimes-used trick in the 1950s.

[turretslug]

Oh, yes, I'm forgetting rather that the 840A is quite bare-bones- the C in the UCH42 doing the local oscillation and no anode feed stabilisation, either. A little like the 640. Still, resolving SSB will hone the safe-crackers fingers and develop patience...

[HamishBoxer]

Think the black can will keep the heat in.

[Radio Wrangler]

There were special corrugated spring steel cans which hugged the glass of valves and reduced temperatures. They turn up sometimes in Racal gear, but the usual bayonet screening can does increase temperature and drift.

David

[Skywave]

Originally Posted by turretslug

On my 750, I found that replacing the EF91 LO's plain aluminium screening can with an identical but matt-black finish type made a surprisingly large improvement to initial warm-up drift. I'd have been very sceptical of someone who told me that it would make a difference - but it did!

The use of matt-black valve screening cans is discussed at length in 'Amateur Radio Techniques' (RSGB), 7th. ed., Pat Hawker: page 339 refers. It makes for very informative reading.

Al. / Nov. 8, '14 //

[Skywave]

Originally Posted by HamishBoxer

Think the black can will keep the heat in.

Just the converse. Consider aluminium heat-sinks for power transistors. For the same surface area and geometry, those that are finished in matt-black anodising have a greater heat-radiating capacity that those that are left untreated and plain aluminium. Classic physics covers all that in the concept of 'black body radiation'.

Al.

[WaveyDipole]

Well the drift its not so bad so as to be unmanageable, especially once it warms up. I also found it does drift a little on band 1 (11-30Mhz) as well, but given the frequencies involved thats perhaps not entirely surprising, especially given that there is no voltage regulation.

I have attached some pictures of the completed restoration. I have also added to some light to the display.

Unfortunately, I made a slight mistake just before putting the vcase back on. I squirted some air duster down the display to remove some dust that had accummulated. Unfortunately, it has left a streak which is quite visible at certain angles and lighting conditions! I'm trying to figure out whether their is a way of cleaning it without disassembling it all again!

[Synchrodyne]

Originally Posted by Silicon

When I look at the schematic I get the impression that when a wander plug is pushed into the 'E' socket it connects the chassis to the Aerial earth.

The Aerial earth is isolated from the HT ground by a 0.01uF mica capacitor.

Even if used with an isolating transformer, the case/cabinet might get a bit 'tingly' when that 'E' plug is inserted.

That the plug in the “E” socket bridges the case the aerial earth is an interesting feature.

One has the impression that the intention was that the receiver be operated with a protective earth connected to the case, quite reasonable since a chassis-to-case fault could result in a live case. But instead of having the case earthed via the power lead, it was done by a separate connection. Not the safest way to do things, but probably in keeping with the way things were often done in the 1950s.

It looks as if it were done that way to allow a single earth connection to be used for both protective purposes and as a functional earth for the aerial system when an unbalanced aerial was used, and for that earth connection to be optimized from an RF viewpoint. Not good in today’s light; someone who thought they were installing an RF “aerial” earth might not fully address protective earth requirements, and anyway an independent protective earth would contravene code in some territories even though independent functional earths might be allowed.

With the receiver fed from an outboard isolating transformer, earthing the case would be unnecessary and I think undesirable. But that would happen if the “E” socket were used to connect an “aerial” earth. With the case earthed in this way, a chassis-to-case fault would effectively earth one side of the isolating transformer secondary, something that would likely go unnoticed, but which would be potentially dangerous in that there would be a shock hazard between the other side of the secondary and earth, something unexpected in what was thought to be an isolated system.

I think that there may be a “get out” though. The aerial shorting plus appears to bridge the second aerial socket to the same earthing point as is connected to the case by the “E” plug. If that is left out, then possibly the second aerial socket, normally used for the “other” side of a balanced aerial, could be connected to an independent RF functional earth (if the local code allows such). That way the case could be left unearthed to allow safe operation from an isolating transformer.

Cheers,

[WaveyDipole]

Thanks Synchrodyne. This might explain why I found the jumper plugged into the centre pin (the second 'A'), while the other end sat unconnected in a hole at the position actually marked with the second 'A', i.e. in storage mode. The antenna wire was left plugged into the first 'A'. I was baffled by that especially when I saw that the instructions suggest that for a random wire antenna the jumper should connect the second 'A' to 'E'.

Since completing the electrical work I have been running it as per the instructions with the jumper bridging the second 'A' with the 'E' but I have also tried it with the jumper removed and could discern no difference in the reception. Given your observations, I will be running it without that jumper connected from now on. If I can get hold of another wander plug then I might perhaps connect a ground earth to the other 'A' rather than the 'E' socket.

[G6Tanuki]

Originally Posted by WaveyDipole

Since completing the electrical work I have been running it as per the instructions with the jumper bridging the second 'A' with the 'E' but I have also tried it with the jumper removed and could discern no difference in the reception. Given your observations, I will be running it without that jumper connected from now on. If I can get hold of another wander plug then I might perhaps connect a ground earth to the other 'A' rather than the 'E' socket.

It'd be worth checking the continuity of the antenna-to-first-RF coils. My 840A worked on all but the highest-frequency range - investigation showed this to be because on that range the local-oscillator didn't.

The connections between the windings of the LO coil and the tags had clearly been soldered by someone with a room-temperature IQ and a soldering-iron only a few degrees hotter.

On this basis I'm minded that you might find high-resistance-tending-towards-infinity issues in the windings of your coilbox.

[WaveyDipole]

Originally Posted by G6Tanuki

1] Because the BFO would appear as a constant 'very strong carrier' and cause the AGC to throttle-back the gain of the receiver even with no signal at the antenna.

2] Because even if it didn't do [1] the effective signal for the AGC to sample is inconsistent when resolving CW or SSB. Normal "AM" AGC works by sampling and averaging the essentially always-present carrier-level of the AM signal; this carrier's not there in either SSB or CW.

Meant to say thanks for that explanation. It makes sense and I should have realised.

Originally Posted by G6Tanuki

On this basis I'm minded that you might find high-resistance-tending-towards-infinity issues in the windings of your coilbox.

Mine does work on all bands but as mentioned there is a little drift on the higher frequency range and the volume does drop off noticeably towards the 30Mhz end of the scale. At these frequencies I would expect any drift to be much more noticeable than on the lower ranges. However, I will check the connections on the coils anyway.

Anyone added a 5th band to the spare set of contacts on one of these, say LW to VLF?