Well it worked great in the beginning, so I don't suppose it is an adjustment problem. It's the same two stations all over the scale so I guess the oscillator just isn't properly controllable. I have a 100 MHz analogue scope so I hope it is enough to make sense of it.

And the valve sockets still give trouble at just about every IF valve I'll start a proper walkthrough with all the test points.

If the same station(s) can be heard all over the dial, the tuning doesn't work (surprise).
Valve sockets and pins should be checked and cleaned, I couldn't agree more, but check also the connections
to the tuning capacitor. There are sliding contacts to the rotating parts of the capacitor and oxidation can
sometimes prevent a good contact.
Note that injecting contact cleaner into the contacts will often bring the tuning far off scale due to the
slightly conductive content of the cleaning agent. It will fall back as the agent evaporates.

Maybe it would be easiest to simply replace those valve sockets.
They can cause the strangest faults.
I found some nice ceramic sockets on Ebay, gold-plated and all. Their pins are a bit too wide for the holes in the PCB's
but can easily be cut down to half width. They have an incredible grip of the valve pins so also suitable
for the valves on the output board (where the valves are mounted horizontally).

Martin

Cleaning the varicap contacts improved the selectivity a little, but it still doesn't work properly. Voltages on both grids of the ECC85 are somewhat off. Can't see anything obviously wrong on the component side of the frontend, but this green ceramic cap looks odd, as if it was missing half of the top layer. Even the markings seem to be cut off at the edge. I didn't find any pieces of it around though, is it really supposed to look like this?

Wonderful photo ! 

Looks odd, I agree. Almost as if it was done on purpose (to adjust its value ?).
I suppose its value could be measured and held up against the schematics.
That type of tuner frontend is rare, it is only found in a few 1964 models so not easily swapped with
a known good one, which would otherwise have helped a bit in diagnosing.

The type of trimming cap to the extreme right in the
photo (with the screw on top, usually two or three in a tuner like this) is known to lose contact
internally, that is usually from the base to the coating on the ceramic insert, occasionally to the
point of physically rattling.
Also worth checking, I suppose, but go easy on them as they are easily disturbed.

Martin

Yes, I noticed that the component values don't quite agree with the schematic and there's not much info on the front end in general.

I realized I can remove the bottom cover of the front end without taking it out entirely... only to find yet another metal cover soldered on the PCB

The anode voltage of the second EF184 was a bit high, 160V (spec says 100V). This is an original "B&O" valve so I tried another supposedly good one - the voltage lowered to 130V, and the reception is different but not much improved... I guess it just shifted the alignment.

This is getting unnecessarily hairy... the annoying thing is that since this happened all of a sudden while the receiver was on (although not on radio), it is most probably a simple component failure. Now only to find it!

EDIT: I inspected those two trimcaps and they seem quite solid. I also worked them back & forth a few turns and they seemed to have effect, so I guess they are doing their job.

That green cap has been constructed from wire wound on the ceramic tube. I found the loose end of it round the back, so I think it has been actually hand tuned - or some too clever repair guy didn't have the correct value at hand?

EDIT^2: read a bit more. That green cap is indeed a 90 pF wire wound trimcap, adjusted once at the factory Oh the inventions of the world...

Having nothing better to do, I desoldered the remaining metal shield from the front end. Looks like the lady(?) at the assembly line didn't have one of her better days. I resoldered everything there as well as some dubious looking spots on the IF section (I know, should only touch one thing at a time...).

At first it worked! But after the receiver warmed up (or I turned the chassis on its end, dunno which one was the cause), the problem resumed. I've been tapping and carefully twisting just about every component and joint but can't find any spot that would respond in any way.

I guess these tubular and disc ceramic cap types used here can't have hidden problems if they are physically in one piece?

Perhaps I'll need to disassemble the front end completely, clean the PCB properly and start over. What a thoroughly annoying job...

That was frustrating so I simply took a long break, thinking it over. After looking over the IF stages which actually seemed to do their job since I heard at least something, I was almost convinced that the fault had to be in the tuner.

It was difficult enough to resolder in place and as the components were impossible to check with everything soldered in, I just decided to take the plunge and rebuild it completely.

No helping it, the dial cord pulley on the variable cap needs to come off. Hopefully this elaborate Blu-Tack® concoction keeps the cord in place and saves me from restringing it.

And here we have the culprit, out of the chassis. The PCB has no placement silk print and I don't have a component placement picture either, so it will need to be documented before and during the disassembly.

All the component legs had been bent flat on the PCB before soldering and the old solder refused to climb up a desoldering wick, so I had brought the project to work where I have a proper rework station.

Still, it took a good three hours before everything except the varicap were off and documented, and the result is seen here. Time to clean and check the PCB as well as all the individual components.

Wonderful project !
You use Blue-tak, I use masking tape.

Martin

Components were found to be good, so I decided to check the PCB for continuity or leakages. I was astonished to read 60 ohms between a couple of separate traces - and there was nothing connected except the variable capacitor.

I disconnected it as well, and one of the halves of it still showed 60 ohms - steady even while rotating it! That was definitely odd, I could've expected a direct short. All the leaves of it were straight and I couldn't see any foreign matter inside it, so I just blew some isopropanol on it followed by compressed air.

All right, measures infinity now. Could this have been the problem all the time? A hard leak like that through the cap would've definitely prevented it from working properly. I had tried to check the cap already during the initial diagnosis, but it is in parallel with a coil so it showed a direct short anyway.

What caused the short, I'll probably never know, but now there's pretty much nothing else to do than to put the tuner back together with high hopes.

That took another two hours - this is actually a rather dense module for a 50 year old design. And if you have been cursing over a dial cord restring, you'd love installing the coils that have hair thin wires coming out in all directions!

After checking, double checking and some more checking, the tuner front end is finally back in the chassis. Time for that disturbing initial power up...

So,

• multimeter over the anode voltage supply...
• connect speakers...
• connect a makeshift antenna...
• mains plug in...

...power on. Valve heaters start as always, voltage seems normal. So I'll just wait a bit until the valves warm up and the receiver should be operational.

Static comes on, but the tuning knob still doesn't do anything! Oh yes, the pulley is off so I need to rotate the cap axle directly

And it WORKS! Beer time

Dillen:

Wonderful project !
You use Blue-tak, I use masking tape.

Martin

I'm happy to say that worked as well, no problem putting the pulley back on!

Hopefully there are no more surprises so I can finally put the covers back on... except that I still have a small project to be done in there, but now I'm just going to enjoy this receiver for a while

Oh yes, while I was checking the IF stages I found a tiny screw inside the detector can! It is definitely not from this receiver so I guess it has been dropped in there a long time ago. Good aiming by somebody, first through the top cover ventilation holes and then through the small adjustment access holes in the can...

I have seen problems with variable caps before but they were almost always a
lack of contact or bent/shorted vanes.
I never saw this problem before and can't help thinking that something from outside
must have entered the cap. Maybe something was spilled down there at some point.
There's no doubt that the intrduction of a 60 Ohm resistance in that circuit will mess things up.

Good work and thanks for sharing !

Martin

Thanks

The short was on the mixer stage of the tuner, where the cap doesn't see any DC - the other half of the cap would've seen anode voltage through a coil, and there a similar short might have burned itself out... or burned something else.

Perhaps I should've bothered to desolder the cap first, which might have saved me from some frustration and hours of work. But that's what hobbies are for, to keep you away from other mischievous activities

***

Due to postal etc. delays with spare parts I had plenty of spare time to educate myself with valve circuits during this project - something I last did as a teenager 25 years ago! I have learned a lot about electronics since then but didn't have much hands-on experience on valve kit apart from replacing an odd valve or a blown capacitor.

I have to mention this site I found that has an amazing amount of scanned copies of old text books, data sheets and other literature on valve and radio technology. They are old enough to have expired copyrights, so they are essentially in the public domain now:

http://www.tubebooks.org/

Prepare to be overwhelmed...