Civic XR4930 radiogram

[Michael Maurice]

This radiogram has several issues:

1: The output stages are oscillating

2; No VHF/FM

I'd never seen a circuit for these radiograms until now when one came with the said radiogram.

The output stage:

I replaced the usual capacitors ie the coupling capacitors and the cathode bypass capacitors. There is also a 470nF capacitor mounted on the transformer which I cannot find on the diagram.

At first it was the LH channel that was oscillating, but there was some arcing coming from the base of the RH channel’s valve so I replaced both valve holders with ceramic bases and gold plated pins, the oscillation then moved to the RH channel!

The oscillation stops when I ground either or both inputs.

There is another oddity with this circuit, the balance control is between the grids of the pentode, not what I would have done, but I cant really change it.

I was thinking of putting some small capacitors say 2.2pF across the 10M ohm grid leak resistor for the triode and say 22pF between the pentodes grid and ground.

I will do my calculations later but that should give roll offs at around 18KHz which is high enough for this application.

Does anyone have any other ideas?

[Radio Wrangler]

I don't like the bias arrangement on the triodes relying on self-rectification and very large grid leak values. Those 10meg resistors aren't likely to have aged well, and they're after every bit of gain they can get out of those triodes.

The pentodes of ECL86 can produce negative resistance impedances, causing oscillations. They've used parasitic stopper resistors in both g1 connections AND both g2 connections. I think they've run into hooting problems from early in development.

One trick worth trying would be a good non-polar 100nF across the reservoir electrolytic. They can look like poor decouplers alone at higher frequencies and allow all sorts of cross-coupling. Beyond that it might be a case of changing those stopper resistor values.

David

[Radio Wrangler]

That 470nF is a heck of a big value for bodging around the signal paths... It would be OK as an added decoupler from HT on the transformers to ground to back up that reservoir electrolytic at high frequencies, so it might be connected there. It's evidence that the manufacturer was bodging hard right up to the last minute and after the schematic/service info was printed.

I'm afraid this isn't one with a definite fix. You just have to continue their bodging process through more 'suck it and see'

Capacitors, inductors and bits of wire don'y dissipate energy, they just reflect it. So most often, playing with Ls Cs and wires just move oscillations aroung and don't act to make oscillation impossible. Resistors, though are good. They turn oscillatory energy into heat and that's strictly a one-way process and you have the laws of thermodynamics on your side! However, resistive stopping reduces signal gain and signal output power.

Trying to meet goals for gain and power can push a design into territory where stability is iffy, and there's no way out without sacrificing something that's a hard goal. All sorts of messing around ensues until they get one which meets the goals and seems stable. "Ship it!" shouts the boss and the design is frozen at that point. So they make a lot. Some of them might even work, but often these sorts of marginal products get themselves a reputation as a stinker in the repair trades. If we're lucky, they have trouble on the production line and low yields so the thing gets taken out of production quite soon.

I'd be inclined to increase the values of the stopper resistors until it started behaving itself, and then live with any reduction in sensitivity or power.

Civic has the sound of an in-house brand for some chain store or catalogue firm.

David

[Graham G3ZVT]

Originally Posted by Radio Wrangler

Civic has the sound of an in-house brand for some chain store or catalogue firm.

David

Civic was a chain of electrical stores.
Comedian Harry Worth did the TV adverts, leading a conga line round a store "Let's all go to Civic, easy terms at Civic, na naa naa naa... (ad lib to fade).

[Red to black]

It looks to be of German origin, the symbols key for the resistors are using the German DIN standard as per Grundig diagrams.

[G6Tanuki]

Originally Posted by Red to black

It looks to be of German origin, the symbols key for the resistors are using the German DIN standard as per Grundig diagrams.

A lot of the stuff sold by Civic was made in Eastern Europe - Poland, Czechoslovakia, Hungary - by companies like Tesla and Unitra. Might be worth looking on radio museum to see if there are any similar looking things there.

[Michael Maurice]

There is nothing on Radio museum, I've checked. I think these were made in Hungary.

As David suggested, I will increase the grid stopper resistors. I think 47K should do it, how does one go about calculating capacitor values for a capacitor between the grid and ground?

I will draw out a circuit and let you know where this 470nF goes.

[Marconi_MPT4]

Videoton radiogram
https://www.radiomuseum.org/r/videoton_zenegepjukebo_r_4930.html

This ECL86 circuit topology is very similar to an earlier 1960's Telefunken design using ECL82 valves. Similarities include wiring a balance pot between output pentode g1 connections, 10M ohm triode grid leak, a tapped volume control pot of unusual value with feedback applied to a resistor in series with the ground connection.
Rich

[Michael Maurice]

Originally Posted by Radio Wrangler

That 470nF is a heck of a big value for bodging around the signal paths... It would be OK as an added decoupler from HT on the transformers to ground to back up that reservoir electrolytic at high frequencies, so it might be connected there. It's evidence that the manufacturer was bodging hard right up to the last minute and after the schematic/service info was printed.


I'd be inclined to increase the values of the stopper resistors until it started behaving itself, and then live with any reduction in sensitivity or power.

Civic has the sound of an in-house brand for some chain store or catalogue firm.

David

The circuit is correct, the 470nF caps are in the feedback ciruit and are C106 and C165.

I shall increase the stopper resistors and see what happens

[Radio Wrangler]

Putting a stopper in the triode grid connection could also help. There's always uncertainty as to just which device is oscillating and which is along for the ride.

If the pentode section is the one oscillating, the connection between valves can route the oscillations backwards to the triode. Connections can carry unwanted signals both ways. Then, there are feedback paths.

Whether a valve or a transistor, if you bias things up, pushing for maximum gain, you're also biasing it up for maximum instability. Similarly with a valve on a VCM, when the bias is adjusted for best Gm, it's also being set up to where it could easiest oscillate, and this accounts for all the ferrite beads these instruments need to be sprinkled with.

David

[Michael Maurice]

I have replaced the 10K grid stoppers on both pentodes with 47K, I have also fitted 47K resistors in the triodes grid’s stage, the oscillations have subsided but are still there.

I’m considering fitting either 10pf or 22pF between the pentodes grid and ground, does anyone concur?

I could also add capacitors between the triodes grid’s stage and ground.

[Radio Wrangler]

Shunt capacitors could make it worse. better try series stopper resistors on the triode grids first.

Shunt capacitors make low impedance RF connections to ground and help circulating RF currents. It's difficult to explain, but shunt paths to ground looking like short inductors at RF can by the action of a biassed-up amplifier be transformed into a negative resistance! The amplifier and its stray-Cs acting as a gyrator circuit. The negative resistance is what powers the oscillation. You'll find gyrator theory in the general area of active filters. Gyrators can convert capacitors into what is effectively an inductor and you can make analogues of LC filters without a single inductor in sight.

David

[Michael Maurice]

I have tried stopper resistors on the triodes, 47K, it still oscillates.

[Edward Huggins]

Without going too far off topic, I don't think there was any real connection between the UK and Hungarian Brands. In my experience, the UK CIVIC company used only UK manufacturers including PERDIO (portable record players) and the then owners of ACE (Radiograms) to built their products for them. Quite happy to be proven wrong!

[Radio Wrangler]

OK, it's used up all my best shots, I'm aftaid. Adding a hundred ohms or two in the pentode anode isn't usually tried because of the loss of power, but if things are getting desperate... It can help if a resonance in the output transformer is dumping a nasty impedance on the pentode anode.

David

[G6Tanuki]

What happens if you disconnect the negative feedback? (disconnect one end of both R107 and R169)? If there's some odd phase change issue going on, it could be that at some point in the frequency-range the negative feedback changes phase and becomes positive.

[gallowfields]

in the first post you said the oscillation was in the left channel ,but after changing the valve bases [both] it moved to the right . are you sure its not just the output valve ?chances are you swapped valve positons after changing bases .
just a thought
brian

[Michael Maurice]

Thanks for all your inputs, (excuse the pun), I'm away for a few days and will look at it again on Thursday/Friday.

[Michael Maurice]

Back again, more tests.

I've replaced the 68nF caps, no change.

I've swapped valves, no change

I've disconnected the feedback loops, no change.

I've fitted 22pF caps across the triode's grid and 10pF across the pentodes grid, no change.

If I remove either ECL86, the oscillations stop.

If you move the balance control away from the centre in either direction the oscillations stop.

I'm starting to tear my hair out. I can't understand what is going on.

[Michael Maurice]

Before completely giving up on this, I managed to find another radiogram, I've stripped the chassis, deck and speakers from it and will service it before testing it.

But, the chassis is filthy, I mean really filthy. If I were to wash it in warm soapy water, then put it in the oven at say 50 degrees, overnight, would it be safe to apply the mains to it?