Ekco U428

[Nuvistor]

Looking at some old circuits I came across this and I am trying to understand the reason for audio from the UL84 grid circuit being fed back to the grid of the UCH81 AM osc. I have tried to sort out from the PCB layout and it looks correct and the Trader sheet is the same.

The manuals are on Pauls DVD so I don't want to print the whole circuit, snippet attached, hope that is OK.

What am I missing?

Frank

[Herald1360]

Odd! Maybe some kind of envelope feedback to reduce detector distortion?

Maybe dc going the other way? The oscillator grid will go negative.

[Nuvistor]

Hi Chris,
Glad it is not just me who finds it odd. I don't have a set to check if it is a drawing error, but the PCB layout seems to be the same. I cannot tell completely because wires to the tone control leave the PCB.

I cannot find the same idea on other Ekco/Ferranti sets, maybe it was a mistake and as it perhaps caused no harm it was left, on the other hand your comments could be spot on.

Frank

[ukcol]

That's interesting! I don't think its a mistake, I think Chris's first suggestion is nearest the answer.

If the amplitude of the local oscillator is varied this in turn will alter the amplitude of the I.F signal via the conversion conductance of the mixer. If you get things like the initial oscillator amplitude and the phase of the feedback correct it seems to me it is possible to reduce any distortion in the detector and I.F stages.

I seem to remember seeing a similar thing in a different context before. A small A.M. transmitter design that used a detector after the output to feedback a proportion of the audio to correct for distortion in the modulator. The design came from one of our members but I cannot remember who it was.

[Herald1360]

Envelope feedback was one way to improve the IMD from a SSB transmitter. It just popped into my head when I saw the circuit.

Maybe Radio Wrangler would have a better idea?

Does disconnecting the line seem to have any deleterious effects?

[Nuvistor]

I don't have the radio, just looking through circuits last night and came across it. I could not find it in any other Ekco circuit, even the AC only ones using the ECH81 range of valves.

To be honest if I had the set with my hearing I probably could not tell the difference.

I could not see a reason, that's why I asked the question. I can see it affecting the LO, but with such a simple circuit I would have thought it would pull the frequency as well as the change amplitude of the LO.

I will keep with the idea of less distortion, unless there are other comments. I am surprised it was not noted in either the Trader or the Ekco service manual.

If nothing else it keeps my grey cells working.
Thanks for the suggestions.
Frank

just looking through circuits last night and came across it

Don't we all, a fascinating way to spend an evening. I tend to do the analogue side of things at work, well someone has to, and the old designs are a great pointer. Try simulating it using Spice, another great way to spend an evening, ltSpice is free (linear.com) and the audio guys have made some valve models for it too.

[Nuvistor]

I have just installed ltSpice but that I don't think I will be trying it tonight, it runs and a demo works.
I will have to learn how to drive it, that may not be as simple for me these days, see how it goes.
Paul is authoring another DVD, that will keep me occupied when it is released.
Frank

[Leon Crampin]

The negative voltage at the grid of the local oscillator is used to provide bias for the output valve.

Leon.

[Nuvistor]

Hi Leon,
I could accept that the negative voltage has some effect on the bias but the cct has a 100K and 150 ohm across the HT (UY85 cathode) with the UL84 cathode connected to the junction. The cct gives the voltage on the cathode as around 10v+. This will give a 10 volts neg from grid to cathode.

The cathode is not decoupled by a capacitor so I presume the small amount of 50hz ripple on the cathode will have an effect of reducing hum.
Thank you for the suggestion.
Frank