Improved Noise Limiter for Eddystone 840A

[G6Tanuki]

The recent thundery weather has made night-time listening on the HF bands using my Eddystone 840A bedside radio a less than life-enhancing experience.

As standard it uses a single semiconductor noise-limiter diode, whose back-bias is obtained from the detected-and-smoothed DC output of the detector so it should have a clamping threshold that automagically follows the average carrier level. All well and good - in theory -for AM reception but the idea fails when listening to SSB/CW neither of which provide a satisfactory reference-carrier.

Looking through an old copy of ART I came across a circuit first published in the late-50s for a manually-controlled noise limiter; originally using valves [6AL5/EB91 was suggested] I reworked it into a version to use semiconductor diodes [in my case these were 'surplus' reject OC71-type transistors with the base and collector strapped].

Built up on a 6-way tagstrip, this mcircuit has the advantage that it sits between the first-audio and the output valves, where voltages are higher and impedances a bit lower; the 840A being a 'live chassis' receiver I was concerned that a limiter fitted in the path between the detector and 1st-audio valve could be at risk of stray capacitive hum pickup.

In this version, C1 is 0.1uF, you probably wouldn't need it if the cathode-bypass capacitor was of zero ESR, but we all know the real world isn't like that.
C2 is 0.01uF, connects to the anode of the first-audio valve.
R1=R2=220KOhm, these set the standing bias at the mid-point of the two diodes to be half that at the top of the upper diode, so the clipping threshold is symmetrical.

C3/R3 are the existing ones feeding the grid of the output valve, as are the cathode resistor and capacitor.

VR1 is a 10KOhm potentiometer to control the DC level fed to the top of the limiter; I used one with a Nylon control-shaft to provide insulation [live chassis, remember]. 10KOhm in parallel with the original cathode-bias resistor is irrelevant to the valve's bias.

The pot's bush mounts nicely on the little bracket where the original limiter's single-pole in/out switch lived.

Of course in use you need to adjust VR1 to match the audio level [which in turn varies with the setting of the AF and RF gain controls] but once you get the hang of it this is not really an issue.

The result wotks rather well. Certainly better than the original (which wouldn't be difficult, it has to be said).. And the mod is easily reversible if you're an originality-obsessive. It definitely makes night-time listening on 60 and 80M less stressful!

[Bazz4CQJ]

Thanks for sharing that. As you will know, my HRO has no noise limiter at all in it's original form, so I'll add that to the "to do list".

Not sure what I've got in terms of reject OC71; will Ge point-contact diodes suffice, or is anything else of interest?

B

[G6Tanuki]

I think you could use pretty much any diodes; in the past in similar clipper/limiter applications I have used everything from gold-bonded OA47 and OA91/AA119 Germanium types to diode-connected OC200 transistors salvaged from the 1960s surplus 'computer panels' that were widely advertised in PW etc.

Good back-resistance would probably be my only real thing worthy of measurement if you want to select suitable diodes - if this is unsuitably low [or varies between the two diodes] it could result in the mid-point not being at half the bias-voltage and that could cause asymmetric clipping.