Murphy A36 double superhet receiver.

[FERNSEH]

Still having problem with instability in the frequency changer and IF amplifier stages.
The circuits around the mixer-oscillator need some explaining. For starters the capacitor C14. In normal practise this capacitor would be returned to ground to decouple the screen grid of the AC/TP and one end of the IF transformer primary. In the A36 C14 has a lower value than usual, it's 0.01microfarad instead of 0.1mfd.
What is interesting is that C14 is returned to the cathode of the frequency changer valve. Capacitor C9 which is connected between the AC/TP cathode and ground has a very low value, 0.00035mfd. (350pF)
This arrangement has to be some kind of neutralising circuit. Murphy used to be very fond of this sort of thing.
Any ideas?

DFWB.

[Radio Wrangler]

The common cathode is the coupling of the LO signal into the mixer pentode. Too much C here will block the LO path into the mixer. THere is a risk of C9 resonating with L9 L10 and creating a right devil's game at all sorts of frequencies. Putting the HT decoupler C14 onto the cathode, not ground looks like an attempt to put the LO voltage onto everything on the pentode except grid 1 - so this is equivalent to adding the LO voltage to the g1 signal voltage.

What a way to do it! It's asking for instability.

I'm afraid this is one of those circuits where it might be easiest to start again.

David

[FERNSEH]

I think the best plan is to examine other radio circuits including other Murphy models which employ the AC/TP. Find out how the frequency changing process is done in other models using the same valve. On the medium waveband there is 60 volts P - P of LO present at the cathode of the AC/TP which means of course there is the same LO voltage on the screen grid of the pentode section.

Ultra used the AC/TP in many early superhets, however, I don't think the valve was used by Ekco.

DFWB.

[FERNSEH]

Almost all of the earlier Murphy superhet radios employed a similar circuit for the frequency changer.
Extracts of the frequency changer circuits from the models A24 and A30.

In the A24 C9 has the very precise value of 0.001373 microfarads. C10 is 0.002mfd.

The FC circuit of the A30 is similar to the A36. C9 is 0.01mfd and C10 is 0.00035mfd.

DFWB.

[turretslug]

Methinks there must have been something of a cheer and sigh of relief on the arrival on the scene of the elegant, effective and isolating triode-hexode....

[FERNSEH]

Or the Mullard octode type FC4 introduced in 1934. AK1 in Europe.

However, Murphy was signed up to Mazda so I guess they were stuck with the AC/TP.

http://www.radiomuseum.org/tubes/tube_fc4.html

DFWB.

[FERNSEH]

News Flash! Lively IF amplifier tamed at last. The fault turned out to be capacitor C29. This is the 10mfd electrolytic capacitor which is wired between the cathode of the V914 double-diode and ground.
The first oscillogram shows the unwanted oscillation at the anode of the IF amplifier. Almost 60 volts peak to peak.
Grounding the signal grid of the AC/TP stopped the oscillations. I also found that when the suppressor grid of the IF amplifier valve was decoupled to ground by a 0.1mfd capacitor the oscillations again stopped.
I was beginning to think the strange frequency changer was blameless and the fault had to be in the IF amplifier somewhere. Perhaps even in the detector and AGC stages. Sure enough it was C29.

DFWB.

[FERNSEH]

The original Mazda AC/ME magic eye is very dim. It's very unlikely one will ever be found, or the Mazda octal version, the ME41.
So I'm considering the American 6G5G or the 6U5. A 6.3 volts heater supply will have to be established for these valves, an auto transformer perhaps, supplied from the four volt valve heater supply.

DFWB.

[Herald1360]

A floating dc heater supply to the eye from a voltage doubler might do it.

Is there any spare room on the main TX for a few extra turns round the winding?

Edit: A quick sim suggests two 1N4007, two 1000u caps will give a ripply 6V or so with a 21R (6.3V 0.3A) load.

[FERNSEH]

There is quite a decent gap on the mains transformer so it might be possible to add a number of turns for the extra 2.3 volts.

Four volt tuning eyes are difficult things to find these days. The Mullard TV4 is one example.
http://www.radiomuseum.org/tubes/tube_tv4.html

DFWB.

[ukcol]

I used a 6E5GT which is more sensitive than the original ME. Here is my mod for my A40C

[Leon Crampin]

I have an excellent ME41 in my A52 - but I've taken it out for posterity. The replacement is an EM71 fitted to an adapter. The display and sensitivity are excellent.

Initially, I used a doubler with capacitors for the heater supply - but the high peak currents caused interference on SW in the form of a rough hum.

Murphys of this era have multiple 4V windings (neglecting the one for the rectifier) and on the A52, I was able to reverse the connections for one winding (pilot lamps, I think) such that there was 8V between two heater windings, with the centre point earthed. The set is unaffected by the phase reversal of one winding.

The EM71 heater floats with each end above earth potential with a small ballast resistor to reduce the 8V to 6.3. This works perfectly.

I'll probably never do it, but it could be arranged for the heater to be extinguished when the set is on "auto" tune with the AFC applied. The ME41 sits in a box inside the set.

Leon.

[FERNSEH]

Many thanks for the useful information. Because the A36 also has short wave coverage I'll not use the voltage doubler solution for the 6.3 volt heater supply.
Either extra turns on the mains transformer or a separate heater auto transformer.
Some interesting information about the RCA 6E5 here:
http://www.radiomuseum.org/tubes/tube_6e5.html

DFWB.