Murphy A40C V8 base transplant.

[ukcol]

I have done quite a lot of work on this set over the years that I have owned it. I decided today to address one of its issues, a noisy V8 Mazda Octal socket.

V8 is described by Murphy as the "ATC (Automatic Tuning Correction) control valve", that is, the AFC reactance stage in modern speak.

The original valve used in this stage was an AC/SP1 which is an RF pentode with a B7 base. The AC/SP1 became obsolete many decades ago and Murphy produced an official modification to address the shortage of this valve. The modification involved changing the B7 base for a Mazda Octal one, fitting a Mazda TH41 and making some minor circuit changes. The TH41 is a triode-heptode frequency changer and the hetode section is used to perform the function previously fulfilled by the AC/SP1.

It looks like this modification had been done to my set many years ago, probably when it was relatively young but the paxolin base now needed replacing.

The pictures show the old base in place, the new base and the new base fitted.

[ukcol]

Next job.

When I bought this set the AC/ME magic eye had gone very,very dim. I replaced it with a 6E5GT, modifying the set to suit. See diagram below.

More recently I obtained an ME41 which has almost identical curves for shadow angle and anode current as the original AC/ME and also has the same 4 volt heater. I have now undone the mod I did for the 6E5GT, and fitted the ME41 with its Mazda Octal base.

I've just been listening to its short wave performance this evening and it is very impressive in spite of the fact that I haven't checked the alignment yet. You will note that i have said "checked the alignment" this is because I will only realign the set if it turns out it is mis-aligned significantly.

[ukcol]

Changing the tuning gang chassis rubber mounts.

Both the main and the short wave tuning gang's mounting bushes were 80 years old and had deteriorated somewhat.

The nearest match I had for size I found in my Halford's grommit kit.

A straightforward but quite fiddly job.

[HamishBoxer]

Colin,I wonder whether you replaced main smoothing caps and did you re stuff them?

If so using which modern equiv.

[ukcol]

Hi Hamish,

Yes I replaced the big electrolytics when I first got the set going some years ago now. I used the screw fixing types from the BVWS.

[HamishBoxer]

Thanks Colin for that.

[ukcol]

The audio pre-amplifier stage.

Valve 5 in the A40C provides audio amplification and inter-station muting and was originally a Mazda AC/SP1. I was advised by Mike Barker (through these pages if memory serves) that when the AC/SP1 became difficult to source Murphy recommended an AC/VP1 as a straight substitute with no modifications or a Mullard VP4B with the grid and anode connections changed over to match the different pin-out of the Mullard valve.

My example of the model has a VP4B fitted and I noticed while I was working on it that the valve was very microphonic indeed when the set was muted by the inter-station noise reduction system. I did not have another VP4B but I did have a couple of AC/VP1s and so I reversed the mod' and tried the Masda valves. They were also both microphonic to an extent but nothing like as bad as the VP4B. I left the quieter of the two in the set, its microphony level is quite low and not a problem.

The original AC/SP1 that the A40C design used was a valve designed to do the muting function (among other things). Muting is achieved by making the suppressor grid (the outer grid) of this pentode 20 volts negative with respect to the cathode. This action cuts off the anode current (but not the screen current) thus muting the audio.

Now I'm speculating - I would imagine that the suppressor grid construction in the AC/SP1 was made as rigid as possible to prevent microphony when this mute function is active. The AC/VP1 and the VP4B were not designed with the inter-station muting function in mind and therefore the outer grid is not as rigidly mounted. Another possibility is that the AC/SP1 was very much more effective at cutting off the anode current and so no microphonic effect could be passed on the the later stage(s) of the amplifier.

[ukcol]

Originally Posted by HamishBoxer

Thanks Colin for that.

Just another point Hamish. I can't remember the exact details off hand, but I found in some cases that I had to buy capacitors that cantained two caps and only use one leaving the other tag blank. It will probably make sense to you when you look at the values you need and compare them with what is available.

[HamishBoxer]

Yes I twigged that having bought 2 x 8/16uF I now need 2 more!Yes I have used BVWS as I believe they are the best.

Thanks again Colin.

[ukcol]

As the ATC system wasn't working I first turned my attention to the discriminator circuit (see attachment 1 below).

I’ll first give a brief description of how the discriminator works. The IF input (top right of diagram) is top capacity coupled (C52 and C53 both 4.5pF) to two tuned circuits. L27/C55 is tuned is tuned to 117.5kHz and L26/C54 to 120.5kHz, (1.5kHz above and below the IF frequency of 119kHz respectively). When the set is correctly tuned producing an IF of 119kHz the voltages across the 2 diode loads R41 and R42 are equal and opposite producing a zero volts output at the junction of R41/R40. If the set is mistuned toward 117.5kHz the voltage across R42 increases and that across R41 decreases causing the output of the discriminator to go positive. This output is passed to the reactance valve V8 (via the LP filter and defeat switching) pulling the local oscillator frequency up toward the correct frequency. If the set is mistuned toward 120.5kHz the opposite happens, the voltage across R41 increases and that across R42 decreases causing the output of the discriminator to go negative. The local oscillator frequency is then pulled down toward the correct frequency by V8.

The fault symptom with the discriminator was that the output would go negative during tuning but it would not go positive. With a 117.5 kHz input to the circuit the C55/L27 tank would not tune. A cold check of components revealed that C53 (4.5pF) was all but open circuit (see 2 & 3 below). I didn’t have a suitable component and so I fabricated a capacitor using a short piece of screened cable as a temporary replacement (see 4 & 5 below). This got the circuit working and finally I fitted a new 4.7 pF silver mica capacitor from RS. The diode load components were out of tolerance and so I replaced them as well. I also replaced the equivalent components in the 120.5 kHz section as a precaution against future problems.

The reactance stage (V8) was also faulty and I will write up on that next.

[ukcol]

OK, I've finally got around to writing up the next part of this story.

Getting the reactance stage working. (The circuit is shown below).

Unfortunately I’m not able to tell you exactly why the reactance stage wasn’t working. This is because I managed to introduce a fault myself fairly early on in the repair process (a dry joint pictured below) but I am able to tell you what work I did do.....

Voltage measurements on the reactance valve V8 revealed no obvious problem with the DC conditions and so I decided to do cold tests on the peripheral passive components. R36 the 90k resistor that passes through the hole in the screening plate around V8’s base had increased to 107k. I replaced it with a 91k resistor. (91k is a standard value in the E24 series). The TH41 valve tested OK. I’ll just remind you that the TH41 was used in an official Murphy modification to replace the obsolete AC/SP1. The 500 ohm pot (R38) was in poor condition and so I replaced it.

When I tested C48 it measured 35pF. The value for this component is listed in the A40C manual as 450pF but the actual component was very clearly marked 45pF 5%. In either case C48 would need to be replaced but which value was the correct one? A look through my A40C documentation didn’t throw any light on this problem. I knew that this reactance circuit had been used (in a nearly identical form) in the Murphy A28 and that manual clearly showed that the capacitor is a 45pF. The markings on the component were correct and the manual was in error. I fitted a 47pF silver mica capacitor.

At this stage I powered up only to find that the reactance stage still didn’t work at all.

The reactance stage is effectively connected across the tank circuit of the local oscillator. The AC/SP1 data sheet says that the components in the feedback path from anode to grid present a signal in quadrature such that the valve presents a negative inductance* across the local oscillator circuit. By varying the DC on g3 the value of this inductance is changed and hence the oscillator frequency.

* My understanding of negative inductance is that of an impedance that increases with frequency but has the same voltage/phase relationship as a capacitor.

Following the oscillator signal from the anode to g1 of V8 I found it present on R36 and C49 but it disappeared at the connection between C49 and the coax cable that feeds g1. I had made a perfect dry joint when I had replaced the valve base. After attending to the dry joint the reactance stage showed signs of working and the next job was to set up R38. Following the alignment instructions for setting R38 in the A40C manual I found that correct alignment was out of range of the pot.

The original valve base was of a different material to the new one and wondering about inter pin capacitance. I speculated that perhaps a difference in bases could be responsible for this alignment problem. The pin to pin capacitance of the original base measured about 0.6pF and that of another base of the same type that I had fitted about 2.0pF. Using a fibre pen I cleaned the pins of the old base and carefully re-tensioned the contacts. I went to all the trouble of re-fitting the old valve base only to find (yes you guessed it) that it had made no difference to the problem. What was going on?

What was going on was that I had made another error. Thanks to the good offices of Mike Barker I had a copy of the Murphy modification instructions for replacing the SP1 in the V8 position with a TH41. It included a change in the alignment instructions for R38 associate with this change of valve type. I was able to correctly align R38 using these different instructions. Finally the automatic tuning system was working correctly.

Was that the end of the story – not quite.

The reactance stage employs a coaxial cable to feed the top cap of V8. For the original AC/SP1 that was the anode connection, for the TH41 it is g1 of the heptode section. The inner conductor of the coaxial cable is separated from the outer by ceramic beads and the outer is a coiled steel spring like structure. The outer is connected to chassis by a metal clamp secured by a small screw. The clamp is first closed around the outer and then sercured in position by this same screw. During the course of the work described above I had found that the outer steel conductor was oxidized causing a poor connection to the clamp. Cleaning the contact area provided a cure only for it to return a few weeks later. After 3 attempts to make the repair permanent I decided to replace the cable. Mike Barker advises not to do this but I suspect it is because the capacitance of the cable is important. I used a short length of low loss UHF coax which had a capacitance of about 2pF less than the original cable. It works well and has proved a cure.

BTW Shortly I shall be starting a discussion in the components and circuits section about a particular aspect of the theory of this circuit that I don’t understand.