Philco A547B

This "Recent Repair" was kindly contributed by Tim Pullin.


Bakelite cased four valve (+ rectifier) superhet. Circa 1947. Valve line-up 6K8, 6K7, 6Q7, 6V6. Rect R52 (5Z4G)

This rather imposing, and to my mind attractive, Bakelite cased set was brought in by a work colleague for repair. It had been hibernating in his loft for many years and he wanted it working for use in his holiday cottage. The case was in good condition, with no chips and only a few surface scratches. The innards appeared complete and original, with the exception of a 6J7, which had at sometime been substituted for the 6K7G. Two spare valves (a 6K7G and another 6V6G) were found rattling around loose inside the cabinet.

Even the fabric covered mains cable was of a decent length and in good nick and although the cores were actually PVC insulated I thought it prudent to upgrade it to a 3-core to provide an earth. This was fine by the customer. The chassis is easily removed and incorporates the loudspeaker (along with its transformer), and is extremely easy to work on and clearly laid out.

Thanks to Mike Horne for the use of the picture.


The usual checks revealed the mains switch, mounted on the back of the tone control, to be open circuit. Interestingly it switches the mains neutral side of the incoming supply, meaning the voltage selector wiring would be permanently live, a fact I found a little worrying to say the least. I made a note to modify the wiring when the mains flex was changed.

The HT capacitors did not respond well to the ohms range of the multimeter, giving virtually no kick at all, so I was a little dubious as to whether they would reform. Nothing sinister was found with the megger on the mains side and the H.T to chassis reading was also ok. The megger was also connected across the mains switch which must just have had dirty contacts, as it read ok after a couple of "belts". This is sometimes worth trying particularly if a set hasnít been used for some time, and certainly saves replacement especially if the switch is on the tone control; the value may not be standard and an exact replacement difficult.††

The output from the rectifier was disconnected, and the set switched on. One scale lamp lit immediately, followed a few seconds later by all valves. A reading of +260v was obtained at the cathode of the rectifier, which meant it could be used (along with a meaty 100k-ohm resistor) to reform the capacitors.

With the set switched off, the connecting wires at the terminals of the smoothing electrolytic were removed (noting their positions carefully of course!) and the tags cleaned. Both halves of the capacitor reformed after about 25 minutes, with almost no drop across the resistor.

Before switching on the set I thought I would check out the valve fitted in place of the 6K7. While my valve data book identified the 6J7 as a pentode (and though has an identical pin-out to the 6K7) it is not a variable mu type, and very probably may have worked but the AGC would not have operated correctly.

I don't have many 6K7 valves, so I decided to give the "spare" one I found in the case a try. It's top cap was missing but there was enough wire protruding to enable a short length of 0.5 mm dia wire to be wrapped around the stub and (quickly!) soldered on, leaving a tail about an inch long. The top cap connection in the set was temporarily clipped to this wire and the set switched on.

This time both dial lamps lit (there proved to be an intermittent connection, further borne out by the blackening of the glass in the bulbs) and the set came up with some whistling and popping from the speaker. A quick twiddle of the tuning knob brought in plenty of stations but they were very quiet and distorted. The fault was not in the output stage though, as a test probe touched to the grid of the 6V6G produced a loud hum and cloud of dust from the speaker!

I had wondered if the output valves grid capacitor was leaky, putting a positive voltage onto its grid. No, amazingly the T.C.C wax paper capacitor was ok, leaving about -3v at the grid. Wondering whether to replace this capacitor as a precaution, I measured the voltage it was blocking (at the anode of the 6Q7G) to give an indication of working voltage for the replacement and found no volts at all.

Hmmmmm. There was normal H.T at the top of the 220k-ohm anode feed resistor to the triode, but none reaching the anode pin. This resistor proved to be open circuit, and a replacement restored normal operations, and full window rattling output. They're not all that easy, but at least it saved a trip to the library for the circuit diagram!

In use this resistor only drops around 60v, so there is no problem with excess current. It's one of those little brown ones with rounded ends and no tolerance band. I have had a few faulty over the years, so it's a good thing to bear in mind and check any you find, especially high value types which seem more prone to going high/open. This fault may well explain the valve changes, and the "spare" output valve, presumably substituted in an attempt to clear the fault.

The set worked well on all wavebands with about a 10' aerial.

With a working chassis all that was left was the case and a few small details.

The original mains flex was changed for a three core. I have a number of lengths of black 3A flex with moulded plugs (cut from scrap modems and the like) and besides black looking more authentic, the moulded plug seems to give the customer confidence. The new cord was secured using a "P" clip attached to the nearby transformer mounting bolt. The wiring modifications mentioned earlier were now carried out.

The live is now taken directly to the switch, and then to the transformer via the voltage selector. Normally I would have bypassed the selector, as it is not needed these days, and wired the live straight into the transformer, not forgetting to disconnect the other primary tappings as they would have acted like an autotransformer. A good idea is to cut them at the selector tag and bend the last half-inch back onto itself and sleeve it. Then the selector can be reused again if required. However I could not do this but for two reasons. One, the transformer has flying leads (not tags) and the voltage selector provides a handy joining point, and secondly the customer wanted to use the set in his holiday cottage in France, so required the 220v tap.

The neutral is now connected directly to the other end of the mains primary, using an insulated crimp connector; itself covered by a rubber sleeve. The earth connection was made to a handy solder tag. ( I always check earth continuity plug-chassis at this stage.)

The 6K7G with the missing top cap was treated to a replacement of the same size taken from a dud PY500 (I don't do tellys any more!) and with the solder sucked away from the top of the cap, it was carefully threaded over the extended wire. The glass "nipple" on the valve was smeared with Araldite, and was left to harden before the wire was soldered to the top cap.

The 6K7G was also slightly loose in its base so it was "caulked" with some more Araldite. A good thing to use is the rounded end of a plastic cable tie. Its stiff enough to apply the glue, but not hard enough to damage anything. And of course there's no need to bother cleaning it, just throw it in the bin! This might seem a bit Heath Robinson, but the customer had asked me to keep the set as original as humanly possible.

Lastly the dial lamp bulbs were removed and the end pips dressed up with a blob of solder. Quite often the soft metal pips get flattened in use which leads to bad contacts. The bulbs can often be "rejuvenated" with a quick dot from a soldering iron.

With the chassis complete, it was left to soak test for a few hours while I cleaned up the case, just in case any underlying fault had caused the demise of the resistor.

The case came up very well after the usual clean with "Brasso", revealing a pleasing mottled brown effect. All the octagonal knobs shined up well too, though I did not attempt to clean the centre metal discs, fearing them to be lacquered.

This set has an interesting waveband indicator. A string attached to a pulley on the waveband selector switch moves a small cursor. The red portion of the cursor shows through some " windows" on the tuning scale to indicate the selected band. The string had slipped off the pulley a few times during testing, so was shortened slightly to renew the tension.

The cursor had to be re calibrated so was moved a little on the string. This done the set was replaced in the cabinet, knobs refitted and tested again. No problems showed after about five hours.††

All in all I was very pleased with the finished set, shame I had to give it back...

Text Copyright © 2001 Tim Pullin

This website, including all text and images not otherwise credited, is copyright © 1997 - 2006 Paul Stenning.
No part of this website may be reproduced in any form without prior written permission from Paul Stenning.
All details are believed to be accurate, but no liability can be accepted for any errors.
The types of equipment discussed on this website may contain high voltages and/or operate at high temperatures.
Appropriate precautions must always be taken to minimise the risk of accidents.

Last updated 14th April 2006.