15-02-2025, 05:38 PM
Hi everyone
For my first post here, I thought you may like to hear the story of a rather sad GEC AC5 which now has a new lease of life.
The owner of the set has kept it in semi-dry storage since acquiring it some twenty years ago from an elderly relative. Unfortunately, it suffered significant damage to the cabinet from a water leak; most of the electronics were spared, though. He told me that from time to time he has shown people that it ‘works’ but that it now only hums. Our conversation came about because he wants to use the set to support WWII themed social events. So, the challenge was to make it safe, serviceable and reliable. It needed to look good but not necessarily perfect, sound good, be robust enough to transport and to ‘just work’ with minimal attention. For some target events, a low power AM transmitter is used, but for others, Bluetooth capability is needed.
Restoration decisions were made with the above expectations in mind. The set was not in original or even 'good enough' condition for preservation of originality to be a consideration. What follows is the summary of the work prepared for the owner:
Initial assessment
Externally, the cabinet is in poor condition with damp-related delamination and missing veneer to the right side, The chrome escutcheon round the tuning dial shows damp-related pitting. All front-facing knobs are present, original and largely undamaged with the usual amount of surface dirt. The rear-facing tone control knob has chipped edges. The shafts of the various controls are rusty. The rear cardboard panel is present and in good condition. The aerial wire is present but in unserviceable condition. The original power connector is present but does not meet modern safety standards.
It was decided to ‘tidy-up’ the cabinet rather than attempt a full restoration. The delamination was treated with PVA, glued and clamped to make sound. Existing veneer was glued where necessary and edges smoothed. The exposed plywood on the damaged side was coloured with walnut stain to more closely match the rest of the cabinet. The original cabinet finish was cleaned and sealed.
The power connector was replaced with a C14 IEC panel-mount inlet. Internal connections were insulated with heat shrink and an embedded 1A slow-blow fuse was incorporated within the shrink-wrapped connection to the mains switch (which was replaced with a double pole switch, switching both live and neutral lines). Earth is connected to the chassis.
The aerial wire was retained but a new flexible ‘tail’ was added to the plug end.
Internally, the cabinet is dirty but undamaged. The speaker cloth has previously been replaced with an (estimated) 1970s common speaker fabric. It is glued in and held with drawing pins. The high voltage cables to the speaker field and output transformer windings are in poor condition representing both a shock hazard (240V at the OP TX) and a fire hazard and will need to be replaced. The high voltage smoothing choke and one high voltage smoothing capacitor have been relocated to a void in the cabinet (they were originally housed securely within the earthed chassis). Because the cabinet back can be easily removed by swivelling three clips (intended to make valve replacement easy), this represents a shock hazard (approx. 75V across the choke and 240V across the capacitor). A solution needs to be found to this (see also comments about the catkin output valve, below). Probably an alternative method of attaching the back cover in a tamper-proof way.
The flexible, multi-strand cables were in poor condition and were replaced, but the solid-core wires were found to be serviceable. Several tags on the speaker connection tag strip were either loose or detached and were resecured. The smoothing choke and capacitor were discarded. The PSU circuit was returned to its design specification with the exception of the choke which was replaced by two 220 ohm wire wound resistors within the under chassis space.
The cabinet back was modified to accommodate three fixing screws. However, all easily accessible high voltages have been removed from the cabinet void or insulated.
The chassis is superficially in good condition with no surface rust apparent. All valves test good for heater continuity. Two valves required rebonding to their bases and one required rebonding of the anode cap. The grid cap and anode cap wires are all unserviceable and will need replacement. The output valve is a Catkin (exposed anode) valve with aged enamel and is thus a shock hazard (nominal 228V). It is easily accessible if the back of the receiver is removed. The tuning mechanism is dirty but operates smoothly. One trimmer on the tuning gang has been unscrewed inappropriately. The tuning scale is badly damaged, possibly by a combination of damp and overheating by the incorrect panel lamp bulbs. It may be possible to stabilise or reproduce. The panel lamp bulbs will need replacing.
At first power-on after restoration, although the mains current draw was not excessive, the HT current draw was nearly twice expected, mostly to the output valve. There was sound for about thirty seconds, then a gradual reduction of volume until silence after two minutes. It is likely that the output valve screen grid was red-plating (can’t be seen in a Catkin). This would be consistent with damage caused by the chronic decline and failure of the power supply noted later. The output valve was replaced with a N42 pentode and operation returned to normal with nominal voltage readings per service manual. The N42 is a glass envelope valve, so its use removed the shock hazard inherent in the original Catkin valve design.
Panel lamps were replaced with bulbs of the correct specification. It is noted, though, that they are significantly under-run (2V vs. 3.5V) and are thus quite dim. This may be why incorrect spec. bulbs had been used previously.
There is evidence of probably four historic repairs. Three repairs are to the power supply circuit (this is not unusual). There is evidence of repair in the 1950s involving the renewal of two of the smoothing capacitors and the smoothing choke. However, the original capacitors were not removed, they were simply bridged, and the replacement choke is of the wrong value (d.c. resistance much too high). There is then a second repair late 1950s/early 1960s when a further smoothing capacitor was replaced with the original being disconnected but left in place. The rectifier valve was replaced probably at this time, but with the incorrect type. It is a serviceable replacement, though. An additional smoothing capacitor was added, probably in the 1970s. It is of an inappropriately high value and was probably an attempt to remove hum. Dates estimated from the date of manufacture of the capacitors. The PSU will be returned to design configuration.
There was found to be evidence of a ‘repair’ to the power switch which had failed internally. It had been bypassed by soldering both wires to the same terminal, thus the radio was switched on simply by connecting the mains lead to power. The switch was replaced as described, above.
The fourth repair involved the replacement of a low-voltage electrolytic capacitor that decouples the output valve. This is to be expected in a set of this age.
On dismantling, this capacitor was found to be original but to have failed.
None of the above is unexpected and will be remedied as part of the rebuild. There are two areas where a decision will be needed, though. The cabinet: do we make it look as good as possible without veneering it or do we replace the missing veneer? It’s a big piece. The tuning dial: do we do nothing and leave it as it is; do we attempt to stabilise it by coating it with a lacquer; do we make a reproduction?
It was decided that the only practical solution for the tuning dial was to make a reproduction. The original was scanned into Adobe Illustrator, and traces and separations made. These were printed to acetate with paper used as a diffuser.
Emerging issues
All tubular capacitors tested are d.c. leaky (as expected). About half resistors are more than 15% off spec. with some showing evidence of heat damage or damp damage. Three read very high with one open.
The power supply has been significantly reconfigured both physically and electrically. All the additional components will be removed and the circuit returned to design spec. Capacitor packs will be rebuilt with modern internals. All tubular and electrolytic capacitors will be replaced. All resistors will be replaced. This is to ensure reliability given expected regular use.
The original filter capacitor pack was rebuilt using modern components (8uF, 600V vs. 6uF, 450V). The smoothing choke was replaced by two 11W, 220 Ohm wire wound resistors in series, mounted within the chassis body (d.c. resistance of specified smoothing choke is 400 ohms). [use of 11W resistors is overkill. A single 2W resistor was initially used but overheated with the original, faulty output valve and was damaged]
Note that C19 (per GEC service manual) is a potential ‘death cap’. It connects the Catkin anode (228v) to the external extension speaker socket.
C19 was found to have low d.c. resistance, allowing the >200v from the output valve anode to appear across the extension speaker terminals. It has been replaced with a ‘X-cap’ safety capacitor, given the intention to use a contemporary, high impedance extension speaker.
The power switch has failed and has been bypassed.
The power switch was dismantled in an attempt to repair. (It was found that the switch hairspring had a fatigue crack and was weak. The mechanism had been further damaged by the knob being forced, bending the fork. It cannot easily be repaired)
The associated volume potentiometer is very high (7meg vs 500k). It is a conductive-paper-and-compression-ring type. Only suitable NOS replacement available is 10k log carbon track. Track is in a.c. audio line with no d.c. potential across it so 10k should work well.
The replacement switch is double pole so will switch both mains lines. A 1A slow-blow fuse has been incorporated in the live line between the switch and the transformer. This cannot be changed by the user (to ensure that in the case of failure, a technician diagnoses the reason it blew. This is to protect rare and expensive components. A further 0.5A slow-blow fuse has been similarly incorporated in the transformer secondary winding centre-tap line).
The 10k potentiometer was found only to alter tone and not volume when connected according to the original schematic. Since it is a carbon track potentiometer, the circuit was reconfigured to bring one end of the track to d.c. ground, the wiper being taken to the first audio valve grid via a capacitor. It then functioned correctly.
Speaker windings all test good, but the anti-heterodyne coil is intermittent – almost certainly a connection tag issue.
Found to be a fractured solder tag.
+ve HT line to output transformer found to be detached from paxolin support strip.
The ‘Tuneon’ tube does glow and gives some indication of the presence of a carrier. However, it is internally blackened. Replacements are very rare. Given how few radio stations are now broadcasting on the AM bands, there seems little value in replacing it.
A Bluetooth module was incorporated into the Gram circuit. 12V a.c. is derived from a double wound mains transformer connected in parallel with the original mains transformer and housed in the original location of the smoothing choke, within the chassis body. 12V a.c. is taken to a terminal strip inside the cabinet void. This is half-wave rectified and smoothed to deliver about 15V d.c. to the input of a buck converter. The buck converter delivers a stable 5V d.c. to a Bluetooth 5 module. The left and right channel audio from the Bluetooth module is summed to produce a mono output which is delivered between the grid and cathode of the first audio valve, via the volume control. It should be noted that the Bluetooth circuitry is electrically floating and signal ground is above chassis ground by around thirty volts. (This is important if connecting test equipment)
This is AM
This is Bluetooth
Enjoy!
For my first post here, I thought you may like to hear the story of a rather sad GEC AC5 which now has a new lease of life.
The owner of the set has kept it in semi-dry storage since acquiring it some twenty years ago from an elderly relative. Unfortunately, it suffered significant damage to the cabinet from a water leak; most of the electronics were spared, though. He told me that from time to time he has shown people that it ‘works’ but that it now only hums. Our conversation came about because he wants to use the set to support WWII themed social events. So, the challenge was to make it safe, serviceable and reliable. It needed to look good but not necessarily perfect, sound good, be robust enough to transport and to ‘just work’ with minimal attention. For some target events, a low power AM transmitter is used, but for others, Bluetooth capability is needed.
Restoration decisions were made with the above expectations in mind. The set was not in original or even 'good enough' condition for preservation of originality to be a consideration. What follows is the summary of the work prepared for the owner:
Initial assessment
Externally, the cabinet is in poor condition with damp-related delamination and missing veneer to the right side, The chrome escutcheon round the tuning dial shows damp-related pitting. All front-facing knobs are present, original and largely undamaged with the usual amount of surface dirt. The rear-facing tone control knob has chipped edges. The shafts of the various controls are rusty. The rear cardboard panel is present and in good condition. The aerial wire is present but in unserviceable condition. The original power connector is present but does not meet modern safety standards.
It was decided to ‘tidy-up’ the cabinet rather than attempt a full restoration. The delamination was treated with PVA, glued and clamped to make sound. Existing veneer was glued where necessary and edges smoothed. The exposed plywood on the damaged side was coloured with walnut stain to more closely match the rest of the cabinet. The original cabinet finish was cleaned and sealed.
The power connector was replaced with a C14 IEC panel-mount inlet. Internal connections were insulated with heat shrink and an embedded 1A slow-blow fuse was incorporated within the shrink-wrapped connection to the mains switch (which was replaced with a double pole switch, switching both live and neutral lines). Earth is connected to the chassis.
The aerial wire was retained but a new flexible ‘tail’ was added to the plug end.
Internally, the cabinet is dirty but undamaged. The speaker cloth has previously been replaced with an (estimated) 1970s common speaker fabric. It is glued in and held with drawing pins. The high voltage cables to the speaker field and output transformer windings are in poor condition representing both a shock hazard (240V at the OP TX) and a fire hazard and will need to be replaced. The high voltage smoothing choke and one high voltage smoothing capacitor have been relocated to a void in the cabinet (they were originally housed securely within the earthed chassis). Because the cabinet back can be easily removed by swivelling three clips (intended to make valve replacement easy), this represents a shock hazard (approx. 75V across the choke and 240V across the capacitor). A solution needs to be found to this (see also comments about the catkin output valve, below). Probably an alternative method of attaching the back cover in a tamper-proof way.
The flexible, multi-strand cables were in poor condition and were replaced, but the solid-core wires were found to be serviceable. Several tags on the speaker connection tag strip were either loose or detached and were resecured. The smoothing choke and capacitor were discarded. The PSU circuit was returned to its design specification with the exception of the choke which was replaced by two 220 ohm wire wound resistors within the under chassis space.
The cabinet back was modified to accommodate three fixing screws. However, all easily accessible high voltages have been removed from the cabinet void or insulated.
The chassis is superficially in good condition with no surface rust apparent. All valves test good for heater continuity. Two valves required rebonding to their bases and one required rebonding of the anode cap. The grid cap and anode cap wires are all unserviceable and will need replacement. The output valve is a Catkin (exposed anode) valve with aged enamel and is thus a shock hazard (nominal 228V). It is easily accessible if the back of the receiver is removed. The tuning mechanism is dirty but operates smoothly. One trimmer on the tuning gang has been unscrewed inappropriately. The tuning scale is badly damaged, possibly by a combination of damp and overheating by the incorrect panel lamp bulbs. It may be possible to stabilise or reproduce. The panel lamp bulbs will need replacing.
At first power-on after restoration, although the mains current draw was not excessive, the HT current draw was nearly twice expected, mostly to the output valve. There was sound for about thirty seconds, then a gradual reduction of volume until silence after two minutes. It is likely that the output valve screen grid was red-plating (can’t be seen in a Catkin). This would be consistent with damage caused by the chronic decline and failure of the power supply noted later. The output valve was replaced with a N42 pentode and operation returned to normal with nominal voltage readings per service manual. The N42 is a glass envelope valve, so its use removed the shock hazard inherent in the original Catkin valve design.
Panel lamps were replaced with bulbs of the correct specification. It is noted, though, that they are significantly under-run (2V vs. 3.5V) and are thus quite dim. This may be why incorrect spec. bulbs had been used previously.
There is evidence of probably four historic repairs. Three repairs are to the power supply circuit (this is not unusual). There is evidence of repair in the 1950s involving the renewal of two of the smoothing capacitors and the smoothing choke. However, the original capacitors were not removed, they were simply bridged, and the replacement choke is of the wrong value (d.c. resistance much too high). There is then a second repair late 1950s/early 1960s when a further smoothing capacitor was replaced with the original being disconnected but left in place. The rectifier valve was replaced probably at this time, but with the incorrect type. It is a serviceable replacement, though. An additional smoothing capacitor was added, probably in the 1970s. It is of an inappropriately high value and was probably an attempt to remove hum. Dates estimated from the date of manufacture of the capacitors. The PSU will be returned to design configuration.
There was found to be evidence of a ‘repair’ to the power switch which had failed internally. It had been bypassed by soldering both wires to the same terminal, thus the radio was switched on simply by connecting the mains lead to power. The switch was replaced as described, above.
The fourth repair involved the replacement of a low-voltage electrolytic capacitor that decouples the output valve. This is to be expected in a set of this age.
On dismantling, this capacitor was found to be original but to have failed.
None of the above is unexpected and will be remedied as part of the rebuild. There are two areas where a decision will be needed, though. The cabinet: do we make it look as good as possible without veneering it or do we replace the missing veneer? It’s a big piece. The tuning dial: do we do nothing and leave it as it is; do we attempt to stabilise it by coating it with a lacquer; do we make a reproduction?
It was decided that the only practical solution for the tuning dial was to make a reproduction. The original was scanned into Adobe Illustrator, and traces and separations made. These were printed to acetate with paper used as a diffuser.
Emerging issues
All tubular capacitors tested are d.c. leaky (as expected). About half resistors are more than 15% off spec. with some showing evidence of heat damage or damp damage. Three read very high with one open.
The power supply has been significantly reconfigured both physically and electrically. All the additional components will be removed and the circuit returned to design spec. Capacitor packs will be rebuilt with modern internals. All tubular and electrolytic capacitors will be replaced. All resistors will be replaced. This is to ensure reliability given expected regular use.
The original filter capacitor pack was rebuilt using modern components (8uF, 600V vs. 6uF, 450V). The smoothing choke was replaced by two 11W, 220 Ohm wire wound resistors in series, mounted within the chassis body (d.c. resistance of specified smoothing choke is 400 ohms). [use of 11W resistors is overkill. A single 2W resistor was initially used but overheated with the original, faulty output valve and was damaged]
Note that C19 (per GEC service manual) is a potential ‘death cap’. It connects the Catkin anode (228v) to the external extension speaker socket.
C19 was found to have low d.c. resistance, allowing the >200v from the output valve anode to appear across the extension speaker terminals. It has been replaced with a ‘X-cap’ safety capacitor, given the intention to use a contemporary, high impedance extension speaker.
The power switch has failed and has been bypassed.
The power switch was dismantled in an attempt to repair. (It was found that the switch hairspring had a fatigue crack and was weak. The mechanism had been further damaged by the knob being forced, bending the fork. It cannot easily be repaired)
The associated volume potentiometer is very high (7meg vs 500k). It is a conductive-paper-and-compression-ring type. Only suitable NOS replacement available is 10k log carbon track. Track is in a.c. audio line with no d.c. potential across it so 10k should work well.
The replacement switch is double pole so will switch both mains lines. A 1A slow-blow fuse has been incorporated in the live line between the switch and the transformer. This cannot be changed by the user (to ensure that in the case of failure, a technician diagnoses the reason it blew. This is to protect rare and expensive components. A further 0.5A slow-blow fuse has been similarly incorporated in the transformer secondary winding centre-tap line).
The 10k potentiometer was found only to alter tone and not volume when connected according to the original schematic. Since it is a carbon track potentiometer, the circuit was reconfigured to bring one end of the track to d.c. ground, the wiper being taken to the first audio valve grid via a capacitor. It then functioned correctly.
Speaker windings all test good, but the anti-heterodyne coil is intermittent – almost certainly a connection tag issue.
Found to be a fractured solder tag.
+ve HT line to output transformer found to be detached from paxolin support strip.
The ‘Tuneon’ tube does glow and gives some indication of the presence of a carrier. However, it is internally blackened. Replacements are very rare. Given how few radio stations are now broadcasting on the AM bands, there seems little value in replacing it.
A Bluetooth module was incorporated into the Gram circuit. 12V a.c. is derived from a double wound mains transformer connected in parallel with the original mains transformer and housed in the original location of the smoothing choke, within the chassis body. 12V a.c. is taken to a terminal strip inside the cabinet void. This is half-wave rectified and smoothed to deliver about 15V d.c. to the input of a buck converter. The buck converter delivers a stable 5V d.c. to a Bluetooth 5 module. The left and right channel audio from the Bluetooth module is summed to produce a mono output which is delivered between the grid and cathode of the first audio valve, via the volume control. It should be noted that the Bluetooth circuitry is electrically floating and signal ground is above chassis ground by around thirty volts. (This is important if connecting test equipment)
This is AM
This is Bluetooth
Enjoy!
