This "Recent Repair" was kindly contributed by Tim Pullin.
Bakelite AC/DC table radio
Introduced January 1950.
Valve line-up: UCH42, UF41, UBC41, UL41. Rect UY41
Features three push button presets and in-built aerial covering MW/LW
I was asked to look at this set after it’s owner had been saddled with it after a mix up at an auction sale. However he and his wife had found the look of this little set had grown on them, and they wanted me to see if it was worth repairing.
The report was “dead/not lit” . I wonder how he knew that....
The set’s owner kindly supplied a copy of the service data from the “Radio and TV servicing” series of books.
Condition is quite good, with no cracks or chips in the Bakelite, all knobs and push buttons present and even a decent length of PVC insulated mains flex. As is commonly found with these sets, the white trim surrounding the tuning scale is cracked. I don’t think the effect of heat helps.
The tuning scale was badly worn with a fair amount of paint and the wavelength markings missing. The Bush logo was also missing from the oval above the grill.
Inside the set was nice and dusty, and
looked complete. The only signs of previous work was a replacement “Radiospares”
on-off/volume control and the tuning knob shaft which had been glued after being
broken sometime in the past. I thought it best leave the knob in situ rather
than trying to remove it, since the knob does not need to be removed for servicing
purposes. I discovered later that the knob was actually glued to its shaft.
In addition to the usual pre-mains checks, a few basic tests were carried out just to make sure the job was worth starting.
In order to save space inside the cabinet the DAC10 uses a very special speaker, which has the magnet and voice coil on the inside of the cone. If this or the output transformer was faulty it would cost more to repair or obtain a replacement (if one is even available) than the owner was prepared to pay.
The mains plug and cord were examined, both were found to be in good order but the 13-amp fuse found inside was changed for a 3 amp. A test across the mains plug showed an open circuit, so it was off with the back cover. The chassis looks a little unusual at first glance but it is just a standard 4 valves + rectifier Superhet with a similar circuit to other Bush sets. It can be placed on its back, front or end while servicing.
The mains flex was continuous to the switch, and the continuity reading across the “set” side of the mains switch (dropper to chassis) was around 1200 ohms as expected. The heater chain was at least continuous, a fact confirmed by removing one valve and watching the resistance reading drop back to open circuit.
The chassis is easily removed, but do watch out for the aerial coils when removing or replacing it. The coils on my set were loose and became jammed between the fixing brackets and case during removal. Careful easing with a plastic tool freed them, the problem caused by loose coil mounting screws.
While I was testing the switch I noticed the mains was connected the wrong way round. The black (Neutral for our U.S and younger readers) wire in the mains cable would have been switched through to the voltage selector on the dropper resistor, while the red(live) would have been switched ,via the dial lamps to the chassis, rendering it LIVE in use. This probably wouldn’t have mattered too much to the user, as the knobs will not pull off when the set is fully assembled and the chassis is well enclosed in Bakelite, but think of the poor radio repairer! I am often amazed but never surprised by the things one finds. Yet another justification for pre-mains checks.
The wires were cut away from the faulty switch and the mains flex temporarily connected CORRECTLY to the set using terminal blocks until a new switched volume control was fitted. The existing wax paper mains filter capacitor tested leaky at about 1MegW, so a new 0.047uF 1000v capacitor was fitted, across the mains at the connector block.
The primary connections for the cabinet mounted output transformer are by screw terminals, so it was off with those, and the audio output from my trusty Advance E2 signal generator applied to the transformer connections. To my (and the owners) relief, a tone was heard through the speaker.
Next, attention was turned to the main smoothing capacitors, which looked in good order, having no bubbles or bulges in the rubber near the terminals. Fortunately the open circuit mains switch would have prevented any damage to the set had it been plugged in. Both sections soon reformed with the Megger, but testing into the H.T line showed a leak to chassis of 20K or so.
With a meter set to monitor the H.T voltage, the set was plugged in via the isolating transformer, Variac and safety lamp, and the voltage slowly increased. The mains had almost reached normal before any signs of life were present in the set. The scale lamps were not working, but valve heaters were glowing normally. HT was only about 120v, and while I was wondering why this might be, a firework display suddenly started in the UY41 rectifier, and the safety lamp flashed and flickered. Off went the mains! Since it was likely that the rectifier itself was faulty it was replaced with another good used UY41.
This time H.T rose to around 170v, the output valves screen grid ( and the HT to the rest of the set) reaching 80v and hiss was heard coming from the speaker. A few stations were able to be tuned in but were quiet. The UL41 output valve’s grid coupling capacitor was leaky, allowing about 3 volts DC on the control grid. Nearly all the other valve electrode voltages were low too.
It was also likely the leak across the H.T was down to the wax paper capacitors. All but one had bad leakage of less than 10k. The resistors used in my chassis are those fat white ceramic ones with the conical or sometimes rounded light brown ends. I have found these to be fairly reliable, and are usually found to be within tolerance.
Performance was much better with all the waxies renewed with HT at almost normal, and the little set gave plenty of volume and a good assortment of stations up and down the MW band. The dial lamps were replaced at this stage as this would make a slight difference to the heater voltages, and of course provides an “ at a glance” indication the set was switched on.
A reading of approximately 1.5 volts was measured at the UL41’s control grid, which was still present even with the new grid capacitor. These valves do have a habit of developing internal leaks/shorts. I did toy with the idea of replacing the troublesome valve with a UL84, (make up an adapter and modify cathode and screen grid resistances accordingly) as I did not have an UL41 amongst my valve stock.
Replacement UL41’s are getting expensive these days. Fortunately I acquired a box of valves from a car boot sale around this time, which included a couple of UL41’s. Both tested OK with the AVO Mk2 valve tester, so one of these was pressed into service for this little set. No DC voltage was found at the control grid of the replacement valve, proving the original valve faulty. Being a stingy soul I kept the faulty valve for an emergency spare.
Volume was fine upon re-application of the mains, and stations received across both bands but the sound distorted at low volume. The signal at the secondary of the output transformer looked OK on the oscilloscope, so any problems were likely to be in either the transformer or the speaker. My basic knowledge of Murphy’s law made the rather special speaker used in this set the chief suspect. It’s a special part unlikely to be found anywhere else, so it’s bound to be faulty isn’t it? Cynical maybe, but often true!
This was duly proved by disconnecting the original speaker and running the set into a spare loudspeaker. It sounded fine, the diagnosis confirmed by testing the DAC10 speaker on the output of VHF set (A Murphy 242, for the curious). Distortion was again present at low volumes.
Since it was unlikely I would find another, at least not to make the repair economic, I decided to examine the unit I had. The dead spider and other debris found inside the cone was brushed and vacuumed out as much as possible. The distortion still persisted though, although I noticed it became less severe when the speaker was handled, disappearing when the front plate was squeezed tightly to the speaker frame.
As one does not want the heads protruding from the speaker flange and cracking the Bakelite case as the fixing screws are tightened, countersunk 4 BA screws, nuts and star washers were used in the unused mounting holes to clamp the two halves of the speaker firmly together.
Although of course these nuts were visible, they would not be too conspicuous when the chassis was replaced, and resulted in a perfectly working speaker so I was prepared to forgive their appearance.
A new switched volume control was fitted, ensuring the mains was connected CORRECTLY to the double pole switch, and the new 0.047uF 1000v mains suppression capacitor fitted between switched live and chassis.
The only suitable control I had was a 1 Meg Log switched pot from R.S. Since this was a customer repair I used a new part. The original had a 500k track so a 1meg 0.5 watt resistor had to be soldered across the track of the replacement control to reduce its value to 500k. The replacement had a plastic shaft of course, but this is no bad thing with an AC/DC set.
Performance was reasonable but I felt it could be better. The set responded well to alignment, (this being a “late” model has an IF frequency of 470 kc/s) performance much improved and stations were found to be spot on when all was re-assembled.
After the dial lamps and crumbling rubber wiring had been renewed the chassis was tested extensively, no other problems showing for the rest of the day.
Case and Fittings
As previously mentioned the tuning scale markings were in poor condition. These sets do get quite warm in normal use, but I think the usual cause of such damage is damp storage.
The customer wanted it improved if at all possible, so I started on my first tuning scale restoration. Fortunately there is a scan of the DAC10 scale available for download on the web. That was the easy part!
The image was then REVERSED, because it was being printed on the back of the transfer, and the side presently nearest the paper will of course be stuck BEHIND the glass. Once the image was safely stored on the hard disk (AND floppy) it was printed out on plain paper a few times to check the size was correct. Slight adjustment was required until the printed image was exactly the same size as the glass and the lettering corresponded exactly between the original and copy, confirmed by laying the original scale over the print and check for blurring or double letters.
I eventually tracked down some waterslide transfer paper locally, although it was only suitable for laser printers or photocopying. (And cost £2 per A4 sheet!!) I tried an ink jet print but the print came out in bobbles that never dried. Fortunately I was able to clean this ink off, and re-think my plan.
Salvation came as I was walking to the bank one lunchtime. I passed a copy-shop, with a board outside advertising the fact they made copies from disk. I stuck my head in and was told they could provide me with a colour photocopy on my special paper from the image on my floppy disk. In fact they duplicated it three times on the same sheet (well, it is my first go!)
I was toying with the idea of placing the transfer over the existing worn scale, but was unsure whether the old paint would provide a reliable base. The paint began to flake off as I was cleaning up the scale so I thought it best to remove the old scale from the glass.
The glass tuning scale was laid flat on a few sheets of newspaper, and a filling knife/scraper used to carefully remove the old paint. The glass was then washed in a bowl of warm soapy water to get it really nice and clean.
I am not necessarily recommending this method because there are probably many people who could suggest a better one. I am merely describing the method used, and you are welcome to try it at your own risk! I’m not really a cabinet man, as far as I am concerned the electronics are the easy bit!
The finished result was a clean piece of glass, which had to be handled by the edge to prevent fingerprints being preserved forever beneath the transfer. Fingers crossed, it’s too late now! After it was carefully cut out, a transfer was floated in a tray of clean water and when ready the transfer peeled away from the backing and applied to the glass.
The first attempt wasn’t bad but seemed rather transparent when tried in the radio.
I talked over various solutions with my wife Jacquie, who is very good with things like this. She even offered to go over the back of the scale and “re-enforce” the black areas with a fine marker pen. Eventually we agreed it was worth trying another transfer applied over the first. Actually this worked very well, (the result can be seen in the picture. Shame about the flash!) but I was anxious to avoid any air bubbles between the layers unsure of what might happen as the set warmed up. The set has only been tested it for four or five hours so far, so I will let you know how it lasts.
The speaker, output transformer and the volume knob were stripped from the Bakelite case and both halves cleaned in a bowl of soapy water. An old toothbrush was used to shift more stubborn muck and cobwebs after which the pieces were dried and polished with Brasso. The back cover came up in a nice mottled pattern, while the main part remained plain brown. I don’t know if this is normal or if one half of the case has been replaced at some time.
The crack in the tuning scale surround was only small, so I thought it best to leave it as it was. Such cracks are common with this model and my theory is they are caused by the different expansion rates of the Bakelite case and the plastic making up the surround. The surround is screwed into the Bakelite, and as the case expands the surround is stressed until it cracks at a weak point.
The three presets were tuned to the stations requested by the customer.
When first assembled there was a louder than expected hum, which remained at a constant level irrespective of the volume control setting, so I concluded it was probably originating in the output amp or power supply.
Levels of ripple on the HT looked fine on the scope, but there was a significant 50Hz at the anode of the UBC41 triode (but NOT at the H.T). I noticed the hum level changed as I brought the oscilloscope probe near the UBC41 grid connection, and moving the mains wiring away from the grid side lead of C20 (0.01uF, grid feed from wiper of volume control) cured the hum. The mains wiring was tied back out of the way with some waxed twine.
I gather that originally all three presets were for MW, but this was changed in later models to provide one LW and two MW presets. Some alterations to the value of the rectifier series and the dial lamp shunt resistors were also made. This set is something of a paradox, as it has the very early serial number of 8691. However it has one LW preset, the modified resistors and an intermediate frequency of 470KHz, which one would expect to find on later models after serial number 39501.
These little sets were often used by the "Wireless for the Blind" organisations as their pre-set buttons meant they were easy to operate for blind or partially sighted users.
There was some case resonance when the
set was being tested with all the gubbins out on the bench, but this almost
disappeared when the set was fully assembled, only audible at high volume. Tone
is surprisingly rich for a small Bakelite box. Performance while nothing special,
is reasonable considering the built in aerial, with lots of stations all across
Text and Photographs Copyright © 2004 Tim Pullin.