Roberts R600: Fades

[Mike Watterson]

Sometimes freezer spray works, though I've not used it since the early 1980s when I was last doing servicing in the day job.
It definitely slowly comes back on its own if it is left on. I'm wondering if it's an electrolytic somewhere in the audio amp. It was fine outside the box after I resoldered the track at the volume control's power switch.
It doesn't seem to cut out or fade when it's been running.

[Murphyv310]

Can you monitor the current drawn and see what happens when it fades out?

[Mike Watterson]

It only ever fades out straight after turning on. At least I think so.

Certainly monitoring the current is a good idea, as would be perhaps the voltages at some points with soldered on probe wires so it can be in the cabinet.

Two versions of the amplifier, hopefully only differing in layout. R4 and C12 (470nF) use a tag of the unused pole of the power switch, R4 being from the tag to the PCB. Initially this was puzzling as the actual switch contacts in use are at the PCB level and harder to see.

There are five electrolytics in that part of the radio. Perhaps C19 or C20 are the most likely culprits if it is a dodgy capacitor. Maybe if it's a capacitor reforming the fault will go away simply by running it.

I ran it over an hour this morning, it did come on, fade out and then after a minute or so gradually returned, a little distorted at first. Then it was fine. I turned it off for 15 minutes and just now turned it on and it's fine.

The Trader & R&TVS seem to use the same designations on this part.

[Murphyv310]

Ah Lokfits..... There's where the trouble lies!!!

[ppppenguin]

Just change the Lokfits and electrolytics and quite likely all will be well.

I've got a load of NOS Lokfits. They are probably as bad as used ones.

[Murphyv310]

I binned all my Lokfits years ago as so many were intermittent.

[Murphyv310]

Duff Lokfits respond to a little heat and a blast of freezer in confirming their status....... Slightly dud, badly dud or totally dud.
Have fun!!

I remember the Philips 210 TV chassis and you could go on a merry dance with them. You'd get a set on the bench with intermittent sound or vision and the customer saying that if you change channel the fault would clear. So the TV would be running away with the PCB exposed and the scope on and ready to go. When the fault appeared as soon as the scope probe was put on virtually any component in the vision or sound stages it would clear.
Eventually we only did a full blanket repair, 90% success rate that way and far less time involved.

[Mike Watterson]

Now that I'm convinced it's not the power switch or the wave change I've been simply gently clicking it on to normal volume.
1) If it's been off for a while there is always an instant of normal sound and it fades to nothing, but in a few minutes gradually comes back, distorted at first.
2) If it's been running recently, then it turns on normally,

So I looked at the current just clicking on. It would be a little over 18mA and drop eventually to a bit under 17mA. At 16.8mA it's always working when the volume is turned up.

Next I took it out and found the jumper wire. Someone has been here before. Lots of burns on wires. The insulation split back on the jumper wire. The service info says 3.5mA and it was 2.5mA, at least it wasn't loads high. Adjusted to 3.5mA with volume at zero.

The capacitors are a pig to get out. The tracks poor adhesion and flimsy. Also neither PCB, R&TVS or Trader note polarity, so mark on the diagram before you take one out.
I shorted cap and tested on 9V on PSU with meter in series to see leakage. Then shorted out twice, to be sure to be sure and tested on the magic Chinese tester.
C20, 16 uF was OK, a bias decoupling.
C23, 640 uF to speaker was OK (I'd not expect it to cause this issue)
C15, 400 uF IF amp psu decoupling was OK
C21, 200uF 6.4V seemed dubious. It would take 1mA rapidly dropping to microamps, everytime it was shorted. ESR 1.9 Ohms, but capacitance OK at 230uF. I replaced it with ancient 220uF 16V that passed tests OK.
C19, 16 uF preamp supply decoupling was OK.

On testing the set the zero volume quiescent current is now 13.5 mA rather than 16.8mA. I'll leave it for a couple of hours and see if this was the fix. I didn't look at the balance preset as when it's good, there is no distortion even when loud. I may check it tomorrow if the repair is real rather than imaginary.

Back in the 70s & 80s I hated the intermittent fault and so did everyone else. Especially if it was in the Belfast workshop and from Sligo or Letterkenny and it was a N1500. I think sometimes we swapped with something not intermittent nearer to base. It's also good if the faulty part goes on fire.

I hope this is the fault, I'm not confident, but it's more like a capacitor symptom than a transistor or dry joint, though there isn't even a his at full volume when it's dead, though taking over 17 mA.

[Mike Watterson]

As I suspected, the C21 was only dubious, not the fault. It's still there.
Next I'll replace the two BC148, because they were a stupid idea and inferior to TO92. Likely they are fine, but badly fitted so one might later suffer an internal fracture. They need a precision PCB.
I'll put meters on three or four likely places and scope and/or amp on the IF output to confirm supply points and that the fault is in the audio amplifier and not the IF module. The zero volume quiescent current seems to be 13.5 mA or 16.8mA when it's working. Perhaps I'd had it on AM vs FM or something.

Maybe tomorrow. I've other more important things to do.

[Mark Hennessy]

One thing that I don't think has been mentioned yet: AF11x

That Mullard module has 3 of them in there - an AF114 and a pair of AF115s usually, but variations are possible.

As we don't appear to have figured out whether this is an AF or RF/IF fault, it is possible that an intermittent tin whisker could be a contender, especially as physical shock from the wavechange switches makes a difference. Also, whiskers can cause leaking - as opposed to dead shorts - which can cause all sorts of bizarre symptoms. I usually eliminate those first, and many extremely obscure faults have been cleared that way.

My first move would have been a 'scope or powered speaker across the volume control to decide where the fault actually was. Certainly, knowing how bad these PCBs are, I wouldn't have chosen to disturb the caps until I had good reason to be suspicious of them, such as actual measurements of in-circuit conditions. They should be Philips blue axial types. Very, very occasionally these can fail dead short, but if that doesn't happen, they seem to plod on forever in cool-running transistor radios. I always power transistor radios from a current-limited bench PSU, which makes it easy to keep an eye on the operating current and avoids collateral damage from shorted caps (I'm sure that would have been the case here, but it's surprising how many repairers don't, so it's a point worth mentioning).

Personally, I also wouldn't replace the Lockfits until I've got in-circuit measurements that confirm their guilt. In my experience, their normal failure mode is noise - if this is in a part of the circuit where it doesn't matter, then you might never know about it. Complete failures happen, and probably more so than the equivalent BC108 or BC548, but later samples were much better in my experience. Obviously, that might be different for TV sets where the operating conditions and applications are different, but I've almost no experience there - but in transistor radios, they are definitely "innocent until proven guilty" IMHO.

Another thing that hasn't been mentioned: headphone socket. They normally respond to a bit of switch cleaner. It's easy to forget about these, especially when it feels like the fault is more complicated, but it could well be choosing to join in with the "fun"

Just one final thought: there's a chance the IF module was built with AF12x transistors. Certainly, the LP1164 used in the R707 moved to those late in production (identifiable by "02" on the label). I don't know offhand if the LP1171 got the same treatment, but if you open up the module and find factory-fit AF12x types fitted, then that could be a useful data point. And if so, a photo of the labels on the case would be useful to see if they identified it in the same way as the LP1164. But if present, we'd need to be sure that they weren't fitted during a previous repair before deciding.

Good luck!

Mark