I don't wish to hi-jack the thread, but several Roberts sets seem to suffer this fate, and maybe other sets of that era too - it's to be expected after more than 50 years. Depending on which caps go short circuit, will determine whether it just flattens the battery, or harms the output stage.
A little while ago I had a Roberts R505 which has three caps in particular which can go dead short. Two will only flatten the battery as in the R600 which is the subject of this thread, but in the R505, if C40 (680uF) has a high leakage rate, the output pair (AC187/188 germanium transistors) will draw excess current and despite being heatsinked, will get very hot. If left in that state for long, they can be written off.
A classic symptom is that when the set is switched on, it works for maybe two seconds then goes dead.
It was covered in a thread here:
https://www.vintage-radio.net/forum/...t=Roberts+R505
If the electrolytics aren't tested, because the output pair get very hot, it would be easy to jump to conclusions and to assume that they've developed shorts. That could lead to fruitlessly removing them to test, and if we're lucky, they'll test OK, but if they're dead, and replaced with substitutes, they'll suffer the same fate too as the fault that caused their demise hasn't been diagnosed.
In my R505, C40 (680µF) on test was identified not as a capacitor, but as a 0.53 Ohm resistor - a dead short from the output transistors via the speaker speech coil to chassis. I tested the other electrolytics and though there were no other shorts, while the set was dismantled and on the bench I replaced the others as it may only have been a matter of time for them to suffer the same fate.
While I was at it, in checking the Robert's Service Data on the tests to be conducted on the audio stage, it stated: 'inject a 1kHz sine wave into S1E and with a 'scope across the speaker, adjust the pre-set pot R36 for symmetry at the onset of clipping'. I checked that, if only out of curiosity, and found that it was clipping quite markedly so I tweaked it.
Firsts pic shows part of the R505 output stage with the three electrolytics that can cause problems, notably C40, which I guess could also damage the 4 Ohm speaker speech coil as that, (rather than the 330 Ohm parallel resistor), would take most of the current until the output transistors became 'three legged fuses'.
Second pic shows C40 under test, identified by the tester as a 0.53 Ohm resistor.
Third pic is the scope trace of a 1kHz sine wave before adjustment of R36.
Fourth pic is the scope trace after adjustment.
Very pleased with the performance of the set after the repairs and tweaks.
I appreciate that not everyone has a 'scope and signal generator, but lots of us do, and maybe we should use them more often - I know I should!
Hope that might help someone.