Roberts R505, am I going mad ?

[martinhog]

Good morning,
The trusty workshop Roberts R505 packed up a few weeks ago, the power supply (running on a wall wart) starting to trip with the audio pulsing. A quick check showed it was drawing excessive current with one of the output transistors running very hot.
As the output transistors were the originals,I changed them along with C40 and C38. I set the mid rail point and the quiescent current , all was well and the radio played on the bench, disassembled, for hours . I put it back together only to be greeted by the same fault, although it played for a few seconds before failing . I took it apart again and in its disabled state, it works perfectly. I re-assembled it slowly, checking for shorts etc, switched it on and the same thing happened again.
Has anyone had this before ? The output transistors were “new old stock” and the caps were brand new.

All answers appreciated.

Regards
Martinhog

[jonnybear]

Try tapping the output transistors removed from any heat sink, could be tin whisker syndrome have had it before, check any leak between E B C to casing on transistor. If the BC transistors are the Mullard squarish type these also give problems.
John

[ajgriff]

When testing (twice) with the chassis out of the case were you using the wall wart as the power source?

Alan

[vinrads]

I too have been caught out with the tin whisker syndrome , if this is the problem you can still use the transistors just isolate them from the heatsink. Mick.

[Ambientnoise]

Indeed, if the transistor heat sinks are isolated when the chassis is out of the case but connected when assembled, you can get this exact problem !

[unitelex]

Could be a hairline crack in a pcb trace or dry joint which is upsetting the bias for the output stage. Disassembled vs assembled can change mechanical stress on the pcb opening up the joint.

Assume you have checked the wall wart is providing the correct voltage, if this PSU uses an iron transformer, its regulation might not be so good, the output voltage could be quite a lot higher than what is stated on the rating plate when under light load.

Best Regards
Chris

[martinhog]

Thanks everyone,I will check tomorrow. The wall wart is a switch mode device with an output of 9 volts. I get the same results using a battery pack. The replacement output transistors are square bodied, the originals being round. The transistors were bolted to the heat sink when disassembled too. I will check again tomorrow and get the magnifying light out !

Thanks again.

[martinhog]

Update: With the radio still assembled, I switched it on with batteries connected, it worked perfectly for 6 hours . I checked the output voltage on the wall wart , it was 9.0 volts., I plugged it into the Radio and the fault appeared again. Back to batteries after a few minutes and it worked well again. A check on a ‘scope showed the wall wart output had a few noises spikes on the output. It had worked on the Roberts for years, however it seems that the noise is now upsetting the radio ? . Thanks for all your suggestions .

[Bob_Moss]

Take a Dremel to the wall wart & change the capacitors, its worked for me in the past.

[Dickie]

Stick a big electrolytic across the wallwart output?

[unitelex]

Originally Posted by martinhog

Update: With the radio still assembled, I switched it on with batteries connected, it worked perfectly for 6 hours. I checked the output voltage on the wall wart, it was 9.0 volts., I plugged it into the Radio and the fault appeared again. Back to batteries after a few minutes and it worked well again. A check on a ‘scope showed the wall wart output had a few noises spikes on the output. It had worked on the Roberts for years, however it seems that the noise is now upsetting the radio ?. Thanks for all your suggestions.

Most likely the capacitors within the wall wart have dried out with age/heat or depending when the unit was made could have been from the capacitor plague era...
Spikes on the output could also interfere with radio reception

As others have said you could open the PSU and replace capacitors then carefully epoxy the case back together and safety test it, or you could add additional capacitance on the outside, from the noise perspective it would be better to keep the capacitors close to the noise source to minimise ESR and emissions.

I would tend towards replacing the wall wart with a newer unit from a reputable brand, and check for correct voltage and clean DC output before connecting it.
An iron cored type will typically emit less RF noise of course, but these are becoming rarer.

Best Regards
Chris

[martinhog]

Hi all, I changed the reservoir cap in the wall wart. Surprisingly, the wall wart case was screwed together and it had an internal fuse too, but only 1 single rectifying diode. The output looked cleaner so I connected it to the Roberts and switched on. It played perfectly for an hour, I switched it off whilst taking a phone call. I turned it back on and the power supply started pulsing. I tried it on batteries, it worked for 3-4 seconds then died with the current consumption shooting up to 2.5 amps. The output transistors were hot (almost implying C40 was short circuit/leaking, this was replaced a few weeks back when the output transistors were changed) . I left it for an hour, connected batteries and it has played for over an hour, drawing around 16-20 mA, its still playing now !.
Thanks again for your replies.

[martinhog]

Update, the dear old R505 failed again whilst running on batteries. a quick tap on the output transistors brought it back to life so I assume its the the old 'tin whisker' problem ?
The transistors were replaced but being N.O.S. they were passed their best ?. I suppose the action of plugging and unplugging the mains adaptor simply gave the whiskers a nudge . The youngsters at work always marvel at "how nice" the "old fashioned wireless" sounds when the drop by for "Popmaster". Thanks for all your help and comments.

[HamishBoxer]

NOS on Germanium is not a good idea as most likely they still fail.