Hacker Mayflower II (RV20) hum, what next?

[Nickthedentist]

Hi everyone,

I've been working on an RV20 chassis which I was lucky enough to get from Nick Lyons. It's an earlyish example (06789) and was in a bit of a state, with generalised light corrosion plus evidence of things around the output stage running very hot for a long time; PVC insulation burned off wires, resistors looking very sad and doubled in value, ECL86 envelopes very brown inside, cathode bypass electrolytics all virtually o/c and so on.

http://www.vintage-radio.net/forum/a...7&d=1486549417

I've replaced all twelve (?) carbon resistors in the output stages, all the leaky TCC Supamolds, all four 50uF cathode bypass electroytics, as well as the 50uF + 50uF reservoir/smoothing can which was leaky as a sieve. The two 250pF Micamolds were also replaced as they'd suffered physical damage and corrosion. The Westinghouse copper oxide bride rectifier has been swapped for a modern silicon bridge plus series R as recommended by Mark Henessey. I've fitted a pair of known-good ECL86s. The charred wiring has been replaced and all chassis earthing tags removed, cleaned and tightened. Layout is unaltered.

Everything works(!), voltages on the ECL86s are sensible and stable, and there is lots of undistorted audio. BUT, there's an intrusive hum which is still present even with the grid of triode V6a shorted to ground. I have 3 other Mayflowers which are silent in comparison.

A few quick scope snaps to follow when I've downloaded them from my camera.

Any suggestions, please, folks? I'm stumped! And yes, I've read all the threads on these

Many thanks,

Nick.

[Nickthedentist]

Snippet of circuit diagram...

[Nukeybrown]

I had a hum on mine at one time seem to remember it was coming from the board with the volume control is on, but this was a while ago now and cant remember exactly which component was at fault but seem to remember it may have been an earth issue along with the earth lead to the speaker

[Nickthedentist]

And some scope traces, details in file name...

Apologies for rubbish photos, dust, and any mistakes (it's late!).

Many thanks,

Nick.

[Goldieoldie]

If you dis c31/32 does the hum go ?
If still there pull v6 v7 alternatively to see if that helps
Have you tried using a 47 uf at key points to see if you can see where it's being generated ?
I did experience hum with one of these ,it was lack of screening around the tape/gram input .I guess this is not relavant with your hum but useful to know perhaps

[TonyDuell]

It appears that the hum across the speaker is at 100Hz. This suggests to me that it is coming from the full-wave rectified HT line, not random mains pickup, which would be 50Hz.

[stevehertz]

You could see what current is being drawn from the HT supply and if it is excessive (sorry, I don't know what it should be, but you could measure on one of your other sets), then that would cause excessive ripple on the HT line. In which case something, somewhere, is taking current down to deck, as it doesn't appear to be due to poor smoothing per se, you having already fixed that. Just a thought.

[ukcol]

Hi Nick

In a healthy push-pull output stage the ripple on the HT line is cancelled in the output transformer, although it is not the case with this Hacker some designs will run the output stage directly from the reservoir capacitor and relying on this hum cancellation for a low hum level. It make be worth checking the output transformer in case it was damaged by the previous fault; it may have a shorted turn or two in one half of the primary.

Also I note that the waveform in your CRO snaps from the output valve control grids has a 50 Hz component; this is indicated by the fact that alternate cycles of the 100 Hz waveform are of slightly different amplitude. This may be a red herring though as I can't see a 50Hz component in the speaker output waveform.

Finally, I don't know what type of valve holders are used in this set, but if they are heat damaged one or both may be leaky.

[Goldieoldie]

I think I remember when I changed the bridge to silicon on my mayflower I had to add extra caps to get rid of the hum ( 47 uf )

[Herald1360]

Extra caps round the silicon might well be the answer if the hum is getting in as RF hash with mains PRF generated by slow diode reverse recovery spikes. 1N4007 and their ilk are so slow that they almost struggle to rectify 50Hz.


10n across each diode in a bridge should kill that source of noise.

[stevehertz]

So for similarly rated rectification purposes, what is a better diode to use than an IN4007 to avoid the problems that you mention?

[turretslug]

BA159 is my HT rectifier of choice, the price difference between it and 1N4007 is so trivial that it more than makes up for the capacitors you won't need to use;

http://www.vishay.com/docs/88536/ba157.pdf

[mhennessy]

The earlier photo shows about 150mV pk-pk across the loudspeaker, and it's pretty sinusoidal. So I don't think it's a problem with the rectifier diodes - these would add a "burst" of RF hash at the switch-off points. Certainly, I've used standard-grade rectifiers - including 1N4007s - countless times in these sets with no ill effects.

I've just done some digging, and found a set of notes from various RV14/20 restorations. Really should scan these in, or write them up in some other form. Anyway, from these, I'd conclude that the PSU ripple levels are absolutely fine.

The fact you have around half a volt of ripple at the inputs to the output pentodes is the clue here. Where is that coming from? Is it the rail supplying the triode section of the valves. The ripple there should be around 100mV, give or take. It's possible that C28 has failed...

BTW, your timebase is rather out of cal

[Nickthedentist]

A very sincere thank you to EVERYONE who's replied to my thread.

Spurred-on by all the brilliant suggestions, I spent a few minutes with the set and my scope this evening.

First, I hung the scope on the grid of pentode V7b and could see ripple with an amplitude of 400mV p-p.

Originally Posted by Goldieoldie

If you dis c31/32 does the hum go ?

Lifting one end of C32 brought it down to 100mV, whilst lifting one end of C31 brought it down to an inaudible 30mV. The other end of C31 had 500mV of ripple on it, and the top end of R29 (the triode's anode load resistor) had a whopping 1.5V.

Originally Posted by mhennessy

It's possible that C28 has failed...

Originally Posted by Goldieoldie

Have you tried using a 47 uf at key points to see if you can see where it's being generated ?

Yes, you guessed it, C28 had indeed gone o/c, even though it tested fine before I applied power and seemed to re-form well. Dabbing a nice new 47uF cap across it eliminated the hum completely. I will do something more permanent tomorrow.

With hindsight, I perhaps should have been more suspicious, as its can buddy, C29, had already gone o/c and been bridged (with a huge 100uF electrolytic dated 1981) by a previous repairer.

Thanks again, chaps, I've learned a lot.

And thanks again to Clint for the donation of the scope. It's been an absolute godsend, even if it needs a bit of calibration...

Nick.

[stevehertz]

Well done Nick and all who contributed, a success.

I generally never trust old electrolytics for this very reason, I just replace them.

[Nicklyons2]

Well done Nick, I'm glad it's back in the 'land of the living'.

BTW I would just like to dissociate myself with the earlier dodgy repair! I was it's 'rescuer' not it's repairer

[Nickthedentist]

Yeah, yeah! Only joking, Nick, you've told me its colourful history already

I'll write up a "Success Story" in due course.

N.