Bush SRG106 motorboating noise

[Michael Maurice]

I am refurbishing a Bush SRG106 radiogram, I have replaced all the relevany capacitors and resistors on both the main amplifier chassis and the radio/preamp section.

It all works except that I get a motorboating sound on the RH channel when the bass control is tuned to minimum or near minimum.

I have attached a circuit diagram of the relevant section.

The motorboating sound stops when you turn the bass control up. it also stops when I remove the ECC83 V2,

The only parts I haven't replaced are the two resistors and capacitor in the cathode circuit, and C65 in the treble control circuit.

[Michael Maurice]

I've now replaced the two cathode bypass capacitors in the tone control section but there is no change, the motorboating is around 3Hz

[Michael Maurice]

Done more prodding around, its oscillating at around 20KHz (just under 1 division at 50uS)

The first four pictures are with the bass control at minimum and the last one is at maximum.

The oscillations are approx 20V P/P on the grid of V6b and approx 2V P/P on the grid of V4b

I cant work out where its coming from

The oscillations are also around V2b,

All oscillation disappears when V2 is pulled

I'm hoping that someone can point me in the right direction.

[Boater Sam]

I have no circuit for srg106. Does this oscillate on gram and radio?
There are feedback circuits on both channels, are they both the same?

[ronbryan]

Michael

Have you checked the frequency compensating C-R networks across the output transformer primary? C98, R100, C100 on one channel, C99, R101, C101 on the other channel (Trader Sheet 1721) all caps 820pF, all resistors 15k.

Ron

[Boater Sam]

The tuning eye comes of one channel, any relevance to your problem?
Is the negative feedback loop the same on both channels or is there a faulty cap in one?
All the interconnecting plugs and sockets OK?
As you appear to have changed every component, is one the wrong value or connected to the wrong place? It is very difficult to recheck your own work for errors, I know, done it often.
When did the oscillation start, after you replaced which part, or was it always there, or has it never worked at all whilst you have had it?

[Michael Maurice]

Originally Posted by ronbryan

Michael

Have you checked the frequency compensating C-R networks across the output transformer primary? C98, R100, C100 on one channel, C99, R101, C101 on the other channel (Trader Sheet 1721) all caps 820pF, all resistors 15k.

Ron

Just had a look at this Ron, the resistors are slightly high, the capacitors are red ceramic, I suppose no harm in changing them if I've got some.

Originally Posted by Boater Sam

The tuning eye comes of one channel, any relevance to your problem?

Actually it doesn't it comes off the same HT line for both channels of the preamp.

Originally Posted by Boater Sam

Is the negative feedback loop the same on both channels or is there a faulty cap in one?

they are all new so should be the same.

Originally Posted by Boater Sam

All the interconnecting plugs and sockets OK?

as far as I can tell, yes


[QUOTE=Boater Sam;1010899]
As you appear to have changed every component, is one the wrong value or connected to the wrong place? It is very difficult to recheck your own work for errors, I know, done it often.[ /quote] I've checked and rechecked and everything appears ok.

Originally Posted by Boater Sam

When did the oscillation start, after you replaced which part, or was it always there, or has it never worked at all whilst you have had it?

I did a blanket change of all components, after doing this the oscillations started.

[John10b]

I had a similar problem with audio amp many years ago and it turned out to be the output transformer!
Cheers
John

[Michael Maurice]

I hadn't thought about the O/P transformer, but I've just done some cold checks on the primaries and I get totally different readings

LH

* 137 ct 398 *


RH

* 233 ct 115 *

Where CT is the centre tap and * is the anode tag.

[Michael Maurice]

Assuming the output transformer has failed, I could replace both of them with this

In which case, could I rewire the amplifier to make it an ultra linear or am I best leaving the taps alone.

Or should I send them to be rewound which is a lot more costly.

[ukcol]

Michael, referring to your your circuit diagram in post #1.

C67 (0.01uF) is well worth checking by substitution if you have not already done so.

[Michael Maurice]

C67 has been replaced with a modern Vishay (yellow) 10nF/400V capacitor as have all the other capacitors and most of the resistors.

[turretslug]

Strong possibility of being a complete shot in the dark, but is it worth doing a quick ballpark sanity check on the output transformer primaries with an analogue ohm-meter (I've assumed that the readings above are from a DMM?) I've occasionally found that a DMM will get itself into a right tizzy with high-inductance components with head-scratching results.

[John10b]

What I did to finally prove it was transfomer, I switched them around to see if fault followed Transformer, and it did!
Cheers
John

[Michael Maurice]

I did swap the transformers and the fault stayed on the RH channel

[Boater Sam]

OK, so assuming all the connections are correct as you have double checked, we can assume a faulty component, new ones can be duff too unfortunately, learnt that in the motor trade.

Any change swapping the ECL83s from one position to another, there is enough of them!? Another ECC82?

I bet you have not changed any of the pots? There are 3 in the front end of each channel. Room for a faulty one?

The HT for the eye is common but the grid control is only from one side, admittedly only on radio, I was just trying to find something that related to only one channel. Does it motorboat on radio as well as gram? The radio is not stereo, but connected to both channels in parallel. But is it?

Once all that is ruled out you will have to get a bit active. I would disconnect both the feed back loops at the output transformers, see what you get. Valve bias should be OK still.
Then you can swap the anodes of V2a and V2b over and see if the fault swaps sides. If so its definitely around V2.
The feed back loops can be swapped and reconnected too.
This may narrow it down for you. In fact, what happens if you disconnect the anodes of the ECC82 rather than just unplugging it? That will rule out HT oscillations.
Any screened cable in the front end of the amps? I am always suspicious of the braid connections to ground at one end.

I found a long while ago that being confident and doing blanket parts changes got me into trouble.

[John10b]

My only other suggestion is to check and ensure all chassis earth connections are 100%, just a slight bit of corrosion or dry joint can, as I am sure you are well aware, cause problems.
Cheers
John

[Michael Maurice]

I haven't got time to check and recheck my work, so I'm going to give it to Ron Bryan next week to cast is very beady eye over it.

[sortedradio]

I have had a similar problem in an amplifier that turned out to be a cracked track in a pot. As the pot was turned down and the wiper passed the crack the valve started to oscillate. As Boater Sam suggests it would be worth checking the bass pot at least.

Martin

[ronbryan]

Michael Maurice brought the Bush SRG106 chassis round on Wednesday this week and I spent some time investigating the fault, i.e. oscillation (which looked like squegging) on one channel of the stereo amplifier, which fortunately was nice and repeatable and dependent on the position of the bass control.

The amplifier is unusual for a radiogram, in that it has its treble and bass tone controls in a negative feedback loop driven from the secondary of the output transformer. Michael had already changed over the output transformers, with no change to the faulty channel, so that eliminated a number of the potentially faulty parts, including the tone correction parts on the OT primary.

I spent some time on the faulty channel checking the earthing tags, cable screens, the feedback network component values and various parts which might have influenced the amplifier frequency response roll-off, but everything seemed to be in accordance with the circuit diagram.

I did not have the speakers, so had fitted 15 ohm power resistors to the OT secondary and confirmed that with amp powered up and a 400Hz tone fed to the input, the unwanted large amplitude LF oscillation could easily be provoked, making observations of amplifier performance difficult. Breaking the negative feedback loop made the amplifier stable.

I planned to see if I could determine the frequency at which the negative feedback had become positive, so decided to start by measuring the open-loop gain of both channels at various frequencies, comparing them using a double beam scope on the dummy 15 ohm loads. I broke the feedback loops at the OT secondaries.

It was immediately obvious that at 400Hz the faulty channel had a gain considerably greater than the good one, so finally I got something that could be investigated. Checking and comparing the gains of both channels, working from the input to output, I found that the concertina phase-splitter feeding the output pentodes on the faulty channel had a gain difference of 10x comparing anode and cathode outputs. You guessed it, a 4k7 resistor had been fitted to the cathode circuit instead of the intended 47k.

I checked the phase-splitter on the other 'good' channel and found in this case that both resistors (R81 and R84) were the same value, but incorrect at 4k7 instead of 47k. So in this instance the gain was the same for both phases, but the valve operating point was more wrong than that of the 'faulty' channel.

The Bush chassis has no tag-strips, but has dozens of those nylon feed-through insulators pressed into the chassis which are used as anchoring points. This method works quite well, but it is fiddly when the component you wish to change is at the bottom of a stack of four other tightly wrapped components on the same post.

The three incorrect resistors were changed to 47k metal film and that restored the amplifier to normal operation, without a trace of squegging.

With hindsight, I probably should have spotted that both of the concertina phase-splitters were incorrectly biased by doing more thorough voltage checks, rather than relying on a gain check to highlight the problem, but I started off chasing a supposed negative feedback problem and had to fight my way back from there.

In some lights, the colour codes on film resistors can be indistinct, especially on those with the matt brown bodies, where reds and oranges can look similar. The resistors with blue bodies are easier to read.

Ron