Arcam Alpha 3 Left Channel broken...

[wprice99]

I have an old Arcam Alpha 3 which has served me well for the best part of 15-odd years now, but unfortunately it's not on it's way out.

Now I'm relatively electronically adept, but more of a specialist in the digital realm than analogue audio stuff, so while I can safely wield a multimeter and test various things, I'm not exactly sure what I'm really looking for with this (possibly worth getting it to a specialist, or just replace it, I know!).

Anyway, the problem is this...

It started off with a very quiet scratchy/crackly noise on the left channel (the one that's blown), which was annoying so I switched the amp off for a while.

After turning it back on the following day, all was working fine, but after a while a pop/thud was heard, and then the output from that speaker was a low-volume hum - the same sort of hum that you get with an earth conflict. After a few seconds, another pop and the music continues... an so on.

I've tested the various combinations of inputs and headphone/speaker output, and the problem remains the same - so the problem lies somewhere in the power amp circuitry.

Following on from this, the amp has now started emitting the usual 'something electronic is getting hot' smell, although on inspection of the board and components, there's nothing obviously burnt yet... but from that point I've left it off until I can work out what to do (I don't really want to power it up and damage anything else beyond repair if I can help it).

I have obtained the service manual and schematics from Arcam, and this is where my knowledge breaks down. While I can stick a multimeter on the various test points labelled on the schematic, I'm not exactly sure what I'm looking for in order to determine what's gone/going wrong.

Does anyone have any hints/ideas, or am I really going to be better off taking it to a specialist?

Wayne

 

[Leolabs]

Start with checking PCB solder joints first.

 

[wprice99]

A close inspection of all the solder joints all appears fine. None look to be dry.

 

[Mooly]

If you can post the circuit, can help more !
Regards Karl

 

[wprice99]

It's too large to attach, but I've uploaded the service manual including the circuit schematics here: Alpha Service Manual

Wayne

 

[Mooly]

Hi, Downloaded manual, will have a look and see what can come up with. If you can do this SAFELY, solder a couple of wires to a 60watt filament mains bulb and solder this in place of the mains input fuse FS201 on your diagram. Make sure the bulb is well out of the way so you can't touch it.
Regards Karl

 

[Mooly]

With no fuse, just the bulb wired in switch on, what happens ?
The bulb should settle to a glow, if something is getting hot try and identify. Measure the supplies to the power amps +&- 37 volts, what are they ? Is R41 and R42 getting hot on the bad channel

 

[wprice99]

Interesting... any particular reason for the light bulb?

 

[Mooly]

The bulb limits the current that can flow, it's crude but very effective, if you had a short circuit output transistor for example it would limit any large current flows that could burn out components and even vaporise any fine PCB tracks. It's damage limitation. I am puzzled when you say that the fault was intermitant to begin with. As Leolabs says, really check for dj's, are any component legs "blackened" through oxidisation. If your not sure just run an iron over any suspect looking ones. If the solder does not flow freely round the joint this could well be the initial cause. If it's permanently duff now it is worth just measuring on ohms the output transistors to check for shorts. You have to start somewhere and if you can not identify problems with ohmeter then you are going to have to switch it on and measure and observe. If something goes pop it's not the end of the world, in fact it can be much easier than trying to guess what may or may not be wrong.
Regards Karl

 

[wprice99]

Novel approach for limiting the current flow. I like it.

The intermittency was how it all started; that was until I smelt the electronic component burning up smell, at which point the amp was switched off. I have no idea if it's still intermittent or whether whatever component was breaking down is now completely gone.

So back to the job in hand... The light bulb dims down to a glow so faint I can't even see it. Voltage between TP1 and TP2 is 74.4V which is what I would expect for +/- 37V.

Resistors R42 and R41 are fine, temperature-wise, but resistor R40 is too hot to touch (there I was casually dapping my little finger on all the various resistors and transistors until I hit that one). Yowch.

The equivalent resistor R140 for the right channel is perfectly comfortable to touch.

Regards,

Wayne

 

[Mooly]

You have to measure the voltage at the speaker output. Is it high? Measure on the speaker fuses, this point should be zero

 

[Mooly]

Post the reading you get, I am going to have to disappear now, but it sounds like the full supply is across R40, probably because of an S/C transistor. With all off measure on ohms R41 and 42 they might be open circuit or very high in value. Also check the output transistors for any low readings and Q11 and 12 as well. Compare your readings with the other channel and remember it matters which way round you connect the leads when measuring semiconductors.

 

[wprice99]

It's 36.4V across the output of the left terminal.

 

[wprice99]

R41 and R42 resistance is very low (pretty much off the scale and in continuity test mode, gives a bleep - about 0.5 ohm). This matches R141 and R142 from the other channel.

None of the transistors Q14, Q13, Q12, Q11, Q10, or Q9 appear to behave any differently than on the other channel, in terms of resistance readings. None are reading short-circuit between the emitter and collector with the probes either way round, and similarly between emitter/base and collector/base.

Perhaps the problem lies somewhere prior to the TP17&19 test point - this could be holding Q12 open allowing the full supply across R40?

Really appreciate you taking the time to think about this, by the way.

 

[Mooly]

Thinking out loud now just puting thoughts down, it is worth measuring the base emmiter volts across the transistors Q11 & Q12. Measure the volt drop across R30, should be .65v. In fact measure base/emmiter drop across all transistors, suprising how this can show up a faulty device. If any volt drops over 0.7v and they are forward biased theres a problem. For npn base 0.7 higher than emmiter, for pnp base 0.7 lower. Check the supply to that opamp pin4 and 8. Still thinking out loud, you say output at 37v (plus 37?) are you sure. If R40 hot, bottom end of resistor must be at negative rail voltage. This would turn on Q14 and pull the output negative- recheck that output is at 37v not -37

 

[wprice99]

I've actually been going through measuring the voltage at the various TP points, and comparing across left and right channels. Here's what I have:

TP1: +37V
TP2: -37V
TP3: -15V
TP4: +14.9V
TP5: -13.8V

The above are pretty much what I expect. The rest, however...

Speaker fuses: -36V (left), 0V (right) - the output is actually -36V, not +36V.
TP12: -36V (left), TP14: 0V (right)
TP13: -36V (left), TP15: 0V (right)
TP20: 14.2V (left), TP21: 0V (right)
TP17: -37V (left), TP19: -2V (right)

I'll just check some of the other points you make...

 

[Mooly]

Whats the volt drop across R41 and 42

 

[wprice99]

R41 voltage drop: 0.008V, R42 Voltage drop: 0.009V

The drop across R30 is 0.65V, as you say.
Q11 and Q12 base-emitter voltage drop is around 0.6v
Q9 and Q10 base-emitter voltage drop is around 0.34v

 

[Mooly]

In a feedback circuit like this a fault can be anywhere. Check those B/E volt drops ALL OF THEM. Whats the voltage across R30 and R21 ?

 

[Mooly]

Q9 and 10 thats fine, these are SOAR (safe operating area protection) . It does sound like a odd fault this, but do check ALL those B/E junction volt drops.