Audio amp circuit only works when specific pin is touched

[yanalesch]

Hello,

I am trying to repair a Yamaha NP-30 keyboard which is completly mute. By tracing the audio signal with my scope I was able to narrow it down to the amplifier(AM) board. The board uses a LA4625 amplifier ic.

When I touch pin 4 (standby) or connect the pin to protective earth from mains socket, the amplifier is turned on and the keyboard works like a charm until it is powered off. When I connect the pin to GND on the AM board nothing happens. What riddles me is that the keyboard uses a Class II DC Powersupply without connection to protective earth. And even when using an additional isolating transformer it shows the same behaviour.
I really don't understand how it is possible that a current flows from pin 4 to protective earth if the keyboard is isolated from protective earth.
The next question would be what could be the cause of this behaviour? Any ideas would be appreciated.
I attached circuit diagramm of the keyboard and the LA4625 circuit plus description of the standby function.
Thanks for your thoughts.

Best regards

yanalesch

[Audiorepair]

I have had similar problems with Yamaha pianos using this power amp IC, which would intermittently mute.

It turned out to be the LA4625 internal standby circuitry that was faulty.
Which is a bit of a bugger, as it is difficult to get these chips now without risking buying fakes/clones.


Look at the datasheet for this IC, and monitor the DC on the standby pin.
You may find it not making sense.

In fact, I now remember that one of those I had would unmute as soon as you probed the standby pin with a scope.
So perhaps the internal fault is, in fact, a lack of a ground reference.


This repair took up a lot of my time, as I never suspected the IC's internal circuitry.



Edit:  Just checked my database.  The one I had that did this was a Yamaha P95.
Same power amp IC.

[yanalesch]

Yes they are hard to get, apart from chinese sellers where it is unclear if they are original or replica. 

In fact I already replaced the LA4625 after I found a seller from spain, but it didn't helped. Maybe the "new" one I bought was also defective? Or the circuit itself caused it? Well bad luck I guess.

But also very interesting, I really can't get behind what happens and why it happens. When there isn't current flow to earth, is it the electric potential shifting what makes the circuit behave so strangely?

Im quiet baffled.

[rob77]

When I touch pin 4 (standby) or connect the pin to earth from mains socket, the amplifier is turned on and the keyboard works like a charm until it is powered off. When I connect the pin to GND on the AM board nothing happens.

from the datasheet....

Pin 4 is the standby switch. A voltage of 2.5 V or
higher must be applied through an external resistor to
turn the amplifier on.

[2N3055]

When I touch pin 4 (standby) or connect the pin to earth from mains socket, the amplifier is turned on and the keyboard works like a charm until it is powered off. When I connect the pin to GND on the AM board nothing happens.

from the datasheet....

Pin 4 is the standby switch. A voltage of 2.5 V or
higher must be applied through an external resistor to
turn the amplifier on.

And current limiting resistor for standby is 10k in datasheet, not 27k like in schematic. Try putting another 22k in parallel with R111..

[yanalesch]

I already tried that and replaced R111 with an 10k resistor.

I also removed R111 and went from 5V instead of the 12V to pin 4 with an resistor calculated to match the max. of 500µA specified in the LA4625 Datasheet.

In both cases same  strange behaviour, the amplifier switched on only after i connected pin 4  for a few seconds to mains earth. I can't measure what is happening because if I connect my scope  to pin 4, even though the keyboard is floating, the amplifier immediately switches on.

[2N3055]

I already tried that and replaced R111 with an 10k resistor.

I also removed R111 and went from 5V instead of the 12V to pin 4 with an resistor calculated to match the max. of 500µA specified in the LA4625 Datasheet.

In both cases same  strange behaviour, the amplifier switched on only after i connected pin 4  for a few seconds to mains earth. I can't measure what is happening because if I connect my scope  to pin 4, even though the keyboard is floating, the amplifier immediately switches on.

Check startup capacitor on pin 5..

[yanalesch]

I will check and be back tomorow, thanks so far.

[yanalesch]

I used an interference supression filter in front of the power supply plus ferrit ring chokes to rule out interference. Didn't helped.

So I checked all seven electrolytic capacitors on the AM Board.. all within spec. Replaced 1µF startup capacitor (C124) on pin 5 with 33µF capicator referring to LA4625 datatsheet. Nothing changed.

Measured voltages on all pins and compared them to the expected voltages according to datasheet.
All voltages are in spec, except the the voltage on pin 7 (POP). Datasheet expects 2.05V, I measure 0.02 V.
Which function has pin 7? Looking at the datasheet I assume it is just the amplifiers case connected to gnd using an capacitor?

[2N3055]

That is interesting. I have a theory.

Datasheet explicitly states that capacitor on pin 7 shouldn't be large, they recommend 0.47uF .
They put in 10uF there. It might have damaged something inside.
This could have been a design flaw...

[yanalesch]

I fear you may be right. Searching the web it seems to be an common error with yamaha keyboards using this amplifier. User audiorepair even described the same issue.
Maybe it is time to search the web for another LA4625. 
And this time replace the capacitor on pin 7 with 0.47µF before replacing the amplifier ones again.

One thing I would like to try before ordering another would be to desolder the amplifier and test it on its own. Basically not a problem, the wiring is stated in the datasheet but where I would need your help is in generating an audio test signal. I'm not familiar with audio engineering at all. How would I approach this?

A sinewave with an frequency in human hearing range? Which amplitude? Would a squarewave from a ne555 be sufficent?

I dont't have an function generator at home. Maybe at work, not quiet sure.

[Ian.M]

Soundcard Scope by Christian Zeitnitz is free for personal non-commercial use and has a two channel audio signal generator. 

Caution: PC soundcard inputs and outputs aren't well protected - you *MUST* limit input levels to avoid damage and should probably have some sort of buffer amplifier or isolation network between outputs and your device under test.  Using a USB soundcard rather than an integrated motherboard one lowers the risk of expensive damage but does not eliminate it.  Beware of sound cards that maintain Audio ground at approx. +2.5V to allow use of single supply amplifiers - capacitively couple the ground as well as the signal, *AFTER* confirming the output doesn't actively drive 'ground' in a bridge-tied load configuration to increase the max. range of the mono component.

Another possible signal source if you don't want to risk your PC is a cheap MP3 player.  Run from batteries as output 'ground' is often actively driven as described above.

[yanalesch]

I can't risk my notebook, it is my only computer, apart from an old raspberry. But I have an old Smartphone which I could use as audio source.
I have to read into the topic, think about the safest way to approach this, get some missing components and then I will get back to you.
I'll keep you updated, thank you all so far.

Have a nice weekend.

[fzabkar]

After the amp switches on, can you mute it by connecting pin 4 to AGND?

[Audiorepair]

Just a bit of speculation here, but the LA4625 datasheet shows 2 diodes in series between transistor base  and ground.

If one of these diodes failed short, then it would be difficult to generate enough Vbe voltage to turn on the transistor and unmute the amp.
Under certain conditions though, and a less than perfect short of the diode, this may be possible.


I'm not sure how easy it would be to determine the integrity of these diodes from the outside world, without having a known good IC to compare.

[NiHaoMike]

It turned out to be the LA4625 internal standby circuitry that was faulty.
Which is a bit of a bugger, as it is difficult to get these chips now without risking buying fakes/clones.

It's a standard analog input amplifier, just disable that part of the circuit and wire in a cheap amplifier module to replace its function.

[magic]

Circuit ground is connected to mains through small capacitors in the PSU and roughly follows mains voltage, but can't source much current because of the high impedance of those capacitors. (Its AC voltage quickly falls down when loaded with some low resistance/impedance to real earth).

When you connect some circuit node to earth, it will be subjected to a fairly high AC voltage limited to very low current.

Basically, what it is doing to this IC, is to alternately sink and source small current from that pin, with whatever voltage (up to tens of V or more, but likely just a few V in practice) which is required to make the pin conductive enough to accept said current. The likely effect is turning the IC on and off and back again.

edit
Actually, if the 10kΩ resistor is still present, most of the injected current may flow into that resistor. So the effect could simply be a small AC voltage appearing on top of the 2.5V DC that ought to be there.

If you want to take any measurements on this circuit without things going funny, you have to connect its floating ground to earth, using the scope's grounding clips or otherwise.

[yanalesch]

After the amp switches on, can you mute it by connecting pin 4 to AGND?

I will try this when I'm back home wednesday.

...the LA4625 datasheet shows 2 diodes in series between transistor base  and ground.
If one of these diodes failed short, then it would be difficult to generate enough Vbe voltage to turn on the transistor and unmute the amp.

Interesting thought, that may be the case. As you stated unfortunately no simple way to determine the integrity of this diodes without decapping the IC or having a fully functional reference IC. When I have tested the IC outside the circuit, perhaps I can exclude this or take it into consideration.

It's a standard analog input amplifier, just disable that part of the circuit and wire in a cheap amplifier module to replace its function.

Yes, probably that's what it will come down to. But until then, above all it's the curiosity about why the circuit behaves the way it does. Like an itch you can't reach.

Circuit ground is connected to mains through small capacitors in the PSU and roughly follows mains voltage, but can't source much current because of the high impedance of those capacitors. (Its AC voltage quickly falls down when loaded with some low resistance/impedance to real earth).

When you connect some circuit node to earth, it will be subjected to a fairly high AC voltage limited to very low current..

Thank you for your detailed explanation magic, but unfortunately I don't quiet understand. I think it maybe my fault by missleadingly expressing myself. When I said that the circuit works when I connect pin4 to mains earth, I meant protective earth, not neutral.
The PSU is 12V DC and is Class II without connection to protective earth. This behaviour even occurs when I use an additional isolation transformer.
You stated that when I connect some circuit node to earth to it will be subjected to a fairly high AC voltage limited to very low current. I don't understand how that happens, when the keyboard is floating. How does the current flow when there isn't a closed circuit? It would be great if you would explain.

[fzabkar]

You stated that when I connect some circuit node to earth to it will be subjected to a fairly high AC voltage limited to very low current. I don't understand how that happens, when the keyboard is floating. How does the current flow when there isn't a closed circuit? It would be great if you would explain.

See this ATX PSU circuit.

http://www.pavouk.org/hw/atxps.png

If the shield is not connected to earth, then it floats at a potential equal to half the mains supply. That's because there are two high impedance capacitors connected to active and neutral (C2 and C3). I expect that the negative DC terminal of your adaptor is connected to similar capacitors.

[yanalesch]

If the shield is not connected to earth, then..

Thanks, that was what I needed to grasp it.


Thank you all so far, you have been really helpfull. I will get back to you after I tried your suggestions and additionaly tested the IC outside the circuit.

[magic]

I don't think it's necessary to pull it out of the circuit. You could simply remove the 10kΩ pullup to gain exclusive access to the suspect pin and of course earth the circuit so that its ground potential doesn't jump around.

[yanalesch]

Hello,
just wanted to give a quick update after I tried your suggestions. I changed the power supply so that I could earth the circuit like magic suggested. After removing the 10k\$\Omega\$ pullup I used an direct connection from power supply to amplifier ic to unmute it when the standby button is pressed. That was it. Now it works like a charm again.
My girlfriend is thrilled. Thank you all for your helpfull replies. And have a great weekend.

[Mister Leverton]

OK....
I am here because I have a Yamaha NP-31 keyboard using the LA4625 with the exact symptoms, no sound, then touching the standby pin with anything metallic and it springs to life. BTW pin voltage when measured is around 5V
Tried changing the pullup resistor on Pin 5 using different values and keeping the current below 500uA- no change.
Sourced a new LA4625 from Yamaha here in Australia, no change.
Removed capacitor from Pin 5 (pop prevention) and the unit now always has sound on startup, the only side effect is a power up popping noise.
Experimented with many different values of capacitor and have settled on 82nF. the unit now only has a faint thud on power up.
I would like to experiment further but at some point you have to move on. I am tipping this is an internal problem as others have stated.

[magic]

If Yamaha sold you a replacement chip then it is hopefully a genuine and unused one, which should fix the problem if it was something internal to the chip.

If problems persist, it may be something wrong on the board. Perhaps a bad capacitor somewhere.

[cascador]

I also have the Yamaha NP-31 keyboard. I tried changing virtually all the resistors / capacitors in a range of 6cm around the LA4625 and nothing changed.
BTW when you use the keyboard with 6 AA batteries, it's exactly the same, so I guess it's not some residual AC and wiring to earth is not necessary.
So my final solution, as it requires only a touch to pin 4 with any metallic object, has been to solder a wire to pin 4, and let is hang behind the keyboard (or you can place the other side of the wire anywhere). When I turn on the keyboard, I just need to touch the wire with some metallic object, and voila.
I don't even trust changing the LA4625, more risky than promising.

[update] SOLVED! Just solder a 1K resistor between the POP pin (7) and the +12V.
Credits to the guy here