A1200 weak crackly sound - both channels went overnight!

[Photon]

I've used my A1700 for three weeks, then one day when I turn it on both channels have gone weak and crackly. The symptom is the same for both channels.

I first checked the voltages at the PSU, which are fine. Following this excellent guide, I looked at the schematic to see what was in common for both channels to fail and started by measuring Vcc and Vee on the opamp.

Vee is as good as it gets, -10.9V at the PSU and almost the same at the chip. But Vcc is -4.05...

Now, to start off, either the opamp or its Vcc input circuitry is faulty.

I can't really see any degree of corrosion of the board, and certainly not near the caps near the opamp. (I will of course replace the 22uF caps when it's working while I have it open.)

I checked the Q341 transistor, marked "2A I":
b -1.5
c -10.5
e -4.0 (connected to Vcc)

I lifted the Vcc leg on the opamp, and then I get
b 0.8 or 1.5
c -10.8
e -0.7 or -1.5

Now, e is floating, but I can't really explain the differences in measurements on b. Looked for power supply circuits in the datasheet, but didn't find any. LF347M is JFET; does that mean Q341 is too?

Either way, it definitely seems to supply the opamp with the emitter (instead of a hidden trace under the opamp supplying the power and Q341 used for something else).

I've traced the faulty -10.9V to another transistor marked "61A", which has
b 0
c -10.8
e 0

c is also connected to a through hole so the tracing must stop there for now. Why is it -10.9V!?

Any ideas on how to solve this or what to look for instead is appreciated

[chiark]

Suggest mailing Anthony Hoffman, whose guide you have already followed. Sorry that doesn't help much, but this might well ring a bell with the Guru himself

[Photon]

Was wondering, since -12V is directly connected to the opamp but +12V strangely is not, if I might connect +12V directly to the lifted leg. The error might affect other parts of the circuitry that aren't fixed by this, of course.

I'll take your advice though, and email him. [Edit: Emailed him.]

I think I'll swap the 22uF caps while I wait for a reply.

Edit:I replaced C303 as it was connected with its - (!) pole to the 12V, putting a new one in the same way. Bit confused by that. + is usually not connected to GND?

Probably not it, though, as it's the same way on my untouched A1200! So all is as it should be.

Hm, so I should use GND from the power connector as reference point?? Which one then please on a postcard. (Or better, here.)

I can feel a "bad GND" if I touch the bare left edge of the keyboard with my skin, if that's relevant. With no peripherals attached.

I did notice that when I unplugged the sound cable, the -10.9V on the Q341 collector went up to -1.5V. I've had it connected to headphones and a PC sound system. All outlets for the Amiga and peripherals are grounded, but of course the C1084S has no ground wire.


Measured from the common between the two caps C303 and C304. Which is === PSU GND. So basically, the opamp power circuit wants to feed it Vcc -1.5V and Vee -10.9V now. The only reason I can think of is a short or a shorted-inside component. I've found neither this far.


The basic problem is that I'm at step 1, check voltages, and don't know enough to find the error to go on. It would be interesting to know what extra things I have to do to connect +12 to pin 4 'directly', if it turns out the other error is harmless.

[Edit 2:] Not sure if the 'bad GND' is a consequence of the malfunction. I can't really easily trace the +12V or opamp Vcc much further without more 'how it's supposed to work' info. So I will replace all the 22uF caps and take it from there, maybe that magically fixes the supply to the opamp...

Awaiting further instructions...

[Photon]

Mods: Another long text, and a few edits in the previous one, hope you don't mind if I keep the posts separate to make clear when I did what.


--------------------------------------
Anthony suggested some simple measurements to do, thanks

The working -Vcc has 12.3 Ohms to -12V, the failing +Vcc supply has 7 Mohm to +12V!

As mentioned I can trace +12V to C301/R301 and no further, after R301 (1 Ohm) it goes to a via, just by the ROMs. (*newsflash, it's the +12V that comes out of the via and is supposed to go through R301 and onto the opamp.)

News:
I made an assumption that Q341 "fed" the +Vcc to the opamp on its emitter. That was wrong. It must come from a via under the opamp, directly from the +12V. Resistance between +12V and +Vcc is 4 Ohms on a working A1200 that I measured on just now.

I made that assumption weaker when I measured base-emitter on Q341, which only has a 2700 Ohm resistor between it. So my adventure tracing from the collector was way off.

And then I...
Replaced the two 22uF caps (for the +-12V) nearest to the opamp with non-SMD electrolytic caps, 22uF 35V. Cleaned up the area with alcohol and shinied up some solder points. Found no damaged vias or traces in the process.

Connected the +12V to the right place (before its 10 Ohm resistor, as for the -12V).

Measured resistance to the +/- Vcc pins, got 11.5 Ohms on both. Good. Powered on, measured +Vcc=+12.15V and -Vcc -10.91V. Good. Powered off, connected the stereo, and I hear noise right from bootup, mixed in with strange intermittent sounds, in both channels. Played some music, absolutely no music is heard, the noise just continues.

Disconnected the +12V patch wire, noise persists & no music.

-~-

Obviously not the right fix, or not enough, or the opamp has been damaged when it had no +12V for a while. But it did play music (weak and crackly) during that time...


My crystal ball is all fogged up by overwhelming feelings of sadness inside, any ideas?

[Edit:]
OK, found the damaged via. It was under C304 (the -12V cap)!

It's next to the -12V via, both aligned on the same "horizontal" as a line going under the opamp. So I think both go inside to under the opamp, then up to the +/-Vcc pins via traces under the opamp.

The patch wire should have done that 'via' job perfectly. But of course I don't know if it goes to any other points - but I could measure if the -12V via goes somewhere else maybe.

If I could test if the opamp is still ok, perhaps by measuring between pins on it on my working A1200, I could save the replacement job... :P

Ideas?

[johnim]

hi i had similer problems see here

http://www.amibay.com/showthread.php?t=19813

[Loedown]

I'll up a schematic I have here of the Amiga audio stage with voltages present to the zone.

[Photon]

As a heads-up, the reason for the noise is likely VREF being dragged down to opamp -Vcc (currently -7V instead of -10.9V before the patch job).

Likely patching the +12V wasn't enough, or there's a second problem to solve, or the opamp was damaged by the no +12V problem. But the opamp +/-Vcc pins are lifted and still the same problem persists. It even drags the PSU -12V to -8V!


The Amiga is working perfectly in all other respects, and there are no shorts between any of the power connector pins or between +5 and -12V.

Steps I can think of is disconnect -12V from PSU and measure at the PSU. And possibly remove the opamp completely. And check Paula output.

[Allen1]

Have both of you checked the little black 1k resistors next to the audio connectors R324 and R334 (the ones behind the keyboard connector) I have had a couple of boards with problems that sound a lot like the problems you are experiencing with the audio output. If one of the resistors is broke then the sound is very crackly and no sound at all when both are broken.

Its the first place I look now if there is any audio problems, these resistors don't always look like they have snapped etc but an multimeter will soon find out if they are ok.

Both of you will have probably already checked these out but just in case

[Loedown]

Originally Posted by Photon

As a heads-up, the reason for the noise is likely VREF being dragged down to opamp -Vcc (currently -7V instead of -10.9V before the patch job).

Likely patching the +12V wasn't enough, or there's a second problem to solve, or the opamp was damaged by the no +12V problem. But the opamp +/-Vcc pins are lifted and still the same problem persists. It even drags the PSU -12V to -8V!


The Amiga is working perfectly in all other respects, and there are no shorts between any of the power connector pins or between +5 and -12V.

Steps I can think of is disconnect -12V from PSU and measure at the PSU. And possibly remove the opamp completely. And check Paula output.

You have a short circuit issue somewhere, which chips are getting hot?

[Photon]

Anthony said shorts are rare. (Especially outside chips, as you hint Loedown.) If there is one, since the Amiga works fine, it's likely the opamp. Still trying to locate U15 on the mobo though, to rule that out.

I'll first measure before the 10ohm resistors on the -12V line and see if those are bad.

But it's likely the mobo "was already bad" with the -12V - it dropped it from 13.4 to 10.9 even weeks ago when the sound was working fine.

When the -12V is fine, I will lift the VREF pins on the opamp and reconnect -12V to the mobo, it's fine at the PSU. If the VREF is fine then the opamp somehow went from bad to crap just by me fixing the +12V. (A new one is on the way to me).

If this fails and the soundwave after the audio filter is okay on my oscilloscope, I will cut some traces, and connect the -12V not to the mobo but directly to an opamp on a separate little PCB that I make and that'll be the end of it.

[Photon]

Again, would appreciate if mods didn't merge, for readability.


Hehe, U15 is the opamp. It's just not marked on the mobo, found it in the parts list...

OK, I traced the -12V to the "VREF/-12V area" between Paula and the CIA by the floppy power connector. Found no bad components, diodes checked out, the small chips there didn't get warm. Traced it on all the way to the opamp.

It's definitely something between -12V and VREF that brings VREF down from +2.5 to negative values...
-~-
When -12V is not connected to the mobo, -12V at power connector and all the way to the opamp is 2.93V. VREF is then +2.4V
When -12V is connected to the mobo, -12V at power connector and all the way to the opamp is about -8.2 / -7.2V. VREF is then -6.3/-7.2V
-~-

"OK" ...

[[[[ NOTE: the rest of this is with NO -12V connected to the Amiga! ]]]]

I couldn't find any other component that is between -12V and VREF, than the opamp.

I already lifted the +/-12V pins on it. I lifted both VREF pins on it now, and STILL the -12V at power connector goes up to +3V...

The opamp is out of the game (unless someone protests). So this didn't give me hope :S(If I get the replacement tomorrow I'll remove the old one completely and test though)

Hm... something else that brings -12V up by ~ 5V, or <unconnected -12V> up to 3V....


OK! Found a clue!

When the Amiga has been turned off a long while, and I measure the unconnected -12V, it's 2.2V for half a minute, then rises in 1.5 minutes to 3.8V and stops there!

Where this strange voltage comes from I have no idea, but slowly rising (slowly shorting to +5V??) to me means a cap or a chip.

Incidentally, Paula gets finger warm in 30 seconds, then distinctly warmer the next 2 minutes. Coincidence or cause?

Paula doesn't seem to be connected to neither VREF nor -12V (only found DIL pinouts though), so this points back to the opamp. But it is disconnected from both VREF and -12V now. Even if the Paula pins are connected to it, that can't affect the -12V and VREF since the opamp is not connected to them. This points to the VREF generator being at fault, if no other chip uses VREF and -12V. I've found the schematic for the VREF generator.

Thoughts...

[8bitbubsy]

Paula pin 31 and 30 is output L/R, which swings from negative to positive (I think).
I assume the voltage is very low, so don't get alarmed if you measure directly from the paula output pins.

[Allen1]

Hi there Photon,

I can't tell from your description if you checked the left and right output channels/stages to ascertain if the components are working after the op-amp, especially R324 and R334 that I mentioned earlier. If there is no load being drawn as in connecting the outputs to an amp etc or if a component has failed as in those 2 resistors, then I would expect to be getting strange voltage readings because the output capacitors are not discharging as they should be and the op-amp is doing what it it can with the signals/voltages that it receives.

[Loedown]

Originally Posted by 8bitbubsy

Paula pin 31 and 30 is output L/R, which swings from negative to positive (I think).
I assume the voltage is very low, so don't get alarmed if you measure directly from the paula output pins.

Ok, let's go through the sound output.

Taken from A4000 audio circuit, why, because it's easier for me to explain.

U400 is Paula, I want you to look long and hard and tell me where it has a negative voltage connection.

Why is there a Vref signal set at 2.5 volts?

I will concentrate on just one channel as they are identical or close enough to it anyway.

U400 outputs left audio on pin 34, this then goes to U402 pin 13, pin 12 of U402 has this magical Vref signal on it. The Vref signal is there so that anything below 2.5V is output as positive and anything above 2.5V is output as negative. It goes into the inverting or - input of U402.

Q440 is used solely as a switch as to whether the filter is active or not through components R441 / R442 / C441

Second part of U402 is used as a buffer, someone will correct me and tell me it's a voltage follower circuit, so what goes in comes out, if measurement on pin 3 is +7.5V then output will be +7.5V

U403 is then used as the final amplification stage using formula for gain as Gain = Rfeedback / Rinput, so R403C / R403B = 0.47 meaning that for every volt present on pin 2 will be roughly halved on pin 1, as the signal is going into the inverting or - pin again it's reversed so that the original Paula output signal is the right way up and anything between 0 - 2.5V is output as a negative signal and anything between 2.5 - 5V is positive.

Amplifier circuitry is not hard to fix or fault find even without a CRO. Do all these tests without anything plugged into the outputs and no audio playing.

Reconnect all pins, measure for Vref at 2.5V, U402 pins 10/12 if not check R401 / R402 or for open circuits, also the pins themselves for open circuits.

Paula pins 33 / 34 0V, if not disconnect before U402 and all other support circuitry and check for 0V, if random voltages check Paula ground and Vcc pins for fluctuations.

U402 pins 8 / 14 check for around 0V, if not check for pins 9/13 ohmage between ground and Vcc, only do this test with the unit off and allow time for capacitors to discharge. Also check for 8/14 to ground for shorts, ohms test.

U402 pins 1/7 should also read close to 0V when no audio but powered on.
Same applies to U403 1/7

Power problems are most likely associated with broken wires or other circuitry dragging the lines down, check back to the socket, meaning back to the actual socket recepticles, sockets can break between pins and board.

The important thing here is not to overcomplicate the original fault which is more likely to be a capacitor problem than anything else and just check continuity between points.

[Loedown]

Originally Posted by Photon

[[[[ NOTE: the rest of this is with NO -12V connected to the Amiga! ]]]]

This will cause distorted and low sound, I had an A500 years ago with this exact fault.

[Photon]

Loedown, thanks for the general description I've done most of the things there though. Which one of U402/403 is the opamp, and the other one is unknown? If it's a voltage follower it's defo interesting! I want to know what chip it is (and where it is on the A1200 mobo)!

I *had* low and distorted sound when the +(!)12V was unconnected due to the corroded via. When I fixed it, the -12V which was at -10.9V previously became -8V and VREF went from +2.5V to -7V. This is the whole problem I'm having now. I think the reason the PSU -13.4V was -10.9V to begin with was due to a short which I'm now trying to fix. A short that has been there all the time on this mobo. The +12V failed a week ago, I fixed it, and

If there's an open circuit problem on the way from PSU -12V to opamp -12V, I have a hard time to see how it could bring it up to -7V. But the -12V is pretty weak, only 0.3A, so not impossible I guess.

So there's the problem in a CLIshell.

Originally Posted by Allen1

Hi there Photon,
If there is no load being drawn as in connecting the outputs to an amp etc or if a component has failed as in those 2 resistors, then I would expect to be getting strange voltage readings

- even on the -12V which "isn't connected to anything"??

But yes, I will def. check the output/dead IC/+5V before swapping Paula I think I checked the resistors but I'll double check. If there's ANOTHER "instant open circuit corroded via" on Paula Vcc, that could be a possible "open circuit drags down PSU" as per Loedowns description there.


Mainly I'm looking for "other chips that use -12V", Loedown hinted at one so I hope you can help me find it...

Edit: OK, got my LF347M. I'm going to make *damn* sure the -12V is okay before I even think about replacing the opamp (even if it's not that hard).

[Allen1]

From the 1200 schematics, I can see that U28 via D576 is on the -12v line (line receiver on RS323 port) , you already know about U15 via R302.

I hope you find the fault but I think that a sudden failure like this is more likely to be the corroding effect on that single track from that capacitor you mentioned or a simple thing like a resistor breaking, all of which are easy fixes, this one though seems to be a lot more complex than it should be. Lets hope that it is the op-amp that has failed as there isn't a lot of the circuit that hasn't already been looked at.

[Photon]

Originally Posted by Allen1

From the 1200 schematics, I can see that U28 via D576 is on the -12v line

I traced -12V to the via near RS232 yesterday, but I can't trace -12V to any pin on U28. If it's connected to it, the trace to the pin is under the IC itself. Is it supposed to/what did you mean?

[8bitbubsy]

Loedown: When you quote my post and say "U400 is Paula, I want you to look long and hard and tell me where it has a negative voltage connection.", then I get really confused. I never said it has a negative voltage connection, but I was indeed wrong with my output volt statement. I thought Paula had an integrated inverter circuit for aligning the DC output.

[Loedown]

Originally Posted by 8bitbubsy

Loedown: When you quote my post and say "U400 is Paula, I want you to look long and hard and tell me where it has a negative voltage connection.", then I get really confused. I never said it has a negative voltage connection, but I was indeed wrong with my output volt statement. I thought Paula had an integrated inverter circuit for aligning the DC output.

No this is true but you implied that it may swing from a positive output to a negative on the audio outputs.

I want to make a correction to my descriptions about the op amps and Paula output, if Paula has a 0V on output in relation to the op amp setup then it would actually drive the output rail at the RCA connections to fully negative. I haven't got a clear bench at the moment but when I do I'll run up my A1200 and give you proper readings from various points in the circuit.

@Photon: Which Amiga model are you actually having issues with an A500 or an A1200 and if A1200 which revision? Ignore me, I SHOULD READ THE BLOODY TITLE

The A1200 is indeed quite a different circuit to the A4000 audio, let me have a look either tomorrow or later, I am fixing up my garage at the moment and will not have a chance to get much done before I finish that.