Hi,

this is the amp schematic:

http://tinypic.com/r/wlxjrm/9

The amplifier works normally, except on Power up and Power down, there is a
huge low frequency DC offset going on which sounds and is very alarming.
Here's a scope capture.
http://tinypic.com/r/5pjslw/9

The bottom blue trace is the Mute signal at the base of Q30.
The Yellow trace is the power amp output, at 20v per division, so pretty
much full swing.

All these traces are without any load, but it does pretty much the same
thing with a load.

The first mute signal positive is at switch on, the second is at switch off.
Horizontal sweep is one second per division, so what you see is turn on,
wait 4 seconds, then turn off.
There is a sine wave running through the input, which you can see after both
mutes unmute, (the fat lines below the zero point), showing the amp is
indeed muting and passing the signal when it should.

If you disable the mute signal by shorting C151, thus connecting Q30's gate
to ground and turning it on constantly, you get the following trace.
http://tinypic.com/r/4szokx/9

There is a switch on spike, but no long drawn out DC stuff going on any
more, which leads me to suspect the muting is causing this to happen.
As you can see by the fat line, the signal is now passed immediately at
switch on, as the mute is now disabled.


I can't find anything obviously wrong, and the amp works normally except for
the huge anomalies at switch on and off.
The negative feedback seems to be in circuit, the Power supplies are all
normal and power up immediately.

I changed Q30 for good measure, no different.

This is mostly SMD stuff, so not that quick and easy to whip stuff off and
on again for testing.


Amplifiers are not my most favourite circuit, and I don't know a lot about
how DC offsets are managed in them, but it seems this one definitely has a
problem managing the DC.




Cheers,


Gareth.

Gareth Magennis wrote:


this is the amp schematic:

http://tinypic.com/r/wlxjrm/9

The amplifier works normally, except on Power up and Power down, there is a
huge low frequency DC offset going on which sounds and is very alarming.
Here's a scope capture.
http://tinypic.com/r/5pjslw/9

The bottom blue trace is the Mute signal at the base of Q30.
The Yellow trace is the power amp output, at 20v per division, so pretty
much full swing.

All these traces are without any load, but it does pretty much the same
thing with a load.

The first mute signal positive is at switch on, the second is at switch off.
Horizontal sweep is one second per division, so what you see is turn on,
wait 4 seconds, then turn off.
There is a sine wave running through the input, which you can see after both
mutes unmute, (the fat lines below the zero point), showing the amp is
indeed muting and passing the signal when it should.

If you disable the mute signal by shorting C151, thus connecting Q30's gate
to ground and turning it on constantly, you get the following trace.
http://tinypic.com/r/4szokx/9

There is a switch on spike, but no long drawn out DC stuff going on any
more, which leads me to suspect the muting is causing this to happen.
As you can see by the fat line, the signal is now passed immediately at
switch on, as the mute is now disabled.


I can't find anything obviously wrong, and the amp works normally except for
the huge anomalies at switch on and off.
The negative feedback seems to be in circuit, the Power supplies are all
normal and power up immediately.

I changed Q30 for good measure, no different.

This is mostly SMD stuff, so not that quick and easy to whip stuff off and
on again for testing.


Amplifiers are not my most favourite circuit, and I don't know a lot about
how DC offsets are managed in them, but it seems this one definitely has a
problem managing the DC.




** SRM450 is the model number of a Mackie powered loudspeaker, the schem is for the amplifier module that drives the 12inch woofer.

Thing to note is that such units undergo frightening amounts of vibration plus the occasional impact with the floor - which can put cracks in the PCB and or loosen components that are not extraordinarily well held down.

No amount of perusing and pondering over a schem will find any of that.

Nevertheless, the muting system is a common one in that the front end of the amplifier is starved of current so it can no longer drive the following class A stage and hence the output stage. Q30 does this by disabling Q29 & Q31 and hence both input diff pairs ( Qs 25 thru 28)

So, one needs to engage the mute circuit ( apply +10 or more to Q30's base) and see why that is NOT happening correctly. The amplifier circuitry is symmetrical so look for a voltage that is not reflected on the opposite side in mute mode.



..... Phil

Oops, my mistake, Q26 CB junction is reverse biased, so it must be Q7
conducting via junction CB and the 15v rail.


Gareth.

Gareth Magennis wrote:




Anyway, it turned out Q7 was leaky, causing Diode string D23/24/25 to
conduct, presumably via Q26 CB junction, D40, and off to the amps output.
This turns on Q4 and Q5 MJE devices, causing the output to swing positive..

This one single SOT device cost just short of £5 on Ebay, via the only
seller I trust outside Farnell and RS.


** Hmmmm, Q7 has an innocent, TO92 looking number ( 2SC3645) but turns out to be a bit special:

https://www.promelec.ru/pdf/2SA1415-2SC3645.pdf

Finding leaky BJTs is a real PITA but you persisted and got there.


..... Phil