One channel on stereo amp blows filter cap on negative rail

[Tiger Luck]

For the record, the amp is in an Advent 300, which I refuse to let die.

Here's a PDF of the schematics:

http://www.davidreaton.com/PDFs/Advent_300_newer_Amplifier_schematic.pdf

The cap that blows is C313. It is a 100mfd 35volt cap It is the filter cap
connecting the negative rail in the lower channel to ground. It is on the far lower
right of the schematic, just above the separate, isolated schematic of the 15 volt
regulator circuit.

The voltage on the negative rail is as it should be. The polarity of the cap was
correct. The corresponding cap in the other channel does not blow. Both caps have
the negative connected to the same source from the power supply and both have their
positives connected to ground. In other words, the negs of both caps are connected
together to the negative rail from the power supply and both positives are grounded,
yet one cap blows and the other doesn't. The cap blew with a sharp crack and the
rubber plug at the bottom blew out. The cap that blew is not shorted. It still acts
like a cap should when measured with an analog ohm meter. I installed another, new
cap of equal value and it started heating up, so I cut the power.

I'm not a pro tech, but fairly knowledgeable and good at logic and such.

Any insights?



--
Einstein forgot to carry the two

 

[Tiger Luck]

Soundhaspriority wrote:

"Tiger Luck" <Tiger_Luck@natures_preserve.afr> wrote in message
news:uy_xk.9006$np7.5413@flpi149.ffdc.sbc.com...
For the record, the amp is in an Advent 300, which I refuse to let die.

Here's a PDF of the schematics:

http://www.davidreaton.com/PDFs/Advent_300_newer_Amplifier_schematic.pdf

The cap that blows is C313. It is a 100mfd 35volt cap It is the filter cap
connecting the negative rail in the lower channel to ground. It is on the
far lower right of the schematic, just above the separate, isolated
schematic of the 15 volt regulator circuit.

The voltage on the negative rail is as it should be. The polarity of the
cap was correct. The corresponding cap in the other channel does not blow.
Both caps have the negative connected to the same source from the power
supply and both have their positives connected to ground. In other words,
the negs of both caps are connected together to the negative rail from the
power supply and both positives are grounded, yet one cap blows and the
other doesn't. The cap blew with a sharp crack and the rubber plug at the
bottom blew out. The cap that blew is not shorted. It still acts like a
cap should when measured with an analog ohm meter. I installed another,
new cap of equal value and it started heating up, so I cut the power.

I'm not a pro tech, but fairly knowledgeable and good at logic and such.

Any insights?

Take a look at the points where the + and - caps are grounded. If that part
of the ground bus is intact, but the bus is broken on the way to the supply
ground, so that it is isolated, that point could float up towards +22 volts,
making the total voltage over the cap that blows 44 volts. Which cap blows
is whichever leaks less, just a matter of chance. So check that ground bus
very carefully for integrity.

Bob Morein
(310) 237-6511

Hi,Bob and thanks

The grounds are intact. By that I mean to say that point in the CB where
electrolytic caps C312 and C313 are grounded is continuous with the ground at the
power supply filter caps. Chassis is the same. All points which are supposed to be
grounded are continuous as read by an ohm meter.


--
Einstein forgot to carry the two

 

[Tiger Luck]

stratus46@yahoo.com wrote:

Tiger Luck wrote:
snip
The grounds are intact. By that I mean to say that point in the CB
where
electrolytic caps C312 and C313 are grounded is continuous with the
ground at the
power supply filter caps. Chassis is the same. All points which are
supposed to be
grounded are continuous as read by an ohm meter.

Those little 100uF caps are _not_ the main 'lytics. The mains would
likely be 1000uF to 10000uF (or possibly even more) and be located
near the main bridge rectifier - which is on the _rest_ of the
schematic along with the power transformer.. I think you'll find the
main negative bus 'lytic to be open or dried out so that you're
putting LOTS of ripple current on those littlle 100uF caps. Those
100uF caps are just local 'clean up' caps and can't handle the load.

G˛

My mistake to refer to them as filter caps.

--
Einstein forgot to carry the two

 

[Tiger Luck]

Chris Hornbeck wrote:

On Wed, 10 Sep 2008 21:23:45 -0700, Tiger Luck
Tiger_Luck@natures_preserve.afr> wrote:

The best odds-on bet is that despite everything posted
the cap is reversed. Sorry, but that's the main line.

Not a chance... cap installed correctly and worked for years.

Are you saying that *ONE* capacitor has failed? Not
a series of replacements? The thing's 30 years old - *all*
of the electrolytic cap's are bad now. That's life.

Replace *all* of the electrolytics and move on. They're
*all* bad. Parts are cheap.

I replaced all the electro caps less than two years ago. Amp worked fine up until
now. Polarity was and is correct. After cap blew today, new replacement cap over
heated. I shut off power before it blew out.

BTW, if your question had been about repeated failures,
the diagnosis would be: the cap got too hot. It got too
hot because it was defective, or it was exposed to too
high a voltage,

I checked voltage across cap terminals with cap out. It read correctly.

or it was exposed to reverse voltage,

Trust me. Polarity was correct.

or it was cooked by ESR.

Equivalent series resistance? Is that a measure of quality?

Caps are no-names, both from the same parts house. hmm...

So, for future troubleshooting, eliminate these four
and you'll have only a mystery remaining!

Good fortune,
Chris Hornbeck

--
Einstein forgot to carry the two

 

Tiger Luck wrote:
<snip>

The grounds are intact. By that I mean to say that point in the CB
where
electrolytic caps C312 and C313 are grounded is continuous with the
ground at the
power supply filter caps. Chassis is the same. All points which are
supposed to be
grounded are continuous as read by an ohm meter.

Those little 100uF caps are _not_ the main 'lytics. The mains would

likely be 1000uF to 10000uF (or possibly even more) and be located
near the main bridge rectifier - which is on the _rest_ of the
schematic along with the power transformer.. I think you'll find the
main negative bus 'lytic to be open or dried out so that you're
putting LOTS of ripple current on those littlle 100uF caps. Those
100uF caps are just local 'clean up' caps and can't handle the load.

G˛

 

[Kevin McMurtrie]

In article <uy_xk.9006$np7.5413@flpi149.ffdc.sbc.com>,
Tiger Luck <Tiger_Luck@natures_preserve.afr> wrote:

For the record, the amp is in an Advent 300, which I refuse to let die.

Here's a PDF of the schematics:

http://www.davidreaton.com/PDFs/Advent_300_newer_Amplifier_schematic.pdf

The cap that blows is C313. It is a 100mfd 35volt cap It is the filter cap
connecting the negative rail in the lower channel to ground. It is on the far
lower
right of the schematic, just above the separate, isolated schematic of the 15
volt
regulator circuit.

The voltage on the negative rail is as it should be. The polarity of the cap
was
correct. The corresponding cap in the other channel does not blow. Both caps
have
the negative connected to the same source from the power supply and both have
their
positives connected to ground. In other words, the negs of both caps are
connected
together to the negative rail from the power supply and both positives are
grounded,
yet one cap blows and the other doesn't. The cap blew with a sharp crack and
the
rubber plug at the bottom blew out. The cap that blew is not shorted. It
still acts
like a cap should when measured with an analog ohm meter. I installed
another, new
cap of equal value and it started heating up, so I cut the power.

I'm not a pro tech, but fairly knowledgeable and good at logic and such.

Any insights?

Part of the circuit might have failed so that it oscillates or pulses.
For example - R313, C305, D301, C309, R335, or any of that bias circuit.
Drag the leads of a big foil capacitor across the leads of C313. It
should spark once to charge it. If it keeps sparking (or smokes),
there's AC current there.

The great thing about stereo amps is that you can put a voltmeter
between the matching parts of each side to find faults.

--
Google is a pro-spamming service. I will not see your reply if you use Google.

 

[Franc Zabkar]

On Wed, 10 Sep 2008 18:29:26 -0700, Tiger Luck
<Tiger_Luck@natures_preserve.afr> put finger to keyboard and composed:

For the record, the amp is in an Advent 300, which I refuse to let die.

Here's a PDF of the schematics:

http://www.davidreaton.com/PDFs/Advent_300_newer_Amplifier_schematic.pdf

The cap that blows is C313. It is a 100mfd 35volt cap It is the filter cap
connecting the negative rail in the lower channel to ground. It is on the far lower
right of the schematic, just above the separate, isolated schematic of the 15 volt
regulator circuit.

The voltage on the negative rail is as it should be. The polarity of the cap was
correct. The corresponding cap in the other channel does not blow. Both caps have
the negative connected to the same source from the power supply and both have their
positives connected to ground. In other words, the negs of both caps are connected
together to the negative rail from the power supply and both positives are grounded,
yet one cap blows and the other doesn't. The cap blew with a sharp crack and the
rubber plug at the bottom blew out. The cap that blew is not shorted. It still acts
like a cap should when measured with an analog ohm meter. I installed another, new
cap of equal value and it started heating up, so I cut the power.

I'm not a pro tech, but fairly knowledgeable and good at logic and such.

Any insights?

Maybe the trace from that channel to the -22V rail is partially open
(ie resistive), allowing C313 to be reverse charged from the +22V rail
via Q317 ???

Just for testing purposes, try running two hookup wires between the
caps in both channels.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

 

[Tiger Luck]

stratus46@yahoo.com wrote:

Tiger Luck wrote:
Chris Hornbeck wrote:
On Wed, 10 Sep 2008 21:23:45 -0700, Tiger Luck
Tiger_Luck@natures_preserve.afr> wrote:

The best odds-on bet is that despite everything posted
the cap is reversed. Sorry, but that's the main line.

Not a chance... cap installed correctly and worked for years.

Are you saying that *ONE* capacitor has failed? Not
a series of replacements? The thing's 30 years old - *all*
of the electrolytic cap's are bad now. That's life.

At work we buy 100uF 25 volt caps in bags of 1000 and have so far
installed 2500 in 20 year old gear, mostly Sony. But we also have some
Ampex 2" quad machines built in 1978 with _all_ the orignal
electrolytics and the machines are in cherry condition. 30 years is
not necessarily the end point for caps.


Replace *all* of the electrolytics and move on. They're
*all* bad. Parts are cheap.

I replaced all the electro caps less than two years ago. Amp worked
fine up until
now. Polarity was and is correct. After cap blew today, new
replacement cap over
heated. I shut off power before it blew out.

What do you see with a scope? Something definitely funny going on. Is
the transformer heating up too?

Transformer has stayed cool. The 3 amp fuses on the positive and negative rails are
not blowing at this point. I don't have a scope, but have a line on a Philips PM3050
for $85. There is more going on here than I can read at the moment. I have the
output pair on the left channel yanked at this point because they shorted out. Now
the output pair on the right channel are heating up when the power is on. The
electro cap that blew is on the right channel. Part of the 15 volt regulator
circuitry failed and I replaced some parts there. But that's it for tonight.

BTW, if your question had been about repeated failures,
the diagnosis would be: the cap got too hot. It got too
hot because it was defective, or it was exposed to too
high a voltage,

I checked voltage across cap terminals with cap out. It read
correctly.

or it was exposed to reverse voltage,

Trust me. Polarity was correct.

or it was cooked by ESR.

Equivalent series resistance? Is that a measure of quality?

Oh yeah, ESR is a big deal.


Caps are no-names, both from the same parts house. hmm...


For little ( 1-4700 uF) caps we were using Panasonic FC series but
after reading some Hydrogenaudio forum posts, I found that the FM
series caps are even lower in ESR, longer llife (2000 vs 1000 hrs at
105 C) and they cost a little less. ($57.54 vs $61 / 1000 for 100uF
25V)


So, for future troubleshooting, eliminate these four
and you'll have only a mystery remaining!



Good fortune,
Chris Hornbeck

Definitely want to hear how this shakes out.

G˛

--
Einstein forgot to carry the two

 

Tiger Luck wrote:

Chris Hornbeck wrote:
On Wed, 10 Sep 2008 21:23:45 -0700, Tiger Luck
Tiger_Luck@natures_preserve.afr> wrote:

The best odds-on bet is that despite everything posted
the cap is reversed. Sorry, but that's the main line.

Not a chance... cap installed correctly and worked for years.

Are you saying that *ONE* capacitor has failed? Not
a series of replacements? The thing's 30 years old - *all*
of the electrolytic cap's are bad now. That's life.

At work we buy 100uF 25 volt caps in bags of 1000 and have so far
installed 2500 in 20 year old gear, mostly Sony. But we also have some
Ampex 2" quad machines built in 1978 with _all_ the orignal
electrolytics and the machines are in cherry condition. 30 years is
not necessarily the end point for caps.

Replace *all* of the electrolytics and move on. They're
*all* bad. Parts are cheap.

I replaced all the electro caps less than two years ago. Amp worked
fine up until
now. Polarity was and is correct. After cap blew today, new
replacement cap over
heated. I shut off power before it blew out.

What do you see with a scope? Something definitely funny going on. Is
the transformer heating up too?

BTW, if your question had been about repeated failures,
the diagnosis would be: the cap got too hot. It got too
hot because it was defective, or it was exposed to too
high a voltage,

I checked voltage across cap terminals with cap out. It read
correctly.

or it was exposed to reverse voltage,

Trust me. Polarity was correct.

or it was cooked by ESR.

Equivalent series resistance? Is that a measure of quality?

Oh yeah, ESR is a big deal.

Caps are no-names, both from the same parts house. hmm...

For little ( 1-4700 uF) caps we were using Panasonic FC series but
after reading some Hydrogenaudio forum posts, I found that the FM
series caps are even lower in ESR, longer llife (2000 vs 1000 hrs at
105 C) and they cost a little less. ($57.54 vs $61 / 1000 for 100uF
25V)

So, for future troubleshooting, eliminate these four
and you'll have only a mystery remaining!



Good fortune,
Chris Hornbeck

Definitely want to hear how this shakes out.

G˛

 

[Eeyore]

Tiger Luck wrote:

Both caps have
the negative connected to the same source from the power supply and both have their
positives connected to ground. In other words, the negs of both caps are connected
together to the negative rail from the power supply and both positives are grounded,

Which is WRONG ! Read the diagram !

The negative of C311 should be grounded together with the POSITIVE of C313

Possibly a pcb legend error. I've known it happen.

Graham

 

[Eeyore]

DC wrote:

"Tiger Luck" <Tiger_Luck@natures_preserve.afr> wrote

Well, this may not be your problem, but things have deteriorated in the
electronic component area over the last few years. Many manufacturers have
out sourced production to China. Most if not all capacitor failures I have
seen in recent years have made in China - or no brand. Try to find something
that says made in Japan!!

Panasonic are reliable. But the op says C311 and C313 have their negatives
connected together AIUI. That's plain WRONG.

Graham

 

[Eeyore]

Arfa Daily wrote:

In many years of repairing this stuff, and replacing electrolytics that have
failed any way from passively to explosively, I have only ever found three
things which cause their demise in the way you have described.

1) Excess voltage across them
2) Incorrect polarity of voltage across them
3) Excess current through them

3) is obviously the most difficult to diagnose, as it really requires a
'scope to see what's going on.

An HF fault causing high local ripple ?

Graham

 

[Eeyore]

William Sommerwerck wrote:

I replaced all the electro caps less than two years ago. Amp
worked fine up until now. Polarity was and is correct. After
cap blew today, new replacement cap over heated. I shut off
power before it blew out.

It's remotely possible the replacement was defective. If the voltage across
the cap's location is correct in level and polarity, then that can't be the
problem.

Question... How safe is it -- other than the cap blowing -- to operate the
amp? I think you need to put a 'scope on a number of points and see what's
happening.

I don't mean to hassle you on the question of polarity, but... Japanese caps
generally have a black line on one side pointing toward the negative lead.

Or a gold one etc etc.

Graham

 

[Tiger Luck]

William Sommerwerck wrote:

I replaced all the electro caps less than two years ago. Amp
worked fine up until now. Polarity was and is correct. After
cap blew today, new replacement cap over heated. I shut off
power before it blew out.

It's remotely possible the replacement was defective. If the voltage across
the cap's location is correct in level and polarity, then that can't be the
problem.

Question... How safe is it -- other than the cap blowing -- to operate the
amp? I think you need to put a 'scope on a number of points and see what's
happening.

I don't mean to hassle you on the question of polarity, but... Japanese caps
generally have a black line on one side pointing toward the negative lead.


These caps have a line of pluses down one side pointing to the plus lead. The CB has

a plus printed next to one hole for the ECs. I installed several of these caps in
this amp. They are foolproof. The amp worked perfectly for a couple of years with
these caps installed.

I don't have a scope, but I may buy a used Philips PM3050 for $85 just to see what
I'm missing.

--
Einstein forgot to carry the two

 

[Ron(UK)]

Tiger Luck wrote:

William Sommerwerck wrote:
I replaced all the electro caps less than two years ago. Amp
worked fine up until now. Polarity was and is correct. After
cap blew today, new replacement cap over heated. I shut off
power before it blew out.

It's remotely possible the replacement was defective. If the voltage
across
the cap's location is correct in level and polarity, then that can't
be the
problem.

Question... How safe is it -- other than the cap blowing -- to operate
the
amp? I think you need to put a 'scope on a number of points and see
what's
happening.

I don't mean to hassle you on the question of polarity, but...
Japanese caps
generally have a black line on one side pointing toward the negative
lead.


These caps have a line of pluses down one side pointing to the plus
lead. The CB has a plus printed next to one hole for the ECs. I
installed several of these caps in this amp. They are foolproof. The amp
worked perfectly for a couple of years with these caps installed.

I don't have a scope, but I may buy a used Philips PM3050 for $85 just
to see what I'm missing.

Couldn`t you just put a meter across the cap to see what polarity it`s
getting?

Ron



--
Lune Valley Audio
Public Address Systems
Hire Sales Maintenance
www.lunevalleyaudio.com

 

[Dave]

"don ward" <dward7@earthlink.net> wrote in message
news:dward7-9A74AE.21552210092008@earthlink.vsrv-sjc.supernews.net...

In article <r41yk.22177$89.13971@nlpi069.nbdc.sbc.com>,
Tiger Luck <Tiger_Luck@natures_preserve.afr> wrote:

Every piece of gear I have that has + and - power buss has the
positive rail cap with the + to the rail and the - lead to ground.
HOWEVER the negative rail has the - to the rail and the + to ground.
Things are not always as they seem
dnw

Then you must not have any pieces of gear that are in working order. On the
negative rail, the - lead of the cap is on the rail and the + lead is to
ground. Can't be any other way.

 

On Sep 11, 10:33 am, Tiger Luck <Tiger_Luck@natures_preserve.afr>
wrote:

William Sommerwerck wrote:
I replaced all the electro caps less than two years ago. Amp
worked fine up until now. Polarity was and is correct. After
cap blew today, new replacement cap over heated. I shut off
power before it blew out.

It's remotely possible the replacement was defective. If the voltage across
the cap's location is correct in level and polarity, then that can't be the
problem.

Question... How safe is it -- other than the cap blowing -- to operate the
amp? I think you need to put a 'scope on a number of points and see what's
happening.

I don't mean to hassle you on the question of polarity, but... Japanese caps
generally have a black line on one side pointing toward the negative lead.

These caps have a line of pluses down one side pointing to the plus lead. The CB has
a plus printed next to one hole for the ECs. I installed several of these caps in
this amp. They are foolproof. The amp worked perfectly for a couple of years with
these caps installed.

I don't have a scope, but I may buy a used Philips PM3050 for $85 just to see what
I'm missing.

--
Einstein forgot to carry the two- Hide quoted text -

- Show quoted text -

You need to put a scope on it. I think you're going to find ac or a
lot of ripples. If you isolate the power supply don't forget to put a
load on it.

 

[Franc Zabkar]

On Thu, 11 Sep 2008 13:51:27 +0100, Eeyore
<rabbitsfriendsandrelations@hotmail.com> put finger to keyboard and
composed:

Tiger Luck wrote:

Both caps have
the negative connected to the same source from the power supply and both have their
positives connected to ground. In other words, the negs of both caps are connected
together to the negative rail from the power supply and both positives are grounded,

Which is WRONG ! Read the diagram !

The negative of C311 should be grounded together with the POSITIVE of C313

Possibly a pcb legend error. I've known it happen.

Graham

The OP wrote that "the corresponding cap in the other *channel* [C312]
does not blow", not the other *rail*.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

 

[Tiger Luck]

William Sommerwerck wrote:

Transformer has stayed cool. The 3 amp fuses on the positive and
negative rails are not blowing at this point. I don't have a 'scope, but
have a line on a Philips PM3050 for $85. There is more going on here
than I can read at the moment. I have the output pair on the left channel
yanked at this point because they shorted out. Now the output pair on
the right channel are heating up when the power is on. The electro cap
that blew is on the right channel. Part of the 15 volt regulator circuitry
failed and I replaced some parts there. But that's it for tonight.

I think you have problems that go fary beyond a bad cap. I suspect the bad
cap is a effect more than a cause.

I would start by pulling the output pair on the right channel and trying to
get everything else working correctly.

Both output pairs are pulled.

I am reading 20 volts AC to ground on each of the stepped down leads from the
transformer and 40 volts between them.

The positive rail is reading 26.3 volts DC to ground and the negative rail is
reading 26.7 volts DC to ground after going through the bridge rectifier and the
filter caps. They are reading 53 volts between them. The replacement for the electro
cap that blew is reading 26.7 volts across its terminals. It connect the negative
rail to ground. Positive side is grounded. It is no longer heating up when power is on.

With power on and about to take some voltage readings across the empty output tranny
terminals on the CB, I noticed some tiny sparking going on at the edges of the trace
leading to the collector terminal of Q317, with occasional riny puffs of smoke (see
schematics):

http://www.davidreaton.com/PDFs/Advent_300_newer_Amplifier_schematic.pdf

I removed as much solder as possible from the traces around all output terminals
with a solder wick and cleaned up the board with switch cleaner. Sparking appears to
have stopped.

Began taking various voltage measurements at same points in right and left channels.
Discovered huge variation on voltage across C-E terminals with output transistors
removed. Right side had about 26 to 27 volts across the C-E terminals of each
output transistor. The left side had no voltage across C-E terminals of one output
transistor and about 52 to 54 volts across the C-E terminals of the other output
transistor.

Also had different readings across C-E terminals of the two pairs of differential
transistors at the input. Right side had about 26-27 volts across C-E terminals of
each transistor and the left side had about 7 volts across the same terminals.
Disconnecting the base of Q302 resulted in both differential transistors on the left
reading identical to the right side across the C-E terminals. I traced it down to
where D304 connects to the base of Q302 thru a 20K resistor, R336. I reconnected the
base of Q302 and disconnecting the cathode side of D304 and got the same results as
disconnecting the base of Q302. Now it's just a matter of going through the same
procedure until I find the faulty component.

Disconnecting the base of Q302 had no effect on the voltage measurements of the
output transistors in the left channel, but disconnecting D304 did. Reading across
the C-E terminals of one output tranny on the right channel with D304 disconnected,
the digital meter read about 27 volts and then began drifting steadily lower,
similar to reading a large cap with an ohmmeter. When I switched the meter to the
C-E terminals of the other output tranny on the right side, it read higher than 27
volts by as much as the first tranny's voltage diminished when it was read.

And that's where it stands. Sorry for not reading and replying to every post, but
when I got enough insight from what I did read to see where I needed to go, I
stopped reading and went to work. Will report all findings and methods used.

Thanks all

--
Einstein forgot to carry the two

 

[Tiger Luck]

Arfa Daily wrote:

"Tiger Luck" <Tiger_Luck@natures_preserve.afr> wrote in message
news:uy_xk.9006$np7.5413@flpi149.ffdc.sbc.com...
For the record, the amp is in an Advent 300, which I refuse to let die.

Here's a PDF of the schematics:

http://www.davidreaton.com/PDFs/Advent_300_newer_Amplifier_schematic.pdf

The cap that blows is C313. It is a 100mfd 35volt cap It is the filter cap
connecting the negative rail in the lower channel to ground. It is on the
far lower right of the schematic, just above the separate, isolated
schematic of the 15 volt regulator circuit.

The voltage on the negative rail is as it should be. The polarity of the
cap was correct. The corresponding cap in the other channel does not blow.
Both caps have the negative connected to the same source from the power
supply and both have their positives connected to ground. In other words,
the negs of both caps are connected together to the negative rail from the
power supply and both positives are grounded, yet one cap blows and the
other doesn't. The cap blew with a sharp crack and the rubber plug at the
bottom blew out. The cap that blew is not shorted. It still acts like a
cap should when measured with an analog ohm meter. I installed another,
new cap of equal value and it started heating up, so I cut the power.

I'm not a pro tech, but fairly knowledgeable and good at logic and such.

Any insights?



--
Einstein forgot to carry the two

In many years of repairing this stuff, and replacing electrolytics that have
failed any way from passively to explosively, I have only ever found three
things which cause their demise in the way you have described.

1) Excess voltage across them
2) Incorrect polarity of voltage across them
3) Excess current through them

3) is obviously the most difficult to diagnose, as it really requires a
'scope to see what's going on. However, given that this cap is effectively
in parallel with the main filter cap for the output stage negative rail, and
the one for the other channel, which is also connected to the same place for
all intents and purposes, it's hard to see how there could be excessive ac
across one of them on its own. Have you tried running the amp without the
offending cap in ? It should almost certainly run normally without it, as
small caps in that position are usually just for 'belt and braces'
decoupling purposes.

I know that you have stated several times that the cap is in correctly, but
are you *absolutely totally completely and without question* sure ? Noted
that you replaced the caps a couple of years back and have not had trouble
in the intervening period, but that is not necessarily an indication that
all was well. I have seen equipment where caps have been fitted backwards by
the manufacturers, and have lasted a long time before finally failing. If
you are basing your latest replacements which are failing, on the way you
took out the first failed one - or even based on the way that the board is
marked, because I've seen them silkscreened backwards as well - I would say
please check very very carefully, go away and have a cup of coffee, then
come back and check again. It's easy to convince yourself that you are
seeing something you actually aren't. It has happened to me many times over
the years ...

I too would like to know what you eventually find, as the circumstances you
are describing are a bit of a conundrum. If it does come down to something
you have done, please tell us. I'm sure no-one will think any the less of
you. If we're honest, we've all been there d;~}

Arfa

Arfa, an update as requested and as promised.

First, a repost of my reply to William Sommerwerck posted today, 9/11/08 at 2:11pm
so that everything is in one post:
----------------------------------

Both output pairs are pulled.

I am reading 20 volts AC to ground on each of the stepped down leads from the
transformer and 40 volts AC between them.

The positive rail is reading 26.3 volts DC to ground and the negative rail is
reading 26.7 volts DC to ground after going through the bridge rectifier and the
filter caps. They are reading 53 volts between them. The replacement for the electro
cap that blew is reading 26.7 volts across its terminals. It connects the negative
rail to ground. Positive side is grounded. It is no longer heating up when power is on.

With power on and about to take some voltage readings across the empty output tranny
terminals on the CB, I noticed some tiny sparking going on at the edges of the trace
leading to the collector terminal of Q317, with occasional tiny puffs of smoke (see
schematics):

http://www.davidreaton.com/PDFs/Advent_300_newer_Amplifier_schematic.pdf

I removed as much solder as possible from the traces around all output terminals
with a solder wick and cleaned up the board with switch cleaner. Sparking appears to
have stopped.

Began taking various voltage measurements at same points in right and left channels.
Discovered huge variation on voltage across C-E terminals with output transistors
removed. Right side had about 26 to 27 volts across the C-E terminals of each
output transistor. The left side had no voltage across C-E terminals of one output
transistor and about 52 to 54 volts across the C-E terminals of the other output
transistor.

Also had different readings across C-E terminals of the two pairs of differential
transistors at the input. Right side had about 26-27 volts across C-E terminals of
each transistor and the left side had about 7 volts across the same terminals.
Disconnecting the base of Q302 resulted in both differential transistors on the left
reading identical to the right side across the C-E terminals. I traced it down to
where D304 connects to the base of Q302 thru a 20K resistor, R336. I reconnected the
base of Q302 and disconnecting the cathode side of D304 and got the same results as
disconnecting the base of Q302. Now it's just a matter of going through the same
procedure until I find the faulty component.

Disconnecting the base of Q302 had no effect on the voltage measurements of the
output transistors in the left channel, but disconnecting D304 did. Reading across
the C-E terminals of one output tranny on the right channel with D304 disconnected,
the digital meter read about 27 volts and then began drifting steadily lower,
similar to reading a large cap with an ohmmeter. When I switched the meter to the
C-E terminals of the other output tranny on the right side, it read higher than 27
volts by as much as the first tranny's voltage diminished when it was read.

And that's where it stands. Sorry for not reading and replying to every post, but
when I got enough insight from what I did read to see where I needed to go, I
stopped reading and went to work. Will report all findings and methods used.
--------------------------------------

Update:

Narrowed down the search to the upper section of the right channel as seen on the
schematics at the link above. Following the logic outlined above, I reconnected
everything and disconnected the base of Q308 and turned the power on to check
voltages discrepancies outlined above. With the unit on its side and probing the
underside of the circuit board, smoke began rising from the top side opposite me. I
cut power and found the following fried components:

Q308
Q310
R326
R328
R324

Pulled all the fried components and checked. Q308, Q310, R326 and R328 were toast
and I replaced them. R324 was still good, but replaced it. Everything except the
output trannies are in place and the voltage discrepancies between the two channels
have disappeared. I will now install the output trannies and give it a go.


--
Einstein forgot to carry the two