"Arfa Daily"

The power supplies in guitar tube amps are universally linear. Simple
transformer, full wave rectifier, reservoir cap followed by L-C or R-C
filtering. Rails up to 500v can typically be found feeding the output
stage, and it's not at all unusual to see ripple voltages of 5 volts or
more on these high voltage rails.

** The ripple voltage on the first filter electro is typically 10% of the DC
level - at ful load.

This means up to 50V p-p for a 500V rail.


The hum on the output is barely noticeable,

** The push-pull output stage rejects supply hum almost completely - until
the amp goes into clipping.

However, it is crucial that the screen supply is hum free or the output will
be amplitude modulated at 100 Hz.

Fender made a blunder with this when they built the Super Twin.


.... Phil

"Phil Allison" wrote in message
...

"Arfa Daily"

The power supplies in guitar tube amps are universally linear. Simple
transformer, full wave rectifier, reservoir cap followed by L-C or R-C
filtering. Rails up to 500v can typically be found feeding the output
stage, and it's not at all unusual to see ripple voltages of 5 volts or
more on these high voltage rails.

** The ripple voltage on the first filter electro is typically 10% of the
DC level - at ful load.

This means up to 50V p-p for a 500V rail.


The hum on the output is barely noticeable,

** The push-pull output stage rejects supply hum almost completely - until
the amp goes into clipping.

However, it is crucial that the screen supply is hum free or the output
will be amplitude modulated at 100 Hz.

Fender made a blunder with this when they built the Super Twin.


... Phil


Yes, all agreed, and of course the reason that the early stages *do* have to
have a smooth anode supply is that they are single ended, unlike the output
stage.

Arfa

I've seen very high capacitance readings on failing electrolytics, on
some cheap meters. About 20 years ago, I went through my inventory of
caps and chuck any electrolytics that read higher than specs. When I
got the ESR meter and built it, I tossed another 20%. Being in my 60s,
some parts were made in the '40s since I had bought out a number of old
shops in the '70s & '80s. ;-)


The cap you described was just starting to fail. First, the ESR
starts to rise, then the capacitance drops as it dries out, even more.
I used to do some failure analysis on components at one plant.

Yes, that was the post, and yes, I think you are right that the cap was in
early stage failure. For sure, it showed no external signs of stress

Arfa

Today's electrolytic failure ...

Teac table-top system - a bit Bose Wave Radio-ish. Nasty hum at power up.
4700 uF main filter cap located right next to the output chip heatsink. So
open that the Bob P meter couldn't read it at all. The little Chinese
component tester managed to identify it as a capacitor, at least. Value ?
44 pF ... :-)

Arfa

Arfa Daily wrote:



I've seen very high capacitance readings on failing electrolytics, on
some cheap meters. About 20 years ago, I went through my inventory of
caps and chuck any electrolytics that read higher than specs. When I
got the ESR meter and built it, I tossed another 20%. Being in my 60s,
some parts were made in the '40s since I had bought out a number of old
shops in the '70s & '80s. ;-)


The cap you described was just starting to fail. First, the ESR
starts to rise, then the capacitance drops as it dries out, even more.
I used to do some failure analysis on components at one plant.

Yes, that was the post, and yes, I think you are right that the cap was in
early stage failure. For sure, it showed no external signs of stress

Arfa

Today's electrolytic failure ...

Teac table-top system - a bit Bose Wave Radio-ish. Nasty hum at power up.
4700 uF main filter cap located right next to the output chip heatsink. So
open that the Bob P meter couldn't read it at all. The little Chinese
component tester managed to identify it as a capacitor, at least. Value ?
44 pF ... :-)


I didn't know the Chinese made Electrolytics that small. ;-)

On Wed, 10 Jul 2013 12:55:38 +0100, "Arfa Daily"
wrote:



"Phil Allison" wrote in message
...

"Arfa Daily"

The power supplies in guitar tube amps are universally linear. Simple
transformer, full wave rectifier, reservoir cap followed by L-C or R-C
filtering. Rails up to 500v can typically be found feeding the output
stage, and it's not at all unusual to see ripple voltages of 5 volts or
more on these high voltage rails.

** The ripple voltage on the first filter electro is typically 10% of the
DC level - at ful load.

This means up to 50V p-p for a 500V rail.


The hum on the output is barely noticeable,

** The push-pull output stage rejects supply hum almost completely - until
the amp goes into clipping.

However, it is crucial that the screen supply is hum free or the output
will be amplitude modulated at 100 Hz.

Fender made a blunder with this when they built the Super Twin.


... Phil


Yes, all agreed, and of course the reason that the early stages *do* have to
have a smooth anode supply is that they are single ended, unlike the output
stage.

Arfa

Yes that is part of it, as well as the fact that there is
amplification of the early stage hum.

peter

"Peter"

Yes, all agreed, and of course the reason that the early stages *do* have
to
have a smooth anode supply is that they are single ended, unlike the
output
stage.

Arfa

Yes that is part of it, as well as the fact that there is
amplification of the early stage hum.


** Yep - that is why there are several stages of filtering between the
raw DC ( at the first filter electro ) and the input triode.

Having a choke ( ie inductor) after the first electro does amazing job
reducing ripple by a factor of a hundred or more with no loss of DC voltage.
Then two or more resistor & electro stages are used.

The supply to the input triode need to have less than 50uV of 100Hz ripple
to ensure no audible hum.

Makes the overall ripple reduction in the order of 100,000 to 1.



..... Phil

Today's tale of capacitor woe ...

Panasonic SA-PTX7 home cinema. Lit up for a couple of seconds then died each
time it was powered. Two caps at fault this time, C784 and C794, both 330uF
at 6.3v working. Both got 3.3v regulated across them. Both part of an L-C
filter. Neither near anything hot (for a change ...) and both got bulging
tops ...

About all you can say for this modern capacitor malaise, is that it's good
for business ... d:-)

Arfa