On Mar 4, 11:09*am, "GB" wrote:
The Natural Philosopher wrote:
Dave Plowman (News) wrote:

Just solder a new 80v cap of almost any flavour over 1000uf across
whats there.

Why would these need to be 80v out of interest? What sort of DC voltages do
these amps typically operate on? I know there's a margin for peaks, but 80v
across a nominal *8 ohm resistance would provide 800w. This is an amp, not a
room heater, surely?

80 volts is the *peak* voltage, at the top of the waveform. The
wattage delivered to the speakers depends on the 'RMS' (root-mean-
square, a sort of average), which is 0.707, (for design purposes; it
varies with the waveform) of the peak.

alexander.keys1 wrote:
On Mar 4, 11:09 am, "GB" wrote:
The Natural Philosopher wrote:
Dave Plowman (News) wrote:

Just solder a new 80v cap of almost any flavour over 1000uf across
whats there.

Why would these need to be 80v out of interest? What sort of DC
voltages do these amps typically operate on? I know there's a margin
for peaks, but 80v across a nominal 8 ohm resistance would provide
800w. This is an amp, not a room heater, surely?

80 volts is the *peak* voltage, at the top of the waveform. The
wattage delivered to the speakers depends on the 'RMS' (root-mean-
square, a sort of average), which is 0.707, (for design purposes; it
varies with the waveform) of the peak.

That's still 400w average power.



--
Murphy's ultimate law is that if something that could go wrong doesn't,
it turns out that it would have been better if it had gone wrong.

In article , GB
writes
alexander.keys1 wrote:
On Mar 4, 11:09 am, "GB" wrote:
The Natural Philosopher wrote:
Dave Plowman (News) wrote:

Just solder a new 80v cap of almost any flavour over 1000uf across
whats there.

Why would these need to be 80v out of interest? What sort of DC
voltages do these amps typically operate on? I know there's a margin
for peaks, but 80v across a nominal 8 ohm resistance would provide
800w. This is an amp, not a room heater, surely?

80 volts is the *peak* voltage, at the top of the waveform. The
wattage delivered to the speakers depends on the 'RMS' (root-mean-
square, a sort of average), which is 0.707, (for design purposes; it
varies with the waveform) of the peak.

That's still 400w average power.

There are life issues too, derate working voltage to 60 or 70% of spec
and they last 10x longer. Similar issues with temp, 105degC caps last
forever when derated.

Spec'd life at max temp and voltage on an electrolytic is rarely
impressive.
--
fred
FIVE TV's superbright logo - not the DOG's, it's ********

"fred"

There are life issues too, derate working voltage to 60 or 70% of spec and
they last 10x longer.

** That is complete nonsense.

Electro life is not related to DC operating voltage, unless the rated limit
is exceeded then it is VERY short.


Similar issues with temp, 105degC caps last forever when derated.


** Also complete nonsense - there is little increase in expected life
using 105C types unless the local ambient is over 75C.

Electro lifespan is all about the quality of the airtight seal in the end of
the cap - cos the end of life occurs when most of the electrolyte has
escaped as vapour through an imperfect seal. The rate of loss of electrolyte
is a function of the local ambient temperature and ripple current ( if any)
heating the electro.

105C types can operate at higher temps, that is their only advantage.

Long life electros come in both 85C and 105 C types - all that is different
is the quality of that seal.



...... Phil

GB wrote:
alexander.keys1 wrote:
On Mar 4, 11:09 am, "GB" wrote:
The Natural Philosopher wrote:
Dave Plowman (News) wrote:
Just solder a new 80v cap of almost any flavour over 1000uf across
whats there.
Why would these need to be 80v out of interest? What sort of DC
voltages do these amps typically operate on? I know there's a margin
for peaks, but 80v across a nominal 8 ohm resistance would provide
800w. This is an amp, not a room heater, surely?
80 volts is the *peak* voltage, at the top of the waveform. The
wattage delivered to the speakers depends on the 'RMS' (root-mean-
square, a sort of average), which is 0.707, (for design purposes; it
varies with the waveform) of the peak.

That's still 400w average power.


only with a split rail supply.

Otherwise its about - with PSU sag and a safety margin the square root
of two times half the voltage minus about 15v, so 24v rms Or in power
terms about 132W into 4 ohms

Unless its split rail with two 80V caps, or an H bridge..in which case
yes, its abut 400W.

"The Natural Philosopher"

Unless its split rail with two 80V caps, or an H bridge..in which case
yes, its abut 400W.

** That is a bit optimistic.

1. If the filter electros are each rated at 80VDC working and one sensibly
allows for higher than nominal AC supply - then the *unloaded* DC rails are
likely to be +/- 74 volts.

2. If the AC tranny is of conventional design ( toroidal or E-core) then the
regulation factor is such that the average DC rail will drop under load by
about 12%.

3. If the capacitance is a typical value like 6800uF for one channel - then
the p-p ripple voltage under steady load is about 4 volts - more if the
audio test frequency is under 100 Hz.

4. An output stage is never able to swing to the full extent of the
available DC rail voltage - particularly so if the output devices are
lateral mosfets. Assuming a typical complementary Darlington output stage
with emitters supplying the load - the output can swing within about 5 volts
of the available supply with an 8 ohm load.

Adding this up we have:

74 - 9 - 2 - 5 = 58 volts peak.

So the "rms" power on sine wave into 8 ohms will be 210 watts.




...... Phil

Phil Allison wrote:
"The Natural Philosopher"
Unless its split rail with two 80V caps, or an H bridge..in which case
yes, its abut 400W.

** That is a bit optimistic.

1. If the filter electros are each rated at 80VDC working and one sensibly
allows for higher than nominal AC supply - then the *unloaded* DC rails are
likely to be +/- 74 volts.

2. If the AC tranny is of conventional design ( toroidal or E-core) then the
regulation factor is such that the average DC rail will drop under load by
about 12%.

3. If the capacitance is a typical value like 6800uF for one channel - then
the p-p ripple voltage under steady load is about 4 volts - more if the
audio test frequency is under 100 Hz.

4. An output stage is never able to swing to the full extent of the
available DC rail voltage - particularly so if the output devices are
lateral mosfets. Assuming a typical complementary Darlington output stage
with emitters supplying the load - the output can swing within about 5 volts
of the available supply with an 8 ohm load.

Adding this up we have:

74 - 9 - 2 - 5 = 58 volts peak.

No. 58 volts peak TO peak.


So the "rms" power on sine wave into 8 ohms will be 210 watts.

No, slightly over 50W.





..... Phil

"The Natural Philosopher"

Unless its split rail with two 80V caps, or an H bridge..in which case
yes, its abut 400W.


** Can you read your own words ??


..... Phil

In article , GB
wrote:
alexander.keys1 wrote:
On Mar 4, 11:09 am, "GB" wrote:
The Natural Philosopher wrote:
Dave Plowman (News) wrote:

Just solder a new 80v cap of almost any flavour over 1000uf across
whats there.

Why would these need to be 80v out of interest? What sort of DC
voltages do these amps typically operate on? I know there's a margin
for peaks, but 80v across a nominal 8 ohm resistance would provide
800w. This is an amp, not a room heater, surely?

80 volts is the *peak* voltage, at the top of the waveform. The
wattage delivered to the speakers depends on the 'RMS' (root-mean-
square, a sort of average), which is 0.707, (for design purposes; it
varies with the waveform) of the peak.

Actually for a sine wave the mean power is half the peak power. It is the
voltages that are in the ratio of 1/root(2).

That's still 400w average power.

Not seen the initial postings in this thread as it has only just started
being xposted to uk.rec.audio. However, the above depends on details like
if the amp is single-rail or dual-rail.

For example, for 100wpc (mean for sinewave) into 8 Ohms you'd need +/- 40V
rails as absolute minimum in a dual-rail design. If it were a single rail
then you'd need 0V and a 80V rail as absolute minimum. In practice a
real-world design would need headroom to drive the devices, so 0 - 80V
single rail would get you somewhat less than 100 wpc (sinewave mean). Then
if you also take into account that the PSU will probably 'sag' under
sustained load you'd find that 80V single rail would be much less when you
are trying to get sustained sinewave power from both channels of a stereo
amp.

So in practice you might find that 0 - 80V single rail ended up giving you
far less than 100wpc mean sinewave stereo. (Although peak levels for music
could be somewhat higher.)

And as I see someone else has pointed out. You also would be safest
choosing a cap whose spec voltage was well above the actual rail voltage
when the amp isn't sagging the voltage. Hence an 80V cap would be sensible
for use with a somewhat lower rail voltage.

Slainte,

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html