F5 power amplifier

[AndrewT]

It's not just mains frequency ripple components that exist on the supply rails.
The signal frequency is also seen as a ripple on the supply rails.
At higher audio frequencies one can observe the saw tooth profile of the mains re-charging the supply voltage followed by the ramp of discharge and superimposed on this is the music wave form.
As the caps get smaller the charge and discharge ramps get steeper, but at the same repetition frequency and the music waveform becomes more prominent.

The tops of these waveforms (maximum supply voltage) are about the same as the quiescent voltage on the PSU when zero audio signal is on the output of the amplifier.
The minimum of the ripple will be at it's maximum during quiescent conditions. At this time the ripple will be small due to the low current demand of the amplifier. The AMP PSRR will determine how much of this quiescent ripple gets through to the speaker. This comes out as very low level hum for the 50/60Hz and 100/120Hz components. But the AMP PSRR is often at it's best at these low frequencies so the hum is suppressed very well.
The higher frequency components come out the speaker as buzz and noise (white). The very high frequencies will be attenuated less than the lower frequencies due to falling slope of the AMP PSRR as frequency rises.
As the output current to the amplifier rises the ripple on the supply lines increases.
The only thing that "hides" the increased PSU supply noise is the noise of the music. But the music cannot hide all of it. We hear some of the increased PSU noise components mixed in with the very varying music signal.
In the limit, as the output approaches maximum power of the amplifier, the troughs in the PSU ripple come down to the level that the output signal clips due to asking for too much output voltage relative to the supply voltage.

Small smoothing capacitors make this series of audio artefacts worse than when using large smoothing capacitance.
It can be ameliorated by drawing very low currents from the PSU.
i.e. play at quiet reproduction levels or use high efficiency speakers.
Peter Daniel goes the latter route. 96dB speakers on ~+-1500uF smoothing and gets sound quality that satisfies his desires and targets.
If one uses low efficiency speakers and/or demands high SPLs then small smoothing capacitors will deteriorate the sound quality that comes from the speakers.

[nicoch58]

nice explanation AndrewT

[AndrewT]

Originally posted by nicoch46
nice explanation

all of that explanation is in the Forum.
But some folk refuse to do the research.
Be informed before you make decisions.
It applies to all your lifestyle decisions.

If I cared to, I could check just how much of that I have learned since I joined the Forum. It would not be a little!

[Renron]

Originally posted by AndrewT
no they don't.
The usual RCCB GFCI run the two phase leads, flow (Live) and return (Neutral) through a current transformer.
The output of the current transformer detects the DIFFERENCE between the flow and return currents.
The trip activates when this difference exceeds the set limit. (in the UK we generally have 10mA 30mA and 100mA RCCBs for domestic use).

AndrewT,
Thanks for catching that, my mind and fingers were on two separate thought tracks at the time............ Don't know how I got that confused. I know better!

Thanks,
Ron

[Renron]

I took too long, and the edit time ran out on my above post, so here it is;

I wrote, "GFCIs are designed to trip their internal breakers when they sense returning voltage on the ground leg instead of the neutral leg"
I should have been exact and left out the word "leg" in the above ground reference. My bad.


Ripped straight from EC&M . Com (Electrical Construction and Maintenance)
"If a ground fault exists with some of the current flowing to ground and not returning on the neutral through the transformer, then the sum of the current flowing on the hot and neutral will not be zero and differential current will be detected. The GFCI's electronic circuitry then measures its magnitude. If it reaches a predetermined level (the GFCI trip threshold) for a given duration, a signal causes the trip coil to energize and the circuit to open. Written by Jack Wells, Vice President of Corporate Development for Pass & Seymour/Legrand.

UL requires the GFCI's to trip at 6mA, although it may take several seconds at this amperage level.

Ron

[AndrewT]

Originally posted by Renron
I wrote, "GFCIs are designed to trip their internal breakers when they sense returning voltage on the ground leg instead of the neutral leg"
I should have been exact and left out the word "leg" in the above ground reference.


Ripped straight from EC&M . Com (Electrical Construction and Maintenance)
"If a ground fault exists with some of the current flowing to ground and not returning on the neutral through the transformer, then the sum of the current flowing on the hot and neutral will not be zero and differential current will be detected. The GFCI's electronic circuitry then measures its magnitude. If it reaches a predetermined level (the GFCI trip threshold) for a given duration, a signal causes the trip coil to energize and the circuit to open.............

your quote from the ECM states it correctly.
What you posted as your interpretation of the EC&M is wrong, with or without the word leg.

Originally posted by Renron
UL requires the GFCI's to trip at 6mA, although it may take several seconds at this amperage level.

is this your interpretation of what UL states?
This seems wrong as well.

[bobodioulasso]

Andrew is right.

This device detects only a difference between neutral and live currents.

The current lost can go directly to ground if the device to protect is grounded

or through your body if the device to protect is not grounded.

This permits to protect persons in contact with the live wire: the current

through their body cannot exceed a determinated value.

[Renron]

AndrewT,
Your knowledge of electronics far exceeds mine, but I can read.

Quote from same article written by Jack Wells, the VP of Pass & Seymore/Legrand. Perhaps he is wrong too.

"A GFCIs. It's important you note that the "let-go" threshold is between 6mA and 18mA. The trip level of 6mA required for personnel-protection GFCIs by UL is based on this let-go current threshold."

"When installing and testing GFCIs, you should know that they do not trip instantaneously. In fact, while they typically trip in 25 ms or so at fault currents exceeding 20 to 30mA, they are permitted by UL to take several seconds to trip at fault currents in the 6mA range."

If you would care to read the entire article in whole, you can find it here:
EC&M article on GFCI

Man, you are one tough cookie!

In practice, if the current is flowing through the ground (wire or earth) then there is a "differential current", and the breaker is tripped. It is sensed by the differential between the hot and neutral, I agreed with you on that earlier. But if a differential exists it has to be going somewhere------ground wire or earth ground, like I said "instead of the neutral leg".

[AndrewT]

Originally posted by Renron
Quote from same article written by Jack Wells, the VP of Pass & Seymore/Legrand. Perhaps he is wrong too.

"A GFCIs. It's important you note that the "let-go" threshold is between 6mA and 18mA. The trip level of 6mA required for personnel-protection GFCIs by UL is based on this let-go current threshold."

"When installing and testing GFCIs, you should know that they do not trip instantaneously. In fact, while they typically trip in 25 ms or so at fault currents exceeding 20 to 30mA, they are permitted by UL to take several seconds to trip at fault currents in the 6mA range."

If you would care to read the entire article in whole, you can find it here:
EC&M article on GFCI

Man, you are one tough cookie!

In practice, if the current is flowing through the ground (wire or earth) then there is a "differential current", and the breaker is tripped. It is sensed by the differential between the hot and neutral, I agreed with you on that earlier. But if a differential exists it has to be going somewhere------ground wire or earth ground, like I said "instead of the neutral leg".

this long reply confirms that you acknowledge that your posted interpretations of what was printed were wrong.

The printed versions are right, it's the way you are reading them and drawing conclusions that is suspect.

[Renron]

I see how my statement was misinterpreted, by my choice of wording was poor. I apologize.
The GFCI does not sense the earth/ground, but the differential current between the hot and neutral when that differential current is flowing through the ground wire or earth ground instead of the neutral.

OK?

UL requires the GFCI's to trip at 6mA, although it may take several seconds at this amperage level.

is this your interpretation of what UL states?
This seems wrong as well.

Comments?