Any wave can (in theory) be shifted to have zero integral over any particular interval, just as a sine wave can be shifted to have non-zero integral. Digital mastering can do this so an entire piece of music has zero integral.Near zero offset with a test signal. Real world signals are not perfectly symmetrical. You could do some measurements and find that out for yourelf.
Using a mathematical model, the integral of a sine wave (or two sine waves for IM measurement) is zero. (Not the integral of the absolute value!) The integral of a real world musical signal is not zero.
What does the integral of the signal give us? The theoretical DC offset of a "perfect" amplifier with zero quiescent DC offset.
But as you say, a musical signal, on any given interval, may actually have a non-zero integral that happens to be a part of the genuine recording. This will change as the signal progresses, and so is actually low frequency AC and may get removed by a cap.
Acoustically, the total over the entirety of a recorded piece should naturally, ideally be zero, unless the air pressure in the room where the recording was made changed from beginning to end
The chip has some impedance you may want to add to the 237 ohms. As well as paralleled with the resistor to ground.with the input capacitor 1uf and series resistance of 237 ohms
i got the RC time constant as
.24 milli seconds for R = 237 ohms and C = 1uf
Any wave can (in theory) be shifted to have zero integral over any particular interval, just as a sine wave can be shifted to have non-zero integral. Digital mastering can do this so an entire piece of music has zero integral.
But as you say, a musical signal, on any given interval, may actually have a non-zero integral that happens to be a part of the genuine recording. This will change as the signal progresses, and so is actually low frequency AC and may get removed by a cap.
Acoustically, the total over the entirety of a recorded piece should naturally, ideally be zero, unless the air pressure in the room where the recording was made changed from beginning to end
Very good points. Thank you for clarifying my explanation.
The DC I referred to is better understood as a "low frequency AC signal" which will be both positive and negative throughout the program. Adding a pole to the feedback network will help greatly in mitigating this phenomenon.
R,
the 237r resistor is half of the RF filter. It should have a following cap going to Signal Ground. Expect this to be <2nF and probably a xxxpF value.
The DC blocking cap whether in the Source equipment (that could be a passive pot in the interconnect cable), or in the Power Amplifier is followed by a resistor to Signal Ground to form a high pass filter. Expect between 10k and 100k.
the 237r resistor is half of the RF filter. It should have a following cap going to Signal Ground. Expect this to be <2nF and probably a xxxpF value.
The DC blocking cap whether in the Source equipment (that could be a passive pot in the interconnect cable), or in the Power Amplifier is followed by a resistor to Signal Ground to form a high pass filter. Expect between 10k and 100k.
Co,
read some posts reporting the changes in output offset for a more thorough analysis of the way the output offset varies, a lot, with adding or removing the DC blocking caps at the input and in the NFB lower leg.
I would expect the range of offset to vary from +100mV to -100mV and even outside these values in a few extreme cases.
I would also expect a properly set up chipamp to have an output offset that is <+-5mV for all warm to hot operating conditions.
read some posts reporting the changes in output offset for a more thorough analysis of the way the output offset varies, a lot, with adding or removing the DC blocking caps at the input and in the NFB lower leg.
I would expect the range of offset to vary from +100mV to -100mV and even outside these values in a few extreme cases.
I would also expect a properly set up chipamp to have an output offset that is <+-5mV for all warm to hot operating conditions.
hi andrew there is 47k as that shunt cap as I have posted the link of the circuit....
http://www.diyaudio.com/forums/atta...6-my-opa549-schematic-does-look-ok-opa549.gif
http://www.diyaudio.com/forums/atta...6-my-opa549-schematic-does-look-ok-opa549.gif
and no RF filter, no dc block in the nfb, no HF decoupling, as well as our DC block in the input that is missing.
High Pass RC = 47k * 1uF = 47ms = rolled off Very Low Bass.
With 2u2F the RC becomes >100ms. I can't hear any further improvement down at this level.
But Salas recently reported better sound when he adopted a 220ms (? not sure) for RC.
What about the RC in the PSU?
Is it removing or affecting your bass?
High Pass RC = 47k * 1uF = 47ms = rolled off Very Low Bass.
With 2u2F the RC becomes >100ms. I can't hear any further improvement down at this level.
But Salas recently reported better sound when he adopted a 220ms (? not sure) for RC.
What about the RC in the PSU?
Is it removing or affecting your bass?
Last edited:
Any wave can (in theory) be shifted to have zero integral over any particular interval, just as a sine wave can be shifted to have non-zero integral. Digital mastering can do this so an entire piece of music has zero integral.
But as you say, a musical signal, on any given interval, may actually have a non-zero integral that happens to be a part of the genuine recording. This will change as the signal progresses, and so is actually low frequency AC and may get removed by a cap.
Acoustically, the total over the entirety of a recorded piece should naturally, ideally be zero, unless the air pressure in the room where the recording was made changed from beginning to end
It's this second by second variation in asymmetrical mains waveform that gives the DC saturation effect in transformers.Very good points. Thank you for clarifying my explanation.
The DC I referred to is better understood as a "low frequency AC signal" which will be both positive and negative throughout the program. Adding a pole to the feedback network will help greatly in mitigating this phenomenon.
The transformer core reads the asymmetrical waveform as an unbalanced +ve & -ve current flow that increases the core flux in one direction.
This also why the DC induced hum varies continuously when you actually listen to the transformer. It constantly varies.
The core is effectively integrating the unbalance in +ve & -ve currents and outputting the results as a vibration in the windings/core.
how about using Lundahl input transformers... the user specifies that it was more transparent sound...
Lundahl 1545a Audio Transformer
Lundahl 1545a Audio Transformer
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Amplifiers
- Chip Amps
- what happens if DC blocking capacitors is removed