Artikel Elektronika [1] Power Supply bukan sumber DC Offset
Di bawah ini adalah artikel dalam sebuah forum diskusi elektronika, yang menurut kami sangat bermanfaat karena berdasarkan experimen seorang member bernama Ulysses. Berikut ini alamat forum dimaksud :
***********
Surprisingly,
this is actually not a significant source of DC offset in analog
circuits. I remember Dan Kennedy telling me so a couple of years ago,
and I just kind of took his word for it. But today I did an experiment
to prove it. I had a Super Stereo compressor on the bench for set-up
and calibration. I hadn't installed the servo amp yet, and was playing
with the DC offset (it's something I measure while trimming the
distortion on the VCAs). Anyway, I had the power supplies trimmed to an
even +/-18.000VDC and due to various factors at one moment I was
measuring -3.69mVDC offset on the output of one channel. So, just as an
experiment I then cranked the positive regulator down as far as it
would go (I've got a trimpot in there for this). I got the positive
rail down to 17.526VDC, almost a half volt drop, while the negative rail
stayed at -18.000VDC. The DC offset changed from -3.69mV to -3.78mV.
For a 0.474V change in supply balance, the result was a 0.00009V shift
in the DC offset. In other words, as I said, this is not a meaningful
cause of offset.
Quote:
Originally Posted by NeilTDeal
On
the analog electronics side DC offset is most often caused by
mismatched supply rails. If the negative rail is say at -14.4 vdc and
the positive rail is at 15.2 vdc, most opamps will establish zero
crossing at +0.8 vdc... That’s offset.
The more significant sources of offset are the DC error of the amplifiers themselves; the amplifier's input bias currents causing differing voltage drops across differing source resistors on the inverting and non-inverting inputs; and bias current offsets between those two inputs. Another big source in my particular circuit (and some other circuits too) is the 2nd-order harmonic distortion in the VCA before it is trimmed. Interestingly, 2nd-order distortion manifests itself as a top-to-bottom assymmetry in the waveform, which results in a net level shift and therefore DC offset.
Some waveforms are naturally rich in the 2nd harmonic and will exhibit some assymmetry on their own. These sometimes look strange on the DAW screen, but probably are not the source of an offset in a recorded file. And if you *always* see the asymmetry as positive-going and never negative-going, it's time to stop suspecting the acoustic source and start looking at the hardware (both analog and digital).
Of course offset may not always end up where you can see it or measure it, because most circuits have some kind of high-pass filter in them someplace. You don't always see the speakers bulge out (or in) because not all amplifiers pass DC. But I sure did see it when I plugged the headphone jack on my Powerbook into my Hafler 9505. Made me VERY nervous for the Dynaudios. (The 9505 is doing about 375 watts into 4 ohms and passes signal down to a small fraction of a Hertz).
One last question would be how much offset is acceptable on the output of a piece of analog gear, or in a file on your DAW. Two different scenarios with different implications. In the DAW, the primary concern is clicks during edits (crossfades can avoid this) and to some extent a loss of headroom. In the analog world there's the worry that a connected piece of gear isn't going to enjoy being fed the volts. In most cases it would take a decent amount of offset to cause any real problems, but avoiding clicks, scratchy pots, and other phantom behavior is a good enough reason to avoid it. My Super Stereo uses a DC servo on each output which, once trimmed, keeps DC down to a small fraction of a millivolt. That's overkill though - the device I use to feed the unit during testing is a very good Benchmark bump box that has about 11mV of offset on one channel, and that doesn't hurt anything at all. Like I said, nearly every device you could feed that into will have some kind of HP filter in it, so you'll rarely see it come out the other side. There are exceptions.
***
Semoga bukti dari experimen ini berguna.
***
Originally Posted by NeilTDeal
Pada peralatan elektronik analog, DC offset seringnya disebabkan oleh ketidakseimbangan pada rangkaian power supply. Jika jalur negatif katakanlah -14.4 vdc dan jalur positif pada 15.2 vdc, kama kebanyakan opamp akan menghasilkan perpotongan tegangan pada +0.8 vdc ... inilah DC offset.
Comments
Post a Comment