When I connect a constant voltage to my soundcard's input, and look at the samples read from it, I get some initial ringing that quickly dies down, followed by a slow exponential decay. This is what I would expect. But the reading, after the ringing has stopped but before the exponential decay gets going, seems to be too high by about 8%. Is this expected too? (I didn't expect it, but I don't really know very much about this stuff.) If so, do you know exactly how much higher than the actual voltage it's supposed to read? If not, do you know what I might be doing wrong?
Some background: I'm trying to calibrate a cheap voltmeter that I bought. It has a separate pot for adjusting its reading of AC voltage, besides the pot for DC. I have a DC voltage reference, but not one for AC. I can use my soundcard to generate a sine wave, but I need to know its voltage if I want to use it to calibrate the meter's AC reading. I can measure the soundcard's output by looping it back to the input. So now I'm left with the problem of calibrating the input, and I'm trying to do that by connecting a known DC voltage to it. But it seems not to be working.
For DC, the meter was already almost exactly correct. So, probably, it's not too far off for AC either. But if I assume it's correct for AC, that means that the soundcard reads 8% high when connected to a DC voltage, compared to being connected to a sine wave. I don't see why it would. But if I suppose that the soundcard reads the same for AC as (initially) for DC, that means the meter is so far out of calibration that even turning the pot all the way won't correct it. That seems pretty unlikely.
So, I'm thoroughly confused.
Some background: I'm trying to calibrate a cheap voltmeter that I bought. It has a separate pot for adjusting its reading of AC voltage, besides the pot for DC. I have a DC voltage reference, but not one for AC. I can use my soundcard to generate a sine wave, but I need to know its voltage if I want to use it to calibrate the meter's AC reading. I can measure the soundcard's output by looping it back to the input. So now I'm left with the problem of calibrating the input, and I'm trying to do that by connecting a known DC voltage to it. But it seems not to be working.
For DC, the meter was already almost exactly correct. So, probably, it's not too far off for AC either. But if I assume it's correct for AC, that means that the soundcard reads 8% high when connected to a DC voltage, compared to being connected to a sine wave. I don't see why it would. But if I suppose that the soundcard reads the same for AC as (initially) for DC, that means the meter is so far out of calibration that even turning the pot all the way won't correct it. That seems pretty unlikely.
So, I'm thoroughly confused.
I can't really advise on the soundcard aspect but one practical aspect... many cheap DVM's suffer with AC accuracy at low voltage ranges. It may be "accurate" measuring mains voltage yet very inaccurate measuring low voltage.
This is a calibration chart for my three meters, an old AVO and two DVM's. You can see how inaccurate one of the DVM's is at low level.
This is a calibration chart for my three meters, an old AVO and two DVM's. You can see how inaccurate one of the DVM's is at low level.
I fairly regularly compare cheap to less cheap DMMs to a long since recalibrated bench DMM.
I consistently find that errors between the voltage readings is very low. Even at 199.9mVac
I don't use the Adc and Aac scales.
The meter of mine that is in error is an early 1980's LED DVM. It is what it is.
I'm not sure what the 8% off would cause, or whether the test method you want to use is accurate (there may be some other DC bias/attenuation on the SC input after the blocking cap).
But maybe you can do it the other way around. Get an AC source and rectify it, no load, measure the DC result, which should be accurate.
Then you can calculate backwards what the AC is where it comes from, taking in acount a very small difference from the diode bias, which will probably be only a few 100mV after charging the cap as there is no load.
Use the found value for the AC to cal the AC range(s).
Jan
a few audio ADC can measure DC - but DC blocking C are common in the signal path, and most audio ADC/sound chipsets have high pass built into the chip
a AC calibration ref can be made with analog switches or CMOS gates/flipflop and a DC Vref - chop the ref V for a squarewave
and the other way - "precision" rectifier op amp circuits with diodes in the loop can be as accurate as your resistor ratios at low kHz
a AC calibration ref can be made with analog switches or CMOS gates/flipflop and a DC Vref - chop the ref V for a squarewave
and the other way - "precision" rectifier op amp circuits with diodes in the loop can be as accurate as your resistor ratios at low kHz
I figured it out. There wasn't any problem with the soundcard measuring a DC voltage, if I just looked at the initial value, before it started its exponential decay. The problem was that my voltmeter's measurement of the AC voltage of the soundcard's output was too low. My soundcard's output goes into an input on my amplifier, and I was measuring the voltage at one of the amplifier's line-out jacks, instead of taking the signal directly from the soundcard.
Almost all sound card ADC & input amplifiers are powered uses a 3...5V . Almost all or all sound card are using cap's on the input (lower AC freq. limit about 10..20Hz).
The last know sound card was the Turtle Beach Tahiti, but having the old ISA (not PCI) PC connecter.
Interesting would be how the Keithly equipment deal with this..
Hp
The last know sound card was the Turtle Beach Tahiti, but having the old ISA (not PCI) PC connecter.
Interesting would be how the Keithly equipment deal with this..
Hp
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.