I calculated for audio frequency the condition on a ground input capacitor C >> 1/((R3+R2)*2*Pi*f) which for audio frequencies would be C >> 80 pF, say 2.2 uF MKP would do... correct?
Of course this will exclude the ability to make power measurements.
Do you think ground protection is neseccary?
EMU 0404 does have line input, but diodes will protect against lower voltages... all above 0.6-0.7V. Given the attenuation of R3 this will occur only for extremely high powers into 8 ohms say 400 V, that is 10000W rms! Is that right?
Of course this will exclude the ability to make power measurements.
Do you think ground protection is neseccary?
EMU 0404 does have line input, but diodes will protect against lower voltages... all above 0.6-0.7V. Given the attenuation of R3 this will occur only for extremely high powers into 8 ohms say 400 V, that is 10000W rms! Is that right?
Last edited:
Oops I meant DC voltage measurements... output power in the audio band will still be measurable... and C1 already blocks DC.
Hi Andrea
I revisited this and tried my method( behringer USB and voxengo span - post#4) but using ARTA instead of span. I couldn't get the y axis in ARTA to be precise enough to compete with span. Span still returned the most precise dbU readings to allow for volatge conversion.
Is there any way of increasing the precision on the y axis when analysing the spectrum?
Ed
I revisited this and tried my method( behringer USB and voxengo span - post#4) but using ARTA instead of span. I couldn't get the y axis in ARTA to be precise enough to compete with span. Span still returned the most precise dbU readings to allow for volatge conversion.
Is there any way of increasing the precision on the y axis when analysing the spectrum?
Ed
Yes, to see the noise.
A 24 bit audio card should be able to make measurements with a noise floor of -100db. I have discovered that it's important to see the noise floor so one knows how to interpret the peaks in the FFT. One needs to see the baseline. Are you just assuming your baseline is a flat line? One needs a reference measurement with no signal, right?
In my own measurement rig, I've decided to optimize the noise performance and live with a little less than 20 KHz bandwidtih. I have a Boonton and other gear that works great for frequency response (don't need 24 bits etc. for this) out to 200 KHz. What I use the FFT for is to track down noise and distortion issues that don't need to be done at high frequency.
I also found that the cost of the A/D converter is not a good indication of suitability for RTA use. Some of the more expensive converters have great THD specs but crappy looking baselines and increasing distortion at below 1VRMS input level. The best performance I have found so far is with an Alesis AI-3 24/48 unit used below 1VRMS input signal.
If you need it all get an AP
PS What's the purpose of using 18 resistors in a series/parallel arrangement instead of a simple resistor?
Cheers,
Michael
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.