Basic DAC Question

Why do you want to know the _total_ _harmonic_ distortion for a sampled audio? system ?.

There is always the classical formula for SNR in dB = 1.76 dB +6.02n, in which n is the number of bits.

Some of the noise components are outside the required audio passband, especially when some form of noise shaping is used and thus filtered out.

In a sampled system, you will only get strong _harmonic_ components, when the produced waveform is a subharmonic of the sampling frequency, at other generated frequencies, the same noise power is distributed among a very large number of frequencies, creating a noise floor.

Look at the spectrum for a DDS system, there are usually a noise floor, but at some frequencies, the noise power is concentrating on a few discrete spurs, while the frequencies in between are very quiet.

Paul

** Funny how so many folk are interested in the linearity of an audio system and wanna know the THD figure - maybe they know more than you.

Good old THD testing is the simplest measure of linearity and when done across the whole audio band is very informative.

Sometimes two high level, high frequency tones are used for DAC and ADC tests (where the difference signal is noted) to avoid the post filter enhancing the THD figure.

....... Phil

A little pointer, (that I guess Phil knows)

Quite often people use 1K to "do" THD tests, but this will not always excercise all the bits, so 997 Hz is often used

martin

I don't think that the OP asked for that, and I certainly didn't have that in mind. I was hoping to calculate the no-filtering THD, which seems a lot easier to calculate. I can always do a SPICE simutation to figure it out for a particular DAC and filter, but having a formula and figuring out the reasons why the formula works is always a big help in getting a deeper understanding of a circuit. And it's an interesting puzzle.

Leaving off the filter, how did you calculate the 0.02% and 3% THD figures? Did you use the "multiply bits by magic number X" method posted elsewhere?

I've a real problem with 'em. Keep policing. It's the only way I'll learn :)

Yes. :)

I have done this too in the tube era, but as you can force the THD figures quite low with feedback (and sufficient open loop gain at higher audio frequencies) it does not tell very much about handling complex waveforms, in which the input stage can be overloaded.

Anyway, a rich harmonic spectrum can be preferable e.g. in guitar amplifiers.

Seems to be standard IMD measurement.

Anyway, IMD sounds nasty all the times, while some even prefer (even order) harmonic distortion.

Again, due to the nature of a sampled system, sometimes the noise is concentrated into discrete spectral lines, while at other times, it is spread into background noise, depending of the frequency, related to the sampling frequency.

Paul

"Paul Keinanen"

** Bollocks.

** Anyone see where the original topic went ??????

** Not in the analogue world.

** Folks - we have ourselves a genuine " live one " !!!!

It thinks non-linearity comes in two distinct flavours.

** Yawn ...............

It musta read that on some web forum full of audiophool freaks.

Or a crackpot sci-fi site.

Beam me up ..............

...... Phil

I suggest that you take a look on RF designing using DDS frequency synthesis.

In simple DDS systems the noise level and the worst case discrete spurs depend on the phase accumulator width, the word count and the bit length of the sine look-up-table and the number of bits in the DAC. Any decent DDS textbook will tell how the noise power is distributed depending of the actual phase accumulator increment (=generated frequency).

In high quality receivers and transmitters ordinary DDS chips can often not be used due to the significant amplitude spurs that can occur at some frequency settings. If such chips are used to directly drive the transmitter, the spurs could be radiated at a significant power levels, potential causing interference to other radio services. On receivers, the local oscillator spurs could cause reverse mixing and spurious signals would be heard in the receiver. For this reason, for instance cleanup-PLLs are used to get rid of these spurs.

While cheap DDS chips such as AD9833 work with say 20 MHz clock frequency and are capable of producing frequencies in the 0-8 MHz range, the spectral behaviour for audio DDS running at much lower clock and output frequencies would be similar. Thus I would suggest that the OP would study the spectral behaviour of RF-DDS chips, if audio specific DDS information is not available.

Paul

I don't have a beam. Will a 10' 2" * 8" do?

-- Service to my country? Been there, Done that, and I've got my DD214 to prove it. Member of DAV #85.

Michael A. Terrell Central Florida

Paul,

Can you recommend any good (and practical!) books on DDS?

We're designing a series of waveform generators, based on just a FPGA and some fast DACs, something like this:

This one works well, because we're running way below Nyquist. Later versions will have other waveforms (including arb) and tricks.

We'll always filter the final outputs, but we are thinking about adding internal channel-to-channel features, like AM/FM/PM. Since these would be all digital and can't be filtered, I can imagine some horrific image and spur problems, so we'll have to be careful about how we spec this, so we don't have mobs of customers outside our place waving torches and pitchforks and spectrum analyzers.

John

That would be "intermodulation distortion".

Maybe, but it's not supposed to be contributed by the amp! ;-)

Cheers! Rich

But

y'alls

does.

John

I have not been working with DDS systems for more than a decade, but I would suggest looking for any DDS white papers from Analog Devices or Harris.

Those living in the UK, might be interested in quite a few articles published in the (Electronics and) Wirelelss World (EWW) magazine.

Some recent versions of the ARRL handook for radio amateurs

will also contain some very down to earth articles about NCOs and DDS systems.

If you can create a clean carrier, adding som modulation to it should not be an issue.

Unfortunately I can not help much more, but I hope that you might find some of the pointers usable.

Paul

You are *not* an authority on the syntax of yat.

John

How about "y'all's", as in: "Y'all's coming over tonight to watch NASCAR, right?"

--
  Keith

"Y'all's coming over tonight to watch NASCAR, ain'tcha?"

--
Service to my country? Been there, Done that, and I\'ve got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida