Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[donnie7]

zmix,

I see you're using a Mac for your measurements. What scope or analyzer tools are you using? Great thread.

[Jim Williams]

Ronzie wrote on Sat, 09 February 2008 17:24

zmix wrote on Sun, 06 January 2008 12:03 Jim Williams wrote on Fri, 04 January 2008 11:49 Looks like one is dominant 2nd harmonic, (and quite high) and the other third harmonic dominant. Those harmonics are well above the residual THD of the chips themselves, I suspect something else is adding to the stew, perhaps the converters?.. Jim,  If you look closely at the initial plots I posted here, you will see that these artifacts are not present in all of the samples. Therefore NOT a constant, and this fact should rule out the converters. There is a known (2nd harmonic) artifact of the TA101 evident in the initial samples which is only present at higher signal levels. The EQ plots are taken from a circuit without a TA101, as were the plots that Klaus provided. My intention in starting this thread was to provide empirical evidence and discourage armchair expert's regurgitation of hearsay. You are welcome and encouraged to contribute by posting any measured results of your own.. -CZ I think a tumbleweed rolled over the cricket I heard chirping. Ron Allaire, Skyline

Send it to me and I'll run it through the AP.
Chirp.

[zmix]

Jim Williams wrote on Sun, 10 February 2008 10:46

Ronzie wrote on Sat, 09 February 2008 17:24 I think a tumbleweed rolled over the cricket I heard chirping. Ron Allaire, Skyline Send it to me and I'll run it through the AP. Chirp.

The tumbleweed or the cricket?

[peranders]

donnie7 wrote on Sat, 09 February 2008 22:35

zmix, I see you're using a Mac for your measurements. What scope or analyzer tools are you using? Great thread.

This caught my interest too. I'm a mac user since 1985!

[Larrchild]

No doubt your readings are much hipper.

[Gold]

Is the MC33078/09 only available as a dual/quad?  Is there a single version under a different model number?

[dcollins]

Gold wrote on Thu, 01 May 2008 20:26

Is the MC33078/09 only available as a dual/quad?  Is there a single version under a different model number?

No single, iirc.  It's a good part.

DC

[KSTR]

FWIW, the TS522/524 from ST are identical to (their) MC33078/79 but have sligtly better guaranteed max offset voltage specs.

There is no single version, it seems.

- Klaus

[MDM,]

if we focus on impulse response instead of bandwidth of steady-state signals, harmonics etc..

how does settling time fit into all of this, and would an opamp system with high slew rate also have a good settling time?

[peranders]

Is settling time an important parameter in audio really?

BTW: LME49710, LME49720, LME49740 are real nice audio opamps.
http://www.national.com/pf/LM/LME49710.html

Have you seen lower distortion figures?
THD 0.000018%

[John Roberts {JR}]

MDM, wrote on Mon, 12 May 2008 04:44

if we focus on impulse response instead of bandwidth of steady-state signals, harmonics etc.. how does settling time fit into all of this, and would an opamp system with high slew rate also have a good settling time?

Slew rate is how fast does the amplifier go when it can't keep up with the input signal (clipping in rate of change). Settling time is, how quickly does it regain control of it's output again after the step input is reached. One can imagine different strategies for how the amp works in this overdriven condition. One where the slew rate is modulated by the degree of overdrive, seems like it would recover more gracefully than one that just toggles to some maximum slew rate during the slightest overload, while the one with the full on slew may get you there sooner.

To properly evaluate settling time, you really need to specify the degree of overdrive. This is more a consideration for data acquisition design involving sample and hold requirements and not for band limited  audio.

IMO a well crafted audio path will be faster than the fastest signal it will encounter. Out of band information, needs to be harmlessly band limited early in the audio path and in such a way that no audible artifacts are generated. This well designed circuit does not have a slew rate, but can be characterized adequately by an impulse rise time, and a full output bandwidth. Properly executed a large and small step impulse will have exactly the same shape (rise time). If allowed to slew limit the large and small step impulse will be reproduced quite differently

JR

[zmix]

 Here is an interesting bit of information from a white paper "Measuring Switch-mode Power Amplifiers" on Audio Precision's website:


"All commercial audio analyzers employ precision low noise and distortion operational amplifiers (op-amps) in their input and signal processing stages.

These devices have typical slew rate capabilities of only 5-10 V/sec compared to the thousands of V/sec that can be present in the unfiltered output of a switch-mode amplifier. Connecting an audio analyzer directly to the unfiltered output of a switch-mode amplifier is a recipe for measurement disaster!

Once the input stage of the analyzer is provoked into non-linear behavior by fast slewing signal components,
all subsequent measurements will be invalid.

The inherent input slew rate susceptibility of audio analyzers is not the result of poor design.
Mother Nature confounds us with practical tradeoffs between fast response and precision.
Within the current state-of-the-art, it is simply not possible to design a high impedance analog input stage having
audio-worthy distortion and noise performance without incurring a practical slew rate limitation.

The input stages of oscilloscopes and lab-grade DVM’s may appear to contradict this statement.
However, and with all due respects, the residual distortion and noise performance of these types of instruments is typically orders of magnitude worse than a low-end audio analyzer. "

[zmix]

dcollins wrote on Tue, 20 April 2010 23:59

Jim Williams wrote on Tue, 20 April 2010 08:12 A 2520 would be a great opamp choice in a fuzz tone. Lot's of harmonics there. I happened to be evaluating 2520's for a project, so lets look at the harmonics. This is at +20dBu into 600 Ohms, and the 2nd is roughly 100 dB down, 3rd around -110.  Even with eight FFT averages the higher order harmonics are barely visible... Myth busted! DC

ed. bp scaled image.

[Jim Williams]

peranders wrote on Wed, 14 May 2008 13:01

Is settling time an important parameter in audio really? BTW: LME49710, LME49720, LME49740 are real nice audio opamps. http://www.national.com/pf/LM/LME49710.html Have you seen lower distortion figures? THD 0.000018%

Yes. The new LME49990 does .00001% THD and has .88 nv noise. I may drop some of those into my AP rig.

[Geoff Emerick de Fake]

Jim Williams wrote on Wed, 28 April 2010 09:21

Yes. The new LME49990 does .00001% THD and has .88 nv noise. I may drop some of those into my AP rig.

At the cost of 9mA quiescent and 2.8pA/sqrtHz (5534 is 0.6pA/sqrtHz).