Tom's measurements of the Modulus-86:
http://www.diyaudio.com/forums/vend...wer-amp-extremely-low-thd-34.html#post5376359
What is your output network like?
http://www.diyaudio.com/forums/vend...wer-amp-extremely-low-thd-34.html#post5376359
What is your output network like?
Yes, but I stop short of kicking what I'd really like to.
Carry on, don't give up, the entertainment is slowly creeping back into this thread.
Yours truly,
Kermit, the Industrial Punk.
Tom did answer that here Modulus-86: Composite amplifier achieving <0.0004 % THD+N. and here Modulus-86: Composite amplifier achieving <0.0004 % THD+N.
Whether you agree that testing at an input slew rate of 1.07V/us as that's the fastest a 192k DAC can put out is right is as much a matter of belief as anything else. If I want slew I have my 300V/us power amps anyway.
It is the Modulus-86, built according to specifications, with the prescribed output network.
Tom as referred to by Bill is right, like Jurgen, that a signal with this rise time cannot come from a DAC as used in stereo. But that does not mean that it is not a good thing for all kinds of other reasons if an amplifier doesn't misbehave on square waves, because. At least, and this is why I posted it, HD as a metric is not necessarily correlated with other good qualities an amp may have.
This is a very naive way of looking at best linearity. The correct way is looking at the FFT spectrum, and decide which spectrum you think is 'best'.
Looking at the time domain waveform can mislead you big time as harmonic levels and phase interact to draw a time domain waveform that looks 'nice' but has more harmonics at higher levels than a time domain waveform that looks 'bad'.
As an extreme example, to illustrate, a square wave - approaching version of a signal has 'equal + and - distortion' but very many harmonics at high levels. Filter out some of the harmonics, making the signal more linear, will make the time domain waveform look 'worse'.
Jan
Last edited:
Your rise time is fast, but are you sure that behavior is normal for a Mod-86?
Not saying it can't be. I realize LM3886 composites may be difficult to stabilize. The test conditions are different than Tom's, but I'm curious.
Last edited:
Forgot to post one more link on this LM3886 & LME49720 composite amp
High slew rate close to clipping can cause things to go a bit funny. I personally don't see an issue with that as long as it recovers again as I neither need the power nor the slew rate in normal listening.
You can always PM Tom, he is very helpful.
It looks like a low pass filter, but with a peak a bit higher up in frequency. Probably a loop stability problem; it is stable but only marginally so.vacuphile said:
To put it another way, when the input has frequency components which don't get filtered out then they appear in the output too.
Number one there is the "musical transient' modeled as an impulse that goes +- full scale in one sample period.
Dec 2017 post:
JC:
"
Now back to what it takes to make a successful audio amplifier design.
While Mark Johnson gave a number of factors, what I am most concerned with here is some almost ignored factors included in the Otala, Lohstroh AES paper 1973 that made one of the first successful sounding power amps in audio history.
This included high open loop bandwidth, as well as slew-rate, open loop linearity, and very high effective bandwidth (for the time).
If Jan Lohstroh alone would have made it, without the changes suggested by Matti Otala, it would have been just another good design, ignored by serious audio listeners. Why this is so, still confounds the critics, but it still appears to be more important than just a 'better' thd measurement, etc. and this has been proven out by a number of successful designs that have that extra 'quality' not shared by the bulk of audio products.
Pavel:
My 12 cents regarding power amplifiers, based on experience:
- stability even with very difficult complex loads, no problems if the load is suddenly disconnected
- high output current capability
- excellent immunity to both air-coupled EMI interferences and supply line interferences
- sophisticated wiring topology immune to grounding scheme of remaining audio chain components
- frequency compensation network that assures high loopgain (feedback factor) even at the upper end of audio band
- very low and as linear as possible output impedance
Reasonably high slew rate and full power bandwidth at least 100kHz as a rule of thumb, no need to emphasize.
"
For posterity and those that don't understand the frustration and short comments.
PS It would be nice to put some general, basic parameters/numbers to these specs. Thanks.
PPS Hmmm, that deserves its own thread and it's a bit of an OT.
http://www.diyaudio.com/forums/soli...s-regarding-power-amplifiers.html#post5378111
Last edited:
Who needs Shannon-Nyquist with friends like that?
Vacuphile--are you tripping the lm3886 protection circuit every rising edge? All your composites share that in common. Internally, with the additional loop gain of the front end op amp, you could easily be over driving the 3886's input.
A protection circuit pulling gain could easily cause those rising edge oscillations.
Rising edge usually does not do this. I do not have much experience with LM3886, but I do have a lot of experience with OPA549. It is SR limited, but behaves nicely on rising and falling edges. If the protection trips, it is because of heat (dissipated power) and it needs some time to trip. I hate those power chipamps, there is no way to cool them down reasonably and thermal coupling between output stage and input differential is a bad thing. I also do not like amplifiers (of any kind) that misbehave at high input SR. Poor design, the designer should have used appropriate RC input filter in case he has troubles. Or just do it different way.
thanks, Stuart was frequently mentioned inside the BT-thread as "no longer with us", so I presumed him to be sick or dead.
Last edited:
- Status
- Not open for further replies.