Any diffs in the output impedances (feedback affects that as you know). Also, the way the input stage is arranged can have a big effect on how cable-ground-loop conducted noise gets into things.
Can't answer that, as I can't find that spec listed, but I should have also mentioned that my Rotel pre-amp has balanced line output, so I was able to take advantage of the Modulus86's differential inputs, while the 3886DR is single ended input only.
Okay, 0.2% or something. Many people can hear 0.19% easily (-54dB in the klippel test). So your amp will be easily distinguishable from many other "hifi amps" and thus not hifi?:
Not directly, but that is even worse. You said "linear voltage relationship between input and output" or that "all competently designed hi-fi amps will sound similar" which will make outliers not hifi.
Strange that you have little experience with (solid state) amps.
Compare it to the more common topology, 2 or 3 emitter follower. Me also don't work a lot with CFP. I was only brainstorming, trying to understand what I heard from CFP amps (bass is the most typical).
I was talking about microscopic level of linearity. I doubt that you can find people who can ABX linearity at this level so ignore that.
Have you conducted ABX for distortion? I would guess that the 7th order will be harder to distinguish for you than 2nd order.
I think you misunderstood my statement.
I think that one too 🙂 We all know what distortion or THD is, but why my statement should be wrong?? (rhetorical)
The LM3886DR is the fundamental LM3886 circuit with all components needed to make it work well and meet its data sheet specs.
The Modulus-86 is a completely different animal. It uses a precision opamp to perform error-correction on an LM3886 output stage. What makes that amp work well is the error-correction. The LM3886 is just along for the ride.
The Modulus-86 and LM3886DR will have slight differences in output impedance as the MOD86 has much higher loop gain. They both have a Thiele network (L||R) in series with the output and the 20-some mΩ of that dominates the output impedance. I would be surprised if the output impedance of the two amps were meaningfully different.
True. That's why I use differential inputs on the Modulus-series of amps. Differential inputs, effectively, removes the ground loop on the input circuit from the signal path.
I suspect the difference between the two amps is that the Modulus-86 has much, much, MUCH lower THD than the naked LM3886. The difference may not look that dramatic when looking at 1 kHz numbers, but the rise in THD+N above a few hundred Hz for the LM3886 is significant (should be +20 dB/dec as it follows the rolloff in loop gain). The Modulus-86 does show a little uptick in THD+N above a few kHz. That's caused by the change in the fundamental amplitude resulting from the Thiele network. The amp's THD is flat vs frequency. Adding nearly 50 dB of loop gain (error correction) at 20 kHz does that... 🙂
I find that I like flat THD vs frequency. The amps I've enjoyed listening to (my DG300B and MOD86) have flat THD vs frequency. The THD of the DG300B is about 0.2 % (1 W, 1 kHz) but it sounds good. I like it. I wouldn't call it precise, but I do like it. The MOD86 is in a different league as far as the precision goes. That sounds both good and precise.
The 300B glows much prettier in the dark, though. 🙂
In the end a lot is left up to personal preference. Also don't underestimate the impact of various cognitive psychological biases.
Tom
The Modulus-86 is a completely different animal. It uses a precision opamp to perform error-correction on an LM3886 output stage. What makes that amp work well is the error-correction. The LM3886 is just along for the ride.
The Modulus-86 and LM3886DR will have slight differences in output impedance as the MOD86 has much higher loop gain. They both have a Thiele network (L||R) in series with the output and the 20-some mΩ of that dominates the output impedance. I would be surprised if the output impedance of the two amps were meaningfully different.
True. That's why I use differential inputs on the Modulus-series of amps. Differential inputs, effectively, removes the ground loop on the input circuit from the signal path.
I suspect the difference between the two amps is that the Modulus-86 has much, much, MUCH lower THD than the naked LM3886. The difference may not look that dramatic when looking at 1 kHz numbers, but the rise in THD+N above a few hundred Hz for the LM3886 is significant (should be +20 dB/dec as it follows the rolloff in loop gain). The Modulus-86 does show a little uptick in THD+N above a few kHz. That's caused by the change in the fundamental amplitude resulting from the Thiele network. The amp's THD is flat vs frequency. Adding nearly 50 dB of loop gain (error correction) at 20 kHz does that... 🙂
I find that I like flat THD vs frequency. The amps I've enjoyed listening to (my DG300B and MOD86) have flat THD vs frequency. The THD of the DG300B is about 0.2 % (1 W, 1 kHz) but it sounds good. I like it. I wouldn't call it precise, but I do like it. The MOD86 is in a different league as far as the precision goes. That sounds both good and precise.
The 300B glows much prettier in the dark, though. 🙂
In the end a lot is left up to personal preference. Also don't underestimate the impact of various cognitive psychological biases.
Tom
Last edited:
When was that?
The first set of quality phones I bought aged 16 in the '70s featured 'diffuse field head transfer function' equalization ie the FR was not flat with the mid boost you mentioned. So the necessity of this was well understood 40 years ago.
Things did get a bit confused at one point when binaural recordings threatened to become fashionable and due to the way they are recorded they actually do require headphones with a completely flat FR to sound right.
Have you put this into ABX test? You know, 0.0004% is too far below what people think is perceivable. I would guess that I would be able to perceive differences in an ABX, but I cannot be sure if it is the THD that make it audibly different.
I didn't get a chance to hear the baby amp, but i agree with Chris that it is a very good amp.
What is the spectrum* of the HD? We all know how poor a summary THD is of the real distortion.
*(i figure a 3d graph with a slider to display the 4th variable is really needed to display this info)
dave
True. But I was talking about Tom's 2 amplifiers, specifically the Modulus-86. Extraordinary claim requires extraordinary evidence, they said.
I thought the THD is too low (even for the basic 3886) such that spectrum is almost irrelevant.
Last edited:
Because music doesn't have any way to measure "THD". It isn't a constant amplitude single sine wave. So you can't listen to "1% THD music", it doesn't mean anything. Nor does listening to an amp rated at 1% mean it will be any more distorted than from one that is rated at 0.02% since the actual nonlinearities when listening to unclipped music through those amps will be at a lower average level. There are other measures that might be made with music (such as residual after nulling), but THD isn't one of them. THD of an amp is more of a way to put a limit on a output power specification than on linearity.
Chrisb
In your comment #40 you wrote that you had been listening to 2 amps by Tom Christiansen. You liked one better then the other. I know I'm being thick, but which one?
Jay
In your comment #40 you wrote that you had been listening to 2 amps by Tom Christiansen. You liked one better then the other. I know I'm being thick, but which one?
Jay
bingo - but you sorta actually knew that, rather than "guessed"
- but not the only reason - the Mod86 is more than a very decent (un)sounding amp - that to my ears beats numerous chip amps I've heard over the years on initial impression by a not small margin.
As I think Tom said somewhere along the line in his documentation or website "the 3886 is just along for the ride"
First the LM4780, now the 8 pin DIP LM49710 - if someone was a conspiracy junkie ....
Last edited:
- Status
- Not open for further replies.