.... and, why is that? We need to get busy. Ultra low distortion withOUT any GNFB ... what does that sound like?
----------------------------------------------------------
Min Z? Lets all only show 4 Ohms... so we can compare with one another. Though 2 Ohms in practice is not unusual and PA should handle it also.
-RNM
Last edited:
Shall I continue? I can post tens of speakers with impedances like these.
The B&W's would be OK rated as 4 ohm. Wilson of course is on a different phase from the rest of this planet.
Just a list of who isn't compliant, perhaps a bit of arm twisting can occur behind the scenes at the next standards meeting!
Last edited:
As Kulinummi does not show complete schematic of his non GNF here is mine
Attachments
Last edited:
Hi,
Those certainly are excel numbers. and 4 Ohms?
Back in the day, to get such low numbers could only be accomplished with very high GNFB VFA circuits.
IMO this is the future direction for amp design....... Towards inherently linear circuits.
Thx, RNMarsh
Last edited:
I'd like to hear it. It sounds like you're about it because you kinda try to do that in a similar way from what you've described with amps before.
Does it sound different? Maybe, or maybe it sounds the same, and it'll be rejected by 50% of audiophiles and loved by 50%. That's just a guess towards this question. It's substantial, that's for sure.
The problem really is that feedback gives you benefits in many ways, but then brings other things that aren't perceived as benefits. You might get better texture in the high frequency range but then in comes the sound of the walls in studio or something else to make it sound sterile. It gets that "petri-dish" sound. Now some recordings will never sound that way, they just sound like magic... But that isn't very many. They usually more "tastefully" processed than going for stringent accuracy. See Steven Hoffman albums.
Lots of feedback at some point doesn't benefit bass, as we covered earlier due to physical problems with the drivers. And it does tend to make bass sound thin, extremely 2d, and kinda like it splats. I think many audiophile would be happy with a curved frequency feedback response.
Follow your own advice and build one yourself. Clearly no one else cares as much about it, as you.
It wouldn't surprise me if some people were watching your actions as a kind of "leading indicator": if Richard Marsh builds a Kolinummi style amplifier and praises its performance, then (and only then) might they start paying attention in a serious way. Right now it's just another piece of online chatter.
T, actually PMA is correct in this case. You are of course correct, IF you only measure with one frequency at a time, BUT when you have multiple tones, then each driver will become an extra load in parallel. Matti Otala found that you have to divide the Lowest impedance by 2 to get a worst case load. He measures this with a transfer function that precisely matches the inverse of the loudspeaker load. This is why we make our amps with so much PEAK output current, far more than what many would judge necessary. Of course, having good peak output current does not lower the distortion at 10W or so when a difficult load is being driven by an output stage that is not properly biased. That is the problem that PMA found with the Parasound A21. Even properly biased, the A21 has a hard time with difficult loads (I must admit) and that is why I never tried it on the Wilson Sasha speakers that I use. It would be ok for the Sequerras, but even a cheaper amp of my design works well for that so I never put it in my system. Now, I use an overbiased JC-5, for the Sashas. It has 6 output pairs, I would prefer 8-12 pairs but I don't have any available. So I just run the amp as hot as I can get away with, (for myself) and I don't leave it on continuously. It is a compromise. I am looking forward to making a lower power amp (with the same heatsink) so I can achieve best performance for me, and most everyone else. Parasound worries about marketing specs and just loves to see that 400W at 8 ohm spec on the data sheet. 250W would be more ideal, with more bias, or at least optimum bias without getting too hot.
Please not. Everybody knows this since decades. Your curves are typical of most of the (bad) speaker's assemblies, with awful impedance/phase curves.
Look at the Sasha W/P impedance curve (and its related phase curve).
Wilson Audio Sasha W/P loudspeaker Measurements | Stereophile.com
How Wilson can pretend this enclosure to be 4 ohms, with 2 ohms where most of the musical contents ask for most of the energy (~100Hz) ?
I was not referring to the usual (bad) practices of the market, that everybody knows (you are beating a dead horse).
I was telling that a correct practice from speaker's manufacturers should be to give the DC value of the speakers as the impedance of their drivers.
That no passive filter should reduce this minimal value as an added charge for the amps.
That it is easy and that offers a lot of benefits to flattening the impedance curve of the speakers with passive networks. Both in the electrical and acoustical domain (phase and linearity, independence from cables serial impedance etc.)
Wondering why so few people take care of this.
Nothing less, nothing more.
I don't even understand why you are arguing: Yes, of course, if you want to know how a given amp will behave with one of those speakers, better to measure its performances on a résistive load equal to the lower impedance point of its future load.
I was not contradicting-you. You misinterpreted my words. Or it's an effect of my poor English.
Well, clearly this is not a subject for your interests and maybe two others. It is in fact a direction I have moved towards for decades and with excellent results.
Now there is a book on the subject. So, many more can quickly try to move in that direction and find out. Probably a DIY'er will be first. Like Dadod.
THx-RNMarsh
Last edited:
Seems to me what you describe hearing there is probably just the very low level remaining nonlinear distortion, IMHO. I don't find it to be a problem with Benchmark AHB2. They describe a little bit about the design here: Feed-Forward Error Correction - Benchmark Media Systems
And here: Laboratory Use of the Benchmark AHB2 Power Amplifier - Benchmark Media Systems
Last edited:
I'm surprised with all the detailed really nice work on DAC's you do you would put forward THD at full power as some kind of lone "benchmark". IMO full power THD by itself is about as meaningless as possible.
Not sure who you are referring to when you say, "you." Hopefully, not me, as I don't put full power THD forward as important to me. I merely posted a graph to compare with graphs other people were posting at the time. If I had other graphs of more interest, I would be happy to post them too, but I don't. I do, however, like AHB2 and find it useful for listening to dac sound quality.
EDIT: Assuming the "you" was me, thank you for the nice words about dacs <blush>.
Last edited:
I've heard this amp and it does sound nice. A little too relaxed even.
But I'm pretty skeptical of your conclusion. I didn't find the AHB2 to defy any kind of normal evaluation. I have no reason to believe mics don't pick up way more than most bargain for when you can actually play it back at a level of such low distortion and with enough control in the speakers to represent it.
You said, "If I could get a lower distortion power amp for around the same price, I would probably want one." Maybe you could qualify "lower distortion" as more than the one metric.
Do you mean you are skeptical that what you are hearing is some low level distortion?
Regarding AHB2, I don't think it defies normal evaluation. I didn't say that. Of course it isn't 'perfect' in terms of low distortion and noise, but it was the best I could find when I bought it (and within my price range, although not without some pain). Hopefully, someone will come up with something even better eventually.
On the subject of relaxed, I bet you might find it very interesting if you ever had an opportunity to listen to a very clean example of a Sabre dac playing back though AHB2 at pretty low volume level. One can adjust 2nd and 3rd harmonic distortion in tiny increments up to rather high levels of distortion via manipulation of the dac control registers. My own conclusion is that there shouldn't be a need to pick amps, preamps, dacs, etc., to control how forward or laid-back the reproduction system sounds. It can pretty much be adjusted to taste, where more forward would probably be related to more 3rd harmonic. If there was a desire to play with higher order harmonics, some DSP could be written to feed the dac with. Don't know what you would think about the whole experience of course, only that you might find it interesting and worthwhile.
Ah, I see. When I referred to a possibly lower distortion amp, I was thinking of much more than THD+N at full power (pretty much everything else, and under all operating conditions). Sorry if I failed to make the sufficiently clear. As you point out, THD+N at full power is pretty useless. I guess I assumed that would be taken as a given when talking about possibly buying an amp based on measurements.
I'm skeptical that low level distortion is what causes petri-dish. We electronic nerds tend to forget the amount of control there is in recording technique, mixing, and mastering. I really think that's where the problems are at kinda in the way we're painters trying to make the carpenter look good. But you'll have varying opinions on what kinda paint is best with what kind of carpentry, when it comes to audio.
If there was any hope of the AHB2 being beyond reproach it would have to start with BenchMark designing all the studio gear, as well. A lot of the process of getting an album out can involve trying to negate less than ideal studio equipment issues. (ones that people may think are not a problem)
If there was any hope of the AHB2 being beyond reproach it would have to start with BenchMark designing all the studio gear, as well. A lot of the process of getting an album out can involve trying to negate less than ideal studio equipment issues. (ones that people may think are not a problem)
That is (to a certain extend) the same topologie as was selected for my power amp, a pre-amp and a low gain l(low feedback) power amp (no shared feedback or any...). The power stage has been configured as a current-conveyor.
P.s. See you (all
) next weekend in Hamburg so you can listen to the implementation of that actual concept
Last edited:
- Status
- Not open for further replies.