Sorry Bruno, meanwhile, seems to me the discussion focused on the SMPS (and not on your amp).
I think you, you should explain why the segment is flat 10-500Hz.
I read your FFT.
why do you say, I do not understand what you're doing?
you said, thanks to the PSRR on the amplifier. well, what have the PSRR?
If I put one of our fft and looks excellent, then you ask me how is it possible? So I have to explain how they solved the problems.
what is the ripple of the psu, what is the attenuation of the hf residual etc..
If my answers have parallels with the theory that we know. all is ok.
so what was your response to justify that your SMPS has no harmonics, and the ripple is low? I want this (I do not doubt what you do)
prestige has nothing to do when you show a measure. the measure speaks to you.
have no doubt if they can understand. our SMPS has developed in side of the pcb amplifier and fft is clean.
Roberto
I think you, you should explain why the segment is flat 10-500Hz.
I read your FFT.
why do you say, I do not understand what you're doing?
you said, thanks to the PSRR on the amplifier. well, what have the PSRR?
If I put one of our fft and looks excellent, then you ask me how is it possible? So I have to explain how they solved the problems.
what is the ripple of the psu, what is the attenuation of the hf residual etc..
If my answers have parallels with the theory that we know. all is ok.
so what was your response to justify that your SMPS has no harmonics, and the ripple is low? I want this (I do not doubt what you do)
prestige has nothing to do when you show a measure. the measure speaks to you.
have no doubt if they can understand. our SMPS has developed in side of the pcb amplifier and fft is clean.
Roberto
The SMPS puts out a 100Hz sawtooth just like an ordinary supply. It is not clean and it has lots of harmonics. The output of the amplifier is the sum of both rails divided by PSRR. PSRR is about 80dB and the SMPS is synchronously rectified so the ripple is symmetrical to begin with. That's all there is to explain! What's so hard about that? Do you know how I got started? I looked at some people's measurements and told myself "okay they get this done. That means it is possible. So what would I do to get at the same result?" Had there been internet fora in those days I might have asked for tips but never would I have dreamt of barging in saying "look matey I don't agree with your measurements unless you explain them cito presto." I certainly shouldn't have expected anything resembling an explanation.
Last edited:
These are the kind of things that could easily be traded in winebars in the southern hemisphere - with or without lid. ;-)
.... just kidding !!!
I don't know where Roberto's problem lies.
Self-oscillating class-d amps generally have very good PSRR. Combined with this huge feedback it looks perfectly plausible. Apart from that - one can clearly see a tiny hint of 50 Hz residual just slightly raising its head out of the noise (so what other proof is needed ?).
Regards
Charles
.... just kidding !!!
I don't know where Roberto's problem lies.
Self-oscillating class-d amps generally have very good PSRR. Combined with this huge feedback it looks perfectly plausible. Apart from that - one can clearly see a tiny hint of 50 Hz residual just slightly raising its head out of the noise (so what other proof is needed ?).
Regards
Charles
Last edited:
I'm trying to work out how many bottles of Rust & Vrede Merlot a pair of NC400 buys. And how I'm going to get those home.
PSRR is initially equal to loop gain (around 53dB) and then there's some forward correction that improves it by another 20 to 30dB. Since the ripple is mostly symmetrical in actual fact you start with a signal that's way smaller than 400mV before you hammer it with the PSRR.
PSRR is initially equal to loop gain (around 53dB) and then there's some forward correction that improves it by another 20 to 30dB. Since the ripple is mostly symmetrical in actual fact you start with a signal that's way smaller than 400mV before you hammer it with the PSRR.
Does your SMPS have power factor correction ??
If not do you plan on implementing it ??
regards
Trev
I'm sorry you think this (I come, saying that it is not possible).
I asked, how can this fft your SMPS?
you said: thanks to the PSRR. You know well that the PSRR can not clean up all the signals present on the rail, especially on the class D. these signals are placed on the point prefilter. Here is a mix with the carrier + modulation of signal. the results are still other frequencies. Removes some of the FB, but the other sum (not being present on this input signal) as you see, there are no mysteries. math is math.
I not have problem (charlies), i just ask why segment 10-500hz is absolute clean. you think this is a stupid question?
Roberto
I asked, how can this fft your SMPS?
you said: thanks to the PSRR. You know well that the PSRR can not clean up all the signals present on the rail, especially on the class D. these signals are placed on the point prefilter. Here is a mix with the carrier + modulation of signal. the results are still other frequencies. Removes some of the FB, but the other sum (not being present on this input signal) as you see, there are no mysteries. math is math.
I not have problem (charlies), i just ask why segment 10-500hz is absolute clean. you think this is a stupid question?
Roberto
No and yes respectively. The SMPS3k would be the first candidate.
That rather takes away the fun of going there to sit in the sun and taste wine
I have posted the full explanation to Charles. Read it. Expect me to ignore further suggestions that both my measurements and my understanding are wrong.
Is the forward correction unique to the nCore and thereby a contributing factor on the improved PSRR over the standard UCD´s ?
If so would there be any plans on letting it drible down to the "mere" UCD siblings in the future?
Bruno, you have already given some very fine reports on the audible improvements of the nCore compared to the UCD, which we know. But I have one last question about the sound of the nCore if I may: How would you describe the ncore´s rendition of edges, dynamics, and sense of scale in its presentation, compared to the UCD´s?
regards,
UcD, Hypex' at least, has always had this forward correction. Loop gain is thirty-ish, PSRR is sixty-ish. Another dutch joint that independently obtained a license for UcD doesn't have this, and as a result they have to sell a regulated supply with it.
I don't really like reporting my listening results because frankly it sounds like tooting my own horn. Measurements are "independent" in that the instrument doesn't care who I am, but my ears are subjective.
Anyhow a well known local audio importer said the NC1200 sounded like it was very, very fast (i.e. like a linear amp with a large slew rate). By this he meant the transients were delivered "as is" without either exaggerating or muffling them (a slow amp can do either). He mostly listened to large orchestral and operatic works which were presented with remarkable poise.
Hi Bruno, there doesn't seem to be anything to wish for with the ncore. Data is stunning. So just out of curiosity:
I remember you published an article some time ago about an amplifier with a digital feedback loop, which I thought quite exciting. It's surely more complicated (and expensive) at the amplifier end, but for the overall chain including DA conversion the difference might not be that big (and I don't have any analog music source anymore). Why didn't it go through? Wouldn't the overall chain be superior? Any chance we'll see something like it at some point?
One more comment: To me, the most exciting application for DIY amplifier like the ncore seems to me the active speaker with digital cross-over (eg. dlcp, pls make it a product!). But then it seems a ridiculous overkill (and way too expensive) to have 3x400W for each channel. A ncore 180 makes total sense! I'd be excited about a stereo version, or an integrated SMPS, too!
Regards
I remember you published an article some time ago about an amplifier with a digital feedback loop, which I thought quite exciting. It's surely more complicated (and expensive) at the amplifier end, but for the overall chain including DA conversion the difference might not be that big (and I don't have any analog music source anymore). Why didn't it go through? Wouldn't the overall chain be superior? Any chance we'll see something like it at some point?
One more comment: To me, the most exciting application for DIY amplifier like the ncore seems to me the active speaker with digital cross-over (eg. dlcp, pls make it a product!). But then it seems a ridiculous overkill (and way too expensive) to have 3x400W for each channel. A ncore 180 makes total sense! I'd be excited about a stereo version, or an integrated SMPS, too!
Regards
Irr4: The digital feedback loop is still ongoing. It is mostly of academical interest though, in that it has led us (Toit Mouton and myself) to explore a few very interesting new theoretical insights. Both of us feel that practically speaking you're always better off using a DAC and a normal analogue control loop though.
Akcheng: Slew rate: have a look at http://www.hypex.nl/index.php?option=com_weblinks&view=weblink&id=82&catid=50. This article explains why slew rate is not a good indicator of sound quality. The power bandwidth of UcD and Ncore is 35kHz, as determined by the low-pass filter. Slew rate follows from that. What you really need to know if an amplifier is fast enough for audio is the HF imd plot (18.5+19.5kHz is what I usually do). If the amp stays clean under those conditions, it is fast enough. Interestingly, many linear amplifiers with much higher slew rates cannot match the IMD result of Ncore.
Judderod: the diy module will be announced with the necessary buzz when it's done.
Charles: The vehicle is immaterial. The orbital velocity of earth dominates.
Akcheng: Slew rate: have a look at http://www.hypex.nl/index.php?option=com_weblinks&view=weblink&id=82&catid=50. This article explains why slew rate is not a good indicator of sound quality. The power bandwidth of UcD and Ncore is 35kHz, as determined by the low-pass filter. Slew rate follows from that. What you really need to know if an amplifier is fast enough for audio is the HF imd plot (18.5+19.5kHz is what I usually do). If the amp stays clean under those conditions, it is fast enough. Interestingly, many linear amplifiers with much higher slew rates cannot match the IMD result of Ncore.
Judderod: the diy module will be announced with the necessary buzz when it's done.
Charles: The vehicle is immaterial. The orbital velocity of earth dominates.
- Status
- Not open for further replies.