It is not a measurement error, it is the measurement floor. Every instrument has a measurement floor below which you can't 'look'. So far, AP has the lowest measurement floor of anything available. The AP 2722 was the world leader until a few years ago it was eclipsed by another AP product, the APx555B.
Jan
Last edited:
It can look funny if you are not familiar with it or see it for the first time. But every instrument has a floor below which you can't see the signal.
In most equipment it is a continuous noise floor, but with really high performance, low noise equipment like the AP the noise is so low that you see it varying with the internal pre-amp gain.
There is no measurement error, unless you call a noise floor a 'measurement error'.
Normally it isn't very useful to re-define terms and it leads only to confusion and misunderstanding.
Jan
In most equipment it is a continuous noise floor, but with really high performance, low noise equipment like the AP the noise is so low that you see it varying with the internal pre-amp gain.
There is no measurement error, unless you call a noise floor a 'measurement error'.
Normally it isn't very useful to re-define terms and it leads only to confusion and misunderstanding.
Jan
Last edited:
Very interesting discussion.
Can ICE amp modules be purchased for DIY? Or someone sells ICE based amps that are not priced as audiophile products? 🙄
I have UcD400 driving woofers and looking to purchase new amps to drive midrange in baffleless dipole, so not huge power needed but it's a sensitive frequency range. I was looking into NCores, but Kurt's comments about ICE AS series sound character got me interested.
FWIW, I have tubes driving midrange and tweeter now. I don't pretend to get the same sound, but "dry" is a term that kind of makes me hesitate. Would like to try good class D vs tubes in my setup and see if I can live with it, and use less energy and space too.
Can ICE amp modules be purchased for DIY? Or someone sells ICE based amps that are not priced as audiophile products? 🙄
I have UcD400 driving woofers and looking to purchase new amps to drive midrange in baffleless dipole, so not huge power needed but it's a sensitive frequency range. I was looking into NCores, but Kurt's comments about ICE AS series sound character got me interested.
FWIW, I have tubes driving midrange and tweeter now. I don't pretend to get the same sound, but "dry" is a term that kind of makes me hesitate. Would like to try good class D vs tubes in my setup and see if I can live with it, and use less energy and space too.
ICEpower Class D Amps Guide | Profusion
ghentaudio --- GA-S250P Stereo Power Amplifier
Personally I use Pascal, but also gave tube class A and a quite good hand built class AB. I don't think the Pascal is particular dry compared to the other amp.
ghentaudio --- GA-S250P Stereo Power Amplifier
Personally I use Pascal, but also gave tube class A and a quite good hand built class AB. I don't think the Pascal is particular dry compared to the other amp.
I´m afraid not.
ICE is purely professional, just as Pascal and the Hypex OEM products.
They just don´t want to support DIY´ers.
Anyways if of any interest, I have a single 1200AS2 including al connectors to spare. You can PM me for info.
Is there a direct or even inverse correlation between power efficiency and distortion in class d designs?
I mean if efficiency goes from, for intance, 90% to 96% at -12dbfs, distortion necessarily decrease (or increase) a certain order of magnitude.
Or both are totally decorrelated?
I.e. a class d design can increase efficiency while keeping the same magnitude of distortion.
I mean if efficiency goes from, for intance, 90% to 96% at -12dbfs, distortion necessarily decrease (or increase) a certain order of magnitude.
Or both are totally decorrelated?
I.e. a class d design can increase efficiency while keeping the same magnitude of distortion.
Class D performance increases with decreased dead time. As you reduce dead time for a given design you start getting small amounts of cross conduction as the output FETs switch. Cross conduction reduces efficiency.
So if you took a class D amplifier that had been optimised heavily for maximum efficiency it's likely that you could increase its performance by decreasing dead time. This would lead to some cross conduction and would lower efficiency.
You could alter the drive strength to the output FETs too. If you increase the drive strength you'll speed up the switching of the FETs. This will reduce the time of the switching window and theoretically allow you to reduce the dead time. This will lead to improved performance without the decreased efficiency due to cross conduction but in this case it takes additional power to switch the FETs on and off faster.
All class D amplifiers can be optimised for performance or high efficiency.
So if you took a class D amplifier that had been optimised heavily for maximum efficiency it's likely that you could increase its performance by decreasing dead time. This would lead to some cross conduction and would lower efficiency.
You could alter the drive strength to the output FETs too. If you increase the drive strength you'll speed up the switching of the FETs. This will reduce the time of the switching window and theoretically allow you to reduce the dead time. This will lead to improved performance without the decreased efficiency due to cross conduction but in this case it takes additional power to switch the FETs on and off faster.
All class D amplifiers can be optimised for performance or high efficiency.
It's very easy to buy icepower modules. There is also one vendor who sells pascal modules, but only incl their own cases.
Thank you for your answers.
Pardon my ignorance, but by “increase drive strength”, do you mean increasing the switching frequency?
I have just seen a 100w (into 4 ohms) class d amplifier developed in my country that claims more more and less 97 efficiency (the claim is that board dissipates 2 or 3 watts of heat) using Sigma Delta modulation.
I presume Sigma Delta pulse density modulation allows higher switching frequencies than pulse width modulation.
But it is not clear to me if the distortion (data not released yet) is low enough to compete with a similarly powered class AB design.
I guess I will have to try it anyway. At least, I gain power efficiency.
Actually that's a good point. Turning a MOSFET on and off takes energy, you have to charge and discharge the gate, and if you do that in 1 second or 1ns the amount of energy is still the same. I wasn't talking about increasing the switching frequency, just the speed that the MOSFETs are turned on and off. But from what I've just written it doesn't matter how fast they are being turned on and off, what matters is how often.
So no, speeding up the rate at which the MOSFETS switch, as a way of decreasing overall deadtime, shouldn't decrease the amplifiers efficiency.
So no, speeding up the rate at which the MOSFETS switch, as a way of decreasing overall deadtime, shouldn't decrease the amplifiers efficiency.
Newbie here. Hope to be excused for my non existing level of knowledge.
I'm looking to build an 8 channel amp for my OB speakers.
Actively driven by an external crossover, would the 125asx2 be a good sounding choice? Limited experience with class d modules but read that late generations are better.
Heard nCore but it sounded a bit too clinical for me.
Class D Amplifiers (Buy Online) | Profusion
Let me know if this thread is incorrect place to ask
Thanks
I'm looking to build an 8 channel amp for my OB speakers.
Actively driven by an external crossover, would the 125asx2 be a good sounding choice? Limited experience with class d modules but read that late generations are better.
Heard nCore but it sounded a bit too clinical for me.
Class D Amplifiers (Buy Online) | Profusion
Let me know if this thread is incorrect place to ask
Thanks
I'm a newbie as well, I am in a similar situation as yours and been doing some digging.
I also have 4-way active stereo system, and also trying out OB now. I use Hypex UcD400 for woofers and can't say they sound clinical, but the frequency range isn't so sensitive either. But I did read some comments that NCores can be somewhat dry so for my midranges I'm inclined to try the ICEpower AS series which some have described as less clinical.
Frankly, it's not clear to me how the AS and ASX series differ. However, I had a conversation with Rouge Audio this week around AS modules and bottom line they highlighted the ICEedge chip (latest generation) in the 1200AS as making the difference in sound vs the other ICEpower modules. ICEpower website and datasheets aren't too clear about which modules have ICEedge. On the Profusion website I only see ICEedge noted on 1200AS1.
So while 1200W is waaaay more than I would need for midrange, I'm further digging to understand if there are other modules with ICEedge.
Please keep us posted with your findings.
I asked what modules are available for a diy:er but unfortunately the as1/as2 modules aren't available to buy. At least not through profusion which would be where I'd buy them.
Does anyone here know where/if as1/as2 modules can be purchased if you're a diy:er?
Pnotus comments about the ASX series made me compare those to AS series. I found some things that might be interesting so I will put out there for knowledgeable members to help me understand the relevance to sound.
Looking into the AS series, and knowing the ICEedge chip makes a difference, I noticed "Output referenced idle noise" drops significantly for the 1200AS vs the others. Checking the ASX I was surprised to find similar levels (even lower) for the 50ASX2, especially the BTL at 20uV. Also the 50ASX2BTL has 125 dB dynamic range, the closest to 129dB displayed by the 1200AS.
Next I looked into THD+N at 1W and 1 kHz also into the graphs to get THD+N at 10W and 6.67kHz. Interestingly, ASX modules seem to perform better at 1KHz, which the AS, and in particular 1200AS, do better at 6.7kHz.
How important are these parameters when selecting an amp for a midrange channel?
