I don't know much about the new stuff.
The only gear I have long term knowledge and respect for were the Nelson Pass designed GFA 5x5 and GFA 5x00 series built in
the 1990 and 1995 lines.
Too bad Nelson took another path in later years.
-
Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Review and Measurements of Hypex NC400 DIY Amp
- Thread starter amirm
- Start date
The only gear I have long term knowledge and respect for were the Nelson Pass designed GFA 5x5 and GFA 5x00 series built in
the 1990 and 1995 lines.
Too bad Nelson took another path in later years.
fpitas
Master Contributor
True. Did those large ones have better reliability?
Killingbeans
Major Contributor
Yes, but you can only do so much, especially when on a tight budget, and when you add "optimistic" convection cooling, the poor electrolytics will get cooked.
Ohh yes. I know. People have rose-colored glasses on when they talk about "indestructible" class AB amps.
O' heck no. They used the exact same caps used in the SMPS smoothing section as the lesser amps from the same company and it took about 1.5 to 2 days of full time work to rebuild one to spec. People paid for that level of service because the amps where rare, cost large as new and fit into a custom installation.
EDIT: I torture tested the amps to 4 Ohms at near clipping to make sure there where no mistakes and my workbench had brown baked marks where the amps sat while under test. So they got realllly hot.
fpitas
Master Contributor
Well, SMPS can be reliable. Now, I can't swear that Hypex or whoever got it right. But it's hardly an impossibility.
The operating temp of the class D with SMPS is much lower than a class A/AB car amp. Hence the longevity should be much better. Like @Buckeye Amps stated today he's had several units fail in ~2 years.
Killingbeans
Major Contributor
It will happen at some point, that's for sure. Especially if we can get the switching frequencies high enough to make electrolytics unnecessary.
I own a pair of diy Hypex monoblock. In the past 2 years they they have given me great pleasure. It is sad that a very failed method of testing them throws a (false) wrench into their flawless performance, claiming they generate distortion of the signal from the low to the high treble. While it might be argued that my ears are mine and they might not detect -60 db distortion, the Hypex took the place of a Mark Levinson, I play music to become emotionally involved and somehow I improved my involvement upgrading from ML to Hypex. Finally, I have always found a correlation between impeccable Amir’s measurements and improved pleasure from reproduced music, be a new IEM, DAC or amplifier.
Of course they can. What do you think all those servers, switches and routers that form the whole Internet (that allows us to have this discussion) run on? Huge transformers and linear power supplies? Some of those systems have been continuously switched on for 10 years or more.
restorer-john
Grand Contributor
The "garbage" as you call it, is NOT generated by the measurement process, it is generated by the amplifier.
The carrier frequency is generated by the amplifier. The garbage is generated by sampling the output with a sample rate that allows some of the carrier through, but doesn't sample it accurately.
Yeah, HF spectrum of the "SOTA" class D amplifiers is always "interesting". Who cares? It is the "industry standard" now, everyone is building, manufacturing and selling class D, so who would expect they would tell it produces a HF "garbage"?
Just to make a couple of things clear as there seems to be some confusion.
The distortion you see originates from the source but at a much lower level, I am not saying the NC400 creates it, however it does amplify it significantly compared to the Rotel amp. It does this in a very audible range and even if I bandlimit the measurement (not the signals) it will still be there.
I don't understand how this happens and as it were you can create a multitone in REW which I've done with 192khz sampling and it shows no distortion which makes it even more of a mystery. It's the same with whatever multiple of a 1khz single or two tone sinus I use.
And for those that think my post is about dissing class D, it's not and I would really like to be able to replace all my traditional amps with these but since its 11 channels the cost is significant and I know better than to do it blind so to speak.
The distortion you see originates from the source but at a much lower level, I am not saying the NC400 creates it, however it does amplify it significantly compared to the Rotel amp. It does this in a very audible range and even if I bandlimit the measurement (not the signals) it will still be there.
I don't understand how this happens and as it were you can create a multitone in REW which I've done with 192khz sampling and it shows no distortion which makes it even more of a mystery. It's the same with whatever multiple of a 1khz single or two tone sinus I use.
And for those that think my post is about dissing class D, it's not and I would really like to be able to replace all my traditional amps with these but since its 11 channels the cost is significant and I know better than to do it blind so to speak.
restorer-john
Grand Contributor
Good try.
Any device that produces spuriae outside the range it is supposed to work with, is, by definition, producing additional garbage.
Please don't attempt to sugar coat it. It doesn't work.
If I feed in a pure 1Khz tone into a 'switching amplifier' and it produces anything other than that 1Khz tone, would you concede it has failed as a "straight wire with gain"?
If the antialiasing filtering of the adc isnot good enough the hf noise of the amp can generate extra noise
As long as the "anything other" is stuff that is inaudible, I would conclude that the amp does what it is supposed to. Have you looked at the output of a DSD DAC?
Sokel
Grand Contributor
- Joined
- Sep 8, 2021
- Messages
- 10,541
- Likes
- 11,401
Why,what's wrong with it?
(DSD256,no filtering,192K bandwidth)
What you are probably referring to is noise shaping of the files,not the machine itself.
Try it with standard-rate DSD (64).
Indeed. DSD only has a SNR of 6 dB, but gets away with it by using noise shaping to shift it to inaudible frequencies.
Similar threads
