Nelson politely pointed to the widespread ultrasonic noise when an SMPS is used. I am not sure what would be the best approach to trying to measure such noise...
However, once the ultrasonic noise is present at the binding posts, it will be coupled to source equipment (DAC) via interconnects' shield and possibly via ground wire in mains cables that are plugged into the same outlet.
However, once the ultrasonic noise is present at the binding posts, it will be coupled to source equipment (DAC) via interconnects' shield and possibly via ground wire in mains cables that are plugged into the same outlet.
=> Scope to measure them
If not, in the case of Meanwell, as pointed kindly by Mark in some recent past, they often provide some tests / measurements and sometimes with scope traces. That's where I got the switching frequency from and the HF residues produced by the VFET amp's SMPS for example.
From there on you can chose a SMPS filter that is likely to filter that HF garbage (or whatever you see BTW) and that way making sure it doesn't prenetrate the rest of your system.
At least that has been my thinking behind...
Claude
If not, in the case of Meanwell, as pointed kindly by Mark in some recent past, they often provide some tests / measurements and sometimes with scope traces. That's where I got the switching frequency from and the HF residues produced by the VFET amp's SMPS for example.
From there on you can chose a SMPS filter that is likely to filter that HF garbage (or whatever you see BTW) and that way making sure it doesn't prenetrate the rest of your system.
At least that has been my thinking behind...
Claude
Hi X,
Impressive noise floor indeed!
Do you by chance have a trace of the noise produced by Sammy's SMPS? I am interested in the 50kHz to 1MHz band...
Just asking, forget if a hassle of course
Thanks
Claude
Impressive noise floor indeed!
Do you by chance have a trace of the noise produced by Sammy's SMPS? I am interested in the 50kHz to 1MHz band...
Just asking, forget if a hassle of course
Thanks
Claude
hmm...
Audio Myth - "Switching Power Supplies are Noisy" - Benchmark Media Systems
I actually more concern about magnetic interference, so I always keep SMPS modules outside of the amp.
Audio Myth - "Switching Power Supplies are Noisy" - Benchmark Media Systems
I actually more concern about magnetic interference, so I always keep SMPS modules outside of the amp.
How is this filtering gonna look like?
I´m going to use a capmultiplier,will that do any filtering?
I tend to favour LRC for the filtering, as complement to Papa's existing filtering stages doing the main job at lower F. Papa is using with good effects RC stages in B1K and VFET...
My measurement gear is audio interface so only up to 22kHz. I would need an RF spectrum analyzer.
Ultrasonic switch noise at 100kHz to 400kHz is in general, not audible provided it is not egregious. Sami’s supplies have built in CMC and LC filters and are FCC and CE cert for emitted noise. You would have to contact him for specifics but in my experience, this is the quietest PSU and amp I have ever heard or tested.
Last edited:
A linear transformer has much larger magnetic interference fields than tiny transformer in a SMPS. On magnetically susceptible amps that have signal trafo like Edcor (M2 etc) the use of SMPS will drastically reduce the EMI picked up by the signal transformer. Here, placing the linear supply outside main amp helps a lot.
Quote: Ultrasonic switch noise at 100kHz to 400kHz is in general, not audible provided it is not egregious.
=> Well, that was my thinking, even re intermodulation etc. as we aren't bats... and then: 200kHz and its harmonics clearly made an impact on the sound of B1K and VFET amp. Same with USB DACs, kind of thing you would hear the difference from the next room. These HF aren't really effectively filtered by conventional RC, you want a L and low ESR etc.
Quote: Sami’s supplies have built in CMC and LC filters and are FCC and CE cert for emitted noise.
=> Based on the schematic I could snap on the net my understanding is most metal casing Meanwell have also at least a LC filter at their output. Small one though. No clue on the bricks. But then we have the scope traces that shows what comes out, regardless the existing LC filter.
The positive I see is that we are on the way to get the same sonic quality for Class A and D from non expensive and easy to use SMPS. Class A/B though I haven't tried and my own amp will be linear with many big caps...
Fascinating...
Claude
=> Well, that was my thinking, even re intermodulation etc. as we aren't bats... and then: 200kHz and its harmonics clearly made an impact on the sound of B1K and VFET amp. Same with USB DACs, kind of thing you would hear the difference from the next room. These HF aren't really effectively filtered by conventional RC, you want a L and low ESR etc.
Quote: Sami’s supplies have built in CMC and LC filters and are FCC and CE cert for emitted noise.
=> Based on the schematic I could snap on the net my understanding is most metal casing Meanwell have also at least a LC filter at their output. Small one though. No clue on the bricks. But then we have the scope traces that shows what comes out, regardless the existing LC filter.
The positive I see is that we are on the way to get the same sonic quality for Class A and D from non expensive and easy to use SMPS. Class A/B though I haven't tried and my own amp will be linear with many big caps...
Fascinating...
Claude
Sami's dual rail smps for audio is very quiet, at least in the audio freq range. I have used them for my M2x, F6, J2, and USSA5. I didn't need to use any cooling fan for the smps.
They can turn on with a huge cap bank without any issue every time, period.
M2x with 264000uf cap bank.
REW reading on a J2.
They can turn on with a huge cap bank without any issue every time, period.
M2x with 264000uf cap bank.
REW reading on a J2.
Last edited:
Hmm... not really impressed by the 50/ 100Hz and its harmonics mains' "residues" on this graph... but perhaps the measurement?
The rest of the noise floor is indeed fine though.
The rest of the noise floor is indeed fine though.
The linear PS i used before converting to Smps are much worst, its between -60db to -80db range, you can hear the humming.
At -110db, you get to see it on the graph, but its totally silent with ear to speakers.
At -110db, you get to see it on the graph, but its totally silent with ear to speakers.
Compared to a commercial set with linear ps. A Sony Vfet TA-4650...
The graph looks kind of horrible, but it sounds so smooth and punchy with high clarity!
I am learning to use REW in a positive way...
The graph looks kind of horrible, but it sounds so smooth and punchy with high clarity!
I am learning to use REW in a positive way...
Thanks for sharing your build with Sami’s PSU. Very neat and tidy! The linear supplies have what I call as a “Forest of noise” - sort of what I observed when running my M2 with linear and it went away when using SMPS.
@meanie
What is the model/part number of the MicroAudio (Cresnet) power supply in your build? I'm assuming that it's a +24/-24 supply but I can't seem to locate in on their website.
Thanks,
Wil
A cap Mx with a linear supply can work very well. Here is mine on LuFo amp running about 3A and 28v. It is really tough to get a linear supply much cleaner than this. What one normally gets is the forest of noise type hash. Amp is putting out about 8Vrms into 10ohm load.
