[801evan]

For RF rejection i also believe a SMPS is better equipped by nature cause it has to filter its own generated switching noise to begin with.

Also yes im aware the current draw of a HP is tiny as <150mA.. but in my opinion it cant be wrong to make the rails voltage even more resistant to varying currents no matter how small.

But its a experiment.. i might have to conclude its not audible

There's a reason no caps (direct connection) sounds so good with ifi icans beating amps 4x their price.just give it $$$ psu.

Modded mine with full carbon fiber chassis

 

[SteveHulk]

Your explanation is just plain wrong. Do you think I do not know what I am doing? The PSU interaction with signal related currents is complicated and is a significant source of distortion and SQ issues - all of which has been designed for and accommodated with Dave. Dave would not show the extraordinary low levels of distortion if I had not completely eliminated this issue.

Dave PSU interactions is complex, but is dominated by the very low impedance linear regulator, and the amp PSU rejection ratio. If you were to use a DCA Stealth at -3dB (way too loud) the HP would be drawing 130 mA. It's easy to calculate the PSU induced error from the amp and linear regulator and that comes out at -198dB. That is negligible compared to the distortion of the output stage, which is a thousand times greater than this, and considering this error would induce simple harmonic distortion only, you could consider it as inaudible so not a problem.

But this error at -198dB is just from the linear regulator, and is nothing to do with the supplied SMPS. What is the error from that? Worst case it comes out at -234dB, or 66 times smaller a problem. So your argument relies on hearing an error that is 66,000 times smaller than the distortion in the output stage?

Like I keep saying - it's the RF noise characteristics that changes the SQ with PSUs. And in my experience, SMPS offers lower random RF noise than a linear PSU, as they do not have RF filters built in.

My comment above was about power supplies in general, not just about DAVE, and is borne out by my experiences with the MF A370 amps I have.

On the subject of DAVE, I don't recall you ever adequately explaining the huge sq increases that are observed by users of the Farad3 and SJ psu solutions.

Do you think we all have tin ears?

 

[alxw0w]

Do you think we all have tin ears?

The thing is our ears are not the absolute tool.
Our impressions can be fooled easily.

 

[SteveHulk]

The thing is our ears are not the absolute tool.
Our impressions can be fooled easily.

Are we building audio systems to be listened to by humans, or a roomful of measuring devices?

 

[alxw0w]

Are we building audio systems to be listened to by humans, or a roomful of measuring devices?

I'm not talking about measurements (not only at least).
I'll give you an example:
Some time ago I tried farad3 with my M Scaler (and I've been using it sicnce then) - I thought it's better than getting power from powerbank.
Lately I tried just for fun my powerbank again and guess what?

I liked it more. Sound was calmer, bass tighter highs were cleaner sounding, sound with farad was maybe more impressive, but for longer run I preferred power bank - it was less fatiguing.
So what now? Hmmm?
Our ears are not the ultimate tool - unfortunately. Without rigors listening tests and unfortunately measurements we can easily fall into endless loop of changing just a noise profile...

 

[801evan]

I'm not talking about measurements (not only at least).
I'll give you an example:
Some time ago I tried farad3 with my M Scaler (and I've been using it sicnce then) - I thought it's better than getting power from powerbank.
Lately I tried just for fun my powerbank again and guess what?

I liked it more. Sound was calmer, bass tighter highs were cleaner sounding, sound with farad was maybe more impressive, but for longer run I preferred power bank - it was less fatiguing.
So what now? Hmmm?
Our ears are not the ultimate tool - unfortunately. Without rigors listening tests and unfortunately measurements we can easily fall into endless loop of changing just a noise profile...

The battle of dirty battery management system vs dirty big caps. Lol. Man, that reminds me of months of benchmarking nightmares a couple years ago of exactly what you are saying. It wasn't fun...it was side-grading and making a judgement call was tough indeed.

 

[alxw0w]

It wasn't fun...it was side-grading and making a judgement call was tough indeed.

Exactly this!

 

[adrianm]

Lately I've tried powerbank with m scaler against farad3 (again), and I think that I liked sound more.
I need further testing but who knows, I could end using car battery for m scaler

I had a 50.000 mah Poweroak battery,which i found to be large enough. They have a wide range of batteries. I know there is no way to really tell what quality the voltage regulators will be, but since it had Panasonic cells I assumed the rest would be equally high quality. I was also doing consulting for Panasonic's energy division at the time so call me biased lol. 2 Years later and I don't notice any degradation though.

 

[801evan]

I had a 50.000 mah Poweroak battery,which i found to be large enough. They have a wide range of batteries. I know there is no way to really tell what quality the voltage regulators will be, but since it had Panasonic cells I assumed the rest would be equally high quality. I was also doing consulting for Panasonic's energy division at the time so call me biased lol. 2 Years later and I don't notice any degradation though.

The issue is the battery management system. All of which aren't geared for audio but for safety and practicability. Then if one uses a step up converter, that just makes for more noise.

 

[adrianm]

Just on this, I have singularly failed to be impressed by any power regenerators I have tried, including some costing near the thick end of 10k. All seemed to give a slight hardness to the top end and all injected electrical noise back into the mains wall sockets. On the other hand what they did do very well is to tame my high mains voltage (250V) which was making my 845 valve amps run a bit too near their limits for my liking. When I got my ss amps the regenerators were sold and replaced by a 3KVA balanced mains transformer to run all my system from it with no need for any additional filters.

Well that makes sense,. It would explain why I like my Genesis one plugged into the Aquarius. Though it does use 2 torroidals, which usually have very low leakage.
FWIW the Genesis is only for source gear that draws less than 100w.. When it detects more than a 100w load it just stops regenerating and becomes a pass through. They recommend passive filters for high current amps, or some crazy transformers at 10k+. Their Mosaic range is pretty nice since you can mix and match the required approach for each component at slightly more palatable prices.
I haven't looked into balanced mains transformers thus far.

 

[Ards]

My comment above was about power supplies in general, not just about DAVE, and is borne out by my experiences with the MF A370 amps I have.

On the subject of DAVE, I don't recall you ever adequately explaining the huge sq increases that are observed by users of the Farad3 and SJ psu solutions.

Do you think we all have tin ears?

Would love Rob's take, too, but wouldn't have put the question quite so antagonistically. Perhaps we can all chip in and pay Rob's petrol costs to get him over to @Triode User for an A/B listen!

 

[Kentajalli]

Your explanation is just plain wrong. Do you think I do not know what I am doing? The PSU interaction with signal related currents is complicated and is a significant source of distortion and SQ issues - all of which has been designed for and accommodated with Dave. Dave would not show the extraordinary low levels of distortion if I had not completely eliminated this issue.

If I had a DAVE, personally, I would never try to re-engineer the circuit in any way:
1- I trust you and Chord
2- I don't understand it fully
3 - What for? Damn thing measures perfectly already!

Like I keep saying - it's the RF noise characteristics that changes the SQ with PSUs. And in my experience, SMPS offers lower random RF noise than a linear PSU, as they do not have RF filters built in.

You are correct in that general assertion, however, having glanced at Farad website and read their Tech Talk , they do implement RF filters throughout. From input stage, all the way to output. Apart from choice of transformer and rectifying diodes, to three distinct filters, they call HF filter at pre, intermediate & post .
They are aware and point a finger at RF noise coming from outside (the mains), and point a finger at ground loop HF noise.

And @Rob Watts ! This is your fault! And possibly together with Chord.
Until you guys come up with a two or three box upgraded Goliath (David upgraded?), one for PSU, another for Femto clock, people with lots of money burning a hole in their pockets, will experiment like this.

Bottom line, as not an expert that I am!, this Farad 3 external PSU looks nice (& expensive) and must be great fun to use and show off to friends!
Not to mention well-made and implemented.
What benefit it may have, I would not spend my money and time to discover.
For reasons of 1- 2- 3- and the fact that I am not that rich.

 

[SteveHulk]

Would love Rob's take, too, but wouldn't have put the question quite so antagonistically. Perhaps we can all chip in and pay Rob's petrol costs to get him over to @Triode User for an A/B listen!

I'm still amazed by this idea that the tiny smps in stock DAVE cannot be bettered, and that people shouldn't even bother trying.

Whether the circuit being supplied is analogue or digital it still seems to be that an ideal power supply is a constant voltage current source. That is to say it holds the rail voltages absolutely constant no matter what current it has to deliver in order to do so. This ideal clearly implies the absence of any noise whatsoever as this would manifest as a variation of power rail voltage.

Real power supplies are obviously a compromise of this ideal.

How much of a compromise is determined by design choices generally based on cost grounds.

The closer to the ideal the real supply is the more expensive it is, and this is a very steep correlation.

Farad3, SJ ARC6 etc are reaching towards this ideal and the prices of these devices directly reflects how close to the ideal they have come.

The DAVE as stock is a wonderful device and a stunning technical achievement in its own right.

I still think there is untapped quality headroom and Farad3, SJ, and even people's home-grown designs are earnest and respectable attempts to realise this potential.

 

[Kentajalli]

Are we building audio systems to be listened to by humans, or a roomful of measuring devices?

Certain devices can almost fully be objectively measured, a power supply, an amplifier, a filter network . . . .
Remember, people in a room full of measuring devices, built, measured and manufactured the DAVE!
It is a mathematical device, with a small analogue output section.

 

[adrianm]

I still think there is untapped quality headroom and Farad3, SJ, and even people's home-grown designs are earnest and respectable attempts to realise this potential.

You are correct in that general assertion, however, having glanced at Farad website and read their Tech Talk , they do implement RF filters throughout. From input stage, all the way to output. Apart from choice of transformer and rectifying diodes, to three distinct filters, they call HF filter at pre, intermediate & post .
They are aware and point a finger at RF noise coming from outside (the mains), and point a finger at ground loop HF noise.

While I did like the Farad as a PSU with the SOTM streamer over their own PSU (clearly much worse) , it was still improved considerably by my mains filter, just like my Genesis. So not quite a one stop shop either.