Fractal Speaker DAC Specialities Masturbatory Conjecture Talk

I made an account just to say this.

A dutch guy, John Brown, I believe his name is, is making a DAC that is capable of outputting about 6W RMS into 8-ohms.

His current product is a fully discrete DAC using some sort of Alien technology he christened "FractalDAC". A guy has ascertained it sounds better than a USB-Fed Terminator though I have no first-hand experience to corroborate his claims.

Logic Board:


DAC board:

(Source: Audiophilestyle)

It has an output impedance of 375 ohms and an output of 2mw, therefore meaning with an adapter connected to an attenuator it should be able to drive relatively high impedance headphones with very high sensitivity.

Would not this essentially transform the necessity of an amplifier into that of a buffered attenuator- or a mere adapter, even, whose specifications would essentially be as low as the tolerance of the components allows?
Therefore, would it be valid if I submitted a buffered wire with a 3.5mm female in one end, and a pair of RCAs on the other, for measuring?

My sources:
https://audiophilestyle.com/forums/topic/24966-ecdesigns/#comments
https://www.diyaudio.com/forums/dig...e-nos-dac-using-tda1541a-769.html#post6293442
https://www.ecdesigns.nl/en/

 

Looks interesting but isn't this just a DAC? Not to mention there are already relevant ecDesigns threads.

 

@purr1n
No. I was attempting to be a smartass.

 

Confessing or bragging?

 

this could be interesting, assuming one has very hi-eff speakers or cans. will probably order when its out

 

Who cares? Where is the electrical spdif or aes and the bnc word clock out to slave the interface?

Most of the shelf USB implementations and drivers = poor sound. You’ll suffer buffer over and under runs in the driver itself. Even with many very expensive products.

 

I tried to read EC designs website and really got nothing out of it except buzzwords

from the manufacturers post in the DIYaudio thread

the use of a power resistor network for I/U conversion is not a new concept or alien, the Japanese have been doing this with burr-brown multibit DACs since the 90s. Chord does this as well with all of their DACs. I think Exogal does this too, and Wadia before them (yes, confirmed)

I have seen several products from Russian firms with this approach, including switching autoformer secondaries following the resistor network to control the volume and match impedance.

these devices operate on the principle that a short, passive path is more 'transparent' (whatever that means) than an active one. in my experience, this is not always true.

the picture in the OP is also confusing, it's not this proposed 6W RMS into 8 ohm DAC but an existing product which has 2mW output into 250 ohms...

the 'powerdac' product specifications according to the manufacturer:

please at least copypasta properly..

 

Why not both?

The perfect amplifier is a wire with gain.

The closest possible with our current technology (which does not allow us to break the laws of physics- just use clever loopholes to "bend" them) is a wire with loss taking in account the original signal is powerful enough to feed the driver. In other words, an attenuator.

The challenge dictates it should:
1: Be completely discrete in the signal path.
2: Measure better than the Schiit in the objective measurement part (SINAD, etc)
3: Bill under $150 for official companies, under $450 for individuals.
4: Sound better than the Schiit in a double blind, volume-matched test.

Therefore, there are no minimal impendance nor power requirements. Therefore, as I mentioned, if the source is powerful enough, an adapter would apply as a sort of "amplifier" (even though it is more of an adapter box) by presenting a load to the source which would output a given power which will then be attenuated to match with the driver, in a similar way to Raal's adapter box. And therefore, the measurements would be as low as the parts allow.

 

Who knows? For all I know, it could be either the next great deal in audio or just flavor of the month. For the moment, I will maintain my eyes open and give the benefit of the doubt since I admit, I am currently too broke to plunk down the $1400 for the stack, especially when the supposedly revolutionary PowerDAC will release early next year.

And I believe he will regulate the power supply of the DAC instead of using a volume control.

The adapter is... exactly that. It might be dangerous to plug in a pair of Grados straight up into some speaker taps.

Essentially, the whole point of this was me proposing to just connect the headphones to a DAC with a wire and call that an amplifier.

Huh. I thought Chord used FPGAs and fancy custom digital filters.

Essentially, it alludes to the shortest way philosophy, where the least "processing" is done to the signal, the better. As theoretically it should be "purer" (if you get what I mean).

And yes, that is the fractal DAC. I apologize if I did not make this clear.

 

yes Chord uses FPGAs for the actual conversion, I'm talking about the analog stage. and actually I could be wrong, Chord is intentionally obtuse about their output stage design, it could be PWM for all I know

what the Fractal guy is proposing is identical technically to the Ion powerDAC from exogal, except the exogal people offer that as a separate box solution in addition to their normal Comet DAC, which uses a D/S PCM chip + DSP, whereas it seems Mr. Fractal wants to offer this in one box

yeah I used to think this way too but then I heard stuff and realized the shortest way ideology is super dumb when taken to extremes. there are many-page-long arguments about 2-stage vs. 3-stage SET amps and the only correct answer is to use as many stages as necessary to achieve the desired sound. spud amps can sound very clear but also dynamically constrained and lacking extension. it's all trade offs.

well it depends on your definition of amplifier. I would call it connecting a headphone to the noise-shaped output of a DAC. it'll work but I don't think it's something I would enjoy listening to

 

I don’t believe anything because the best converters were made by people who have spent years designing converters and improving their own designs.

We’re much more likely to see a new crazy Bricasti module that does something ridiculous like the dac and reverb, a new endgame Lavry converter that kills all, or RME get their act together with the analog parts and steamroll everyone. They’ve been getting better and RME is halfway there and now moving into AVB to compete with MOTU, who kinda leapfrogged them for multichannel interfaces while undercutting everyone else.

 

Ive ordered the U192ETL and the DA96ETF

I'll report back. might also loan the combo @k4rstar so he can compare with his DAC's