New $350 R2R DAC

[sandston]

Massdrop just released this R2R DAC with some pretty impressive spec's. 

 

https://www.massdrop.com/buy/massdrop-x-airist-audio-r-2r-dac?utm_placement=0&referer=B725EP&mode=guest_open&tk=IsLmskCSkgcprfEp8lizEw&utm_campaign=Automated Daily Promotional 2018-06-06&utm_source=SparkPost&utm_medium=email&utm_term=Daily Promotional&utm_content=1528271271999.259150416514067821831544

[tmtomh]

Very interesting! How can they produce an R2R DAC at that price point?

[Ron Scubadiver]

R2R is inherently expensive owing to the need for many high precision resistors.  This could be a deal or a dud.

[Ralf11]

perhaps there is a new technology allowing resistor trimming?

 

or a loss leader to get awareness in the US for this brand

 

or maybe it sux..

 

https://forum.headphone.com/t/massdrop-x-airist-audio-r-2r-dac-review/1653

[mansr]

1 hour ago, Ralf11 said:

perhaps there is a new technology allowing resistor trimming?

That won't help. Look at the size of that thing. You'll easily get a temperature difference of 0.1 °C or more from one end to the other. Now look up the temperature coefficient of the best resistors and do the maths. The variation is enough to throw the accuracy way off what is claimed.

[Ralf11]

maybe it is full of heat pipes?

[mansr]

11 minutes ago, Ralf11 said:

maybe it is full of heat pipes?

With a giant solid silver heat spreader? Not at that price.

[rickca]

this sign-magnitude DAC has two independent 24-bit ladders with 48 resistors apiece per each of the two channels, which cancel out errors and distortion

 

Nice theory.  Would it yield the precision required for great SQ?

[Ralf11]

17 minutes ago, mansr said:

With a giant solid silver heat spreader? Not at that price.

 

heat pipes are quite cheap - they use a working fluid contained in an impervious solid

 

but if you still don't like it... how about Peltier modules?

[PorkChop]

Offer says it doesn't ship until December, so maybe that's like a crowdfunding price.  Interested, but will wait for some "real" reviews after release.

[arglebargle]

Some advance reviews over at SBAF. 

http://www.superbestaudiofriends.org/index.php?threads/massdrop-rdac-airist-audio-r2-r-dac.6415/

[mansr]

30 minutes ago, rickca said:

this sign-magnitude DAC has two independent 24-bit ladders with 48 resistors apiece per each of the two channels, which cancel out errors and distortion

 

Nice theory.  Would it yield the precision required for great SQ?

Define great. As for numbers, they don't add up. If it was just about spectrally flat noise, averaging two D/A units would increase the effective resolution by 1 bit. Now we're not just dealing with simple noise but also various forms of signal-correlated non-linear distortion. This doesn't neatly cancel out as they'd like you to believe.

[mansr]

41 minutes ago, Ralf11 said:

heat pipes are quite cheap - they use a working fluid contained in an impervious solid

 

but if you still don't like it... how about Peltier modules?

The problem isn't one of moving heat from one place to another. The challenge here is to maintain 96 resistors (and the wiring between them) spread out over several square inches at very close to the same temperature. The only way to do this is by bonding them to a large chunk of something with very high thermal conductivity, such as diamond, and I'm not convinced even that would be enough to maintain the tolerances required for a 20-bit DAC (forget about 24-bit, it's impossible by any means).

[mansr]

34 minutes ago, PorkChop said:

Offer says it doesn't ship until December, so maybe that's like a crowdfunding price.  Interested, but will wait for some "real" reviews after release.

This is apparently a limited release, get in or get lost. No word on plans for production beyond this initial run.

[asdf1000]

8 minutes ago, mansr said:

This is apparently a limited release, get in or get lost. No word on plans for production beyond this initial run.

 

Potentially some issues (or not), reading from this post onwards:

 

https://www.head-fi.org/threads/massdrop-x-airist-audio-r-2r-dac-a-discrete-resistor-ladder-dac-for-350.881315/page-3#post-14284822

 

[arglebargle]

^ The issue being the origin of the design. SBAF  has a dedicated thread on that as well. Curious. 

 

http://www.superbestaudiofriends.org/index.php?threads/md-x-airist-x-sosolar-rdac-design-mysteries.6420/page-3

[mansr]

17 hours ago, mansr said:

That won't help. Look at the size of that thing. You'll easily get a temperature difference of 0.1 °C or more from one end to the other. Now look up the temperature coefficient of the best resistors and do the maths. The variation is enough to throw the accuracy way off what is claimed.

The best resistors Vishay has to offer in standard SMD package have a temperature coefficient of 5 ppm/°C. To achieve 20-bit resolution, the resistors need to match better than 1 ppm. Assuming otherwise ideal resistors, the temperature variation across the entire PCB must be no greater than 0.2 °C. Judging by the available photos, this can't possibly be maintained. The heat from that FPGA at the end and the power supply underneath will cause far greater variations. Getting better than 14-bit precision from a design like this is a fool's errand.

[mansr]

13 hours ago, arglebargle said:

^ The issue being the origin of the design. SBAF  has a dedicated thread on that as well. Curious. 

 

http://www.superbestaudiofriends.org/index.php?threads/md-x-airist-x-sosolar-rdac-design-mysteries.6420/page-3

Looking at the photos of this and the other board, I would say it is definitely a copy. No question about it. Yes, there are only so many ways to build a ladder DAC, but this goes beyond that. Every last detail on those boards is identical. Even if two people started with the same schematic, they'd be extremely unlikely to produce the exact same PCB layout.

 

As an analogy, imagine two people deciding to do a film adaptation of a book. The similarity here is equivalent to the two films being frame by frame identical except for using different actors.

[opus101]

27 minutes ago, mansr said:

Getting better than 14-bit precision from a design like this is a fool's errand.

 

There are measurements over on SBAF - http://www.superbestaudiofriends.org/index.php?threads/rdac-measurements.6418/#post-210258.

 

The THD at full scale looks around -70dB which looks in the right ballpark for 0.05% resistors.

[opus101]

20 hours ago, tmtomh said:

How can they produce an R2R DAC at that price point?

 

Evidently by skimping rather on R&D costs....

[mansr]

4 minutes ago, opus101 said:

There are measurements over on SBAF - http://www.superbestaudiofriends.org/index.php?threads/rdac-measurements.6418/#post-210258.

 

The THD at full scale looks around -70dB which looks in the right ballpark for 0.05% resistors.

That's roughly equivalent to 12 bits.

[Ralf11]

Bottom line, it's an illegal copy of a flawed design - n'est pas?

[mansr]

52 minutes ago, Ralf11 said:

Bottom line, it's an illegal copy of a flawed design - n'est pas?

Something like that, yes.

[esldude]

Maybe they are using new technology that uses a laser to heat each resistor to the perfect temperature all the time.   Sounds like EE's letting those pesky blindspots they learned at university unwilling to open their eyes to the possibilities once again.  Just believe it and make it so.  When will you guys learn?

 

My guess is it has 27 subjective bits of resolution.  24 bits via the resistors, 1 bit from doubling, and 2 bits from non-objectively verifiable magic.   

[Ron Scubadiver]

21 hours ago, PorkChop said:

Offer says it doesn't ship until December, so maybe that's like a crowdfunding price.  Interested, but will wait for some "real" reviews after release.

December, yikes.  Hell could freeze over by then...