• The Vendor's Bazaar forum is for commercial offers and transactions. Only unmoderated members can post here.

    diyAudio provides this forum for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members. Use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

Reference DAC Module - Discrete R-2R Sign Magnitude 24 bit 384 KHz

Marketing, DSD is needed, especially for end user products, so it will come, first for the dac1101, then the code will be moved to the dam1021.

There are no technical reasons for DSD, most DAC's are multibit anyway. In fact the whole point for a R-2R DAC is to avoid delta-sigma modulation as much as possible, and DSD is per definition delta-sigma modulation....
 
Forta,

Excellent post. But I bet you will have to write it a few more times before your logic sinks in.

Take care,

I play it with Vortexbox, and the DSD2PCM is included in the software, so when i play a DSD file, the Amanero give a signal at my led's that it is 352.
An externally hosted image should be here but it was not working when we last tested it.
 
Member
Joined 2004
Paid Member
In rev2, I think there are only 2 versions, 0.05% and 0.012%.

You're right, I asked Soren himself, and he replied:

"See post #1, I cut the SKU's down to 0.05% and a combined 0.01%/0.02% versions."

And indeed, there it says:

...

Update on September 8, 2015:

We're out of stock of the 0.01% and 0.05% versions, we still have a little stock of the 0.02% version.

Boards from next production run will be rev 2 and will be available Mid September 2015, we finally received the precision resistors so the boards are being mounted right now. Rev 2 boards have the improved vref circuit and added parts to improve power on/off muting.

Pricing for new batch, board rev 2 is:

dam1021-05: USD 195 / EUR 174 (0.05% resistor version)
dam1021-12: USD 295 / EUR 263 (0.01%/0.02% resistor version)

Pricing for first batch, board rev 1 is:

dam1021-05: USD 170 / EUR 137 (0.05% resistor version, sold out )
dam1021-02: USD 235 / EUR 190 (0.02% resistor version, still a few left)
dam1021-01: USD 325 / EUR 263 (0.01% resistor version, sold out)

Shipping cost and any sales tax and/or other local taxes to be added.

...

Is the 0.01%/0.02% version worth it sonically?
 
Soekris, i'm really interested in your R2R dac... i have some questions...

It will look 100% like this?

kozel-rmaf-2015-image42.jpg


Can you let us know the more of features? how we can install new filters; what filters will be implemented and the like?
And finally, when it will be shipping?

All the best and i'm really hoping to get one on my hands really soon.

JP
 
1. The nature of the D/A conversion type used by the board. R-2R and PCM are obverse and reverse of the same coin. If you like DSD go for LPF with fancy output stages. I presume there are a few good examples available on this forum, let alone commercial ones.

Not that I intend to use DSD, but the DAM would be a fine device for native DSD playback ... with the right FPGA firmware.
The FPGA would have be configured as a one-stage LPF running at the bitstream frequency of 2.8Mhz. The output of the LPF (even of 0-1-stream) is multibit so perfectly suited for the R-2R-DAC.

The imperative one-stage approach would make it hard to implement it along side at side with the current PCM design, I think. So I have no hope to see the native DSD in some later firmware.
The one-stage approach, by its nature, does not allow the IIR filtering at intermediate level and the FIR filters would be relatively short (2-4 time the length of the the current FIR2 at most) so also more FPGA horsepower would be desirable for that.

Edit: as the FIR filter applied to a 0-1-stream is only adding filter coefficients, longer filters than in the PCM case can be realized, so perhaps there would be enougth horsepower.
 
Last edited:
Soekris, i'm really interested in your R2R dac... i have some questions...

It will look 100% like this?

kozel-rmaf-2015-image42.jpg


Can you let us know the more of features? how we can install new filters; what filters will be implemented and the like?
And finally, when it will be shipping?

All the best and i'm really hoping to get one on my hands really soon.

JP

It will look just like that, just with text on the case, didn't have time for that on the prototypes. More info on http://www.soekris.dk.

It's an end user product and not really supposed to be opened.... But it will have the same serial interface so maybe I'll post a note somewhere around here.... Filters are currently the same as dam1021.
 
Not that I intend to use DSD, but the DAM would be a fine device for native DSD playback ... with the right FPGA firmware.
The FPGA would have be configured as a one-stage LPF running at the bitstream frequency of 2.8Mhz. The output of the LPF (even of 0-1-stream) is multibit so perfectly suited for the R-2R-DAC.

The imperative one-stage approach would make it hard to implement it along side at side with the current PCM design, I think. So I have no hope to see the native DSD in some later firmware.
The one-stage approach, by its nature, does not allow the IIR filtering at intermediate level and the FIR filters would be relatively short (2-4 time the length of the the current FIR2 at most) so also more FPGA horsepower would be desirable for that.

Edit: as the FIR filter applied to a 0-1-stream is only adding filter coefficients, longer filters than in the PCM case can be realized, so perhaps there would be enougth horsepower.

No filters are possible on a DSD stream, neither are volume control. So it will be conversion to 352/384 Khz so the filters and volume control can be used, will never be anything else on my R-2R DACs. Which happens to be what most others are doing as there are almost no 1 bit D/A converters nowadays....
 
You're right, I asked Soren himself, and he replied:

"See post #1, I cut the SKU's down to 0.05% and a combined 0.01%/0.02% versions."

And indeed, there it says:

Is the 0.01%/0.02% version worth it sonically?

As dimdim wrote:

It's very hard to make an apples-to-apples comparison since I'm pretty sure that there are no two DAM implementations that sound 100% the same.

My DAM implementation manages to sound different from a friend's implementation, even though we are using 99% same parts (same USB interface, same Vref mod, 95% same Salas BiB PS, same filter, etc).

Still, I hear that his implementation easily beat an Yggdrasil the other day. And he has a 0.05% board.

As I have stated before, thanks to the sign magnitude principle the 0.05% and 0.02/0.01% versions will probably sound the same, the advantage of more precise resistors is slightly lower THD and better long term performance due to lower drift.
 
I think the DSD hype is only a marketing gag from the companies only to sell new units!
Since you cant edit the DSD stream, all mixing and mastering of the record is done in PCM and in the end converted to DSD.
No filters are possible on a DSD stream, neither are volume control. So it will be conversion to 352/384 Khz so the filters and volume control can be used, will never be anything else on my R-2R DACs. Which happens to be what most others are doing as there are almost no 1 bit D/A converters nowadays....
And so i think DSD makes no sense for DAM.
 
Perhaps things get less controversial if I reformulate my "DAM DSD proposal" and put it in relation with what will be the "factory solution".

My proposal of feeding the DSD bitstream in a digital LPF running at the same frequency as the bitstream can, of cause, also be viewed as a DSD to PCM conversion, in fact it must be as we want to feed the R2R ladder with it. After that LPF filter we have a (64*44,1) kHz PCM signal, which thus admits further processing, especially volume control.

The factory approach is to feed the DSD bitstream in a decimation filter resulting in a (8*44,1) kHz PCM signal, the intermediate frequency level of the DAM.
The obvious advantage of this approach is that it fits seamless in the current signal processing with the IIR, FIR2, and volume control running at this intermediate frequency level. Therefor I understand the desiccation of Soren to do it (for the moment) like that.

A decimation filter, here, is a digital low pass filter followed by throwing away 7 of 8 samples after filtering.
The, in my eyes, at least esthetical, asset of my proposal is that you do not throw away 7 of 8 samples that you later have to interpolate again (as one of the duties of the FIR2 filter).

P.S.
I hope the LPF in the decimation filter can be altered by the user like the other filters of the DAM. This feature is one of the assets of the DAM and has proven itself as extremely successful.
 
Dynamic issue solved but bass....

I made 0.01R/499R/5600UF, + 470uf ps cap mod

what i hear now it is a definitive improvement on dynamics. But there is something wrong with bass, it is too much punchy and annoying. Is there any way to measure output is fine or not. Maybe there is something wrong with my ears i always listen delta sigma dacs before so I am not used to listen R2R before.
 

Attachments

  • IMG_0461.jpg
    IMG_0461.jpg
    770.6 KB · Views: 801
  • IMG_0462.jpg
    IMG_0462.jpg
    783.9 KB · Views: 764
DSD is comming to DAM1021

Marketing, DSD is needed, especially for end user products, so it will come, first for the dac1101, then the code will be moved to the dam1021.

Any time frame? (end of November ? )
Cheers
Danny
Btw. The only reason I was holding of to compare DAM with ES9018 is luck of DSD support..... hopefully soon there will be chance to compare BIII with DAM (both dual XLR build)

An externally hosted image should be here but it was not working when we last tested it.
image38244.jpg
 
I would like to understand why people are moaning about DSD support to be included. I can see a couple of reasons why it shouldn't.

1. The nature of the D/A conversion type used by the board. R-2R and PCM are obverse and reverse of the same coin. If you like DSD go for LPF with fancy output stages. I presume there are a few good examples available on this forum, let alone commercial ones.

2. In practice DSD support, I guess, would be reduced to implicit conversion to PCM at FPGA. The same could be achieved right now using software player. You could even do that once offline and forget about DSD at all [1].

3. Funny thing is that music industry embraces DXD when they have DSD release of new materials in mind. So a traditional audiophile mantra 'wanna the best" doesn't fit here.

4. Hi-rez bubble. As an exercise I let a couple of audiophile guys listen the same tracks in different format supported already by this board. Red book up to 192/24; binarual recordings. Results: their ability to discriminate low/hi-rez tracks was on par with performance of Jamaica bobsleigh teams during Olympics (btw. I really like them!).

5. Given above as well as limited hardware/labour resources it would be nice to see development elsewhere. Like further polishing filter stack or more ambitious audio over ethernet (broadcast/multicast/unicast).

[1] https://github.com/MaxKellermann/MPD/tree/master/src/pcm/dsd2pcm


Hi,

DSD support is what they committed when 0.99 FW come.
I just don't understand why their new product support DSD but not for their old product DAM1021.

Anyway, I have DSD/CD album 50/50. That is why I want a DSD DAC.

Regards