Soekris DAC-s dac1101 and DIY dam1021

[ergo]

There is an intro thread about the DAC-s here but as this has 'on the bench' content also I thought I'll start a thread here too. At one point when the forces will allow I'm sure Jon will chime in with AP measurements of his cad1101 as well.... will be interesting to check if they match in end.

I spent most of today's easter day assembling the Soekris DIY version. I added the WaveIO for USB/I2S and used Jan Diddens Silent Switcher for the power supply.

Goal was to get sound out of it and do some measurements and of course get to listen to it too. Still contemplating if I'll put it into same case as the Balanced Preamp as it has more than enough room or to order one more of those from Italy and keep the DAC in separate housing. Leaves more freedom for upgrade paths. One never knows when the next itch comes.

The ratnest wiring will of course will be cleaned up once in final box. This is just to allow messing with it and getting it to work right.

[ergo]

Some measurements...

I don't have a newer AP, so no nice low level sine wave waveforms. But some standard 1kHz FFT-s have been done.

44.1kHz / 0dBFS


44.1kHz / -60dBFS


96kHz / 0dBFS


96kHz / -60dBFS


96kHz / -60dBFS

[JonMarsh]

Something to keep in mind with this sort of measurement- I'll basically have to do something like this, using a MacBook Pro to drive the Soekeris, is considering the residual level and noise floor of the system- which certainly appears to be quite low. Distortion on these R2R ladder DACs is usually mostly due to resistor match- not that it isn't "real" but it is more "passive" and generally less objectionable.

Of course, the 44.1 has a higher noise floor, but that's still really very good- and the -60 dBFS signal is quite clean. I'm curious to see what things look like using a MacBook Pro as the signal generator on USB and then measuring with the AP in FFT. I also plan to look at dynamic signals and try to assess the filter settings.

This weekend has still be largely spent working, getting caught up on some other time sensitive stuff now that the webinars are done (i.e., stuff due this coming week!) but I did get to have breakfast with my daughter and her family, who'd come up from Mission Viejo for the weekend - first time to see the grand daughter! That was fun, and I got some pictures, too. All of them - Elaine, Jon, and Stella - look to be doing well.

[ergo]

Very good to hear that your'r having some fun family time too in between the crazy work hours. The baby will start developing so fast now for next 2..3 years, so better make sure you don't fully miss all that fun

Yeah, I used Windows 10 PC and tried ARTA / REW / HpWworks demo and Adobe Audition 3 for signal playback. Windows audio is unfortunately indeed much more convoluted with WASAPI, ASIO etc. HpWworks also generates the Jitter test signals, but I have to still learn some more how to plot and use those results.

I also tried Amanero OEM Combo384 Module as the USB -> I2S converter but had much less success with that. The driver is from 2013 and seems to not play fully well with Win10. It fully crashes ARTA in ASIO mode for example.

Do you know good signal generator software for MacOS. I could try that with the MacBook from work later this week. I also plan to test the Consumer Soekris as this weekend it was on loan to my friend whom we did the speaker upgrades recently.

[JonMarsh]

Hi Ergo,

AudioTest.APP on the app store has a lot of features and flexibility- and it's hard to beat the price at $4.99 USD. From Katsura Shareware, on the App Store.

[ergo]

I did play with Consumer dac1101 a bit today. The below measurements are with it connected to the MacBook Pro and using the audiotest SW.

One of the first learnings with the Consumer DAC is that one needs to read the manual and know how to use/set the playback volume from the volume control. Namely after power up the DAC starts with a very low level. When turned up it goes not upto 0dB but to +10dB instead.

So if one maximizes the volume with 0dBFS signal the result does not look too good.
The result looks like this from 0 to -9steps down
96k_32bit_0dBFS_vol-9steps.png


Lowering the volume to -10steps one can see a much cleaner result
96k_32bit_0dBFS_vol_-10steps.png


96k_32bit_-60dBFS_vol_-10steps.png


44.1k_16bit_0dBFS_vol_-10steps.png


44.1k_16bit_-60dBFS_vol_-10steps.png

[ergo]

The low graph on this image is the residual noise of the FX100 analyzer I'm using... all measurements used 50x averaging also this time.

[ergo]

But now the most interesting news I found from http://www.soekris.dk/products.html
dac1321 / dac1421 / dac1541 - three brand new ready made units

http://soekris.eu/shop/dac_modules_consumer_line_new_en/

The top one is 1110 EUR and has

USB Input Type B, Isolated, Full/High Speed
USB Input Mode Selectable Audio Class 1.0 or Audio Class 2.0
SPDIF / AES / Toslink Inputs Up to 24 bit / 192 Khz
USB Input PCM Up to 24 Bit / 384 Ksps
USB Input DSD Up to DoP-128 and DSD-256

[Evil Twin]

Not Nice!

Mr Ergo,

In the condition I'm in, that was not very nice to post to that impressive looking unit! It was all I could do to restrain myself from placing an order this afternoon for the top one... and I've still got the evening to get through intact without giving in! Talk about being strong in the Force... It is not will power that is critical, but Won't Power...

For someone who has been living total deferred gratification the last several months, this is not easy to deal with. Of course, with June Delivery, it would still be deferred.

For those who don't click through the web site, see below...

[ergo]

Sorry to cause unplanned expenses to the Palpatine Heavy Industries

The urge is there for sure... for me just few reasons that keep me back are
* I had just decided to go for the RTX6001 audio analyzer over at DIYaudio for approx same price
* I already bought both the cac1101 and the DIY one which is still playing next to the preamp case it should go into... if I'd buy another one as ready unit this will in all likelihood end up shelved. So better to keep myself back and finish what's started.

I have been listening to the DIY dac now more than the consumer one. I do like it. Perhaps the best aspects are the feel of speed and precision of various attack sound etc. and secondly the depth of the soundstage. I do not have super high end DAC-s to compare it to but for now it does well against NAD C390DD all digital amp and RME UCX soundcard.

[JonMarsh]

This is a pretty interesting looking piece- I was saved from spending the money because the order stalled at the PayPal point- they seem to have some glitch or issue with their online ordering process. Also, note that Soekeris USA is going/has gone out of business. Too much competition from China on the computer networking products, I think.

[Evil Twin]

Pre-Ordered.

Never underestimate the power of the Dark Side...

[DFAL Minion]

should measured & subjective evaluation of my Dark Master hint at equivalence to the "Rebel's Yiggy device, this will most likely be an acquisition for upgrade to our local device.

[Evil Twin]

Indications are that there will be a somewhat different set of tradeoffs- no closed form filter, but also no MSB zero crossing glitch, as this is a sign-magnitude R2R ladder DAC, who's basic structure seems similar to the TotalDAC devices. IF measurements are similar to the ones Ergo has made, then it should have an interesting characteristic of a very low noise floor and a distortion floor that tracks downward with the number of bits/signal level.

This may be subjectively quite acceptable.

The most compelling comparison might be to the AURALiC Vega DAC, which is a similar form factor, similar inputs and outputs, but a very different technology using one of the 32 bit Saber chips. The Vega has a very quiet output stage also...

Investigating the FFT behavior with CCIF signal tones should be most interesting- depending on the digital filter behavior.

[Evil Twin]

An inquiry to Soekris Europe resulted in a quick response that due to some component delivery delays, they expect they will be building units next week, and shipping soon thereafter, for the DAC1541. It is with considerable anticipation that I await the delivery... pray do not disappoint in the results!

[Evil Twin]

There is a saying on this benighted planet that all things come to those who wait...

What isn't discussed is how long that waiting may take... but in some cases, eventually in spite of materials logistics delays, things come through. This was the case yesterday, for the new "high end" Soekris DAC/Preampo/HeadAmp.

Striking and useful features include:

• AES/EBU inputs
• S/PDIF inputs in both RCA and BNC format, as well as optical coax
• USB input, switchable Class 1 (compatible with Windows without drivers) and Class 2 (compatible with Mac without drivers, and higher sample rate options)
• Balanced XLR outputs and unbalanced RCA outputs
• Headphone outputs both on 6.3mm phone jack and 4 pin XLR
• Variable cross feed controls for headphone listening
• Rotary encoder volume control with up to +10 dB gain
• Multiple digital filter options
• DOP and DSD capable on USB input
• R2R ladder DAC using 0.01% resisitors- no delta-sigma converter or CHIP DAC
• Wide sample rate range, with front panel indicators of operating range
• Very compact; self contained power supply. Makes an AURALiC Vega look porky. Could be considered portable; would hardly take any room in a suitcase, but is AC powered. Not as portable as their original DAC boards, which were powered off the USB input (only input on those DAC boards)
• Galvanic isolation on all inputs to main DAC

Available now in stock. The received unit was SN00014.





With only brief headphone listening, it's clear it is a very smooth sounding unit with articulate extended bass and a clear detailed midrange, and no apparent digititus in the treble.

After a bit of run-in time, we will have to find a local source to do some bench testing on this unit- perhaps Mr. Marsh can be persuaded to tear himself away from his new vehicle next weekend, to establish the baseline capabilities on the AES/EBU input.

[Evil Twin]

One necessary follow up note- the listening has been done with the newer ATH-ANC9 Quiet Point headphones- they are a surprising value, and significantly more transparent across the band than the "audiophile" ATH-MSR7GM, which cost about 35% more. And this in spite of the noise cancelling amplifier built in.

This is apparent on cuts as diverse as Nickel Creek's "Lighthouse", You+Me "Capsized", to Lorde's "Tennis Court" and "Royals".

[DFAL Minion]

- perhaps Mr. Marsh can be persuaded to tear himself away from his new vehicle..............

flux capacitor powered & sourced from the supply depot on the asteroid Elon Poe in the Sil'Musk asteroid belt, are we then to assume?

[JonMarsh]

Well, I'm sure that would work, but to make for a level playing field, ET figures we should test them with good old American 115VAC instead... no power regeneration, no batteries or flux capacitors, etc, and for the moment, without re-clocking. the idea is establishing some parameters for a solid "budget" system.

It will also be interesting to see how the Vidar's test with their revised bias system (compared with the Ragnorak that drove the Stereophile people crazy on the bench)

All that can come later in a follow up.

Vidar's havne't shiopped yet, though they were quoted as being in stock when the order was placed.

[JonMarsh]

UPDATE:

Got shipping notice on the Vidar's just a few minutes ago. Shipping via FedEx today, numbers have been issued, but not yet active. Should have them Saturday or Monday, I think...


It will be interesting to see how these do compared with the Benchmark AHB. I'm wondering if these use the same Sanken Ring Emitter transistors; I hope so, they're the best audio parts I know of in this voltage class. It will also be interesting to see how the current feedback gain topology works out- this normally can provide more loop bandwidth and higher slew rates.


Those without balanced outputs won't be able to take advantage of the bridged mono block mode, but IMO this is the right way to do this- one has a fully balanced signal path to the speakers without using anyone extra inverter/converter stage.

BTW, this is how to make an NAD M22 into a mono block, too- make an input adapter that goes from balanced to two singled ended outputs out of phase, and feed each channel separately.






This may be a very interesting combination at this price point.

[wkhanna]

just went to the website to order the Soekris DAC 1541

wouldn't you know it, 'out of stock' and not even an option for backorder.

any advice for an alternative in a similar price range?

@JonMarsh

[Norm]

just went to the website to order the Soekris DAC 1541

wouldn't you know it, 'out of stock' and not even an option for backorder.

any advice for an alternative in a similar price range?

@JonMarsh

Not Jon but I have been VERY happy with this one if you are up to DIYing case and power supplies. I like the integrated USB and SPDIF interfaces.

https://soekris.modhouseaudio.com/soekris-oem-products/dam-1941-and-1921

[wkhanna]

hi, Norm!
thanks for the feedback!
if i had time in my life right now i would consider a DIY with the board, since it appears to still be available.
but alas, working full time and running a side business along with coping with the time suck managing my disabilities requires leaves no bandwidth for such an endeavor.
i would love some pics of your unit, though!
and glad to heard you are still enjoying it!:T

[Norm]

Hi Bill,
Time suck of (insert 2020 BS on top of existing life here) I can certainly understand.

Here are some pics of the Soekris dam1921 I built.

Providing source for Audio Crafters Guild demo system at LSAF 2019



The dam1921 is the component far left top shelf.

Here is a close up, the PCB is mounted as essentially a breadboard to black painted MDF base with black acrylic front panel.



The dam1921 is the single-ended output version thus the many empty PCB locations where the negative polarity DAC are not loaded. To the right note the direct connect of wires between RCA jacks and PCB. This unit is configured with no output buffers. I have the board allowing op amp output buffers but have yet to play with that configuration. Also note the skinny rectangular PCB along bottom of this picture. That is the optional front panel control board. It allows input selection, sample rate display, and volume control. When this is not used the dam1921 defaults to max volume and the active input.

This photo from soekris.dk (http://www.soekris.dk/dam1941.html) shows the full dam1941 balanced out board set. The dam1921/1941 is sold as a loaded and tested PCBA. The accessory PCBs are empty and builder has to source parts and solder, not a newbe project to use these.



I will note the power supply has a layout error. The heatsinks along bottom edge (only one installed in build pictured) have solder pins for mechanical mounting. These are connected to ground. The TO220 regulators used do NOT have their thermal pads at ground potential. Thus the regulator needs to be electrically insulated from heatsink with mica or silicon thermal pads. But then what's a DIY project without at least one short circuit the power supply to ground moment?
:E :M 8O :rofl:

[wkhanna]

that is great stuff!!!
thank you so much, Norm!

you make it look so easy, but your not fooling me, lol!!!
:T:T:T