The Best DAC is no DAC

Jump to Latest
#501
nautibuoy said:
... I have been thinking of using a TVC as a combined attenuator/LP filter, however, this implies several centimetres of wire from the flipflop to the TVC, which now seems to represent a significant compromise - any thoughts?

Ray
Click to expand...

I can't find the original post, but it was suggested to keep Flipflop <-> transformer distance as short as possible.

How critical is this and why?
 
#502
Stijn, I'm assuming this applies to use of a TVC just as it does with the R and C layout of the LP filters?

gentlevoice said:
@nautibuoy: What really matters...
... Also, the distance from the output resistors to the output capacitors (including ground return - i.e. the loop distance) should be as short as possible. At these frequencies inductances even in 1 mm length of track may cause HF attenuation - if the HF attenuation happens in the track between the output resistors and the filtering capacitors the HF noise will end up being passed on to the electronics down the line in the audio system.
Click to expand...

The windings within the TVC also represent a long track but I have no idea whether, being a transformer, the behaviour is different to a simple length of wire or PCB track?

Hopefully there'll be someone able to help with my understanding.

Ray
 
#503
I got the JLSound USBtoI2S board to put out Native DSD with the XMOS trial Windows driver. There is a a 4R7 resistor required between pin 1&9. After that the driver regonizes the board as DSD capable.

So, first impression. There is no way this, no-dac thing-a-me-bob has any right to sound this good (Not to critize your board Ray 🙂), its quite something. I have none of the noise I heard in Doede's prototype.

I don't normally put a new build onto my main system this quick, but curiousity got the better of me and it seems to behave well enough. This thing is razor like analytic.

Ray, one issue I have is that with certain tracks the relay keeps cutting the signal. Any idea? & playing anything other the 5.6448 MHz, (DSD128), it cuts out continuesly.


An externally hosted image should be here but it was not working when we last tested it.
 
Last edited:
#504
RESULT! Now you understand the enthusiasm from myself and others.

Next you need to get DSD256 capability.

stijn001 said:
Ray, one issue I have is that with certain tracks the relay keeps cutting the signal. Any idea? & playing anything other the 5.6448 MHz, (DSD128), it cuts out continuesly.
Click to expand...

The relay is triggered when the JLSounds board detects a Codec change. I would suspect the prototype driver in the first instance. I haven't had any of the issues you've described; I'm using Debian Stretch on an old Intel atom computer as an HQPlayer NAA with everything upsampled to DSD256 - could you replicate something similar and see how it behaves. Or, how about getting a cheap Intel NUC and using it as an NAA; BavMike recommends this one and the price is pretty reasonable

http://www.amazon.co.uk/gp/product/...ue&ref_=ox_sc_act_title_1&smid=A21ISB9LUWIYXR

Ray
 
#507
from stijn001

"I got the JLSound USBtoI2S board to put out Native DSD with the XMOS trial Windows driver. There is a a 4R7 resistor required between pin 1&9. After that the driver regonizes the board as DSD capable."

I totally missed this in the manual for the board.

Would the same mod enable DSD for OSX?

Thanks and best
 
Last edited:
#512
On the aforementioned driver download page it suggests that there will be periodic beeps or some such after an hour. Does this happen with you? I would love to get native DSD out of my JLSounds card. Personally I feel they needs to do something about this. Their site states it is capable.
stijn001 said:
I got the JLSound USBtoI2S board to put out Native DSD with the XMOS trial Windows driver. There is a a 4R7 resistor required between pin 1&9. After that the driver regonizes the board as DSD capable.

So, first impression. There is no way this, no-dac thing-a-me-bob has any right to sound this good (Not to critize your board Ray 🙂), its quite something. I have none of the noise I heard in Doede's prototype.

I don't normally put a new build onto my main system this quick, but curiousity got the better of me and it seems to behave well enough. This thing is razor like analytic.

Ray, one issue I have is that with certain tracks the relay keeps cutting the signal. Any idea? & playing anything other the 5.6448 MHz, (DSD128), it cuts out continuesly.


An externally hosted image should be here but it was not working when we last tested it.
Click to expand...
 
#515
stijn001 said:
I made the same comment..

The XMOS chip with DSD256 under Windows has been shown to work on the Wave IO board. The only one who seems to have bothered to develop this is Lorien. (Maybe DIYINHK has hacked Lorien work, some suggest)
Click to expand...


The problem is you need the Theyscon driver for native DSD with Windows. And it's not free. They must pay a licensing fee. They simply haven't paid the licensing fee for the driver.
 
#520
So, patiently waiting for the flip-flop PCBs to arrive I thought I would share my probable next steps on this journey.

Currently, I have my initial build of a direct-DSD project working very effectively and consistently. I'm using HQPlayer to upsample to DSD256 on a Windows 10 workstation, outputting to an NAA device running Debian Stretch that works with native DSD256 with my JLSounds USB board. Music sounds great and I'm not experiencing any extraneous noises.

So where next? The first thing is, because everything is now DSD256, to raise the -3dB point of the LP filters further from the audio band; I'm going to 80KHz to start with.

I'm also planning to run with the HQPlayer volume control, removing the cost and complexity of catering for this in the hardware projects. Less is more as the saying goes.

I am planning two projects so I can compare the results.

The first is to use the DSD output from the JLSounds board in a single-ended configuration feeding a Broskie Aikido MFB kit, which includes an active second-order LP filter on its input. The plan is to use the DC offset present on the DSD output for tube biasing (I still wonder whether it is stable enough?). BTW, the Aikido MFB board isn't listed on the Glass-ware website so if its of interest you'll need to contact John Broskie and ask about it.

The other project will use the flip-flop board that I've just put together, outputting to a Broskie BCF buffer. The BCF is a cathode follower configured for balanced input to unbalanced output. I need unbalanced for my power amps.

Will be interesting to see how the different approaches compare.

Any thoughts/observations?

Ray
 
Last edited:
You must log in or register to reply here.
Top Bottom