Signalyst DSC1

...maybe offer the shift registers mounted for (apparent) less skilled solderers... I have no prob with resisters etc...

Yes, sorry for my inexactitude but I meant to say 'preferably pre-populated (with good quality resistors)'. SMD active devices are what I find hardest to solder but if those were to be pre-soldered perhaps it would be as well to do the resistors etc. too?

I have considered the Chinese DSC1 boards available from ebay but I wonder about their use of only 1% tolerance resistors?
 
Found a little time today - have music. Using DSD512 it sounds really very good.

But I have some noise in the background that seems uncorrelated to the signal, varies in intensity but at its loudest is very detrimental. It starts when I hit Play in HQPLayer, even before the music starts. Is this something characteristic of this approach to "decoding" or can I hope to drive it away and let the good music flood through.

Need to figure how to auto the mute too - almost drove my Utopia drivers through my eardrums on one occasion. Not good for me or my cans!

More later.
 
But I have some noise in the background that seems uncorrelated to the signal, varies in intensity but at its loudest is very detrimental. It starts when I hit Play in HQPLayer, even before the music starts. Is this something characteristic of this approach to "decoding" or can I hope to drive it away and let the good music flood through.
In the differential version, this noise is absent.
Need to figure how to auto the mute too - almost drove my Utopia drivers through my eardrums on one occasion. Not good for me or my cans!
More later.
This is a big problem when playing DSD. But it is completely solved in Amanero with the new firmware version of 2003.
 
Hi PPY

Thanks for the information. I (perhaps foolishly) sold my Amanero i/f sme tme ago. Maybe should consider buying another.

I considered the differential board but am not a big fan of transformers and my headphone rig is single-ended end to end so it would not be an ideal match. I'm fiddling with different filters in HQP, not sure if it is really making a difference but does seem to vary.

I need to listen to it over some time and decide if this noise (which is really no worse than typical vinyl background in level) is going to be something I can't live with.

I've only sat down and listened for a couple hours so far, but the music itself is sounding very good, kind of organic. I should have some time over the weekend for more listening and I've had it running overnight to get some signal through the caps and op amps.

Mark
 
I considered the differential board but am not a big fan of transformers and my headphone rig is single-ended end to end so it would not be an ideal match. I'm fiddling with different filters in HQP, not sure if it is really making a difference but does seem to vary.
This noise, similar to the murmur of water, can not be removed in an unbalanced version. It is most strongly manifested in quiet music - jazz, classical music. You can reduce noise only by lowering the frequency to DSD256 or DSD128.
 
hmm

The accuracy of the resistors is not the main parameter here. There is no difference between 1% and 0.1%. The noise parameters and the impulse response of the resistors are more important.

Perhaps not the main parameter, but if you read the thread here, people have confirmed better performance (measured) with higher precision resistors. I suspect that at least .1% should be used, some have gone to .05% and reported improvements. Susumu thin film at this precision at not that much money (we are not talking about foils or anything)
 
Random Thought of the Day

Don't want to push the thread off topic but following my recent experience with the chinese DSC derivative (mostly a clone of Jussi's design) which sounded good except in the really quiet passages, an improved implementation of the DSC1 principle would be nice. I know there is the balanced version from the same ebay team and I will probably get one of those in time.

However, as an enthusiastic owner of the Soekris R-2R DAM1121, it occured to me that it might be possible to create an alternative firmware for that device which would repurpose the ladder as a DSC1 type filter. This is very similar to what Holo Audio have done with their Spring (and one version of the forthcoming Cyan dac).

The Soekris sounds great on PCM, precision of resistors is not such an issue on the DSC1 due to the averaging nature of the process, many of us already own a Soekris dac, what do you guys think? We could speak nicely to Soren...
 
No, with such resistors it will be very bad. It will be strongly influenced by the wiring between the resistors and on them.
I did many measurements with different SMD resistors (MELF Vishay 2%, MELF 0.1%, Susumu 0.1%, Yageo 0.1%). All measurements are almost identical and are practically limited by the capabilities of my E-MU 0204:
  1. THD 0.0008-0.00037%
  2. SNR 98-109db
 
I have the unbalanced ebay kit - pretty close to the Signalyst DSC1 design.

It works well, sounds great - BUT suffers a certain background noise that is audible in quiet patches. Without that it would have probably replaced my Soekris DAM1121 as main dac. The music flows really well through this kit, detailed and relaxed sounding, just a shame you cannot ride the low level of noise (like vinyl surface noise).

The maths seem to indicate that resistor precision is not such an issue as with a typical ladder r2r dac as it averages the level of each bit through the ladder.

I think it was PPY here who has the balanced version and that solves the noise issue.
 
Last edited:
I have the unbalanced ebay kit - pretty close to the Signalyst DSC1 design.

It works well, sounds great - BUT suffers a certain background noise that is audible in quiet patches. Without that it would have probably replaced my Soekris DAM1121 as main dac. The music flows really well through this kit, detailed and relaxed sounding, just a shame you cannot ride the low level of noise (like vinyl surface noise).

The maths seem to indicate that resistor precision is not such an issue as with a typical ladder r2r dac as it averages the level of each bit through the ladder.

I think it was PPY here who has the balanced version and that solves the noise issue.

Thank you so much for answering my concern. The seller also confirm that the low level of noise has been removed by modding something in schematic, damn the seller is liar.
For the transformer, I saw some use 1:1 (as 600:600 or 10k:10k) specially 1:5 transformer, so what I will need for this board?
One more thing, the output is Balanced, so is this possible to change it back to unbalanced because I can't find unbalanced version?