hello,
i am interested to build this DSC project, while i browsing to the signalyst website
the DSC1 is still there at the website to be used with the HQP software
and there are no update information about this hardware, so i think signalyst still using DSC1 until now?
so far i learned, DSC1 has it's I/V and Filters and Output Buffers
while DSC2 and DSC3 using only output transformers,
are there any significant difference between DSC1 until DSC3?
thanks..
i am interested to build this DSC project, while i browsing to the signalyst website
the DSC1 is still there at the website to be used with the HQP software
and there are no update information about this hardware, so i think signalyst still using DSC1 until now?
so far i learned, DSC1 has it's I/V and Filters and Output Buffers
while DSC2 and DSC3 using only output transformers,
are there any significant difference between DSC1 until DSC3?
thanks..
Last edited:
This is difficult to respond not knowing what you find significant.
DSC1 is nice but unbalanced and has several reported issues. Jussi Laako does not use his own design anymore. What he later used for a while was the transformer design from Pavel. But now he is back to commercial DAC's. I have the impression that Holo Audio is on his list of favorites.
DSC2 has in itself several versions. v2.5.2 and 2.6.2 are nice and balanced, and work on the clock provided by the interface. The latest DSC2 design as published on puredsd.ru uses onboard clock and can work with both Beaglebone and USB boards.
In this tread Pavel published an untested further evolution of this board. Geosto and I build this design, correcting some minor mistakes, testing it and corrected some more issues. Pavel helped with releasing new pure firmware which restores the mute function of pin28 of the beaglebone. And then Geosto named this latest version DSC3.
This latest version, designed in principle still by Pavel, really sounds the best. It is probably not so easy to build for people with relatively small experience with SMD, but DSC3 is a very good PCB design, with very nice components, to get the most out of this concept. As it stands, it can compete with the best commercial dacs.
DSC1 is nice but unbalanced and has several reported issues. Jussi Laako does not use his own design anymore. What he later used for a while was the transformer design from Pavel. But now he is back to commercial DAC's. I have the impression that Holo Audio is on his list of favorites.
DSC2 has in itself several versions. v2.5.2 and 2.6.2 are nice and balanced, and work on the clock provided by the interface. The latest DSC2 design as published on puredsd.ru uses onboard clock and can work with both Beaglebone and USB boards.
In this tread Pavel published an untested further evolution of this board. Geosto and I build this design, correcting some minor mistakes, testing it and corrected some more issues. Pavel helped with releasing new pure firmware which restores the mute function of pin28 of the beaglebone. And then Geosto named this latest version DSC3.
This latest version, designed in principle still by Pavel, really sounds the best. It is probably not so easy to build for people with relatively small experience with SMD, but DSC3 is a very good PCB design, with very nice components, to get the most out of this concept. As it stands, it can compete with the best commercial dacs.
@wengtech: Hi ... Since I do not know exactly why you are interested in the DSC1 this may be a bit off ...
But there is also MarcelvdG's RTZ FIRDAC - more information in this thread:
https://www.diyaudio.com/community/threads/return-to-zero-shift-register-firdac.379406/
Bohrok2610 with a somewhat different layout than Marcel's and using MELF resistors reached (to my memory) around - 127 dB THD with this design and using a good clock.
And there is Andrea Mori's DSD DACs (balanced and single-ended) that can be found here:
https://www.thewellaudio.com/lite-series/
Just in case it may be interesting.
Cheers, Jesper
But there is also MarcelvdG's RTZ FIRDAC - more information in this thread:
https://www.diyaudio.com/community/threads/return-to-zero-shift-register-firdac.379406/
Bohrok2610 with a somewhat different layout than Marcel's and using MELF resistors reached (to my memory) around - 127 dB THD with this design and using a good clock.
And there is Andrea Mori's DSD DACs (balanced and single-ended) that can be found here:
https://www.thewellaudio.com/lite-series/
Just in case it may be interesting.
Cheers, Jesper
Yes, these measurements are nice but DSC3 measures better.
https://www.diyaudio.com/community/attachments/rtz_1khz-jpg.1199176/
You see THD of -112dB and that is good, but what bothers me is that the fourth harmonic is at the same level as the second one. With DSC3, only the second and third harmonic have some significance, higher harmonics are lower.
My 1 kHz measurement of DSC3
https://www.diyaudio.com/community/attachments/1-khz-png.1192222/
Of course, these measurements are nice and all that, but I do not know how the boards from the previous post sound. They could sound better then DSC3, I do not know. Only people who build them both could say something about that.
https://www.diyaudio.com/community/attachments/rtz_1khz-jpg.1199176/
You see THD of -112dB and that is good, but what bothers me is that the fourth harmonic is at the same level as the second one. With DSC3, only the second and third harmonic have some significance, higher harmonics are lower.
My 1 kHz measurement of DSC3
https://www.diyaudio.com/community/attachments/1-khz-png.1192222/
Of course, these measurements are nice and all that, but I do not know how the boards from the previous post sound. They could sound better then DSC3, I do not know. Only people who build them both could say something about that.
I'm using DSC2 and to be honest I'm getting trought it and with UCD180HxR unbelievable enob of 17.9bits and thd over 110db which is very good result on my bad first implementation : ) I'm using opa1637 as an diff i/v for dsc2, than ina849 as a diff to se with THP opamp as dc servo, dc coupled all, and furter in chain is mfb filter, also bass/treble tone control and finaly se to diff using opa1637 for ucd180 diff input, results with DSC2 is very good. Can you tell me where is a schematic or anything else in relation to DSC3? I didn't know that DSC3 is exist : ) I would do revision with new opa1633 and also some tune inside tone control digital pot, want to make relay based 6 step pot for booth bass and treble, and maybe DSC3 next time but I didn't know until now that dsc3 is exist : )
more info/problems until full inmplementation about dsc2 out frontend implementation I worked on -> https://e2e.ti.com/support/amplifie...-simulation-bug-with-ina846-or-thp210-in-tina
more info/problems until full inmplementation about dsc2 out frontend implementation I worked on -> https://e2e.ti.com/support/amplifie...-simulation-bug-with-ina846-or-thp210-in-tina
Last edited:
Hi jeroen_d: Thanks for your feedback ... interesting to see measurements of other approaches to discrete DSD DACs 
...
And before writing more I would like to emphasize that I do not place decisive emphasis on measurements, however, they are part of what I use to evaluate any kind of audio electronics including DACs. I think the measurements may give some information about the "structural" performance of a given circuitry although IME they may only hint at how this circuitry may actually sound. IMHO of course.
As to the measurements I do think that bohrok2610 achieved something quite remarkable with Marcel's RTZ FIRDAC as can be seen in this post (#1090):
https://www.diyaudio.com/community/...t-register-firdac.379406/page-55#post-7419240
Cheers, Jesper
...And before writing more I would like to emphasize that I do not place decisive emphasis on measurements, however, they are part of what I use to evaluate any kind of audio electronics including DACs. I think the measurements may give some information about the "structural" performance of a given circuitry although IME they may only hint at how this circuitry may actually sound. IMHO of course.
As to the measurements I do think that bohrok2610 achieved something quite remarkable with Marcel's RTZ FIRDAC as can be seen in this post (#1090):
https://www.diyaudio.com/community/...t-register-firdac.379406/page-55#post-7419240
Cheers, Jesper
What was the input level? To get comparable results you should use -3dBFS level.
In relation to post 2186 , thats very good replacement for dsc2 audio transformers, in the same time you have possibility for tone control with good result close to sound while using audio transofrmers. My next plan for next half analog audio project will be Hypex Nilai frontend using the same dsc2 (maybe dsc3 if I get any info about it) dac definitely, dsc is good sounding dac with great dynamic and is definitely worth to try, in my case its on my main system, just need to revisit it in next project using better digipot with less thd, in my case I have used one of the Maxim Integrated digital switches with discrete pot resistors which was my bad because analog signal trought Max add noise to my tone control, next plan is using the same Max but this time to control relay based resistor lader with 6 steps for treble and 6 steps for bass, it will sound finaly right : ) I will probably open source that project.
Last edited:
Just to share revised concept, work in progress, hope usefull. Configuration will be fully dual mono. Dac dual mono, smps dual, power amp.... etc for dual mono configuration : )
Main care was done to dsc2 frontend. All the gnd signal line gnd is fully separated from supply line, canaling those grounds so that every part of the frontend is on their gnd canal, whole pcb canaled, with their ref gnd at ina849 which doing the magic things here. : ) And also relay based tone control is another improvement. Dual mono in comparation to previous design as stereo variant. New opa1633 instead of opa1637 as i/v. Shileding smps and power amp from fine circuity. All the tings when sumed in comparation to my single pcb stereo variant it will be definitelly better and sound better, measurement too : )
Main care was done to dsc2 frontend. All the gnd signal line gnd is fully separated from supply line, canaling those grounds so that every part of the frontend is on their gnd canal, whole pcb canaled, with their ref gnd at ina849 which doing the magic things here. : ) And also relay based tone control is another improvement. Dual mono in comparation to previous design as stereo variant. New opa1633 instead of opa1637 as i/v. Shileding smps and power amp from fine circuity. All the tings when sumed in comparation to my single pcb stereo variant it will be definitelly better and sound better, measurement too : )
Attachments
Last edited:
Transformer_1P_SS, Description: Transformer, single primary, split secondary as one like on dsc2. I do not see much diference between dsc2 and dsc3 in my case with ct7302 so I will stay with diff dsc1 : ) Thanks for link!
My setup on dsc2 was +-16mA diff using 5k resistors, it sounded realy good with yellow transformers https://www.audiodesignguide.com/DSC2/index.html
resitor value calculation for Idiff=16mA
5V / ((0.016A / 32) + (0.016A / 32)) = Rvalue
My setup on dsc2 was +-16mA diff using 5k resistors, it sounded realy good with yellow transformers https://www.audiodesignguide.com/DSC2/index.html
resitor value calculation for Idiff=16mA
5V / ((0.016A / 32) + (0.016A / 32)) = Rvalue
Last edited:
This is the transformer we use for the DSC3
https://nl.aliexpress.com/item/4001146194450.html?spm=a2g0o.store_pc_groupList.8148356.50.3b307856EJ5zzl&pdp_npi=4@dis!EUR!€ 96,22!€ 96,22!!!101.40!101.40!@212c01e917021275833357163ed239!10000014890914892!sh!NL!2470306294!
The footprint is similar to the yellow transformer used in the DSC2
https://it.aliexpress.com/item/32810421236.html
Last edited:
Audio transformers may benefit from being degaussed when new. For line level transformers, they can be driven with high-ish line level audio, or with something like pink noise, for maybe a couple of days (while loaded nominally, say maybe with 10k resistors).
Also, for dac use its probably better to have transformers with copper tape shields between primary and secondaries. Its harder for the manufacturer to get wide bandwidth that way, since its easier to rely on capacitive coupling between primary and secondary to extend HF bandwidth. However, having a shield connected to the output side ground should help reduce coupling of RF common mode noise from the dac output into the transformer secondaries.
Also, for dac use its probably better to have transformers with copper tape shields between primary and secondaries. Its harder for the manufacturer to get wide bandwidth that way, since its easier to rely on capacitive coupling between primary and secondary to extend HF bandwidth. However, having a shield connected to the output side ground should help reduce coupling of RF common mode noise from the dac output into the transformer secondaries.
Pdm modulated it will setup very fast to the optimum level : ) Yes, thats unlear if yellow or black have shield between primary and secondary, maybe manufacturer can do it by reguest? Or maybe it allready have shileld? To be honest yellow sound very relax and flyid, I realy had no need to try anything else.
Work in progress on diff DSC dac monoblock with their diff frontend, main care to canaling DGND, signal GND and REFGND from the supply GND at a center at REFGND, I'm expecting pcb size around 200x180 with all the things inside, ct7601 as a isolated usb bridge, ct7302 as a heart of pcb with their 4x optic inputs, 3x i2s inputs. Power supply is torus transformer 50VA with 5X secondaries, two for upper monoblock two for lower monoblock, and one 5V for rpi4 and DGND things. Pre regulators is lm317/337 with postregulators mini J.D. super regulators. Looking for low dropout linear regulator 3A for rpi4 and J.D. 5V will be for dsc, digital things also their regulators. Booth will form main pcb for an power amplifier 2x monoblock with their smps supply.
Attachments
Last edited: