Not too long ago I ordered a set of the transformers used in DSC3. They arrived a couple of weeks ago, and have been tried in various configurations.
One configuration was with the primary and secondary floating with respect to each other. There are two windings 10k windings, one of which is center-tapped. Tried them with the CT winding as primary, and then tried it again with the CT winding as secondary.
Also tried the transformer with primary and secondary windings sharing a common ground.
All variations sounded different.
Don't know if anyone wants to know how it turned out in any more detail, but will say the point of view here is that maybe they aren't the best option for dac output stage use.
Also, would say these transformers are not all that close to what is possible with some prototype SOA transformers used here for comparison.
Not what I was hoping for. It would have been nice if they turned out to be a good solution for DSD dacs in general. Will report back if I come across anything good to use instead.
One configuration was with the primary and secondary floating with respect to each other. There are two windings 10k windings, one of which is center-tapped. Tried them with the CT winding as primary, and then tried it again with the CT winding as secondary.
Also tried the transformer with primary and secondary windings sharing a common ground.
All variations sounded different.
Don't know if anyone wants to know how it turned out in any more detail, but will say the point of view here is that maybe they aren't the best option for dac output stage use.
Also, would say these transformers are not all that close to what is possible with some prototype SOA transformers used here for comparison.
Not what I was hoping for. It would have been nice if they turned out to be a good solution for DSD dacs in general. Will report back if I come across anything good to use instead.
Got my device fully working, all bugs resolved expect got one banal mistake on inputs pcb, forgot one more i2s input : (, need to redesign it a bit and than devoce will be fully done. Today got some measurements on regulators https://www.diyaudio.com/community/threads/super-regulator.247281/post-7886003 , soon I will do some audio measurements.
Here I fit a sc-pure clock.
Happy new year and happy listening!
Happy new year and happy listening!
in DSC3 resistors were replaced with some microcircuits with a bunch of terminals - what is this?
Any concern changing Vccb on 8T595 from 5.5v to 3.3v to lower the output -4.4db?
I see bandwidth wide enough from the spec.
Reason about that is the transformer performs better with lower input (-8db to -10db from HQP)
With PCM material, HQPlayer can control the volume.
While for DSD material, with direct SDM volume control is disabled.
I see bandwidth wide enough from the spec.
Reason about that is the transformer performs better with lower input (-8db to -10db from HQP)
With PCM material, HQPlayer can control the volume.
While for DSD material, with direct SDM volume control is disabled.
Last edited:
Dear 2A3SET I am doing a DSC3 too, can you tell if teh sc-pure clock is a big improvment ? I search also some information to plug the DSD'IT on DSC3 . Kinds regards
In my opinion reclocking with flip flop as done in dsc3 is enought, realy not needed an extremely precise and not a cheap clocks for this type of dac. I will make fixed ct7302 -> dsc2 dsd512 mode in next revision, only which I will do is two cheap clocks 49 and 45 and put two flipflops as like done on dsc3, that was done very good, excatly what dsc is needed. My next project is power dsc dac, want to create torus transformer with 32 primaries with their center tap, and 64 mosfets, double n type, source to source in serie, , push upll variant, the principe the same as dsc but instead of driving resistors I will try to drive mosfets and their primaries. Volume control by variable power supply. : )
Last edited:
This issue I believe is only related to Crystek clock,I sorted out the issue by using a 1m coax feeding the mlck to Amanero combo768,
Later swapping R12 with 1K resistor also worked out.
View attachment 1356408
With sc-pure clock on the adapter, I got issue now with 48k material @DSD512
Falling back to original 33ohm resolved the problem.
I also don't have problem on 33ohm with NDK clock.
its 0.015% its frol whole chip or for single resistor?Resistor arrays,
it this "Resistor arrays" better that regular resistors?
Since the resistors are switched at RF frequencies, one might also ask about crosstalk and or substrate-coupled noise effects.
Substrate coupling will be very small due to the alumina substrate (good insulator as opposed to the doped semiconductor substrate used for almost all integrated circuits). Magnetic package coupling might be worse than magnetic coupling between separate resistors; I'm not sure if it is worse, but it might.
You should try put 33pF to those 33R to form an RC low pass filter. I have dsd512 working very good, ct7302 -> dsc2, I thinking to fix dsc to fixed dsd512, it work the best on dsd512 expecialy when I play an song with soo many instruments, noticeable improvement while on dsd512. ct7302 is very practical as an asynchronous src bridge and as a detector of the input carrier frequency, so it is very simple to implement an reclocking mechanism for a fixed dsd512 on dsc, that is my plan for revision 3., similar to what is done with dsc3.This issue I believe is only related to Crystek clock,
With sc-pure clock on the adapter, I got issue now with 48k material @DSD512
Falling back to original 33ohm resolved the problem.
I also don't have problem on 33ohm with NDK clock.
My choice for rsistors on revision 3 is https://www.lcsc.com/product-detail...Elec-RTAN0603BTE10K0_C6792862.html?s_z=n_rtan
Last edited:
What about prons like temperature coupling and resistor match in those soic8 package?
Those are bound to be better.
Whether any of it (matching, thermal coupling, magnetic coupling, substrate coupling) matters much for the DAC performance is a whole different question. It's a single-bit FIRDAC, so it should be far less sensitive to these things than a multibit DAC.
I have merged 3 pictures, it tells how things related to reclocking dsc dac is simple while using ct7302, for detecting input carrier freq and setting their output carrier freq, enought for an mcu to decide which oscilator to tun on in reclocking circuity on a fixed dsd512 dsc2
Attachments
My experience of music is almost original when I play music through ddpd, there is a bit of noise, I admit it is audible, but it is a pleasant noise, but I can tell you that there is nothing even close to what ddpd does in relation to digital music! My opinion. In most case revision 3 I will not do but directly switching all my focus to direct digital power dac, this year focus to dsc based direct digital power dac, looks promising because of dsd wide format in dsc and easy work on reclocking 595 shift registers. In comparation to ddpd dsd narow format this dsd wide format migh sound even better!
Last edited:
- Home
- Source & Line
- Digital Line Level
- Signalyst DSC1