Return-to-zero shift register FIRDAC

Jump to Latest
#3,721
Well, the v14 reclocker board works and I have now mounted it in my 2nd RTZ DAC case:

IMG_20241117_103939.jpg


I have gone away from using the Lundahl output transformers. I just feel now that I was losing to much finer detail with them. That does open up a bit of space in the case for something else....hmmm - Flux Capacitor maybe? 🤓
 
  • Like
Reactions: rockies914, pgour, LTK and 3 others
#3,722
Hi Cestrian,

After it runs for about a week, I would suggest to double check the value of the regulator load resistor. You may find its optimal value varies with type of clocks you are using. Also, its probably getting about to the point where trying an Iancanada clock should show some real benefit. Unless there is a problem with radiated coupling of EMI from the USB and PCM2DSD boards into the clocking circuitry, the clocking function should be able to sound very good.

You also may find some benefit from running the dac in SE mode. That can also help take it to the next higher level.

IME, you still can't have heard what Marcel's dac is capable of doing at its finest.
 
#3,723
Hi Mark,

I haven't yet fitted a load resistor, I wanted to hear what it sounds like without and what change different values make to the sound. As you say, I will give it time to run in properly before making judgements.

Trying SE mode is also on my "roadmap" of things to experiment with. Just because this is all in a case and looking "finished" means very little - it has happened previously and been ripped up and modified hugely. All part of the "fun"! 🙄 🤣
 
#3,725
Hans,

I have abandoned Andrea's DSD dac. It got to a point where I couldn't make it keep up with Marcel's dac and no more practical modifications were possible without a redesign.

Regarding Andrea himself, I understand he has been very ill for some time. Don't know what the outcome will be.

Mark
 
  • Like
Reactions: Hans Polak
#3,726
Currently running in the V14 reclocker board from Cestrian , I've briefly listened to it . I must admit things are sounding really excellent now . The high's are really fast and detailed , cymbals probably sound the best I've heard so far .

I'm using separate regged PSU's for all the stages
Marcel FIRDAC
OPA1632 output stage based on Bohroks design
Cestrians V14 Reclocker including ideas and advice from MarkW
Pjotr PCM2DSD V4 fw without ODDR
Jlsounds I2SoverUSB

This is definitely the best DIY DAC I've tried so far and theres still more things to try , got to say a huge thanks to all those thats contributed to this thread so far .
 
  • Like
Reactions: pgour, rockies914, Spacejack and 3 others
#3,728
My latest versions are linked to in post #1, but there are people who made their own variants.

Quoting post #1:
see post #1931, https://www.diyaudio.com/community/threads/return-to-zero-shift-register-firdac.379406/post-7483396 , for the latest BOM and schematics, and see post #241 https://www.diyaudio.com/community/threads/return-to-zero-shift-register-firdac.379406/post-7175580 and its links for the latest layouts. The latest schematic versions are DAC3_10_holdtimefix.pdf and DAC3filter_6.pdf.
 
  • Thank You
Reactions: savan
#3,729
Markw4 said:
Will give DSD512 from HQ Player a try in maybe a few days. Need the dac in its normal DSD256 state for a few days first.
Click to expand...
A few days ago the 22MHz Acko clock failed. We are working to understand the failure mechanism and get the dac functioning at its best again. Maybe I can do some DSD512 testing pretty soon using the 24MHz clock. Its still on my list of things I need to get done.
 
#3,730
This might be interesting? https://www.diyaudio.com/community/...ef-2-5v-need-your-opinion.420058/post-7849356 This is concept for ultra fine phase locked oven control for an sc cut crystal, and eventualy an fine votage reference can be placed in the same oven. Xor is used for phase lock in an range around 10Hz, that way an fine temperature can be gained probably in range > 0.001C ! I would be glad if as many of you as there are on the forum together try to make this concept of a fine ocxo master clock!
 
Last edited:
#3,732
Some of the manufacturing methods that can lead to low phase noise in an OCXO are:

  1. Designing the crystal resonator to produce low phase noise through a combination of voltage control, buffer elements, and stable frequency setting.
  2. Choosing the right materials during OCXO production to ensure there is no frequency interference brought about by the presence of a material that impedes signal frequency.
  3. Extensive in-house testing of the oscillator to ensure the device conforms to the highest standards of excellence in its performance.
  4. Building the crystal oscillator with as high a Q factor as possible, which results in much cleaner signal processing and consequently, low phase noise.
So one of the conditions for good phase noise is directly related to frequency stability, I didn't included other factors. The good candidate is an SC cut crystal. I thinking to put ct7302 as a crystal driver, ct7302 have 190db dynamic range and an thd+n=-175db, that will give an nice phase noise when combined with sc cut crystal stabilised in an ultra fine oven control. Combining 12MHz sc cut crystal for example with ct7302 which reguire an 12MHz crystal only, ct7302 is fully asynchrone, and you no more need an pcm2dsd, you have now 3 things in the same time, an fine mclk out trought ct7302, an pcm2dsd, dsd2pcm, and not just that, you have also 4 spdif inputs and 3 i2s inputs!

Putting an fine voltage ref inside this oven too and you can make an fine voltage regulator to supply booth things itself.
 

Attachments

Last edited:
#3,733
You should go read Andrea Mori's clock thread. Driscoll oscillator topology measured best phase noise (CT7302 would not be good for that). Heaters aren't needed for audio clocks because slow/long-term frequency drift is inaudible, etc. At least you go SC cut crystals right 🙂

Also, for in-house testing you need PN (phase noise) measurement equipment. Something like this: https://www.miles.io/TimePod_5330A_user_manual.pdf At least of a couple of companies still make such instruments. There is also a lower cost Chinese design but it can't measure the lowest phase noise audio clocks.
 
Last edited:
#3,737
savan said:
https://www.audiodesignguide.com/DSC2/index.html removing dsc2 and putting firdac I think result would be much better than 94db snr?
Click to expand...
Not ideal, far from it IMHO. You should do like LTK did. Ask Cestrian for gerbers and a BOM to build his V14 clock and reclocker board. Once you hear that you will never want to go back to not having it. Use it to drive Marcel's dac. Probably you will find it to be in another league above DSC2.
 
  • Like
Reactions: savan and LTK
#3,738
No need reclocker since I want to use ct7302 in src mode 3, asynchrone, in asynchrone mode I no need reclocker since it act as a reclocker, and not just that ct have dejiter option in registers for booth spdif and i2s, nice ic, this is my 3rd diy build using ct7302, even I have implemented ct7601 for usb audio by eliminating amanero copletely. E1DA Cosmos adc uses ct7601 for their audio usb out, probably with reason. : )
 
You must log in or register to reply here.
Top Bottom