So you're saying R2R is the best but is more difficult to make 24 bits? How about 192 or above sample rate is that difficult? I can live without DSD.
I'm sorry if stupid questions, I know very little about DAC DIY.
Some BurrBrown from ifi sounded pretty magical to me. Had very nice qualities that my topping with ES9039PRO doesn't. But overall was worse. But anyhow, if I want to have a real step up, you're saying that R2R is where I should be looking? Nothing bellow 1k eur sounds like much of an improvement, but super expensive DACs sound out of this world. Much better than mine. Maybe it's all on the FE, PSU etc, idk.
No, I was saying new is not always newer basement. It is difficult w/o chip correction if you talk about discrete R2R, discrete can drift with time and precision becomes worse. Even dust on the pcb can be a problem with time.
The subject is not realy in this topic about bits, but maybe more about Sigma Delta vs the older ones which are not and if all is about clarity and THD to get musical result.
I gave an illustration of an expensive SMSL or topping DAC made of PCM1704 that is the max non delta sigma went with 24/192. It was sounding less good imho than the well layouted 20,18,or 16 bits. And I don't say it is about bit, but anyway the expensive one at SMSL or topping is whith old chip. A more expensive discrete design migth exist because discrete is also a reason to say it is better marketing wise. It is more about the whole design that matters. But we need to beginn somewhere, so the dac chip is central in all the discussion.
Upsampling or reconstruction filter is another discussion subject. Many modern dac chip has a NOS filter option but people noticed it is often better made passive with discrete parts, etc ! Idem for the upsampling, some like better computer soft where they also can compute original filters.
It is not about dac chip only, there is poor or cons with newest chip if we are talking about DIY and have choice as the OP has and wonder.
The subject is not realy in this topic about bits, but maybe more about Sigma Delta vs the older ones which are not and if all is about clarity and THD to get musical result.
I gave an illustration of an expensive SMSL or topping DAC made of PCM1704 that is the max non delta sigma went with 24/192. It was sounding less good imho than the well layouted 20,18,or 16 bits. And I don't say it is about bit, but anyway the expensive one at SMSL or topping is whith old chip. A more expensive discrete design migth exist because discrete is also a reason to say it is better marketing wise. It is more about the whole design that matters. But we need to beginn somewhere, so the dac chip is central in all the discussion.
Upsampling or reconstruction filter is another discussion subject. Many modern dac chip has a NOS filter option but people noticed it is often better made passive with discrete parts, etc ! Idem for the upsampling, some like better computer soft where they also can compute original filters.
It is not about dac chip only, there is poor or cons with newest chip if we are talking about DIY and have choice as the OP has and wonder.
Last edited:
what you willl winn with most modern well made designs is ultimate transparency and noise floor. Is it musicaland more conform to the reality is another subject. Some notice a too big transparency w/o weigth according the aera of the audio band. An unbalanced level of clarity and presence on micro dynamic vs macro dynamic which are more noticable with acoustical instrument is often talked. Not saying the output stage still matter. So it is not about bits or SINAD imho.
1) a few examples of sentences I really do not grasp and make both my CPU and RAM seize:
2) No. There are none, zero (0). There are no Group Buys/PCBs for a reproducable/easy to build "true DIY" DAC with adequate illustrated manual or an illustrated thread and recent technology like ES9039Q2M, AK4497SEQ suitable for the average builder. Let alone that they can compete with SMSL products both in price and performance but this is a moot point as they do not exist. There are some people pointing to their designs but these are not available to build yourself AFAIK.
I hope Grunf perseveres with a nice project as the time is about right. Regardless of what we all may think about what is best that is. Enough, no too much threads about that already.
Last edited:
Seems to me I read in another thread that to get the cost down on an assembled board the minimum order is for about 50 pieces. Are there 50 people here interested in a ES9039Q2M dac board that is more well designed and thus more costly to make than a Topping D50? How about if the cost for a board is €350+ ?
Last edited:
I have already a ess9039q2m dac with balanced output for $100. I will not build that myself. Current building a dual pcm1794 with modular in/output. Publish it here when tested. Price range $200 for diy..
Reads a lot like the old, "Value = SINAD/$" belief system.
Would you say you don't believe a significantly better sounding and higher overall quality ES9039Q2M dac could be made for somewhat more cost?
How about a MarcelvdG RTZ dac board combined with a Cestrian Clock and reclocker board for €350+ ?
Last edited:
@jean-paul , rewritten for your cpu:
1) which section, aera, of my post is not understood from you (but grammar)
a)...Mola-Mola, RTZ, Hollow May,...
b) this seems clear to me you just talk about bits and sampling rate only. It doesn't explain your allegation about why old dac chips are not good to you. I have an exemple with your ES9028, that worked better modded but still way below in all the dpt vs all the pcm I have...also for transparency. 99% of the libraries are 16/44. I don t care about streaming or soft that doesn't work when disconnected from internet like the most famous of it.
c)Layout and ground interuption at the dac chip, loss of ground impedance...should matter with the number talked here with new chips. Tht chips can have an edge about power supply and management of the intigrity of signal according the design...
d) Vias are tht, they see the bottom side in a multilayers pcb. And according kickad you can not manage but making different NETs. Burried vias and micro vices are still expensive.
e) 64 bits are announced, if I hear you, one should choose the last technology. Read spl volume management for the last. sentence.
2) Look harder, i see 2 or three project...not finished but existing.
1) which section, aera, of my post is not understood from you (but grammar)
a)...Mola-Mola, RTZ, Hollow May,...
b) this seems clear to me you just talk about bits and sampling rate only. It doesn't explain your allegation about why old dac chips are not good to you. I have an exemple with your ES9028, that worked better modded but still way below in all the dpt vs all the pcm I have...also for transparency. 99% of the libraries are 16/44. I don t care about streaming or soft that doesn't work when disconnected from internet like the most famous of it.
c)Layout and ground interuption at the dac chip, loss of ground impedance...should matter with the number talked here with new chips. Tht chips can have an edge about power supply and management of the intigrity of signal according the design...
d) Vias are tht, they see the bottom side in a multilayers pcb. And according kickad you can not manage but making different NETs. Burried vias and micro vices are still expensive.
e) 64 bits are announced, if I hear you, one should choose the last technology. Read spl volume management for the last. sentence.
2) Look harder, i see 2 or three project...not finished but existing.
Last edited:
The point might have been that its easier to solder THT than it is for small pitch SMD. Vias are not soldered so they don't matter in the same way as THT pads.
That was not my point talked, I talked about layout and signal not soldering. I assume if one is talking about new design, he must know about smd soldering antway...and here also for electrical and RF reasons.
Aha. It is clea…. syntax error
Let’s not forget this is a thread by grunf about a new DAC he will design. Maybe it is more polite to wait for it instead of carpet bombing the thread with comments how things should be done.
Let’s not forget this is a thread by grunf about a new DAC he will design. Maybe it is more polite to wait for it instead of carpet bombing the thread with comments how things should be done.
Last edited:
For me SINAD is just one parameter which is important, low noise and low distortion.
Tells nothing about if I like it. Same for boutique parts or special power supplies or clocks.
Yes those things can make a design sound different, but when the standard parts are chosen well no need for those special modifications. If have tried it all for me still works less is more listening fun and spend more money on different projects.
At the moment I get good reviews on my lat fet composite amp design from people who are long time users of class a as d designs. It is a very simple amp.
Same for my hybrid phono preamp.
I am expect a lot of potential when using the pcm1794 my simple way as done in 2009 by koifarm with his shanling cd-t100se. I have heard this player on several listening demonstrations for me the reference what I like best. It has dual pcm1794 and tube output stage.
Tells nothing about if I like it. Same for boutique parts or special power supplies or clocks.
Yes those things can make a design sound different, but when the standard parts are chosen well no need for those special modifications. If have tried it all for me still works less is more listening fun and spend more money on different projects.
At the moment I get good reviews on my lat fet composite amp design from people who are long time users of class a as d designs. It is a very simple amp.
Same for my hybrid phono preamp.
I am expect a lot of potential when using the pcm1794 my simple way as done in 2009 by koifarm with his shanling cd-t100se. I have heard this player on several listening demonstrations for me the reference what I like best. It has dual pcm1794 and tube output stage.
Anyway, the op was answered and my entries were about what I know from his Dac he showed pictures and which is 100% diy pcb...which can be a problem with modern chips that needs for many several layers, very low capacitance to acheive the good numbers of the datasheet. There was not OT post, at the opposit I given what I think to be poors and cons related from what I know from his work till now about dacs.
It is of course all in his hand or the people that could help. For a first DAC design with smd and RF management. And whatever he choose, industrial avaliability is not his concern as I said, so NLA or not is not a real problem for dac chips but some not seen yet at Mouser... like the voltage ESS one that what rated class A in a cheap Cambridge about its sound which could be spot on with tubes for instance. Up to him.
It is of course all in his hand or the people that could help. For a first DAC design with smd and RF management. And whatever he choose, industrial avaliability is not his concern as I said, so NLA or not is not a real problem for dac chips but some not seen yet at Mouser... like the voltage ESS one that what rated class A in a cheap Cambridge about its sound which could be spot on with tubes for instance. Up to him.
Last edited:
So then, what performance is needed in order to compete, a declaration of audible transparency in an ASR review? Great SINAD? Or, what exactly should constitute "performance?" Maybe listening tests conducted to ITU standards?
Otherwise, to compete in terms of performance is left without definition. That leaves only price.
On this point, it completely ignores all the people buying easy to build plug-together dacs, including manuals, all made by Iancanada, and popularized by internet influencers like Gabster. People drop $$$ on those things because, according to some customers, they sound better that most commercial streamer dacs at the same approximate price point for a finished unit.
It also ignores all the group buys once put on by Andrea Mori, with lots of people lining up to spend some $$ on things like good clocks. Did those clocks make a difference in SQ? IME and in the experience of a lot of other people, the clocks did make a worthwhile difference.
Last edited:
Perhaps, but SQ is clearly not the only driving force behind the slew of discrete dacs over the past few years. Novelty, at least before the Chinese joined the race, was a great selling point. The protected, unreproducible firmware in most dacs of this type is another distinguishing marketing point.
Yet, some first tier companies still seem to do great with off the shelf, sometimes NOS, dac chips.
From what I am told, some of those don't sound all the great compared with some of present day discrete designs. A few or several years ago was a different situation.
Never understood why tubes are needed in a dac. An i/v convertor with tubes in a GG connection is perhaps unique, if technically compromised, but what does a voltage out dac do with tubes?
If the various tube effects are indeed desirable in a system, and they often are, what's wrong with applying them in the line stage? Why start adding the strong spices right at the source?