Hey man if you ask me I disagree as well with you. And I can illustrate that. You chose the Ad811, as for the moment of your hifi journey as the best I/V you heard (my understanding with your prefered chip that is the PCM1702 you find better than ad1862 or tda1541A, the last one being one of your fav if I understood you well.
It's all about trade offs and at the end a one shot according the rest of our system and lisstening room. We all understood it is about trade offs and I agree some shematics are better than others... but it is still about trade offs.
So you chose as a try with good result (which I do not discuss but as the best of one can makes) the ad811 acording W Jung paper and the good reg of the same man as a to do all, whatever most can notice a current buffer can be so so as an I/V stage. OK it has the good low input Z, but a >= 500 R in serie is needed, that is adding their own problems (electronically and layouty ). OK, noone says it is bad at the end, testimonials are saying it is good, and you said better than your own tubes tries if I folllow. (not clear if it is about pcm1702 or TDA1541A... and for the last with a common cathode scheme for the first i/v stage)
BUT, and it is a big but, blah etc ... you came with the servo Super reg (which is one of the best reg, (hey, Monsieur W Jung is not a John Doe of the audio electronic) till now still. Because of the low PSRR and CMR of that particculat chip (the current OPA 811). But hey a D1 like Pass has even (I am not sure but you should know) less Z input with the hexFet than the 811 current chiop (and btw there is a better one NLA that this one, although at Rochester yet)
But there are another paths. We work on red book dac here as far I follow. Ultimate low noise is not needed (while TDA1541A is still spectacular for a 16 bits). And don't get me wrong I work to allow till 192 k hz materials, cause they just are 5% of my library.
A low noise enough voltage "stabilizer" can adds a snap you have no idea untill you experiment it. It is so not a reg, but has not the servoo drawbacks, and are good enough here as far ears can judge of the poors and cons, imho.
Don't get me wrong I am sure your DAC sounds great and are optmized to your whole hifi setup (as mines are).
cheers
It's all about trade offs and at the end a one shot according the rest of our system and lisstening room. We all understood it is about trade offs and I agree some shematics are better than others... but it is still about trade offs.
So you chose as a try with good result (which I do not discuss but as the best of one can makes) the ad811 acording W Jung paper and the good reg of the same man as a to do all, whatever most can notice a current buffer can be so so as an I/V stage. OK it has the good low input Z, but a >= 500 R in serie is needed, that is adding their own problems (electronically and layouty ). OK, noone says it is bad at the end, testimonials are saying it is good, and you said better than your own tubes tries if I folllow. (not clear if it is about pcm1702 or TDA1541A... and for the last with a common cathode scheme for the first i/v stage)
BUT, and it is a big but, blah etc ... you came with the servo Super reg (which is one of the best reg, (hey, Monsieur W Jung is not a John Doe of the audio electronic) till now still. Because of the low PSRR and CMR of that particculat chip (the current OPA 811). But hey a D1 like Pass has even (I am not sure but you should know) less Z input with the hexFet than the 811 current chiop (and btw there is a better one NLA that this one, although at Rochester yet)
But there are another paths. We work on red book dac here as far I follow. Ultimate low noise is not needed (while TDA1541A is still spectacular for a 16 bits). And don't get me wrong I work to allow till 192 k hz materials, cause they just are 5% of my library.
A low noise enough voltage "stabilizer" can adds a snap you have no idea untill you experiment it. It is so not a reg, but has not the servoo drawbacks, and are good enough here as far ears can judge of the poors and cons, imho.
Don't get me wrong I am sure your DAC sounds great and are optmized to your whole hifi setup (as mines are).
cheers
Last edited:
Before Walt Jung released his shunt regulator in my DAC I used a Salas shunt but the first version and I have to admit it was good compared to series regulators either discrete or integrated.
I believe that uBib is even better, but we still can't call it a real regulator, it lacks a feedback loop which limits it.
Cost, size, fact that latest serial ultra low ldo regs are actually quite good, especially for the size and price point. When making ultimate dac (anything really), it costs money, quite a bit, it takes significant portion of shelf space, and not everyone needs/can afford that. To me, bang for buck is a well placed price category for diy, after which you enter serious diminishing returns.
look at Linear Audio test bench...
Can we agree w/o feedback loop it is not a regulator ?
It seems that everyone has their own criteria for what is DIY, for me DIY is that regulator and not a solution that I could have copied from any commercial device. Why shouldn't our DIY be above that commercial device, that's exactly why I see it as a challenge and it constantly pushes me forward.
We can all agree that the Power Supply management is VERY important. I can add, passive parts are also to setup the best results at theend according each own hifi whole ssetup (alas no universal receips that the minimal needed electronic needs, many good "regs" are good enough with.)
Sure W Jung one of the best. Oh it is not about DIY VS commercial in the discuss. We have all huge expectations. I am very respectfull for the sharing and others experiments as well I am wit users personal benchmarks. Indeed we "figth" for the diminishing returns cause again about our expectations (what I like to call "last drosophiles outrage")
Sure W Jung one of the best. Oh it is not about DIY VS commercial in the discuss. We have all huge expectations. I am very respectfull for the sharing and others experiments as well I am wit users personal benchmarks. Indeed we "figth" for the diminishing returns cause again about our expectations (what I like to call "last drosophiles outrage")
Last edited:
Don't get me wrong. My pcm63 dac is fully discrete shunt based regulated. It's a beefy mofo. But that thing costs a pretty penny. Not anyone can afford such a device, and i can tell you that it has diminishing returns to what it could cost for several percent lower performance. Heck, my tda1541a will cost a lot more, and new pcm63 board with fully discrete is on the way. Diy device costing several hundred bucks, that beats A LOT of commercial products costing many, many times as much is where the sweet spot is for most of the people.
You can see a part of the beefy supply. Not saying you are wrong, but saying you are not right either. Power supply is important (to me as much, if not more than anything else in chain), and if budget is a constraint, modern ultra low ldo are a really good solution (and those are hella not used frequently in commercial devices due to cost cutting, where you will see decades old stuff in top shelf stuff too).
I think carbon R is really cool for the job as per the picture. I wonder if real carbon comp are not even better due to some cool no bouncing behavior, though not sure as it is very low power and drift here. But we do not care about the noise with so low Z value (aka, let the carbon comp alive, they became NLA)
And oh, you guys like The tempreance Movement, rock band grup ? Very good musicians, especially with guitars (but we all listen to all sort of music on our systemes, yeah).
And oh, you guys like The tempreance Movement, rock band grup ? Very good musicians, especially with guitars (but we all listen to all sort of music on our systemes, yeah).
Thanks @diyiggy I am aware of that write-up, but thought that was an older version of the Salas shunt and wanted to ask.
I currently use Miro's PS2 with a PCM63 board and Salas shunts with an AD1862 board, but am thinking of trying something different. Just to tinker and see how different power supplies impact the sound
here is a good beginning this gentleman gave as very handfull gift. (@Mark Johnson ) :
Ever-since I got "addicted" to @miro1360 's diy DAC, after rummaging through the dust filled boxes in my lab finding a few, not finding a few, shocked to see the current prices of ICs like TDA1541, OPA627 TO can type I came to the following conclusions:
Hold on to those chips that you have (AD1865N-SOIC, AD1865NJ, and NK)
Buy those which you can (AD1862 2 pairs)
Salvage those which you can (TDA1541, TDA1541A, PCM56 and recently PCM1704) from old working/dead CD players as soon as possible
Because someone/somewhere (with a more active brain than mine) will realize the business opportunity in the old stocks of these long gone out of production chips and start selling them at a premium, which we are now experiencing.
I failed in salvaging AD1862 but when it comes to TDA1541, TDA1541A and PCM1704 I really got lucky; found a number of old dead cd players (lens or supply issues) with working chips for as low as $5, salvaged them used them/using them.
Compared these the DIY DACs with those, that I already have like AK4490, AK4499, ES9018, ES9038Q2M, ES9038Pro and of course Chord DAVE with many combinations of operational amplifiers and honestly there wasn't a single time I was disappointed in these DIY DACs, so miro thanks.
Unfortunately I am not into the habit of testing my musical equipment through various performance charts, because my ears cannot see those charts all they could do is listen and if sounds good then it is good.
Haven't tried Vunce's IV stage or the tube IV stage gentlemen like @Michelag shared, hopefully will do it in the future.
One DAC chip which continues to amaze me is AD1865 particularly the NK variant, haven't seen it in the data sheet but I purchased them way way back when they were still in production and bought them from AD directly.
Would anyone be kind enough to share experience/knowledge on the "NK" version
Hold on to those chips that you have (AD1865N-SOIC, AD1865NJ, and NK)
Buy those which you can (AD1862 2 pairs)
Salvage those which you can (TDA1541, TDA1541A, PCM56 and recently PCM1704) from old working/dead CD players as soon as possible
Because someone/somewhere (with a more active brain than mine) will realize the business opportunity in the old stocks of these long gone out of production chips and start selling them at a premium, which we are now experiencing.
I failed in salvaging AD1862 but when it comes to TDA1541, TDA1541A and PCM1704 I really got lucky; found a number of old dead cd players (lens or supply issues) with working chips for as low as $5, salvaged them used them/using them.
Compared these the DIY DACs with those, that I already have like AK4490, AK4499, ES9018, ES9038Q2M, ES9038Pro and of course Chord DAVE with many combinations of operational amplifiers and honestly there wasn't a single time I was disappointed in these DIY DACs, so miro thanks.
Unfortunately I am not into the habit of testing my musical equipment through various performance charts, because my ears cannot see those charts all they could do is listen and if sounds good then it is good.
Haven't tried Vunce's IV stage or the tube IV stage gentlemen like @Michelag shared, hopefully will do it in the future.
One DAC chip which continues to amaze me is AD1865 particularly the NK variant, haven't seen it in the data sheet but I purchased them way way back when they were still in production and bought them from AD directly.
Would anyone be kind enough to share experience/knowledge on the "NK" version
I don't even know exactly how it came to me from Rocehester, through a friend of a friend in the EU. I got the 4xAD1862 that way, then I got the 4xPCM63 via the forum, and I got the AD1865 in several pieces on ebay as well as the 2xPCM1702, 2xPCM56, 2xAD1860 and 4xAD1856. I've practically turned into a collector of old DAC chips and a warehouseman.