Measurements
Are you 'scoping the signal? E.g. o'scope I2S lines at the input of the DAC. Or spectrum analyzer at the DAC's output.
If you're lookin' for stuff to check out, see the Measurements section of Stereophile D/A Processor reviews. ]
WARNING!!!!!!!!!!!!! NOS dacs measure like ****!!!!!!!!!
Example: The $1900 tda1543-based Border Patrol DAC.
See:
BorderPatrol Digital to Analogue Converter SE Measurements | Stereophile.com
Take some screenshots and post in this thread.
How about quantitative metrics?
Are you 'scoping the signal? E.g. o'scope I2S lines at the input of the DAC. Or spectrum analyzer at the DAC's output.
If you're lookin' for stuff to check out, see the Measurements section of Stereophile D/A Processor reviews. ]
WARNING!!!!!!!!!!!!! NOS dacs measure like ****!!!!!!!!!
Example: The $1900 tda1543-based Border Patrol DAC.
See:
BorderPatrol Digital to Analogue Converter SE Measurements | Stereophile.com
Take some screenshots and post in this thread.
Last edited by a moderator:
Yes, they do. However, in doing so, they fit in nicely with just about every other audio component group where some objectively measure like crap, yet subjectively sound great.
Having a component measure "poorly" -- regardless of subjective preference -- is still a valid excuse for gathering metrics, "good' or "bad".Yes, they do. However, in doing so, they fit in nicely with just about every other audio component group where some objectively measure like crap, yet subjectively sound great.
IME, the best measuring DACs I own (Asus Xonar sound card or Musical Fidelity DAC, both heavily DIY'd ) do sound worse than the highly-tweaked Philips CDP (with NOS tda1387) I reported on earlier in this thread. But a heavily-tweaked tda1541a/saa7220 Philips CDP sounds best (it tests so-so).
As of this writing, by far most of the universe is scientifically undiscovered. And that includes information theory and psychoacoustics. Please give us objectives as few more decades ... while our big-budget telescopes and colliders scan the cosmos.
Such as? You've not shown there's a problem with the cooking-grade HC logic yet. What problem exists which the Potato logic is supposed to solve?
No surprises there, exactly what you would expect from the passive TDA1543.
It´s possible to do a lot better (measurement and soundwise) by using one TDA1543 per channel, using only the right channel of the chips, and by selecting chips for low MSB error. I have found this a necessity when I ditched the volume control (started using digital volume). Particularly when I listened late at night with absolute silence outside the flaws became very audible.
The manufacturer´s comment on the next page makes no sense to someone who understands what is going on. He says that R-2R dacs are superior in accuracy; well, NOT the tda1543! First of all, it is not even R-2R. It is pretty inaccurate. He tries to imply that there´s some "objective rightness" to it, and I don´t think that is true at all. Just marketing BS.
But we have to make the distinction: objective vs. subjective. My criticism is the marketing BS and trying to imply an "objective rightness" where there is little of that.
The tda1543 passive sound (relaxed and soft) is more because of flaws and inaccuracies. I still like it, even knowing this. Somehow the psychoacoustic effect works and you may perceive it as sounding right. To me it is reminiscent of cassete tape sound. But again, there is little technical rightness to it. I like having other options.
Last edited:
Well, at least Border Patrol price is not outrageous. Not that I would pay that price, ever.
These ones on the other hand ask £11k (british pounds) for basically the same thing (tda1543 passive).
Vertex AQ Aletheia DAC-1: Looking For a Quiet Life? - The Audiophile Man
Insane. Such is the world of "high end" audio.
These ones on the other hand ask £11k (british pounds) for basically the same thing (tda1543 passive).
Vertex AQ Aletheia DAC-1: Looking For a Quiet Life? - The Audiophile Man
Insane. Such is the world of "high end" audio.
Such as what?
Such as WHAT???!
Uhhh ... your DAC. Measure the pulse-train ... etc., etc. Did you look at the Stereophile page?
It's not insane. Look how much time Braxy "invests" in his project. Some jobs pay by the hour. And not all audio fanatics are DIY po' boyz.
So ... for the avg. Border Patrol buyer, solderin' n' tweakin' (ad nauseam) is "insane".
Last edited:
Precisely old bean, such as what?
Repetition as poetry perchance?
You're not making any sense whatsoever here. WTF does 'measure the pulse train' mean in practice?
As a result of your linking it? Nope. I'd already "invested" some time reading that article prior to your mention of it here.
So after all this digression - back to my earlier question. What is Potato bringing to the party which plain vanilla HC omits?
Last edited:
Measurements
All these long, drawn-out threads full of SUBJECTIVE discussions about Ling this and Phi that ... and how an ancient, jellybean DAC (that Philips themselves DIVESTED interest in) sounds wonderful ... and not a single graph, table, or o'scope screen capture, to document some performance science.
Even Stereophile -- which has $$ at stake when one of their advertiser's products looks metrically "bad" -- realizes the importance of objective Measurements.
BOTTOM LINE:
Toots ... you're in denial. I suggest enrollment in a DIYer Anonymous help group.
If you do not understand THAT -- or (more generally speaking) the validity of metrics for a project you have devoted a SIGNIFICANT portion of your conscious hours to -- your mental clarity is in question.
All these long, drawn-out threads full of SUBJECTIVE discussions about Ling this and Phi that ... and how an ancient, jellybean DAC (that Philips themselves DIVESTED interest in) sounds wonderful ... and not a single graph, table, or o'scope screen capture, to document some performance science.
Even Stereophile -- which has $$ at stake when one of their advertiser's products looks metrically "bad" -- realizes the importance of objective Measurements.
BOTTOM LINE:
Toots ... you're in denial. I suggest enrollment in a DIYer Anonymous help group.
There's no 'if' involved here as I've clearly stated that to me its nonsense.
That's another matter entirely and beyond the context of the present discussion which, if you recall began with your proposal to substitute a Potato logic chip for the HC I've specified. But most likely you don't wish to be reminded of that context
Leaving aside the fact that 'validity of metrics' is a distraction I'm content to go down the path of their validity for a moment as it might shed light on related matters.....
You seem unclear (because you used 'if' at the start of your sentence) that I consider metrics valid. There's no doubt in my mind that I do, however they're not valid as guarantors of audio satisfaction, they're valid as ways to demonstrate the circuit is operating as intended and no major design errors have been committed. Given your lack of clarity about my own view of metrics I consider it instructive briefly to note some evidence you've been avoiding.
An objective metric of frequency response (FR) is of relevance to NOS DACs - the Stereophile review you linked has two figures showing FR of the Border Patrol DAC - fig2 (white noise stimulus) and fig3 (swept tone stimulus). Eagle-eyed observers will note that there's inconsistency between the FR in these two - fig2 shows the classic NOS droop (~3dB@20k) whereas on fig3 the droop is far less pronounced (~0.5dB).
One of the aspects of lingDAC which I'd like to improve on is the dependence of the FR on pre-selecting the inductors. The aim with my CLCLC in lingDAC is is to get a reasonably accurate FR, with the NOS droop corrected. I mentioned that I'm trying a different approach in phiDAC - multiple inductors and a lower Q filter design. Only someone who wasn't paying attention to these remarks (or alternatively read them but didn't understand their significance) would come to the conclusion that FR was unimportant to me.
I do doubt my own clarity sometimes - having someone whose own mental clarity is shown to be lacking comment on the absence of mine is ticklingly hilarious
Oh you mean the ones that you're repeatedly showing you're not paying attention to? Those long, drawn-out threads? So far zero subjective discussion about phiDAC, very little about lingDAC.
Whatever happened to 'COMPROMISED' ?
Yet more evidence you're not following along, or perhaps your memory is selective?
Certainly the measurements are important to their readers. And of course the present discussion isn't about the importance of objective measurements, its about the importance of supporting claims with reasoning and data - the Potato claim being uppermost.
Projection perchance?
phiDAC's v0 PCB layout has been finished, it'll go off for manufacture now. Turn-around is normally under a week. Here's a 3D rendering of the blank board. Next up - work on the BOM
hey man one question
does any/all of the output active stages in Ling/Phi or the ones in the other tda1387 thread have some gain? how many dB?
I'm not really clear on what you mean here. The DAC chip is current out, the finished DACs are voltage out so there's transimpedance in the I/V stage (so many volts out per milliamp input). Is that what you wanted to know? Or are you simply asking for the output level? In which case they're (ling and phi) both 2VRMS max.
Its gain is a transimpedance - a single DAC chip running at 6V has 1.2mA output range for which there needs to be about 5.6V voltage range. So the figure is 5.6/1.2m or about 4700ohms.
In phiDAC that gain is distributed across two stages, with the bulk of it in the first (flat FR) stage. The filter stage has a relatively modest gain of 7.4X (17dB).
In phiDAC that gain is distributed across two stages, with the bulk of it in the first (flat FR) stage. The filter stage has a relatively modest gain of 7.4X (17dB).
hi again
I meant if the output stage was taking the typical 1.2V RMS of the DAC into smth higher. Sorry for the confusion in terminology.
Yes, the output is higher at 2VRMS (the standard for CD players).
Here's a pic of the first completed prototype on the PCB. Listening relative to lingDAC it gives up a little in HF quality, most noticeable on orchestral violins. If the reference is a typical S-D DAC though, then the HF quality is much improved (not to mention the dynamics are in another league against ES9023/PCM5102 etc). Against lingDAC, the soundstage isn't quite as well lit-up over the full width, it seems rather that the edges are 'brighter' and the centre is a little dimmer. But considering the considerably reduced BOM cost and time spent building, its a trade-off I reckon many won't mind making
Here's a pic of the first completed prototype on the PCB. Listening relative to lingDAC it gives up a little in HF quality, most noticeable on orchestral violins. If the reference is a typical S-D DAC though, then the HF quality is much improved (not to mention the dynamics are in another league against ES9023/PCM5102 etc). Against lingDAC, the soundstage isn't quite as well lit-up over the full width, it seems rather that the edges are 'brighter' and the centre is a little dimmer. But considering the considerably reduced BOM cost and time spent building, its a trade-off I reckon many won't mind making
